V5 Process Technology 3

Volume 5
Process Technology III
1. Emission Control....................................................................................................... 2
2. Dedusting................................................................................................................. 91
3. Maintenance........................................................................................................... 171
4. Gas Analysing Systems........................................................................................ 267
5. High Level Control ................................................................................................ 338
6. Practical Work ....................................................................................................... 390
Cement Manufacturing Course Version 2005 Volume 5 - Page 1

Chapter 1
Emission Control
1. Sources and Reduction of NOx-Emissions ................................................................. 3
2. Sources and Reduction of SO2-Emissions ............................................................... 45

Sources and Reduction of NOx-Emissions
Arnaud De Luca
TPT 01/21068/E
(Replaces report no. PT 96/14160/E)
1. Introduction.......................................................................................................................6
2. Chemical And Physical Properties And Environmental Aspects Of Some Nitrogen
Compounds And Ozone...................................................................................................6
2.1 Nitrogen (N2) ........................................................................................................6
2.1.1 Physical Properties ..........................................................................................6
2.1.2 Chemical Properties ........................................................................................7
2.2 Nitrogen Oxides (NO, NO2 N2O) ..........................................................................7
2.2.1 Physical Properties ..........................................................................................7
2.2.2 Toxicology .......................................................................................................9
2.2.3 Environmental Aspects ..................................................................................10
2.3 Ammonia ............................................................................................................11
2.3.1 Physical Properties.........................................................................................11
2.3.2 Chemical Properties .......................................................................................11
2.3.3 Toxicology ......................................................................................................12
2.4 Ozone.................................................................................................................12
2.4.1 Physical Properties.........................................................................................13
2.4.2 Toxicity ...........................................................................................................13
2.4.3 Formation of Trophospheric Ozone................................................................14
3. NOx Emission from Cement plant .................................................................................15
3.1 Emission Limit ...................................................................................................15
3.2 Conversion .........................................................................................................16
4. Nitrogen Input into the Kiln System..............................................................................17
5. Behavior of Nitrogen in the Process.............................................................................17
5.1 NO Formation.....................................................................................................17
5.1.1 Formation of "Thermal NO" ............................................................................17
5.1.2 Formation of "Fuel NO" ..................................................................................20
5.1.3 NO-Decomposition Mechanism in the Combustion Process ..........................21
5.2 Formation of Nitrogen Monoxide in a Cement Kiln ............................................21
5.3 Main Influencing Variables for NO Formation ....................................................24
5.3.1 Temperatures .................................................................................................24

5.3.2 Temperature Peaks .......................................................................................24
5.3.3 Excess Air.......................................................................................................25
5.3.4 Residence Time..............................................................................................26
5.3.5 Burner Operating Parameters ........................................................................26
5.3.5.1 In-flame Air Level at Ignition....................................................................26
5.3.5.2 Total Axial Momentum.............................................................................26
5.3.5.3 Tangential Momentum and Swirl Level on the Swirling Channel ............26
5.3.5.4 Swirling and Axial Air Amount, Distribution and Velocity.........................27
5.3.5.5 Influence of Cooler System on NOx Emissions.......................................27
6. NoX Emission Reduction Possibilities..........................................................................28
6.1 Reduction of Nitrogen Input ...............................................................................28
6.2 Primary Measures ..............................................................................................28
6.2.1 Reduction Technologies .................................................................................28
6.2.1.1 Kiln optimisation / Linkman......................................................................28
6.2.1.2 LowNox Burner........................................................................................29
6.2.1.3 Flame Cooling .........................................................................................30
6.2.1.4 Free Lime/ burnability..............................................................................30
6.2.1.5 Lump fuel.................................................................................................30
6.2.1.6 Slag addition............................................................................................31
6.2.1.7 Secondary Firing / Multi-Stage Combustion............................................31
6.2.2 Possible Negative Effects of Primary Measures.............................................32
6.2.3 Effectiveness of NOx Reduction Measures ....................................................34
6.3 Secondary Measures .........................................................................................34
6.3.1 Reduction Technologies .................................................................................34
6.3.1.1 Selective Non-Catalytic Reduction ..........................................................34
6.3.1.2 Selective Catalytic Reduction (SCR).......................................................38
6.4 Effectiveness and cost of NOx reduction measures ..........................................40
6.4.1 Primary Measures ..........................................................................................40
6.4.2 Secondary measures......................................................................................41
Literature.............................................................................................................................42

Summary:
NOx is produced to different degrees in all stationary and mobile combustion sources.
Because of the high flame temperatures the NO generation in cement kilns is relatively high.
To reduce NOx emission, 2 kinds of measures are available:
Primary Measures
Kiln Optimisation / LINKmann
Low NOx burner
Flame cooling
Multi Stage Combustion
These primary measures have a NOx reduction between 0% up to 50% that depends on
main parameters like the measures used or the kiln type. It is therefore very difficult to
estimate their efficiencies.
Secondary Measures
SNCR
SCR (not yet available in cement industry)
The efficiency of these secondary measures can be estimated and they do not influence the
clinker process.
SNCR can reach all current legal NOx emission limits.
Message:
1. NOx is not a kiln control parameter.
2. If legal NOx emission limit cannot be reached with keeping a stable process, SNCR
has to be introduced.

1. INTRODUCTION
The reaction between oxygen and nitrogen at high temperature produces mainly NO with a
small proportion of NO2. The sum of these components is generally called NOx (nitrogen
oxides).
Thus, NOx is produced to different degrees in all stationary and mobile combustion sources.
Because of the high flame temperatures the NO generation in cement kilns is relatively high.
2. CHEMICAL AND PHYSICAL PROPERTIES AND ENVIRONMENTAL ASPECTS

OF SOME NITROGEN COMPOUNDS AND OZONE
2.1 Nitrogen (N2)
2.1.1 Physical Properties [1]

At atmospheric pressure and room temperature, nitrogen is a colorless, odorless, non-
combustible gas. Nitrogen condenses to a colorless liquid at -195.80C and 101.3 kPa and
forms a white solid at -209.86C.
Mr 28.0134
Triple point
Temperature 63.15 K
Pressure 12.463 kPa
heat of fusion 25.8 kJ/kg
Boiling point (101.3 kPa) 77.35 K
heat of vaporization 199 kJ/kg
Critical point
Tcrit 126.2 K
pcrit 3.39908 Mpa
Qcrit 314.03 g/L
Properties at 0C, 101.3 kPa:
Relative density (air = 1) 0.967
Specific heat capacity 1.039 Jg-1 K-1
Dynamic viscosity 15.9 x 10-6 Pa s
Thermal conductivity 23.86 mWm-1 K-1

2.1.2 Chemical Properties [1]
Nitrogen has an extremely high heat of dissociation:
N2 2 N H0 = 943.8 kJ/mol
No marked dissociation takes place even at 3000C and standard pressure. The strength of
the NN bond is responsible for the inertness of N2.
Important reactions of nitrogen with non-metals are those with hydrogen, yielding ammonia,
and with oxygen. The latter, an endothermic reaction, gives nitrogen monoxide:
N2 + O2 2 NO H0 = 180 kJ/mol
2.2 Nitrogen Oxides (NO, NO2 N2O)
2.2.1 Physical Properties [1]

Compounds of oxygen with nitrogen are considered as a class and called nitrogen oxides
(often denoted as NOx). The known oxides and their equilibrium reactions are as follows:
+I
N2O Dinitrogen monoxide
+II +II
Nitrogen monoxide NO N2O2 Dinitrogen dioxide
+III
N2O3 Dinitrogen trioxide
+IV +IV
Nitrogen dioxide NO2 N2O4 Dinitrogen tetroxide
+V
N2O5 Dinitrogen pentoxide
+VI +VI
Nitrogen trioxide NO3 N2O6 Dinitrogen hexoxide

Table 1: Physical properties of nitrogen oxides
Compound N2O NO NO2 / N2O4 N2O3 N2O5
Oxidation state 1 2 +4 / +4 3 5
Tcr, C 36.41 -93 157.85
pcr, Mpa 7.245 6.485 10.132
Qcr, kg/m3 452 520 550
mp, C -90.86 -163.65 -11.20 -100.70 32.4*
bp, C -88.480 -151.770 21.150 -40 to +3
Specific heat cp, kJ kg-1 K-1 0.879 0.996 1.326 0.862 0.778
Standard enthalpy of formation 1864.190 3007.684 721.199 1101.435 104.589
HF, kJ/kg
Heat of vaporization at bp, 376.070 459.031 414.257 517.416
kJ/kg
Density, kg/m3
Gas (0C, 101.3 kPa) 1.9775 1.3402 3.4 (20C) 1.447 2.05
Liquid (20C, 101.3 kPa) 793 1446.8 (2C) (solid)
Dynamic viscosity, mPa -s
Gas (25C, 101.3 kPa) 14.874 19.184 12.838
Thermal conductivity, W m-1 K-1
Gas (25C, 101.3 kPa) 0.01718 0.02573 0.1124
Liquid (20C, 101.3 kPa) 0.1336
* Sublimation point
N2O
Under normal conditions (i.e. room temperature and atmospheric pressure), dinitrogen
monoxide, also called nitrous oxide, N2O, Mr 44.01, is a colorless gas with a weak,
pleasant odor and a sweetish taste. If inhaled, it can bring about a spasmodic inclination
to laugh and a condition resembling drunkenness hence, its historic name, laughing gas.
NO
Nitrogen monoxide, also called nitric oxide, NO, Mr 30.01, is a colorless, toxic,
nonflammable gas at room temperature. As soon as it comes in contact with atmospheric
oxygen, it is oxidized to nitrogen dioxide, a brown vapor.
NO2
Nitrogen dioxide, NO2, Mr 46.01, is a brownish red, toxic gas with a pungent odor; for
physical properties, see Table 1.

2.2.2 Toxicology [1]
N2O
Dinitrogen monoxide (laughing gas) does not irritate the mucous membranes. It has a
powerful analgesic action but is only weakly narcotic. The gas displaces nitrogen from air-
filled body cavities (middle ear, sinuses, intestines, brain ventricles) resulting in an
increase in pressure. After chronic exposure, polyneuropathy and myelopathy have been
observed. TLV-TWA value is 50 ppm (90 mg/m3).
NO
Pure nitrogen monoxide does not have any irritating effects. It reacts, however, with
hemoglobin to form methemoglobin, resulting in cyanosis and possible death. The TLV-
TWA value is 25 ppm (31 mg/m3).
NO2
Nitrogen dioxide is an irritant gas. Its MAK value is 5 ppm (9 mg/m3). TLV-TWA 3 ppm (5.6
mg/m3), TLV-STEL 5 ppm (9.4 mg/m3). Inhalation of nitrogen dioxide causes pulmonary
edema which may result in death (lethal dose 200 ppm). The substance is only slightly
water-soluble but highly lipid-soluble. It therefore penetrates the alveoli where it damages
the capillary walls resulting in exudative inflammation. The respiratory tract is obstructed
due to formation of foam.
Concentrations exceeding 60 - 150 ppm produce coughing and a burning sensation in the
chest. Pulmonary edema becomes apparent after 2 - 24 h. The patient suffers respiratory
distress and insomnia. Chronic exposure to low doses results in coughing, headache, loss
of appetite and gastrointestinal disorders. Patients should be kept under clinical
observation. Inhalation of ammonia from ammonium hydrogen carbonate is
recommended.

2.2.3 Environmental Aspects [2]
Conditions which produce nitrogen oxides do not occur solely in stationary combustion
sources. The exhaust gases from internal-combustion engines, particularly those of the
gasoline-burning spark-ignition variety, contain traces of nitrogen oxides, and because of the
sheer number of them they are estimated to contribute about 50% of the total anthropogenic
NOx burden.
Vehicle exhaust also contains the other ingredients needed to produce the effect known as
photo-chemical smog; this has long plagued large conurbations in badly ventilated situations
such as Los Angeles, Tokyo and Mexico City, and it has more recently become a feature of
many other urban centers. The nitrogen oxides, activated by solar ultraviolet, react in the
urban atmosphere with the unburned hydrocarbons from vehicle exhausts to produce a
noxious cocktail of corrosive, oxidizing and irritating chemicals such as organic peroxides.
They also interact with atmospheric oxygen to produce ozone. It is ironic that, at a time when
one form of atmospheric pollution is destroying the ozone in the stratosphere, which
provides vital protection against excessive solar radiation at the surface, another is creating
it near ground level, where it is an economic and health hazard. Vehicle exhaust is, of
course, discharged virtually at ground level, and its effects are "local" rather than global: the
smog pall from Los Angeles sometimes extends 100 miles or more inland.
But nitrogen oxides are now known to be a key component of a much more widespread
problem which is even more serious in terms of its international implications than is
photochemical smog. It is now plain that a sizable proportion of the blame for the very topical
environmental concern of acid rain which was formerly perceived as an effect only of sulfur
oxides, can in fact be attached to nitrogen oxides. Acid rain which ought really to be termed
"acid precipitation" since the most spectacular manifestations tend to occur in snow rather
than rain is no observer of national boundaries. Pollutants released in one country may
come to earth in precipitation not just in neighboring countries but even in other continents.
As the extent of the problem and its causes have become better understood and
documented, it has become the subject of international dispute and negotiation. Ecological
damage to the forests and lakes of Eastern Canada has been ascribed to (amongst other
things) acidity originating from industrial sources in the US Mid-West. Similar damage which
has ravaged the forests and lakes of Scandinavia has been blamed on pollution exported
from the nearer EEC countries, especially the United Kingdom and Germany, and former
communist central European countries such as Poland, Czechoslovakia and Romania. And
since the communist regimes in these countries collapsed the appalling extent of the
devastation of their own forests has become only too apparent.
By what mechanism acid rain causes forest damage is not known for certain. Two theories
are currently given greatest credence. One proposes that the acidity leaches cationic
nutrients such as calcium, magnesium and potassium away from the root zone and down
into the subsoil, causing starvation. The other postulates that the acidity solubilizes normally
immobile aluminium in the soil; this interferes with the normal uptake of other cationic
nutrients and, if it reaches a threshold concentration (which varies according to the species
of tree), it is directly toxic. This could also explain the dire effect on the aquatic life of lakes
and rivers in the affected regions.
Other pollutants reckoned to damage forests include heavy metals which are given off in
coal smoke as well as in emissions from metal smelters and ozone which - as already
mentioned - is generated in interactions between waste nitrogen oxides and atmospheric
oxygen.

Studies made in the United States for the National Acid Precipitation Assessment Program
(NAPAP) - a statutory body set up under the acid Precipitation Act of 1980 - have shown that
the acidity is worse in high-level rain clouds than in lower-level clouds, suggesting that the
pollutants causing it are carried up into the higher levels rather than diffusing up from ground
level. So, while it is necessary to tackle exhaust emissions from vehicles to alleviate the
smog problem and any strategy for curbing the contribution of nitrogen oxides to acid
rain must center on the control of emissions from the large stationary sources.
2.3 Ammonia
Ammonia, NH3, occurs in nature almost exclusively in the form of ammonium salts. Natural
formation of ammonia is primarily by decomposition of nitrogen-containing organic materials
or through volcanic activity. Ammonia and its oxidation products, which combine to form
ammonium nitrate and nitrite, are produced from nitrogen and water vapour through
electrical discharges in the atmosphere.
These ammonium salts, as well as those arising from industrial and automotive exhausts,
supply significant quantities of the nitrogen needed by growing plants when eventually
deposited on the earth's surface. Ammonia and its salts are also by-products of commercial
processing (gasification, cooking) of fuels with vegetable origins such as coal, lignite and
peat.
2.3.1 Physical Properties
Mr 17.0312
Liquid density (at -33.43C, 101.3 kPa) 0.682 g/cm3
Gas density (at -33.43C, 101.3 kPa) 0.888 g/L
Melting point (triple point) -77.71C
Vapor pressure (triple point) 6.077 kPa
Boiling point (at 101.3 kPa) -33.43C
Heat of vaporization (at 101.3 kPa) 1370 kJ/kg
Standard enthalpy of formation (gas at 25C) -45.72 kJ/mol
Net heating value, LHV 18.577 kJ/g
Gross heating value, HHV 22.543 kJ/g
Ignition temperature acc. to DIN 51 794 651C
Explosive limits
NH3 - O2 mixture (at 20C, 101.3 kPa) 15 - 79 vol % NH3
NH3 - air mixture
(at 0C, 101.3 kPa) 16 - 27 vol% NH3
(at 100C, 101.3 kPa) 15.5 - 28 vol% NH3
2.3.2 Chemical Properties

Gaseous ammonia reacts very violently to explosively with nitrogen oxides to form nitrogen,
water, ammonium nitrate or nitrite. The reaction with N2O does require ignition.

2.3.3 Toxicology
Ammonia is a strong local irritant. On mucous membranes alkaline ammonium hydroxide
forms which dissolves cellular proteins and causes severe necrosis (corrosive effect).
Ammonia or ammonium hydroxide can penetrate the cornea rapidly, leading to keratitis,
damage of the iris, cataract and glaucom.
Oral ingestion of aqueous ammonia can corrode the mucous membranes of the oral cavity,
pharynx and esophagus and cause the shock syndrome, toxic hepatitis and nephritis.
Ammonia is absorbed rapidly by the wet membranes of body surfaces as ammonium
hydroxide, converted to urea and excreted by the kidneys.
Human Exposure: Concentrations of 50 ppm are perceived easily; 50 - 72 ppm does not
disturb respiration significantly. Levels of 100 ppm irritate nose and throat and cause a
burning sensation in the eyes and tachypnoe. In addition to the symptoms described above,
200 ppm induce headache and nausea; 250 - 500 ppm, tachypnoe and tachycardia; 700
ppm, immediate onset of burning sensations in the eyes; 1000 ppm causes immediate
coughing.
The nitrogen metabolism is not significantly changed after exposure to 500 ppm of ammonia.
The TLV value has been set at 25 ppm with a short-term limit exposure value at 35 ppm; the
MAK is established at 50 ppm.
2.4 Ozone
Ozone is thermodynamically unstable and spontaneously reverts back into diatomic oxygen.
This process is promoted by the presence of transition metals or their oxides.
An irritating pale blue gas, ozone is explosive and toxic, even at very low concentrations. At -
111.9C it condenses to form a dark violet liquid which freezes at -192.7C. In the Earth's
stratosphere, it occurs naturally (5-10 ppm), protecting the planet and its inhabitants by
absorbing ultraviolet radiation of wavelength 290-320 nm.

2.4.1 Physical Properties
Mr 48.0
bp (101 kPa) -111.9C
Mp -192.7C
Critical temperature -12.1C
Critical pressure 5.53 MPa
Critical density 437 kg/m3
Critical volume 1.471 x 10-4 m3/mol
Heat capacity, gas

0C 794 Jkg-1 K-1
25C 818 Jkg-1 K-1
Heat of vaporization 15.2 kJ/mol
Heat of formation 144.8 kJ/mol
2.4.2 Toxicity
Ozone, being an extremely powerful oxidizing agent, readily oxidizes a variety of functional
groups in biochemicals. Studies indicate that free radical formation, lipid peroxidation,
carbonyl and aldehyde formation, and oxidation of SH groups, are some of the major sites of
attack. Ozone readily reacts with olefinic compounds, particularly polyunsaturated lipids,
forming unstable ozonides. Their decomposition results in the formation of toxic free radicals
which can in turn amplify the primary cytotoxic or tissue damage. While ozone is considered
to be a toxic gas, there are factors which mitigate the immediate danger to individuals
working with it. Toxicity is dependent on concentration and length of exposure. OSHA has
set an 8-h TWA-PEL of 0.2 mg/m3 (0.1 ppm) for ozone.
Fig. 1 illustrates the relationship between various exposure levels and exposure time for
humans. The odor threshold concentration for ozone is approx. 0.02-0.04 mg/m3 (0.01 - 0.02
ppm).

Figure 1: Human toxicity limits for ozone exposure:
2.4.3 Formation of Trophospheric Ozone

CO + OH H + CO2
H + O2 H M HO2 + M
HO2 + NO OH + NO2
NO2 + Light ( < 420 nm) NO + O
O + O2 + M O3 + M
Net: CO + 2 O2 + Light ( < 420 nm) CO2 + O3

3. NOX EMISSION FROM CEMENT PLANT
There is no normal or average NOx emission from cement kilns. Many factors like kiln
system, fuel characteristics and burning conditions are influencing the NOx emission.
The cement kiln NOx emissions vary largely between 300 and 2500 mg NO2/Nm3.
Figure 2: NOx Emission values from various kiln types (USA plants)
3.1 Emission Limit [16]

All combustion processes primarily produce nitrogen monoxide NO with a much smaller
proportion of nitrogen dioxide NO2 (of the order of 5%). In the free atmosphere, i.e. at
relatively low temperature, however, nitrogen monoxide is oxidized further to form nitrogen
dioxide NO2. Owing to this oxidation, no difference is made between the two gases when
emissions are concerned and they are referred to generally under the formula NOx (= NO +
NO2 expressed as NO2), or nitrogen oxides. However, decisive for the NOx emission is the
formation of nitrogen monoxide (NO) in the kiln system.
Emission Limits: Nm3dry = m3 at 273 K, 101300 Pa and 0% water

Europe 200 - 1800 mg / Nm3dry
Germany:
New plants / modification 500 mg / Nm3dry
Existing plants 800 mg / Nm3dry
100% waste burning 200mg / Nm3dry
(17. BlmSchV)
USA 300 - 1100 mg/Nm3
Note: for comparison reason, all emission limits have been converted to mg/Nm3dry

US plants burning hazardous waste are regulated under BIF (Burners and Industrial
Furnaces). Other plants do have a state permit defining certain parameters like NOx, SO2,
CO and THC emission. The limits for these emissions are called emission standards. These
standards are individually defined for each plant and usually represent the operating
situation under certain conditions. Therefore, the US standards are different from the
emission limits in Europe where emission limits are valid for a whole state or country.
All the above explained emission limits do include definitions how and when the compliance
tests have to be carried out. It is, e.g. a very important difference whether the emission has
to be measured continuously or not.
3.2 Conversion
[ppm NO] to [mg NO2/Nm3]

Usually the value given by the analyser is expressed in [ppm NO] but the legal emission
limits, at least in Europe, are expressed in [mg NO2/Nm3]:
[mg NO2/Nm3] = 2.05 * [ppm NO]

Conversion for a specific oxygen content
The emission limits are usually defined for a specific oxygen content (usually O2 ref=10%).
Therefore, the value given by the analyser has to be converted to the specific oxygen
concentration:
(21 O2ref )
Cref = Cmeasured
(21 O2measured )
Cref [mg/Nm3] => Concentration for a specific O2 content (i.e. as defined in emission limit)
3
C measured [mg/Nm ] => Concentration as measured
O2 ref [%] => Oxygen concentration as defined by the emission limit
O2 measured [%] => Oxygen concentration during the measurement
Other conversions
The emission standards in the USA are usually not using mg/Nm3 but ppm, lb/tdry feed, gr/dscf,
lb/1000lbgas, lb/hr, etc. To convert them into mg/Nm3 certain assumptions are necessary.
Example:
Values given by the analyser at stack: NO = 221 [ppm] and Oxygen = 7%
Assumption: Preheater kiln exhaust gas: 2.4 Nm3/kg cli @10%O2
221 [ppm NO] * 2.05 = 453 [mg NO2/Nm3 at 7% O2]
453 [mg NO2 Nm3 at 7% O2] * (21-10)/(21-7) = 366 [mg NO2/Nm3 at 10% O2]
366 [mg NO2/Nm3 at 10% O2] * 2.4 [Nm3/kg cli] = 0.88 [g NO2/kg
cli]
0.88 [g NO2/kg cli] * 2.204 = 1.94 [lbs NO2/t cli]

4. NITROGEN INPUT INTO THE KILN SYSTEM
Nitrogen is introduced into the kiln system as molecular nitrogen (N2) in the combustion air
(primary, secondary, tertiary) and as nitrogen compounds in the fuel. The quantity of the
relevant N2 introduced into the kiln system is defined by the stoichiometric air required for the
combustion plus excess air.
The concentration of N2 in the air is always 78%.
The quantity of fuel-N is defined by the fuel input and the nitrogen content in the fuel:
Heavy oil: 700 1000 ppm by mass of N
Coal: 1000 4000 ppm by mass of N
5. BEHAVIOR OF NITROGEN IN THE PROCESS
5.1 NO Formation
NO formation only occurs at elevated temperatures (> 800C). It is always connected with
the combustion process.
The formation of nitrogen monoxide is not a simple process that can be described by a few
equations. The complexity of the reactions involved has hitherto prevented the formulation of
a conclusive theory regarding the formation of nitrogen monoxide NO. However, it appears
to consist essentially of two phenomena, the products of which are referred to as "thermal
NO" and "fuel NO".
"Thermal NO": The "thermal NO" is produced by the oxidation of molecular nitrogen in the
combustion air according to the formula:
N2 + O2 2 NO (0)
"Fuel NO": The "fuel NO" is produced by the oxidation of organically bound nitrogen in
the fuel.
5.1.1 Formation of "Thermal NO"

In the zone of combustion products that is after the flame front (oxidation zone), "thermal
NO" is produced if the combustion gases remain for a sufficiently long time at temperatures
above about 1600C. This is determined by what is known as the Zeldovich mechanism
according to the reactions:
N2 + O NO + N (K1) ; N2 + O NO + N (K2) (1.1)

N + O2 NO + N (K3) ; N + O2 NO + O (K4) (1. 2)

Here the speed is governed by that of the fastest reaction, i.e. the reaction between atomic
oxygen (radical) and the nitrogen molecule. The rate of formation is therefore proportional to
the concentration of atomic oxygen and molecular nitrogen. Assuming that combustion takes
place in the presence of an air surplus, the following equation is obtained:
d [NO ]
dt
~ 2 K 1 [N2 ] O [ ] (4)
The pronounced dependence of K1 on temperature can be seen in Fig. 3. Thus, at high

temperatures the equilibrium of the reaction is on the side of NO formation.
In the flames of rotary cement kilns the atomic oxygen primarily comes from the thermal
dissociation of O2:
O2 2 O (5)
Therefore, for the formation of NO the following equation is obtained:
d [NO ]
dt
[ ]
= K [N2 ] O
12 (6)
Figure 3: Dependence of K1 on temperature

The amount of nitrogen monoxide actually produced during technical combustion processes
is a long way below equilibrium concentration (cf. Fig. 4) owing to the marked dependence
on temperature of the NO reaction and the relatively short time that it remains at this
temperature. In air 50% of the equilibrium value is attained at temperatures around 2000C
but only after about 2.5 seconds. The dwell time of the gas in the flames of rotary cement
kilns are obviously a whole order of magnitude below that. Thus, if the gases remain for a
long time at high temperatures, the formation of NO is favoured.
Figure 4: Equilibrium N2 + O2 2 NO
Hence, the formation of "thermal NO" is governed by the following parameters:

Temperature
Gas composition
Dwell time at high temperature

5.1.2 Formation of "Fuel NO"
Fossil fuels contain varying proportions of nitrogen compounds:
Heavy oil: 700 - 1'000 ppm by volume of N
Coal: 1000 - 4'000 ppm by volume of N
Already during the pyrolysis of these fuels, i.e. at relatively low temperatures, the nitrogen
compounds break off as secondary compounds, such as amines and cyanides, which are
then oxidized by OH radicals or O2 to form nitrogen monoxide, or which react with nitrogen
compounds (mainly NO) to form molecular nitrogen. The formation of NO from fuel nitrogen
occurs almost regardless of how it is bonded, but is largely influenced by the amount of
nitrogen contained in the fuel. It can be assumed that all fuel-bonded nitrogen compounds
finally adopt a composition "I", where "I" may be either NH2 or atomic N. These are then
converted into either NO or N2:
I + OX ... NO + ... (7)
I + NO ... N2 + ... (8)
Hence, the formation of "fuel NO" is governed by the following parameters:
The formation of NO depends on the quantity but not of the nature of nitrogen
compounds contained by the fuel
With increasing air surplus the proportion of nitrogen compounds in the fuel converted
into NO increases
Under sub-stoichiometric conditions it is possible for NO to be converted into N2. This
takes place by combination of the NO with atomic nitrogen present in large proportions in
the flame according to the formula:
NO + N N2 + O (9)
The conversion of nitrogen compounds in the fuel to NO depends to a large extent on the
design of the burner of the precalciner. Here such operations as
heating and distribution of drops or particles in the combustion air
driving off the volatile part
heterogeneous combustion
etc.
play a very important role.

5.1.3 NO-Decomposition Mechanism in the Combustion Process
Homogeneous Reaction
NO reacts with hydrocarbon radicals according to equation 10. The product HCN reacts later
like fuel-N and can form NO as well as N2.
CH x +NO
HCN + OH( x 1)
k10 (10)
For a significant NO-decomposition according to (10) high hydrocarbon concentration and

low air factor (~0.6 - 0.9) are required.
Also reactions between NO and HCN as well as ammonia from the fuel do have an influence
on NO-decomposition.
Heterogeneous Reaction
NO can also be reduced on catalytic active surfaces of solids outside of the flame. As
catalyst can act coal and ash particles and metal oxides. Prerequisite for this reaction is the
presence of CO or H2 in the exhaust gas. In case of coal particles also a gas-solid reaction
between NO and the coal can decompose NO.
The gas-gas reaction proceeds according to equations 11 and 12.
k (11)
NO + CO
11
0.5 N 2 + CO2
NO + H 2
0.5 N2 + H 2O
k12 (12)
5.2 Formation of Nitrogen Monoxide in a Cement Kiln

If the formation of nitrogen monoxide (NO) in a kiln has to be considered, it has to be
distinguished between the formation of NO in the main firing system of the rotating part and
its formation in a secondary firing system, if present.
NO Formation in the Main Firing System (rotary part)
In the firing system of a rotary kiln turbulent diffusion flames are used in which the greater
part of the combustion air is only fed in after the fuel has been ignited (secondary air,
proportion about 90%). As a result a severely sub-stoichiometric pyrolysis zone and flame
front are produced. The nitrogen compounds in the fuel are therefore converted into NO to
only a small extent. This means that only few "fuel NO" is produced, the nitrogen in the fuel
is converted into N2. In the zone of the combustion products (flue gas) temperatures up to
1800C are attained and there "thermal NO" is produced. Thus, in the flames of the main
firing systems it is mainly "thermal NO" that is produced.
NO Formation in Secondary Firing Systems
In a secondary firing system the fuel fed in burns at much lower temperatures (800 to about
1200C) than in the main firing system. This means that the NO produced in the secondary
firing system cannot have formed thermally, i.e. from N2 and O2 of the combustion air. In
other words, it is primarily "fuel NO" that is produced in a secondary firing system.

Figure 5: Regions where NO formation occurs in a rotary kiln plant with cyclone
preheater, calciner and tertiary air duct
Kilns without secondary firing

All kilns without secondary firing, i.e. wet process kilns, long dry process kilns and preheater
kilns without riser duct firing, have one thing in common: NO emission is determined
exclusively by the conditions in the kiln burning zone. After leaving the burning zone, the kiln
gases drop rapidly in temperature to a level at which the NO in the gas decomposes very
slowly.
Wet kilns are characterised by a specific heat consumption which is 1.5 to 2 times the
specific heat consumption of a preheater kiln. This means a high specific amount of exhaust
gas. On the other hand, the high specific combustion air consumption will also mean a
somewhat lower secondary air temperature. This combined with the long material retention
time in the burning zone should reduce the NOx concentration in the exhaust gas from the
burning zone.

SP kiln systems with riser duct firing
In many SP kiln systems 10 - 20% of the fuel is fired into the riser duct. Measurements at
several riser duct fired kiln systems indicate that firing coarse fuel (e.g. old tires) into the kiln
riser duct will reduce NOx emission from the kiln system.
This is probably explained by the fact that a large part of the fuel directly falls down into the
kiln charge, creating a reducing atmosphere in the bottom part of the kiln back-end in which
NOx from the burning zone is reduced.
Conversely, when firing finely ground fuel into the kiln riser duct, the specific NOx content in
the exhaust gas will often increase on passing through the riser duct.
As the NOx emission from the kiln may also increase slightly due to an increased excess air
rate the total NOx emission from the kiln system will often increase when starting up riser
duct firing with finely ground fuel.
Precalcining kiln systems
In precalcining kiln systems with tertiary air duct, firing into the rotary kiln typically accounts
for only 40 - 50% of the total heat consumption and the specific amount of combustion gases
from the kiln burning zone is reduced proportionally. On the other hand, the NOx
concentration in the kiln gas may be considerably higher than in preheater kilns.
This is probably due to the shorter material and longer gas retention times in the precalciner
kiln burning zone combined with a very high secondary air temperature.
When examining the contribution from the calciner firing to the emission of NOx, we must
distinguish between two basically different types of precalcining kiln systems, viz.. the In-
Line (ILC) type in which the kiln gas passes the firing region of the precalciner and the
Separate Line (SLC) type in which the kiln exhaust gas bypasses the firing region of the
precalciner.
ILC systems
In these systems the fuel combustion in the calciner takes place in a mixture of kiln exhaust
gas and hot air from the cooler (tertiary air). Some of the nitrogen in the fuel reacts with NO
from the kiln exhaust gas while another part reacts with oxygen (from the tertiary air) to form
NO.
The result may be a net production as well as a net reduction of NO in the calciner.
However, in most cases the calciner contributes a little to the NO emission.
SLC systems
In these systems the combustion in the calciner takes place in pure hot air. In the case of oil
firing, NO production in the calciner is negligible, but when applying fuels containing nitrogen
up to 50% of the nitrogen compounds in the fuel may be converted into NO. The specific NO
production in the SLC calciner may be as high as 4 lb NO2/st (1400 mg NO2/Nm3). This was
measured in a calciner fired with pet coke which has a high content of nitrogen and a low
content of volatiles.
The NO in the calciner exhaust gas is added to the NO in the gas from the rotary kiln which
leaves this type of kiln system without being reduced. When fired with solid fuels, SLC
systems must therefore be expected to generate somewhat higher NOx emissions than ILC
systems.

5.3 Main Influencing Variables for NO Formation
5.3.1 Temperatures
In Chapter 5.1 it is shown that the NO forming reaction is accelerated exponentially with the
temperature. The temperature of the combustion gas is defined by heat generated in the
flame and the heat radiation from the burning zone. Temperature has a major influence on
NO formation.
Figure 8: Influence of the sintering zone temperature on the NO concentration in

the waste gas
5.3.2 Temperature Peaks [6]

Fig. 9 shows the increase in thermal NO formation due to temperature peaks in an ideal
chemical reactor for different amplitudes of a given temperature behavior. For calculating the
curves, first the NO concentrations for the temperatures T + T and T - T were calculated,
averaged and then referred to the NO concentration for T .
The subscript "L" in the diagram characterizes the combustion air temperature before the
start of combustion. It appears from the diagram that the increase in NO formation becomes
greater with increasing amplitude T and decreasing average combustion air temperature.
For practical purposes it is important that NO formation can increase by as much as 20 - 40
% even for very small amplitudes of the gas temperature, e.g. 100 K.
Temperature peaks arise more particularly if, in a given combustion chamber, the
momentum and angle of exit of the swirl air and axial air of the rotary kiln burner cannot be
optimally adjusted to each other or if short-term fluctuations in the fuel feed occur. Modern
burners should therefore offer the greatest possible scope for varying their settings.

Figure 9: Calculated increase in NO formation based on local or time-dependent
temperature differences with the average combustion air temperature as
parameter
5.3.3 Excess Air

Fig. 10 shows the behavior of NO concentration in the waste gas. The graph can be
subdivided into two areas:
A) Air excess factor < 1.2. The NO concentration increases exponentially with the O2
concentration. This is according to the explanation in chapter 5.1 NO formation.
B) Air excess factor > 1.2. The NO concentration is not anymore a function of O2. The
high amount of excess air is cooling the flame, reducing the residence time of the gas
in the kiln and diluting the exhaust gas. These 3 factors are reducing NO generation
and concentration in the exhaust gas.
For most kiln systems the excess air factor is below 1.2. Therefore, oxygen concentration
has a major influence on NO formation.

Figure 10: Influence of excess air on the NO concentration in the waste gas
5.3.4 Residence Time

The longer the combustion gas remains in the very hot part of the kiln (burning zone), the
higher is the NO formation yield.
5.3.5 Burner Operating Parameters

The following burner operating parameters are influencing the flame characteristics (see
[13]). For specific information about formulas and definitions of burner aerodynamics see
chapter 6 report TPT 01/21063/E.
5.3.5.1 In-flame Air Level at Ignition

A minimum of air shall be entrained into the flame prior to ignition. This can be achieved by a
low primary air input and the optimum use of an internal recirculation zone generated by a
bluff body and/or swirl.
5.3.5.2 Total Axial Momentum

Total axial momentum affects the overall entrainment into the flame jet. In general higher
axial momentum result in enhanced mixing and higher NOx emission levels. For mono
channel firing the axial momentum can be observed as a single parameter but for multi
channel types a change in axial momentum also influences the tangential momentum.
5.3.5.3 Tangential Momentum and Swirl Level on the Swirling Channel

The second parameter affecting the entrainment into the flame is the tangential momentum.
In general higher tangential momentum results in more rapid heat release in the near burner
zone and higher NOx emission levels.

5.3.5.4 Swirling and Axial Air Amount, Distribution and Velocity
One of the main parameters affecting the tangential momentum is the swirling air and axial
air distribution and injection velocities. In combination the product massflow (kg/s) times
velocity (m/s) forms the momenta (N) on the different channels. It can be noticed that both
swirling air velocity and amount may have a different influence on the flame characteristics.
5.3.5.5 Influence of Cooler System on NOx Emissions

The following graph from reference [14] is showing the important effect of the cooler system
on NOx emissions.
The grate cooler generating a hot, comparatively clean secondary air has considerable
higher NOx output than the planetary cooler with its dusty secondary air, supposing identical
kiln system for both cooler systems.
Mean value and standard deviation of measured NOx emission from different types of kiln
systems.
Figure 19: NOx emission for different cooler types

6. NOX EMISSION REDUCTION POSSIBILITIES
To reduce the NOx emission from a pyroprocessing system to a certain controlled level,
three basically different methods are available:
Maintain the existing process while reducing the nitrogen input into the system
Modify the existing process (primary reduction measures)
Maintain the existing process while adding a separate gas cleaning unit for the exhaust
gas (secondary reduction measures)
6.1 Reduction of Nitrogen Input

It is not economical to reduce the nitrogen content in the combustion air. However, a
reduction of fuel-N for the secondary or precalciner firing may be feasible under certain
conditions.
6.2 Primary Measures

Primary measures are all actions, which directly influence the burning process. These
measures have a strong impact on the process, positive or negative, and thus have to be
introduced carefully, especially if a low NOx limit is required. (see chapter 6.2.3).
The primary measures can be divided into 2 categories:
1. Measures that reduce the NOx formation:
- Kiln optimization / Linkman
- LowNOx Burners
- Flame cooling
- Free lime / burnability
- Slag addition
2. Measure that reduce the NOx created by the main burner:
- Preheater with Multi Stage combustion (MSC)
6.2.1 Reduction Technologies
6.2.1.1 Kiln optimisation / Linkman

NOx formation is very sensible to all process modifications: the more stable the system,
the less NOx will be produced.
constant fuel and kiln charge flow (short and long term)
To satisfy the first point an accurate dosing system and fuel transport is required.
A constant fuel quality is difficult to achieve if waste fuel is used. Therefore, specific
effort is needed for homogenization, preparation and analysis of the fuels.
constant fuel and kiln charge composition

constant secondary air flow and temperature
The amount of heat from the secondary air depends mainly on operating
characteristics of the clinker cooler. The new grate plates (for the recuperating zone)
developed during the last years, help to stabilise the cooler operation and the heat
input into the kiln.
minimum temperature fluctuations in kiln and cooler
To reduce the temperature of the secondary air the tertiary air should be extracted
from the cooler at the kiln hood.
low excess air factor
A constant kiln and cooler operation is a prerequisite for optimizing of the excess air
factor without reducing conditions for the clinker burning. The excess air factor () for
the main burner is always above 1.
6.2.1.2 LowNox Burner

The so-called LowNox burners are actually low primary air burners as opposed to burners
from direct firing system, whose primary air percentage is much higher.
Nowadays most of burners available on the market are LowNox burners nevertheless
experience has shown that the percentage of primary air has to be kept above 10% in order
to get a good flame, and thus to avoid clinker quality problems or production reduction.
Therefore, a new burner has to be chosen for its qualities related to a good combustion and
not only for a low NOx production (burner specifications, report TPT 01/21063/E).
lower primary air i.e. indirect firing instead of direct firing
flame front closer to the burner by
bluff body
higher volatile coal
optimum distribution of primary air i.e. customised multi channel burner
NOx reduction measures at the burner should achieve a reduction of temperature
peaks and low oxygen content in the flame. Temperature peaks can occur if at a multi
channel burner air volume and exit angle of swirl and axial air are not adjusted properly
and if the fuel flow is fluctuating. Therefore, the burner should have a wide adjusting
range. The optimum operating point has to be determined with systematic long time
tests.
The most important factor for NO generation is the ignition distance (= distance
between burner exit and ignition of the fuel). An extension of the distance is increasing
the NO generation. The reason for this behaviour is the degree of total air (O2) mixed
with the fuel which is higher for longer distances.
short retention time of the exhaust gas in the flame and the kiln
To shorten the retention time of the exhaust gas in the hot kiln, the gas speed in the
kiln should be as high as possible. The limits for the gas speed are dust circulation
and mechanical stress for the refractory materials. The upper limit for the specific fuel
heat input is 7 MW / m2 which is about equal to a specific exhaust gas flow of 2.5
kg/m2s.

6.2.1.3 Flame Cooling
To lower the flame temperature these measures can be used:
dust injection
(polluted) water injection
alternative fuels with high H2O content
Comment
The injection of pure water just for the sake of less NOx does not make much sense. It
would be more attractive to combine this with another advantage as the use of polluted
water or the use of alternative fuels with high water content.
However we do not recommend water injection (fresh as well as contaminated) into the
flame just for NOx reduction due to the possible important influences on the burning process
by changing the flame characteristics.
Quantitative rules :
1. Each ton of water injected into the flame (or anywhere in the high temperature zone >
800C) causes an additional heat consumption on the cement kiln of 4.6 GJ per ton of water.
Deducted example : At 10 g water per kg clinker the increase is 46 kJ/kg cli or +1.4% in heat
consumption.
2. Each ton of water injected into the flame causes a reduction of the kiln capacity of 2 tons
of clinker. That can hurt, e.g for the above example with 10 g/kg cli it means minus 2% of the
production.
3. On a precalciner kiln the injection of 1 g water per kg clinker into the main flame causes a
reduction of the flame temperature by some 10C.
4. There is a limit where the flame temperature gets that low that it is not possible any more
to produce clinker (maintain sufficient low freelime). This point is reached at approx. 42 g
water per kg clinker on a precalciner kiln (which is the absolute limit on a precalciner kiln if
no other disturbing factors would occur).
6.2.1.4 Free Lime/ burnability

The raw mix composition and the clinker quality have also an impact on the NOx production.
If the raw mix is difficult to burnt (i.e. high LSF) or if the free lime in clinker is low, a very hot
burning zone is needed, which leads to more NOx:
lower burning zone temperature i.e. higher free lime in clinker
lower LSF i.e. lower combinability (sintering) temperature
6.2.1.5 Lump fuel

Lump fuel, like tyre, injected in the kiln inlet can reduced NOx emission by the formation of
local reducing conditions. The volatile content is also a main factor for NOx reduction. (see
6.2.1.7)
This technology can also be considered as a secondary measure. Because of the reducing
condition of the waste fuel combustion, NO is reduced to N2. This reaction is called NSNCR
(non-selective non-catalytic reduction).

6.2.1.6 Slag addition
Some tests have been made to reduce NOx by adding slag into the raw mix. The most
known system is the so-called Cemstar process (patented) that injects unground steel slag
into the back of kiln. The impacts on the clinker quality, the cost and the NOx reduction
mechanisms have still to be investigated.
6.2.1.7 Secondary Firing / Multi-Stage Combustion

Mainly fuel-N is contributing to the NO formation (fuel-NO). It is possible to design the
secondary firing in such a way that only little NO is produced and that a portion of the NO
from the kiln is reduced. This can be achieved with a multi stage combustion [8].
Multi-stage combustion has turned out to be a suitable method for conventional and also for
highly sophisticated clinker burning processes. In the first combustion stage which extends
over the sintering zone and the transition zone in the rotary kiln, combustion takes place in
an oxidising atmosphere (excess air coefficient > 1) to ensure good clinker quality. The
nitrogen oxides which are inevitable formed in this high temperature zone are partially
decomposed in a second combustion stage which has, at least locally, a reducing
atmosphere (excess air coefficient < 1).
Figure 20: NOx reduction by Multi Stage Combustion

This reaction pattern is shown diagrammatically in Fig. 20 for a KHD Low-NOx precalciner
(Pyroclon). Multi-stage combustion is logical in precalciner plants in order to avoid renewed
formation of nitrogen oxides in the calciner from nitrogen in the fuel.
A part of the precalcining fuel is introduced above the kiln inlet to form a reducing zone by
arranging an air deficiency (< 1). This fuel burns here under sub- stoichiometric conditions
and generates CO and others components that act as reducing agent for NO created in the
sintering zone and at the same time prevent the formation of more NO.
The remaining precalcining fuel is introduced into the tertiary air duct and is fully oxidised.
The gases are slowly mixed together and at the top of the precalciner, the pyroclon
generates an intense turbulence that mixes the gasflows together and thus fully oxidises the
remaining unburned particles.
Impact of Combustibles
The combustibles used in the precalciner have also a strong influence on the NOx reduction.
Besides the nitrogen content in the fuel, the percentage of volatiles is an important
parameter for NOx reduction. Based on plants experience and theoretical studies, a high
percentage of volatiles in the combustible used in the precalciner has a positive impact on
NOx reduction.
For example, at Rdersdorf plant (KHD-Pyroclon, not Holcim plant), they achieved a NOx
reduction from 900 to 400 mg/Nm3 by changing the fuel introduced into the precalciner from
hard coal(%volatile low) to lignite (%volatile high) [18].
Similar results have been found in LD (MSC-Polysius). The following table gives the
qualitative impact on NOx reduction of different AFR injected into the precalciner.
Table 1: Composition of AFR used in LD and impact on NOx reduction (in calciner)
Combustible PCI H20 Ash Volatile. C H N NOx reduction*

in MJ/kg in % in % in % in % in % in %
Coal Dust 28,5 1,4 12,6 25,2 71,9 3,93 1,5 +
Waste oil 39,1 6,2 ++ (?)
Tire chips 26,4 0,9 18,0 58,3 63,7 6,4 1,0 ++
Fullers Earth 8,4 23,6 45,9 47,3 20,6 2,6 0,2 ++
Paper residues. 3,5 47,6 22,3 53,7 15,0 0,2 0,1 ++
Ferrocarbon 8,5 21,8 63,3 16,6 25,1 3,0 2,5 -/+
Fly ash 1,0 0,1 95,3 1,2 5,6 0,03 0,05 -
*:+: Reduction of NOX emission
-: No impact on NOx emission
6.2.2 Possible Negative Effects of Primary Measures
Reduction efficiency difficult to predict

Because of numerous and complex formation/reduction reactions occurring in the clinker
process, the reduction efficiency of primary measures is difficult to predict.

More over it is very difficult to predict the total influence of several primary measures.
Indeed, the total reduction potential of primary measures applied simultaneously is not the
sum of the potential reduction of each measure.
Not always applicable
Not every primary measures is applicable for all kind of kiln or at least at an acceptable cost.
Limited efficiency
Depends on the process but low NOx limits are not always achievable with primary
measures only (see example here below).
Does not eliminate emission fluctuations
As the primary measures can not react fast enough to the increase/decrease of NOx
emission, the NOx short-term fluctuations (i.e. hourly average) can hardly be solved by
primary measures. The primary measures have rather a positive impact on long term NOx
emission.
Disturbance of process
The process modifications to reduce NOx emission (e.g. water injection, low primary air,)
can greatly perturb the system in such a way that clinker quality or the required production
cannot be reach anymore.
Loss degree of freedom
A main objective for a kiln operator is to produce clinker with a defined quality. If NOx has to
be reduced only with primary measures, it adds a new process parameter, which is not
always compatible with a good quality and high production, especially if low NOx limits are
required.
Cost can increase excessively if high efficiency is required
Example
In order to reach the NOx emission limit of 500 mg/Nm3 with primary measures only, LD
plant (MSC Polysius) had to run the kiln with low oxygen content at kiln inlet. Together with a
high chloride input through AFR, this kiln operation led to a high chloride and sulfur cycle
and consequently to severe corrosion. After 3 years, the damage was so important, that it
has been decided to run the kiln with 3% oxygen at kiln inlet: NOx emission limit could not be
reach anymore and it has been decided to use SNCR to control NOx emission.
Learning:
After this costly experience at LD, we learnt that NOx has not to be a kiln control
parameter. If NOx emission limit cannot be reached with keeping a stable process
(e.g. enough oxygen at kiln inlet, good flame shape, clinker quality), SNCR has to
be introduced.

6.2.3 Effectiveness of NOx Reduction Measures [9]
The following graph gives the NOx potential reduction efficiency for the primary measures.
The NOx reduction efficiency of two or more measures carried out simultaneously is not the
sum of each NOx potential reduction!
6.3 Secondary Measures

Secondary measures are exhaust gas treatments mainly SNCR (selective non-catalytic
reduction) or SCR (selective catalytic reduction). The main difference compared to primary
measures are that secondary measures are not a modification of the clinker process. They
have thus no influence on the clinker quality or on the clinker process.
6.3.1 Reduction Technologies
6.3.1.1 Selective Non-Catalytic Reduction

Selective Non-Catalytic Reduction (SNCR) is the reduction of NOx to N2 by NH3 or other
substance that can liberate NH2-radicals.
SNCR is applied for about 10 years, especially in Europe. For example, in Germany, where
the emission limits are low, there are 20 plants using SNCR.
Reaction Mechanism for NH3
The desired process of reducing nitrogen monoxide (NO) by ammonia (NH3) is initiated by
dissociation of the injected ammonia. At room temperature ammonia is stable, but at roughly
600 to 800C a marked thermal decomposition is initiated by the reaction:
NH3 + O H NH2 + H2O (13)
As the reaction proceeds, the NH2 radical is responsible for the reaction of nitrogen
monoxide (NO).
N H2 + NO N2 + H2O (14)
=> 1 mole NH3 reacts theoretically with 1 mole NO

Reaction Mechanism for urea (CO(NH2)2
The thermal urea decomposition occurs at the same temperature as for NH3 and reacts with
NOx as followed:
NH2-CO-NH2+ O2 2NH2 + CO2 (15)

N H 2 + NO N2 + H2O (16)
=>1 molecule urea gives 2 NH2-radical, which means that 1 mole of urea can
theoretically react with 2 moles of NO
The reduction of the nitrogen monoxide (NO) is effected via the NH2 radical. This means that
all substances capable of supplying a NH2 radical may be used for reduction of NO.
Hereafter is a list with substances that can reduce NO. Nevertheless, not all these
substances have been used in cement plants, and no emission report about the by-products
that could be emitted at stack (N2O, NH3,) is yet available.
Table 2: Nitrogen Hydrogen containing Substances useful for SNCR- Procedure
Substance Formula Aggregate

Ammonia NH3 Gaseous
Ammonia water NH3 aq Solution, 25%
Inorganic ammonia salts (NH4)n Xm Solid
Ammonia chloride NH4Cl Solutions
Ammonia bisulphate NH4HSO4
Ammonia sulphate (NH4)2SO4
Ammonia bicarbonate NH4HCO3
Ammonia carbonate (NH4)2CO3
Organic ammonia salts (NH4)nXm Solid
Ammonia formiate NH4OOCH Solutions
Ammonia acetat NH4OOCCH3
Urea CO(NH2)2 Solid
NH2CONH3.aq Solution
Isocyanic acid OCNH Little stabile
Cyanid acid HOCN Little stabile
Isocyanuric acid (OCNH)3 Solid
Cyanuric acid (HOCN)3 Solid
Cyanic amide H2NCN Solid
Cyanuric acid amid (H2NCN)3 Solid
Amine (NH2)nXm Gaseous, liquid
Nitrile XCN Gaseous, liquid
Influencing Factors
Temperature Window
Reduction of nitrogen monoxide (NO), defined as the decrease in NO relative to the original
NO, is a function of the temperature where the ammonia is injected and the reaction is
made. As Fig. 13 shows, a high NO reduction is only achieved in a narrow temperature
band. This is generally referred to as the "temperature window". In a normal flue gas from a
cement kiln the maximum reduction of nitrogen monoxide (NO) is attained at about 950C.
In a cement kiln at the most the temperature zone between 900 and 1100C is technically
"accessible". This means that the temperature zone is located in the lower stage of the
preheater and this is where the ammonia can be injected. For this injection only the "lower"
(left-hand) part of the NO reduction curve in Fig. 13 can be used.
Dwell Time
The formation of the NH2 radical from the ammonia is dependent on the time. When
ammonia is injected in to the cement process, this dwell time is fixed by the process. In any
case the injection points in a cement kiln must be chosen that this dwell time is as long as
possible.
Figure 21: NO-Reduction (Reaction-kinetic Model Calculation)

Ammonia Slip
The proportion of injected ammonia that does not decompose to NH2 radicals is called the
NH3 slip. This proportion leaves the reaction zone and travels with the flue gas to colder
regions, during which only a small proportion can decompose to NH2 radicals. The greater
part of the slip, though, will remain stable as ammonia and, following various adsorption
processes in the preheater, the cooling tower, the raw mill, or the filter, will finally be emitted
through the chimney as gaseous ammonia or adsorbed by aerosols.
Based on the practical experiences carried out in many plants, SNCR is sufficient to reach
actual NOx limits. This ammonia slip could be nevertheless the main limiting factor for SNCR
use if very low NOx concentrations are required. Ammonia slip can have also an influence
on opacity (special USA regulation).
Injection Rate
The nitrogen monoxide reduction rate rises with increased NH3 injection rate, but work tests
have shown that a greater injection rate will produce a greater ammonia slip.
Ratio [NH3]/[NO0] usually between 0.4 - 0.8 NO0 NO without reduction
Note: ammonia slip has to be added to ammonia emission from the raw mix, whose typical
range is 20-50 [mgNH3/Nm3, 10%O2].
Table 3: NH3 slips for different NOx level (VDZ tests made in 3 German plants) [20]
NOx emission level : NOx emission level

800mg/Nm3 :200mg/Nm3
+ +
Plant description Reducing agent NOx original Molar ratio NH3/NH4 Molar ratio NH3/NH4
level [NH3/NO] emission [NH3/NO] emission
3 direct/comp. direct/comp.
[mg/Nm ] [mg/Nm3]
[mg/Nm3]
Plant A 2700 t/d, 25% NH3 800 1000 0.8 1.1 7 / 11 1.5-2.0 96/110
DS4 solution
Plant B 2000t/d, 25% NH3 1300 1600 0.6 1/18
DS4 solution (for test
3
@ 800mg/Nm )
16%NH3 (for test 1.5/1.8 18/105
3
@200 mg/Nm )
Plant C 1800 t/d, 25% NH3 1100 5/17 31/49
DS4 solution
Ammonia slip can by minimised by:

Correct injection temperature
Enough dwell time to assure a complete mix between the ammonia and the gas
Optimised injection system, especially spraying nozzles
Injection rate based on the stack NOx analyser
Reducing agent

6.3.1.2 Selective Catalytic Reduction (SCR)
Selective Catalytic Reduction (SCR) is a catalytic reduction of NOx. NH3 is still injected to
reduce NOx, but in presence of a catalyst, the reaction occurs at lower temperature (300 -
400C) with a better NH3 exploitation. As the reaction occurs at low temperature, the NH3
emitted from the raw material can also react with NOx.
The catalyst material used is commonly a mixture of titanium dioxides (TiO2) and vanadium
pentoxides (V2O5). The catalyst used depends on the gas composition, i.e. dust content,
SO2, SO3, HCl and HF. For cement plants dust is a main problem and the catalyst design
have to avoid any build-ups to keep enough surface for the catalytic NH3/NOx reaction.
Two locations for SCR installations are possible:
High-dust SCR: before the filter, with high dust load. The problem is to keep the
catalyst clean of build-up. This solution seems to be currently preferable for
technical and economical reasons.
Low dust SCR: located after the filter, in a clean gas. This system required
reheating of the filter exhaust gas, resulting in additional cost.
Figure 21: SCR installation for high-dust and low-dust

Besides dust, SO2 is also a main issue: on the catalyst surface, SO2 is converted to SO3,
which can react with NH3 to produce ammoniac hydrogenosulfat (NH4HSO4). This
component sticks on the catalyst wall and deactivated it (catalyst poisoning) which reduced
the catalyst lifetime.
SCR technology is widely used in other industries for NOx abatement, e.g. in waste
incinerators plants, but only few tests have been carried out in cement plants and one full-
scale unit is currently tested in a Austrian plant.
The main uncertainties are related to the high dust concentration in the gas, the catalyst dust
removal techniques, lifetime of catalysts and total investment.
The main advantages of SCR could be the high NOx reduction efficiency (>90%) with a low
NH3 slip (<10 mgNH3/Nm3). SCR could also reduce PCDD/F, VOC and CO (depends on
catalyst).
The investment and operating costs, which depend greatly on the catalyst lifetime, are still
very high. (Cembureau estimation: Investment of ~3.5 4.5 Mio USD for 3000 t cli/d).

6.4 Effectiveness and cost of NOx reduction measures
The following figure shows a survey from CEMBUREAU carried out in 1999 in several
European countries. It appears that the technologies to reduce NOx differ greatly for each
country. A explanation could be the low NOx emission limits that push the plants to use
SNCR technology (e.g. Germany, Sweden or Switzerland) and also by the types of kiln in
operation in each country (e.g. SNCR for wet kiln are yet no available for wet kiln)
Figure 3: NOx abatement in the European cement industry

NOx Flame Mineralised Staged SNCR SCR
cooling clinker Combustion
Austria 3 1 1
Denmark 1
France 10 1
Germany 4 7 15 1
Italy 3
Sweden 2
Switzerland 1
TOTAL 13 7 11 18 1
6.4.1 Primary Measures

Table 4: Investment cost and operation cost for Primary NOx measures
Method Efficiency Cost

[%] Achievable Invest. Operation
emission level Mio USD USD/year
mg/Nm3
High Level Control (eg.Linkman) 0 20% 0.3 18000
LowNox Burner Burner 0 30 0.3
Firing system 1.0
conversion
Flame cooling 25 50% 0.1
Lump fuel firing (incl. Midkiln) 10 20% 1.0
Multi stage Combustion 20 50 800 - 500 A.S. 1.1
AT: 1.8
Note: the figures given above are investment cost only and they dont take into account the
positive impact or negative on the process.

6.4.2 Secondary measures
Investment cost
Investment cost for SNCR depends mainly of the substance used to reduce NOx, especially
its toxicity class.
Nevertheless, the tendency is to build a storage installation that can accept different types of
substances in order to buy the cheapest product available on the market.
Investment cost: 0.5 1.5 x 106 USD (price based on European experience)
Operating cost
Operation cost depends mainly on the price of the product, the energy for injection and also
the NOx reduction.
Specific cost of reagent:
Ammonia solution 25%NH3: 0.16 USD/t cli
Urea-solution (45%) : 0.24 USD/t cli
Photochemical solution: 0.05 USD/t cli

LITERATURE
[1] Ullmann's Encyclopedia of Industrial Chemistry, Vol. A25 1994 VCH, 3-527-
20125-4
[2] Controlling Nitrogen Oxides, Nitrogen No. 197, May June, 1992
[3] NOx Reduction in the Cement Industry by Application of Multi-Stage
Combustion (MSC) and Selective Non-Catalytic Reduction (SNCR) Techniques,
Dr. L. Bretrup, Krupp-Polysius, Cemtech April, 1991
[4] An Overview of the Formation of SOx and NOx in various Pyroprocessing
Systems, F.L. Smidth, Peter Bechtoft Nielsen, Ove Lars Jepsen, IEEE, May
1990
[5] Stockstoffoxide NOx, Bildung im Zementofen und Reduktionsmglichkeiten,
NOx gerechte Konstruktion fr den Ofen LD 11, U. Fankhauser, VA 92/6086/D
[6] Reduction of NOx Emission in Cement Clinker Burning, A. Scheurer, VDZ, ZKG
No. 3/1988
[7] NOx Minderung durch Einsatz eines Stufenbrenners mit Rauchgasrckfhrung
vom Vorwrmer, H. Xeller, ZKG 40 (1987) H.2, S. 57 - 63
[8] Brennstoffstufung ein wirksames Mittel zur Nox-Emissionsminderung, ZKG 42
(1989)
[9] Cement Kiln NOx Control, A.T. MacQueen and others, Radian Corporation
California, 0-7803-0960-X/93, IEEE 1993
[10] Flames "Semper Sursum", Tom "La Flamme" Lowes
[11] International Flame Research Foundation (IFRF), Ijmiden Cemflame
Consortium, VA 92/50/D, F. Schneider, 1992
[12] Activities of VDZ Committee "NOx reduction", ZKG No. 1/88, J. Kirsch, A.
Scheurer
[13] The effect of burner design and operation and fuel type of cement kiln flames,
IFRF Research report CEMFlAME1, W.L. van de Kamp / J.P. Smart
[14] Die SO2 - und NOx - Emissionen bei modernen Zementdrehofenanlagen mit
Blick auf zuknftige Verordnungen, P.B. Nielsen
[15] Die Bedeutung der 17. Verordnung zur Durchsetzung des Bundes- Immissions-
schutzgesetzes (17. BImSchV) fr die Zementindustrie (Verordnung ber
Verbrennungsanlagen fr Abflle und hnliche brennbare Stoffe in der
Bundesrepublik Deutschland)
J. Waltisberg, HMB, Verfahrenstechnische Abteilung; "Holderbank" NEWS 2/91;
1991
[16] Die Bedeutung der 17. Verordnung zur Durchsetzung des Bundes - Immissions-
schutzgesetzes (17. BImSchV) fr die Zementindustrie (Verordnung ber
Verbrennungsanlagen fr Abflle und hnliche brennbare Stoffe in der
Bundesrepublik Deutschland)
J. Waltisberg, HMB, Verfahrenstechnische Abteilung; "Holderbank" NEWS 2/91;
1991
[17] Cost of NOx Emission Reduction of Holderbank Cement Kilns in European

Union Countries, Th. Lang, ETPS 99/7065/E, 1999
[18] Rdersdorf Emission Solved, European Cement and Construction, p. Kehl at
al., September 1998,
[19] Reference Document on Best Available Techniques in the Cement and Lime
Manufacturing Industries, Integrated Pollution Prevention and Control (IPPC),
European Commission-IPPC Bureau, mars 2000
[20] Application of Best Available Techniques (BAT) in the German cement
industry, M. Schneider at al., Zement-Kalk-Gips, 2000, volume 53, pp.1-11

Sources and Reduction of SO2-Emissions
Arnaud De Luca
TPT 01/21069/E
1. Introduction.....................................................................................................................47
2. Chemical and physical Properties and environmental aspects of some Sulfur
compounds .....................................................................................................................47
3. SO2 Emission Limits / Normal Emission ......................................................................53
4. Sulfur Input into kiln system [4] [6]...............................................................................55
5. Behavior of S-Compounds in the Process...................................................................56
6. SO2-Emission Reduction Possibilities .........................................................................59
6.1 Reduction of Sulfur Input into the System..........................................................59
6.2 Modification of the Existing Process ..................................................................60
6.3 Secondary Reduction Measures ........................................................................61
7. Examples of SO2 Emission Reduction .........................................................................70
8. Literature .........................................................................................................................89

Summary:
The SO2 emission of a cement plant depends on the
type of the kiln
quality of the raw material and fuel
operating conditions of the kiln and the raw mill system
secondary SO2 reduction measures
To reduce the SO2 emission from a pyroprocessing system, three basically different methods
are available:
Maintain the existing process while reducing the sulfur input into the system
With today's technology it is possible to reduce the SO2 emission to an acceptable level. The
most efficient method is the wet scrubber that has a very good SO2 removal efficiency (90-
95%) and produces synthetic gypsum. The drawback of this method is the high investment
cost.
Others methods can be used like SO2 absorption with Ca(OH)2 in the preheater. The main
drawback of these methods is their limited SO2 removal efficiency.
A third method is of course the adjustment of the raw material.

1. INTRODUCTION
SO2 was one of the first air pollutant with which the public made very negative experiences.
Mainly SO2 but also other pollutants from coal fires in the city of London were responsible for
the death of 4000 people between December 5 and 9, 1952.
Today, the SO2 emission is known to be responsible for a part of the acid rain which is
attacking plants and animals, for increased pH of lakes and for corrosion and decomposition
of buildings.
Today, the total worldwide SO2 emission is declining because of reduced sulfur content in
the fuels and scrubbers for large SO2 sources like coal or fuel fired power stations.
The SO2 emission of some cement plants is still on the high side and should be reduced in
the future. The formation of SO2 in kiln systems and methods to reduce them are presented
hereafter.
2. CHEMICAL AND PHYSICAL PROPERTIES AND ENVIRONMENTAL ASPECTS

OF SOME SULFUR COMPOUNDS
Sulfur (S)
Sulfur is the 15th most common terrestrial element and the 9th most abundant element in
the universe. It is widely distributed in nature in different compounds as well as in pure form.
The main sulfur content of the lithosphere is estimated to be approx. 0.05%.
Physical Properties [1]
The melting / solidification point of normal sulfur is 119.3 - 114.5C, depending on the mode
of heating.
Density of solid at 20C, Rhombic 2070 kg/m3
Melting point Natural Rhombic 110.2 C
Density of liquid
115C 1808 kg/m3
445C 1614 kg/m3
Boiling point 444.6 C
Density of vapor /density of air, 470C 1.837
Chemical Properties
Molecular weight: 32,1 g/mol
Sulfur is one of the most reactive elements; it reacts directly with most elements except
iodine, gold, platinum and the noble gases. In humid air it is weakly oxidized, forming traces
of sulfur dioxide and sulfurous acid. At approx. 250C, sulfur ignites in air and burns with a
blue flame but the presence of sulfur dioxide significantly raises the ignition point.

Pyrite / Marcasite (FeS2)
Common and important iron and sulfur mineral
Physical Properties [13], [2]
Pyrite Marcasite
Cristalline form cub. rhomb.
Density t/m3 5 4.87
Melting point C 1'171 450
Solvent HNO3 HNO3
Chemical Properties [2]
Pyrite Marcasite
Molecular weight g/mol 119.98 119.98
Ignition temperature C 350 - 550
Complete combustion C 850 - 940
Four main reaction steps (Eqs. 3 - 6) make up the overall pyrite roasting reaction
(represented by Eq. 7). The specified enthalpy values [7] refer to 298 K:
2 FeS2 2 FeS + S2 (g) H = + 293 kJ/mol (1)

S2 (g) + 2 O2 2 SO2 H = - 723 kJ/mol (2)
2 FeS + 3 O2 2 FeO + 2 SO2 H = - 948 kJ/mol (3)
2 FeO + 0.5 O2 Fe2O3 H = - 282 kJ/mol (4)
2 FeS2 + 5.5 O2 Fe2O3 + 4 SO2 H = - 1660 kJ/mol (5)

Sulfur Dioxide (SO2)
Sulfur dioxide is produced industrially in greater quantities than any other single sulfur
compound. It is generated as the first stage in the manufacture of virtually all the sulfuric acid
used by industry, irrespective of the basic raw material.
Physical Properties [1]
Sulfur dioxide, SO2, is a colorless, non-flammable, toxic gas with a characteristic pungent
smell and acid taste.
Melting point (101.3 kPa) - 75.5C

Dynamic viscosity at 0C 368 Pas
Critical temperature 157.5C
Boiling point (101.3 kPa) - 10.0C
Latent heat of vaporization (at bp) 402 J/g
Standard density at 0C (101.3 kPa) 2.93 kg/m3
Standard enthalpy of formation - 4636 J/g
Specific heat capacity cp (101.3 kPa):
0C 586 J kg -1 K -1
500C 816 J kg -1 K -1
Chemical Properties [1]

Molecular weight: 64.06 g/mol
Sulfur dioxide is very stable; thermal dissociation becomes significant only above 2'000C. It
can be decomposed by shock waves, irradiation with ultraviolet or X rays or by electric
discharge.
The reaction of sulfur dioxide with oxygen to form sulfur trioxide is industrially the most
significant of all its reactions because of its importance in sulfuric acid production. In the gas
phase, it only takes place at elevated temperatures and, for a satisfactory yield of sulfur
trioxide, it requires the presence of a catalyst. In aqueous solution, sulfur dioxide is oxidized
to sulfuric acid at low temperature by air in the presence of activated coke or nitrous gases,
or by oxidizing agents such a hydrogen peroxide.

Environmental Aspects and Toxicology [1]
A substantially larger amount of sulfur dioxide than utilized industrially is produced by the
combustion of sulfurous fossil fuels and is discharged into the atmosphere with the flue
gases. The calculated total emission of sulfur dioxide from power stations, traffic,
households, industry and trade in the Federal Republic of Germany in 1975 amounted to
approx. 3.6 x 106 t. In contrast, only about 2.7 x 106 were used for that year's sulfuric acid
production of approx. 4.2 x 106 t H2SO4.
In the last years the SO2 emissions from power plants in Germany were reduced
continuously with the installation of flue gas desulfurization plant and with the utilization of
low-sulfur fuels, but emissions from traffic, households, industry, etc. in 1990 still account for
620 x 103 t/a. Emissions from power stations amounted to 320 x 103 t/a.
For worldwide atmospheric emissions, only rough estimates are available. For 1970, while
global emissions were estimated at 157 x 106 t SO2, only about 61 x 106 t were consumed in
the production of approx. 94 x 106 t of sulfuric acid. The worldwide cement production (1985)
is estimated to about 700 x 106 t/a. Based on an average SO2 concentration of about 1000
mg/Nm3, the portion of SO2 emitted by cement plants compared to the total anthropogenetic
SO2 emission (200 x 106 t/a) is about 0.7% [5].
Large-scale emission of sulfur dioxide close to ground level has indisputably been the cause
of some spectacular environmental problems in the past. A large area around Sudbury,
Ontario, was completely sterilized by sulfur dioxide from primitive ore roasting operations
around the turn of the century and much of it is still barren as a result of the ensuing soil
erosion. Calamitous sulfuric acid fogs occurred in Donora, Pennsylvania, and London,
England, in 1948 and 1952, when adverse climatic conditions prevented sulfur dioxide from
industrial sources and domestic coal fires from dispersing.
The causes of these episodes were correctly identified and the action taken to control
ground level smoke and sulfur dioxide concentrations has effectively prevented them from
recurring. In the case of industrial sources, however, this has often been achieved by merely
building taller chimneys to disperse the sulfur dioxide over a wider area and as the
consumption of fuels has increased over the years it has been argued that the problem has
merely been shifted from the location where the sulfur dioxide is produced to other locations
downwind.
Considerable publicity has been given to the apparently increasing problem of acid rain and
there does, indeed, seem to be a correlation between the incidence of acid rainfall and the
atmospheric transport of pollutants from major industrial locations.
Establishing in any degree of certainty to what extent sulfur emissions are responsible for
acid rain is, however, complicated by incomplete knowledge of the magnitude of sulfur
emissions from natural sources, the atmospheric chemistry of sulfur and the importance of
other potential acidulates such a nitrogen oxides.
Nevertheless, the evidence linking sulfur dioxide pollution to acid precipitation has been
mounting gradually. A recent report by the U.S. National Academy of Sciences, for example,
has stated that there appears to be a direct proportionate relationship between sulfur dioxide
pollution and the amount of acidic sulfates in precipitation.

Ill-effects of sulfur dioxide on humans and animals are mainly related to irritation and
damage of the moist mucous membranes by the formation of sulfurous acid. The odor
threshold of sulfur dioxide in air is between 0.3 and 2.5 ppm. In most human beings,
concentrations of 5 - 10 ppm will lead to irritation of the respiratory tract; in sensitive people
they may produce spasms of the bronchi. Higher concentrations will cause heavy irritative
coughing, while breathing sulfur dioxide at concentrations above 400 - 500 ppm, even for
only a few minutes, is dangerous to life. The maximum workplace concentration is defined
as 5 ppm (14 mg/m3).
Long-term exposure may possible lead to bronchopneumonia and, in extreme cases, to toxic
pulmonary edema with dyspnoea, cyanosis and cardiac-circulatory failure.
Sulfurous acid absorbed by the body is converted to sulfate and discharged in the urine, as
evidenced by an increase in urine acidity.
Plants are impaired by sulfur dioxide at even lower concentrations than human beings.
Sulfur dioxide levels of 1 - 2 ppm are enough to cause acute damage to the leaves of plants
in only a few hours, followed by necrosis, probably resulting from impaired photosynthesis.
Figure 4: Atmospheric Chemistry of SO2 [3]

Sulfuric Acid
Physical Properties
Pure sulfuric acid H2SO4 is a colorless, water-white, slightly viscous liquid, mp 10.4 C, bp
279.6C. It can be mixed with water in any ratio.
Chemical Properties
Molecular weight: 98.08 g/mol
Sulfuric acid is a strong acid with characteristic hygroscopic and oxidizing properties. Sulfuric
acid, like the sulfate ion, is chemically and thermally very stable.
SO2 + 1/2 O2 SO3 H = - 99.0 kJ (6)

SO3 (g) + H2O(l) H2SO4 (l) H = -132.5 kJ (7)
Environmental Aspects and Toxicology [1]

Sulfuric acid has a highly corrosive effect on the eyes, the mucous membranes and the skin,
even in low concentrations. Because it completely destroys living tissue, concentrated
sulfuric acid causes burns that penetrate deeply and heal only slowly. Swallowing sulfuric
acid produces extreme pain in the digestive tract, vomiting and shock and there is a danger
of perforation.
Sulfuric acid vapors or mists irritate the eyes and the mucous membranes of the nose,
pharynx and respiratory tract, causing heavy coughing and breathlessness.
Sulfuric acid releases are highly deleterious with respect to ground and surface waters. The
substance is toxic to both fish and algae, both directly and as a result of reaction with other
materials in the water. Any concentration 1.2 mg/L is considered lethal to fish: 6.3 mg/L or
more causes death within 24 h.
Sulfur Trioxide (SO3)
Physical Properties
Gaseous SO3 Liquid SO3
Nominal density, g/L (0C, 1013 mbar):3.57 Density (25C): 1.9 g/cm3
bp, (1013 mbar): 44.8 C
-3 -1
Specific heat Cp, kJ m K Vapor pressure,
100C 2.543 20C 0.26 bar
500C 3.191 100C 8 bar
Heat of evaporation (boiling point): 538 J/g
Critical temperature: 217.7 C

3. SO2 EMISSION LIMITS / NORMAL EMISSION
Emission Limits
Germany 400 mg/Nm3 dry (no waste fuels)

Switzerland 500 mg/Nm3 dry
Europe 100 - 2400 mg/Nm3
Normal Emission
There is no normal SO2 emission level for cement kilns. The SO2 emission depends mainly
on quantity and quality of "S" input, kiln system and SO2 reduction systems. It is known that
SO2 emission of cement kilns can be between very low values of less than 50 mg/Nm3 and
very high values up to 3500 mg/Nm3.
Figure 5: Average SO2 Emission according to a Holcim Survey
35
30
25
NUMBER OF KILNS
20
15
10
0
>2100
0 - 300
300 - 600
600 - 900
900 - 1200
1200 - 1500
1500 - 1800
1800 - 2100
NOT AVAIL.
SO2 CONCENTRATION AT THE STACK [mg/Nm3 dry]

Figure 6: SO2 Emission for different kiln types
Result of PCA SO2 emission survey 1982 [4]

Assumption: wet: 2.1 Nm3 dry / kg cli (18 kilns)
long dry: 1.5 Nm3 dry / kg cli (12 kilns)
SP: 1.4 Nm3 dry / kg cli ( 5 kilns)
Precalciner: 1.4 Nm3 dry / kg cli (12 kilns)

4. SULFUR INPUT INTO KILN SYSTEM [4] [6]
Contained in the raw materials (mainly as FeS2) and the fuel, sulfur enters the process
mainly in the form of sulfates (gypsum CaSO4*2 H2O), sulfides (Pyrit: FeS2) and organic
sulfur compounds. In the process, the sulfur compounds may either be reduced or oxidized
to form gaseous SO2.
Table 1: Sulfur content of kiln feed and fuels of the Holcim Plants in 1994
% SO3
Average Minimum Maximum
Kiln Feed 0.46 0 1.93
%S
Fuels Average Minimum Maximum
Coal 0.86 0.02 3.51
Pet Coke 4.22 1.01 8.3
Diesel Oil 2.61 0.02 3.5
Heavy Fuel Oil 2.86 0.12 11.6
Alternative Fuels Average Minimum Maximum

Liquid 0.56 0.11 2.01
Solid 0.92 0 3.9
Natural Gas 0.52 0 3.17
Tires 1.63 0.8 3.9
Waste Oil 1.16 0.5 2

5. BEHAVIOR OF S-COMPOUNDS IN THE PROCESS
Wet Kilns
A schematic representation of the sulfur circulation in a wet kiln is shown in fig. 4. Depending
on the raw materials, the conditions in the kiln burning zone and the internal circulation, 30 -
80% of the sulfate compounds entering the kiln burning zone may decompose and form SO2
which leaves the kiln burning zone with the exhaust gas together with the SO2 formed by
oxidation of the organic sulfur in the fuel fired into the kiln.
On passing through the calcining zone, the preheating and the drying sections of the kiln,
some of the SO2 will be reabsorbed in the raw materials but due to the poor contact
between the kiln gases and the raw materials, especially in the calcining zone, 40 - 60% of
the SO2 from the kiln burning zone will be emitted from the kiln.
In the preheating section of the kiln, the SO2 concentration in the gas may even be
increased if the raw materials contain sulfides which will burn here in accordance with
equation:
+II-I >550C III -II IV -II

______
4 FeS2 + 11O2 2Fe2O3 +8 SO2 (8)
Having left the kiln, the exhaust gas usually only passes through a precipitator before being
emitted to the atmosphere through the stack.
Depending on the type of precipitator between 10 and 30% of the SO2 in the exhaust gas
may be absorbed in the raw materials at this stage. Taking into account internal circulation
and the evaporation and absorption rates mentioned, about 30% of the sulfur entering the
kiln system will be emitted as SO2. The emission will increase roughly in proportion to the
total input of sulfur compounds in raw materials and fuel.
Since the specific heat consumption of wet kilns is high, SO2 emission depends very much
on the sulfur content in the fuel. If the excess air is reduced below a certain level, a sharp
rise in SO2 emission from the kiln may occur, as local reducing conditions will increase SO2
formation in the burning zone and make SO2 re-absorption in the kiln back-end more difficult.
Long dry kilns
Except for a possible slurry preheating section, the design of the long dry kiln is virtually the
same as that of a wet kiln.
Consequently, the SO2 absorption factors are comparable to those of wet kilns as mentioned
above.
However, with the same type of raw material and fuel the SO2 emission from the long dry kiln
system will be lower than that from the wet kiln. This is due to the lower specific heat
consumption and the fact that the exhaust gases from the long dry kiln are often used for
drying in the raw mill in which 20% - 50% of the SO2 might be absorbed by intimate contact
with freshly ground raw meal particles.

Figure 7: Behaviour of sulfur in a wet or long dry kiln system without raw mill
Preheater kilns
A schematic representation of the sulfur circulation in a dry-process preheater or precalciner
kiln with bypass is shown in Figure 5.
In 4 and 5-stage cyclone preheater kilns complete preheating of the raw meal takes place in
intimate contact with the exhaust gas from the kiln and in the lower stages the temperature
reaches 850C at which point part of the raw meal starts to calcine.
The SO2 coming from the kiln is thus brought into contact with free CaO and CaCO3 at a
temperature at which the following reactions proceed relatively fast:
+II II IV -II II IV -II
CaO + SO2 + O2______ CaSO4 (9)
+II IV II IV -II II IV -II IV -II
CaCO3 + SO2 ______ CaSO3 +CO2 (10)
In this way, nearly all SO2 formed in the kiln is absorbed by the hot meal and reintroduced
into the kiln and bound into the clinker. Only if the sulfur circulation between the kiln and the
lower preheater stage reaches extreme levels or in the case of local reducing condition in
the kiln back-end and the riser duct SO2 may escape via the lower stage. This will often be
the case if coarse waste fuel (e.g. rubber tires or pieces) is fired into the riser duct or kiln
inlet.

Except for this situation, considerable SO2 emission from preheater kilns (without by-pass)
will only occur in case the raw materials contain non-sulfates such as pyrites which form SO2
already in the upper stage cyclones where the temperature and the concentration of free
lime and alkaline materials are too low to ensure complete reabsorption of the SO2 formed,
according to above two reactions (9, 10).
Normally, 30 - 50% of the sulfur present in the raw meal in the form of pyrites will leave the
preheater as SO2. Part of this is absorbed in the raw mill and the precipitator. Still, SO2
emission from the stack will amount to 15 - 30% of the sulfur entering with the raw materials
in the form of non-sulfates.
This means an SO2 emission of 500 - 1000 mg / kg cli for each per mil sulfur present as non-
sulfate in the raw materials.
Figure 8: Behaviour of sulfur in a preheater kiln system

Precalciner kilns
The precalciner kiln system in which the kiln gases pass the precalciner offers ideal
conditions for absorption of SO2 from the kiln due to the high amount of free CaO and a
temperature of approx. 900C in the calciner.
But also other precalcining systems where kiln gases enter into contact with sufficient free
CaO in the kiln riser duct and the lower cyclones absorb SO2 from the kiln.
The only source of SO2 emission from the preheater of a precalcining kiln system is
therefore any content of non-sulfates in the raw meal which will generate SO2 in the upper
preheater stages as explained above. Many precalciner kilns in the USA have a kiln gas
bypass. On leaving the kiln riser duct the kiln gas is cooled by air dilution and/or water
injection to 400 to 230C (depending on the filter system). The bypass outlet is placed to
ensure the lowest possible dust concentration in the bypass gas.
This means limited possibility of absorption of the SO2 extracted with the kiln gases in the
bypass duct. Very often, more than 50% of the SO2 contained in the by-passed kiln gases
will thus enter the atmosphere via the stack. This might explain why the average SO2
emission from the precalciner kiln systems is higher than that from the SP kiln systems.
It should be borne in mind that the evaporation factor for sulfur components in the kiln and
consequently the SO2 emission from the bypass may increase drastically if coal combustion
in the precalciner is incomplete. In this case, residual carbon from the calciner will enter the
kiln inlet where it will create reducing conditions, resulting in greatly increased sulfur
"evaporation". Accordingly, improving the combustion in the precalciner is a mean of
reducing the SO2 emission from a system with bypass.
6. SO2-EMISSION REDUCTION POSSIBILITIES

To reduce the SO2 emission from a pyroprocessing system to a certain controlled level,
three basically different methods are available:
Maintain the existing process while reducing the sulfur input into the system
6.1 Reduction of Sulfur Input into the System
Raw Material
A reduction of the total sulfur input or very important for preheater kilns the total input of
sulfides is often possible. A sulfur-selective quarrying may not only reduce the total input of
sulfur but also reduce its fluctuations. The maximum SO2 retention capacity of a kiln system
as described above is reached at the lowest SO2 fluctuation. Because at very low sulfur input
the absorption capacity of CaCO3 or Ca(OH) 2 is not used completely whereas during very
high sulfur input the SO2 generation is over the absorption capacity of the raw meal. In some
cases the SO2 emission stays below limit if the sulfur content in the raw material is not
fluctuating too much.

In the case of the Hver plant in Germany, the desulfurization with Ca(OH) 2 injection into the
preheater is working at its limit because of its very high sulfide input. As explained below,
this method is only working up to an SO2 emission of about 1000 mg/Nm3 dry. If Hver would
not have a very homogenous sulfur content in their raw meal it would not be possible to stay
below the emission limit. Not even with Ca(OH) 2 injection.
The costs of selective quarrying or opening a new quarry should be compared with the costs
of the installation and operation of a system for secondary reduction measures.
A selective elimination of sulfur in the raw material is not yet developed.
Fuel
Reduction of the sulfur input via the fuel can usually only be attained by accepting a
considerable increase in fuel cost. Except for the case where the fuel can be replaced by
alternative fuels.
A reduction of the fuel-sulfur will under normal conditions not reduce SO2-emission from
preheater kilns without bypass.
6.2 Modification of the Existing Process
Non-Preheater System Preheater System

As documented and explained above, preheater systems do have the lowest SO2- emission
of all clinker production systems.
This modification is reducing the SO2 emission significantly if:
the SO2 emission is not mainly caused by non-sulfates input via the raw material
the total input of circulating elements (S, alkali, Cl) is low enough to avoid a kiln bypass
This modification is reducing the SO2-emission because of:
retainment of sulfates and fuel-sulfur by embedding sulfur in the clinker
SO2-adsorbtion during compound operation in the raw mill
lower heat consumption
Modification of fuel preparation and feed [8]

A reducing atmosphere in the kiln is increasing the formation of CO and the volatility of
"SO2".
Therefore, the SO2 emission of preheater kilns with bypass, wet and long dry kilns and to a
lesser degree also of normal preheater kilns is increased if a reducing atmosphere is present
in the kiln:
>1000C
2 CaO + CO 2 + 2 SO 2 (11)
This can be avoided by better fuel preparation and feed as well as with an increased O2
concentration in the kiln.
Coal, Coke:
no oversized particles (residues on 90 m sieve = 0,5 x % volatiles for coal)
continuous feed (feeder, transport)

Oil:
good atomization: burner nozzle e.g. Pillard, Unitherm, temperature, pressure
Alternative fuels:
solid: no oversized lumps: e.g. chipped instead of whole tires, small bundles
liquid: see "oil" above
Besides fuel preparation and feed also the O2 concentration in the kiln should be optimized
to reduce the sulfur volatility. To control the O2 in the kiln, a kiln inlet probe is recommended.
O2 control with gas extraction after the preheater is also possible but because of fluctuation
of the false air intake into the preheater it is less accurate.
In most cases, a compromise in the O2 concentration (reducing atmosphere) has to be
found, because the NOx-emission, fuel and power consumption can be increased by
increasing the O2-concentration in the kiln.
SO2 adsorption in the raw mill
The exhaust gas from preheater and long dry kilns are usually used in the raw mills to dry
the raw meal. The freshly ground raw meal in the raw mill has a large and very active
surface where SO2 can be adsorbed. The highest degree of adsorption (100%) can be found
at very low SO2 concentration whereas at very high concentrations (2500 - 3000 mg/Nm3
dry) like at the Untervaz plant only an adsorption of about 30% is possible. In systems with
an emission of about 1000 to 1500 mg/Nm3 dry an adsorption of 50% is normal.
The degree of adsorption in the raw mill is also influenced by the temperature and the
humidity. A lower temperature or a higher humidity is increasing the adsorption capacity.
Raising the proportion of compound to direct operation, time is reducing the time of high SO2
emission and therefore also the effort of secondary measures (e.g. amount of Ca(OH) 2
added to the kiln feed).
6.3 Secondary Reduction Measures

Catalytic or non-catalytic elimination of SO2 (as it is known for NOx) is not reasonable
because oxidation would lead to SO3 and reduction to SH2, both would be more harmful
than SO2. Therefore, only absorption or adsorption of SO2 can be used as secondary
reduction measures.
Wet scrubber
Wet scrubbers or flue gas desulfurization (FGD) plants remove sulfur oxide from the gas
using scrubbing slurry containing calcium compounds. The SO2 removal efficiency is up to
90-95% and the final product is the so-called synthetic gypsum that can be used in cement.
Power stations or waste incinerator plants use this technology for many years.In cement
industry, the first scrubber was installed at the early 90s. The first wet scrubber within
Holcim group is in operation at ML plant since March 2000.
The basic chemical reactions of SO2 absorption with limestone
SO2 + H2O H2SO3 [1] Absorption
CaCO3 + H2SO3 CaSO3 + CO2 + H2O [2] Neutralisation
CaSO3 +1/2 O2 CaSO4 [3] Oxidation
CaSO4 + 2H2O CaSO4*2H2O [4] Crystallisation

Figure 9: Basic Reaction inside the Vessel
Process
Step 1: Reaction between Calcium - slurry and SO2 in gas stream
The SO2 in the gas stream is scrubbed with slurry of finely ground calcium. This reaction
gas/liquid takes place inside a vessel where nozzles spray the slurry in counter current into
the gas stream. The rain drops of slurry fall to the bottom of the vessel and are recycled.
The fine mist carried in the ascending gas stream is removed by mist eliminators before
being discharging to the atmosphere.
The reagent is usually kiln feed or high-grade limestone ground separately. The reactivity of
the reagent depends, among other, on the percentage of CaCO3 and its finesse: the finer,
the more reactive.
To insure a high SO2 removal efficiency, the ratio liquid/gas (L/G) has to be kept within a
determined range, specific for each scrubber type. If L/G is too low, the contact between
calcium and SO2 is poor (low SO2 removal) but if L/G is high, the energy consumption to
spray more slurry will be too high.

Step 2: Oxidation - crystallisation
Once the reagent has reacted with SO2, the product, calcium sulfite (CaSO3) is oxidised to
the so-called synthetic gypsum. The final oxidation takes place in the upper part of the slurry.
In most cases, blowers are installed to provide enough oxygen for a complete oxidation.
In this upper part of the slurry also the gypsum crystallisation takes places. The fine particles
(< 10 m) act as crystallisation seeds where gypsum crystals can grow.
Step 3: Dewatering
Once the gypsum concentration in the vessel reaches a determined density, a certain
amount of slurry is sent to the dewatering installation. During this last operation, the slurry
moisture is reduced from about 90% to 10% H2O.
Different dewatering systems can be used like hydrocyclone, vacuum belt or centrifugation.
The final product is called synthetic gypsum and can be used with clinker in the same way
as natural gypsum.
Main Process Parameters

There are only 3 main process parameters to operate a scrubber:
pH: SO2 removal depends on pH value. Therefore, pH in the vessel has to be kept within a
certain range, which is controlled by adding fresh limestone into the slurry. Normal pH range
is between 5 and 6. Nevertheless, this pH value depends on the level at which the pH is
measured since, at the top of the slurry, the pH is low (acid) and at the bottom the pH is
higher due to the dissolution of CaCO3. Thus, this pH range is specific for each scrubber.
Slurry density: slurry density control keeps the suspended solids at optimum level to
prevent scaling and to provide crystal growth for maximum dewatering.
Water balance: due to the intense water evaporation, fresh water is injected to keep the
slurry height in the vessel within a defined range. This fresh water is usually sprayed through
the mist eliminator in order to clean it.
Synthetic Gypsum:
The synthetic gypsum quality depends on the reagent composition, especially the
percentage of inert particles (for kiln feed: SiO2, Fe2O3, Al2O3,), and the remaining
moisture after the dewatering process.
Based on the experience gathered up to now, the conversion of CaCO3 in gypsum is very
good (less than 5% unreacted CaCO3 in gypsum) and almost no CaSO3*H2O has been
detected. Therefore, without considering the inert particles, the synthetic gypsum is very
similar to natural gypsum and can be mixed with clinker.
Experience
The first wet scrubber in Holcim is in operation since March 2000 at Midlothian (Holnam-
USA). The reagent used is filter dust from the main bagfilter. This reagent is very fine and its
chemical composition is similar to the raw mix.
On a technical point of view, the system is reliable. Actually, the main issue is the gypsum
water content after the dewatering of about 40%. This high water content seems to be
caused by the filter dust used as reagent. In this case, the reagent is too fine and there are
too many fine particles that can act as seed for the crystal growth. Therefore the gypsum
crystals are very small and water is strongly attracted around these crystals. Hence, the

current dewatering system (hydrocyclones and vacuum belt filter) is unable to separate
enough water from gypsum. Fortunately, the gypsum quality seems to be good enough to be
mix with clinker.
Figure 10: MLs Scrubber: process

Figure 11: MLs Scrubber: general view

Figure 12: MLs scrubber: dewatering system
SO2 absorption with Ca (OH) 2 (slaked lime) [9]

A other common measure against SO2 emission is SO2 absorption with Ca (OH) 2.
Ca (OH ) 2 + SO 2 CaSO 3
It can be added to the exhaust gas:
at the upper cyclone stage of the preheater together with the kiln feed
in a scrubber after the kiln system
The addition of Ca(OH) 2 to the kiln feed is very successfully used in three Holcim plants with
preheater (RK, HV, UV) (see chapter 7). However, this technology has its limits. For instance
only concentrations not exceeding about 1000 mg SO2/Nm3 can be reduced. At the same
time, the emission limit value should be less than 400 mg SO2/Nm3.
If the normal SO2 emission is very high (> 1000 mg/Nm3) as it used to be in the Untervaz
(UV) plant in Switzerland, a separate SO2 scrubber after the preheater has to be considered.
The only slaked lime scrubber for a cement plant was built in the UV plant. There, the kiln
exhaust gas flows through a Venturi reactor with an expanded fluidised bed formed by the
absorbent which consists of a blend of hydrated lime (slaked lime) and raw meal [9]. A more
detailed description of the UV reactor is attached (see chapter 7).
Experience showed that the temperature for SO2 absorption with slaked lime is very
important. Efficient absorption was only observed at temperatures above 350C and below
70C. Therefore, systems with Ca(OH) 2 injection into the conditioning tower or Ca(OH) 2 in
bagfilters do not have a sufficient SO2 absorption efficiency [10].

Figure 14 SO2-Reduction =f(Mol-ratio Ca/S)

Figure 15 SO2 REDUCTION WITH Ca (OH) 2

Replacement of Ca(OH) 2 power by other Ca-compounds
To save on the high cost of the fine slaked lime, other desulfurization agents were tested at
the Hver plant of Nordcement for their suitability. These were coarser slaked lime,
powdered chalk and partly calcined kiln meal. However, the results of these experiments
were mostly unsatisfactory. The effectiveness of the slaked lime currently in use was in no
case anywhere near achieved. Powdered chalk has the lowest effectiveness. The reactivity
of the coarse slaked lime was also barely detectable due to the small surface area. In the
case of the partly calcined raw meal which had been pneumatically withdrawn from the
lowest cyclone stage of a preheater, a slight desulfurization effect could be discerned.
However, the technical problems in handling were so great that its further use was
abandoned.
Tests with unslaked lime (CaO) were not carried out since on account of its lower reactivity
compared with slaked lime, the difference in cost does not justify the increased quantities
required. Optimisation, e.g. also for absorption of larger quantities of SO2, is only likely to be
achieved by selecting a finer slaked lime. Thus a material with a BET specific surface of 18
or 36 m2/g could certainly lead to fixation of larger quantities of SO2.
Untervaz started to use a Ca(OH) 2-mud from a former dump of a carbide manufacturer. The
mud has a negative price and can be added to the raw material. This is reducing the
Ca(OH)2-feed to the scrubber.
At the Fort Collins plant precalcined material (CaO) from the precalciner is added to the
bypass exhaust gas to reduce SO2-emission. In addition to that, bypass dust is recirculated
in the bypass to increase the concentration of SO2 absorbed in the dust. These measures
are sufficient to stay below the SO2-limit.
Activated Coke Absorber
Activated coke is under presence of water a very efficient SO2 absorber. The only filter of
this kind is installed in the Siggenthal (SG) plant of "HCB. The SO2 emission is reduced by
the filter from above 500 to below 50 mg/Nm3.
However, this type of absorber is very pricy. The Polvitec in SG is only economical because
of the contribution from the village of Zurich for burning their sewage sludge.
Sodiumbicarbonate Absorber
Obourg.
Passamaquoddy Absorber
SO2 scrubbing is only a side effect of the Passamaquoddy scrubber. As described in chapter
7 the main purpose is to recover CKD to avoid dust dumping. The SO2 is first absorbed by
the CaO in the CKD and reacts afterwards with the alkali oxides in the CKD to form an alkali
salt. The alkali salt is extracted and if possible sold as a fertiliser.

7. EXAMPLES OF SO2 EMISSION REDUCTION
Below a collection of short descriptions of plants and their measures against SO2 emission
is given. Everybody is welcome to contribute to this collection.
SO2-emission reduction
Plant: Fort Collins (FC)
Capacity: 1500 t/d (design) / 1345 t/d (standard)
Raw material: Containing kerogen and sulfur (~10% as Pyrite)
Fuel type: Nat. gas, coal (0.44 % S), pet coke (4 to 5% S)
Plant description: 1-stage-preheater with precalciner
Supplier/equipment: Polysius preheater, Allis-Chalmers kiln
Initial SO2-emission problem: The SO2-limit of 150 ppm (i.e.~ 320 mg/Nm3) can not
[1], [2] always be maintained, (sometimes up to 300 ppm) and limits
the kiln output. The SO2 emissions originate from the bypass
and not from the precalciner. Kerogen in the raw material is
burned in the calciner, combusted into CO2 and SO2. At the
same time CaO reacts with SO2 if sufficient Oxygen is
available to form CaSO4 (SO2 scrubbing). The bypass is the
control valve for the sulfur level in the clinker. The bypass
rate is 100 %!
Solutions investigated: The reason for the emission from the bypass is believed to
be due to a limited quantity of CaO being available in the
[2]
bypass dust to combine with SO2 from the kiln. So adding
some dust to the bypass-duct would help. Since a lot of sulfur
input into the system is from coal, the specific fuel input has
to be reduced.
The sulfur content of the kiln feed is rather high with 3.6%.
In order to control the SO2-emission, a splitter gate should be
installed, in order to deviate a portion of return dust to the
bypass scrubber.
Solution realized: Dust return (mainly precalciner dust) from the bypass /
precalciner bagfilter into the bypass duct has been installed
[4]
before 1994, however the return dust system needs further
modifications.
Emissions reached: The necessary SO2-limit has been reached. Most of the time,
no additional scrubber dust is used for stack sulfur control.
[3]
Consequently the low stack SO2 (average 40 - 50 ppm)
would allow a lower waste dust quantity.
Further Optimization: The plant capacity should be increased up to 1600 t/d with
several modifications (and reduction of losses), without
[3],[4]
increasing the emissions. A modification of the kiln feed inlet
is planned for 1995.

Literature:
[1] POLYSIUS: Telefax (1989), VA-Dok: FC, Reg. 2D
[2] Kupper, A.K.: HMC Report VA 91/5866/E, VA-Dok: FC, Reg.2A
[3] Bachmann, R.: HMC Report VA 94/4263/E, VA-Dok: FC, Reg.2A
[4] Kupper, A.K. and Brki, Dr. P.: HMC Report VA 95/4288/E, VA-Dok: FC, Reg.2A

Plant: Hver (HV)
Capacity: 3'000 t/d (design), 2'900 t/d (standard)
Raw material: Sulfur components: Pyrite 0.05 - 0.45 %w/w
Fuel type: Coal (0.4%S), fuel oil, nat. gas; alternative fuels: whole tires
(max. 10% of total heat input)
Plant description: Kiln with 4-stage twin line suspension preheater and planetary
cooler
Supplier/equipment: KHD kiln and preheater
Initial SO2-emission problem: SO2 emissions of up to 1300 mg/Nm3
[1],[2] (TA Luft = 400 mg/Nm3) In direct operation, up to 900 mg/Nm3
in compound operation
Solutions investigated: The following investigations were proposed:
[3],[4],[5] Reduction of SO2 emissions by lowering the exhaust gas
temperature.
Lowering the temperature in the raw mill in compound
operation.
Desulfurization with Ca(OH)2: different trials have been carried
out. To reach an emission < 400 mg/Nm3, the SO2
concentration in the stack should be held at about 350 mg/Nm3.
The calculated amount (based on trial results) of Ca(OH)2
needed to stay below the limit would have been 2.4 kg/t clinker
in 1990 and 6.4 kg/t clinker in 1991.
Solution realized: Desulfurization with Ca(OH)2 addition to the kiln feed (bucket
elevator). The Ca(OH)2 is proportioned by speed control rotary
valves with three independent pneumatic injection pipes, which
can be used simultaneously when required. The Ca(OH)2
addition was started up in November 1991.
Emissions reached: The addition of Ca(OH)2 is controlled by the measured SO2-
[1] concentration at the stack, in order to stay below the TA-Luft
limit of 400 mg/Nm3.
Further Optimization: Since the costs for slaked lime are quite high, different other
[1] desulfurization agents have been tested:
powdered chalk: low absorption efficiency
coarse slaked barely detectable effect due to
lime: small surface area
partly calcined slight desulfurization effect, but
raw meal: additional technical problems.
Tests with unslaked lime (CaO) have not been carried
out, due to its low reactivity compared with Ca(OH)2. A
further optimisation could be reached, using a finer
slaked lime (higher surface area).

Literature:
[1] Boes, K.-H.: Measures to reduce the SO2 emission during clinker burning at
Nordcement AG's Hver works.
Zement, Kalk, Gips (1993), P.514-518
[2] Waltisberg, J.: HMB Aktennotiz VA 86/66/D,
VA-Dok: HV, Reg. 17
[3] Hasler, R.: HMB letter to Nordcement AG,
VA-Dok: HV, Reg. 17
[4] Waltisberg, J.: HMB Bericht VA 87/5437/D,
VA-Dok: HV, Reg.18
[5] Waltisberg, J.: HMB Bericht VA 91/5904/D,
VA-Dok: HV, Reg. 1
Figure A Slaked lime chemical and physical data [1]

"Bltenweiss slaked white lime from Fels-
Werke GmbH
Chemical and physical data
CaO 71.99%
MgO 0.65%
SiO2 2.00%
Fe2O3 0.45%
Al2O3 0.47%
SO3 0.27%
CO2 1.23%
Combined H2O 21.60%
Moisture 0.90%
BET approx. 10 m2/g
Ca(OH) 2 90.0%
Fineness:
R > 0.063 mm 4.0%

Figure B SO2 reduction efficiency with slaked lime addition

Keine KH-Zugabe = no slaked lime feed
KH-Zugabe = slaked lime feed
Zeit = time
SO2-Reduktion = SO2 reduction
Kalkhydratzugabe = slaked lime feed rate

Figure C Slaked lime silo and dosing

Plant: Rekingen (RK)

Capacity: 2'200 mtpd (design), 2130 mtpd (standard)
Raw material: Sulfur component: 0.55 % Pyrite
Fuel type: Coal 0.4%S, fuel oil 0.8%S, waste timber
Plant description: 4-stage-suspension preheater kiln
Supplier/equipment: Polysius kiln and preheater
Initial SO2-emission problem: SO2 emissions of up to 1200 mg/Nm3 (LRV-limit =
500 mg/Nm3) in direct operation [LRV = Swiss
clean air act]
Solutions investigated: Reduction of SO2 emissions by lowering the
exhaust gas
[1],[2] temperature. Reduction from 180C to 150C
results in a reduction of 50 mg/Nm3, the same
reduction is obtained by further lowering the
temperature to 140C.
Desulphurization with Ca(OH)2, added to the kiln
feed only during direct operation leads to a
reduction of 400 - 450 mg/Nm3 at an initial
emission of 750 - 1200 mg/Nm3
Solution realized: [3] Desulphurization with Ca(OH)2 addition to the kiln
feed
3 kg/tcli (max. 17kg/tcli) during direct operation
(20% of the operating time)
Costs are about 60'000 CHF/year
Emissions reached: The addition of Ca(OH) 2 is controlled by
measuring the SO2 concentration in the stack, in
order to reach the limiting values of LRV, i.e.
below 500 mg/m3.
Literature:
[1] Utzinger, K.: Reduktion der SO2-Emissionen durch Senkung der Abgastemperatur,
Versuchsbericht (1986), VA Dok: RK, Reg.17
[2] Scheuch, J.: Entschwefelungsversuche mit Ca(OH)2 im Ofendirektbetrieb,
Versuchsbericht (1986), VA Dok: RK, Reg.17
[3] "Holderbank" Cement Course: Source and reduction of emissions gaseous, VA
91/5882/E, P.7

Figure A Dosing of slaked lime to the kiln feed

SO2 -emission reduction
Plant: Siggenthal (SG)
Capacity: 2000 t/d (design), 1900 t/d (standard)
Raw material: [2] Sulfur: 0.7% S
Fuel type: [2] Fuel oil (2.3%S), dried sewage sludge, complete tires
(1.3%S)
Plant description: Kiln with 4-stage preheater
Supplier/equipment: Polysius kiln and preheater (Dopol), Fuller grate
cooler
Initial SO2 emission problem: SO2 emissions of up to 1500 mg/Nm3 in direct
operation,
[1] up to 900 mg/Nm3 in compound operation (LRV-limit =
500 mg/Nm3) [LRV = Swiss clean air act]
Solutions investigated: SG had a project to burn sewage sludge containing
Hg and at the same time the SO2 and NOx emissions
where too high. A pilot plant with an active coke filter
as well as an ammonia injection into the riser duct was
tested successfully. Mercury, SO2 and other pollutants
were absorbed on the active coke. NOx was reduced
to N2 and H2O with ammonia injection.
Solution realized: A POLVITEC (Polysius Environmental Technology)
active coke scrubber behind the kiln EP was
successfully commissioned in 1994. Beside SO2 and
Mercury it is also filtering out other pollutants like
heavy metals, NH3, organic compounds and dust. The
loaded coke is injected into the kiln as a fuel
substitution.
Investment cost: 30'000'000 CHF
Operating cost: 3.3 CHF/t cli
Emissions reached: [3] < 12 mg/Nm3 dry
Literature:
[1] Dr. R. Stenger, : HMB Bericht MA 92/93/D, VA Dok: SG, Reg. 2

[2] ATR, Annual Technical Report 1994,"Holderbank"
[3] Effektive Emissionsminderung bei Einsatz von Sekundrbrennstoffen im
Zementdrehofen von "HCB Siggenthal, Dr. 1D. Rose, Dr. L. Brentrup, Krupp
Polysius

Plant: Untervaz (UV)
Capacity: 1900 t/d (standard)
Raw material: Sulfur components: Pyrite and Markasit, usually 10 - 16
g SO3/kg clinker
Fuel type: [7] Coal (0.07%S), fuel oil; alternative fuels as dried
sewage sludge, waste oil (0.6%S), destillation residue
(0.35%S), plastic (0.02%)
Plant description: Kiln with 4-stage-suspension preheater and planetary
cooler (kiln III)
Supplier/equipment: KHD kiln, preheater and cooler; Lurgi SO2 -scrubber
Initial SO2-emission problem: SO2 emissions of up to 3600 mg/m3 (LRV-limit = 500
mg/Nm3) in direct operation [LRV = Swiss clean air act]
Solutions investigated: In 1984/85 trials to lower the SO2 emissions by
changing
[1],[2],[3],[4] the operating parameters have been carried out.
Addition of Ca(OH)2 to the raw meal: Reduction of up to
50% of SO2, still insufficient.
Because no possibility of lowering the SO2 emission of
the Plant to the required 500 mg/Nm3 by simple means
as the one explained above could be found, the use of a
scrubber had to be considered. Because a wet SO2
absorber (washer) was too expensive (investment and
operating costs) a dry absorber was chosen.
Solution realized: A system consisting of a circulating fluidized bed
supplied
[1],[2] by Lurgi, Germany, was started up in 1988. It works as
follows:
The kiln exhaust gas flow through a venturi reactor with
an expanded fluidized bed formed by the absorbent that
consists of a blend of hydrated lime (93% Ca(OH) 2) and
raw meal (0.8-1.0t/h Ca(OH) 2 and 2.5t/h raw meal). The
latter is mainly added to prevent caking of the very fine
hydrated lime. Water is injected to operate the reactor
as close as possible to the water dew point and to
promote the reaction with the SO2. Due to the intensive
contact of the exhaust gas with the absorbent in the
fluidized bed at a temperature (approx. 65C) close to
the water due point (approx. 58 to 61C), the SO2
becomes very effectively combined with the absorbent,
whereby predominately calcium sulfite (CaSO3) is
generated. The absorbent then passes with the exhaust
gas into an electrostatic precipitator. The bulk of the
precipitated absorbent is returned to the venturi reactor
(approx. 175t/h), while the remainder is discharged and
either returned to the kiln feed or added to the cement.
The stoichiometric factor of Ca(OH) 2 to SO2 is about 2.3
to 2.6.
The pressure loss in the system is 2200 Pa. This results

in a power demand for the fan of 315 kW. With the
addition of 155 kW for the electrostatic precipitator and
the transport systems, the total power consumption of
the system amounts to 470 kW.
Adding all the material to the kiln feed results in an
increase of the SO3 content in the clinker from 1.0 to
1.3%. This is not desirable, but so far not of much
concern, because it has turned out that the additional
sulfates in the clinker act as a gypsum substitute, so
that the SO3 content of the cement can be kept at 2.7%
SO3 as before.
The investment for the whole desulfurization plant,
including auxiliary installations, amount to 15 million
CHF. The operating costs are about 2.2 CHF/t cli
(Slaked lime 1.6, electrical power 0.4, maintenance 0.5,
gypsum substitution 0.3 CHF/t cli).
Emissions [5] During the acceptance measurements in May 1989, the
reached: following values have been reached:
Direct operation: 431 mg/m3
Compound operation: 418 mg/m3
Further [6] In 1990 the circulating fluidized bed was operated with 1
Optimization: t/h hydr. lime and 2.5 t/h raw meal. It has been
investigated to use hot meal instead of hydr. lime. Due
to analyses of hot meal, it seemed not to be possible to
replace hydr. lime completely, but it should be tested
whether a part of the hydr. lime could be replaced
Literature:
[1] Bonn, W., Hasler, R.: Verfahren und Erfahrung einer roh- stoffbedingten SO2-
Emission im Werk Untervaz der Bndner Cementwerke. Zement, Kalk, Gips (1990),
P. 139-143
[2] "Holderbank" Cement Course: Source and reduction of emissions gaseous, VA
91/5882/E, P. 5 - 6
[3] Hasler, R., Wickert, Dr.H.: HMB Bericht VA 86/5281/D, VA-Dok: UV, Reg. 17A
[4] Berclaz, Ch.:HMB Bericht VA 85/73/D, VA-Dok: UV, Reg. 17A
[5] Strahm, E., Waltisberg, J.: HMB Bericht VA 89/5665/D, VA-Dok: UV, Reg. 2
[6] Waltisberg, J.: HMB Aktennotiz VA 90/6/D, VA-Dok: UV, Reg. 2
[7] ATR, Annual Technical Report 1994,"Holderbank"

Figure A Burning of dried sewage sludge in a cement kiln
Figure B Circulating Fluidized Bed Absorber

Plant: Aalborg Portland (not in Holcim Group)
Capacity: 0.5 Mio. t/year of white clinker with 4 kilns
Raw material:
Fuel type: Fuel oil and pet coke
Plant description: Wet process plant
Supplier/equipment: FLS
Initial SO2 emission problem: Due to the white color of white clinker, and since in
the
[1] white kilns very little sulfur is retained, only special
fuel quality could be used. Most of the sulfur was
emitted as SO2 in concentrations of 500 - 800
ppm.
Solutions investigated: Since the fuel consumption is generally higher for
white clinker and for a wet process, in 1980 a
conversion to a semi-wet process was considered.
Due to high capital costs and technical risks this
investigation was not followed further.
The recovery of some of the waste heat seemed to
be a more attractive solution. In combination with
the SO2-problem, a combined heat recovery and
desulfurization process seemed to be very
attractive, and it was decided to build a pilot plant.
From March 1989 to February 1990 a test program
was carried out. The degree of desulfurization was
found to be 75 %.
Solution realized: The heat recovery and SO2 absorption unit for the
four wet kilns consisting of 4 gas/water heat
exchangers and 2 scrubber systems was
commissioned in 1991.
The exhaust gas from the kiln passes first the
gas/water heat exchanger before it enters the SO2
scrubber at a temperature of about 115C. In the
scrubber SO2 is absorbed in a slurry loaded with 6
10% solids consisting of 98% CaSO4.2H2O and
2% CaCO3. The slurry is sprayed in counter
current to the exhaust gas and collected in the
recycle tank at the bottom of the scrubber where it
is oxidized with air (CaSO3 + 0.5O2 CaSO4). A
part of the slurry is pumped to a centrifuge where
water and gypsum are separated. The rest is
reinjected through a circulation line into the
scrubber. A chalk slurry of 30% moisture is
injected into the circulation line before the spray
nozzles to replace the used and extracted
absorbens (CaCO3+SO2 CaSO3 + CO2). The
optimum pH for the circulated slurry has been
found to be between 5.4 and 5.6 and is controlled
by the amount of chalk slurry added to it.
The circulated slurry passes on its way from the
recycle tank to the nozzles a heat exchanger to

preheat the water for the gas/water heat
exchanger.
The exhaust gas leaves the scrubber with 75%
less SO2 and a temperature of about 70C.
Emissions reached: The scrubber reaches an SO2 reduction of 75-
80%, i.e. the remaining SO2-concentration is below
160 200 ppm.
Further Optimization: Further optimization concerning the desulfurization
is not foreseen.
Literature:
[1] H.E.Borgholm: A new heat recovery and desulfurization plant for 4 wet kilns in
Aalborg Portland. 35th IEEE cement industry technical conference in Toronto, IEEE
catalogue (1993), 395 - 409.
Figure A Heat recovery and desulfurization plant of Aalborg Portland / Denmark

Plant: Dragon Products, Thomaston (Maine, USA), (not in

Holcim group)
Capacity: [1] 4000'000 t/year
Raw material:
Fuel type:
Plant description: Wet kiln
Supplier/equipment:
Initial SO2 emission problem: SO2 emissions never were a problem (500 mg/Nm3).
[1] During the time when Martin Marietta owned the
plant (i.e., before 1980), the US EPA designed
Thomaston, an upper class coastal residential area,
a non attainment area (i.e. an area which had not
attained the required level of emissions). This was
due to dust emissions from both kiln and quarry
(CKD). The Dragon plant discarded dust to produce
a lower, but not a low alkali cement.
Solutions investigated:
Solution realized: [1],[2] The Passamaquoddy scrubber was commissioned
in 1990. Conversion is accomplished in the
Recovery Scrubber by reintroducing into CKD the
CO2 that was released during calcining. Exhaust
gas from the kiln is cooled in a heat exchanger (A)
and bubbled through a reaction tank (B) containing a
slurry of CKD and water. CO2 in the gas reacts with
CKD in the slurry to reclaim CKD for kiln feed, while
SO2 reacts with potassium sulfate in the CKD to
form potassium sulfate in solution.
CaO + SO2 0.5O2 CaSO4

CaSO4 + 2KOH + CO2 CaCO3 + K2SO4 +H2O
Reclaimed CKD is separated in a settling tank (C),
rinsed in a secondary settling tank (D) to remove
potassium sulfate remnants, and returned to the
cement plant as raw feed. Potassium solution is
pumped to the crystallizer (E) for recovery as dry
potassium sulfate. Heat for the crystallizer is
obtained from the exhaust gas heat exchanger (F).
Emissions reached: [3] The exhaust gas SO2 elimination is claimed to be
over 90%. Beside SO2 also HCl as well as ammonia
and some less volatile organic compounds are
absorbed.

Literature:
[1] Dust and Other Secondary Materials Management Using the Passamaquoddy
Recovery Scrubber
[2] The Recovery Scrubber , Passamaquoddy Technology
[3] Clean emissions valuable by-products, International Cement Review March
91
Figure A Recovery scrubber from Passamaquoddy

8. LITERATURE
[1] Ullmann's Encyclopedia of Industrial Chemistry, Vol. A 25, 1994 VCH,

3-527-20125-4
[2] Handbook of Chemistry and Physics, 60th Edition, CRC-Press,
0-8493-0460-8
[3] Surehaltige Niederschlge - Entstehung und Wirkung auf terrestrische kosysteme,
1984, VDI, Kommission Reinhaltung der Luft
[4] F.L. Smidth, An overview of the formation of SOx and NOx in various pyroprocessing
systems; P.B. Nielsen, O.L. Jepsen, No. 96
[5] Der Anteil der Zementindustrie an den anthropogenen SO2- und NOx-Emissionen,
H. Wickert, "Holderbank" News 6, 1985, p. 15-16
[6] ATR-Databank 1994
[7] Kiln Optimization Seminar, Circulation Phenomena VA 90/5714/E, U. Gasser
[8] Kiln Optimization Seminar, Fuel Preparation / Firing Systems,
VA 89/5653/E, F. Schneider
[9] Environmental Protection Seminar, Sources Reduction of Emisisons (gaseous),
VA 91/5882/E, J. Waltisberg
[10] Parameter study on desulfurization in baghouse filter,
VA 90/5687/E, A. Edlinger, R. Hasler
[11] Massnahmen zur Minderung der SO2-Emission beim Klinkerbrennen im Werk Hver
der Nordcement AG, K.H. Boes
[12] Emissions of NOx and SO2 from cement clinker burning, V. Johansen, A.H. Egelov,
Denmark

Figure 16 Influence of oxygen content in the kiln gas

Chapter 2
Dedusting
1. General ..................................................................................................................... 92
2. Bag Filter................................................................................................................ 100
3. Electrostatic Precipitator...................................................................................... 121
4. Dedusting Concepts for Cement Kilns................................................................ 160

General
Arnaud De Luca
TPT 01/21071/E (Replaces report no. PT 96/14047/E)
1. Introduction.....................................................................................................................93
2. Standard Technology for Dust Emission Reduction...................................................93
2.1 Comparison between bag filter and electrostatic precipitator ............................93
3. Present State of Cement Kiln Emission .......................................................................95
4. Present Legal Situation..................................................................................................96
5. Dust Charateristics.........................................................................................................98

Introduction
Dust filters were the first secondary emission reduction measure in the cement industry. The
motivations for dedusting of exhaust gas and vent air are:
Compliance with environmental regulations
Reduction of product loss
Protection of employees and equipment from harmful dust impacts (irritation plugging,
erosion)
This paper is focused on the dedusting of kiln/raw mill exhaust gas and clinker cooler vent
air. They are the largest dust filters of the entire clinker production line. Because they have
to provide both very high reliability and efficiency under extremely severe conditions those
filters have to fulfil the most difficult task among all cement plant dedusting equipment. The
basic working principles of the presented filters are also valid for other applications.
1. STANDARD TECHNOLOGY FOR DUST EMISSION REDUCTION

Two types of filters have been widely used in the cement industry: bag filters (BF) and
electrostatic precipitators (EP). BF use the filtration technology and EP use electricity to
collect the particles.
Some of the older dedusting technologies like gravel beds or multi-cyclones are still in
operation but are not build any more, mainly because they have difficulties to comply with
todays more stringent dust emissions regulations.
1.1 Comparison between bag filter and electrostatic precipitator

Electrostatic Precipitator
EP used to be a very popular filter because of the low pressure drop, the little maintenance
required and the high efficiency under normal operating conditions. Today, the main problem
of the EP is its efficiency under not normal operating conditions. Transitions phases
(direct/compound operation), or kiln start-up/shut down always causes increased dust
emission. Moreover, during high CO peaks the EP must be shut down to prevent CO
explosion.
Some decades ago, increased dust emission during such incidences where generally
accepted by the neighbours and the authorities. Nowadays, with the stronger environmental
awareness of the people and with more stringent emission limits (20 mg / Nm3, continuous
dust emission control) it can be difficult to achieve the required long term efficiency with an
EP.
Nevertheless, an advantage of the EP can be the classification of the dust particles. The
coarse particles are collected in the first electrical field, the fine ones in the last.
Condensables, like alkalis or chlorine are enriched in the fine fraction of the dust. With an EP
less dust has to be extracted from the filter to reduce the concentration of the condensables
in the system.

Bag Filter
Bag filters used to be very popular mainly in USA, especially the reverse air filters. For about
10 years, this technology has spread all around the world and nowadays, for most of new
plants or for new filters, BF are chosen.
The main BFs advantage is their very high efficiency that is unaffected by the process
conditions. Their main disadvantage is the high pressure loss, the additional maintenance
cost for the regular replacement of the bags.
More over they produce a not neglectible amount of waste (used bags). Depending on the
local regulations for waste elimination and the quality of bags their elimination can be
expensive.
Figure 13: Dust emission comparison between EP and BF

Figure 14: Filter types for 9 new kilns
Plant Group Country Year of Kiln/raw mill Cooler filter

Company Commissioning Filter
Saraburi Siam City Thailand 1995 EP EP
Lgerdorf Alsen Germany 1996 EP --
Davao UCC Philippines 1996 EP EP
Chekka SCL Lebanon 1997 Bag filter Bag filter
Devils slides Holnam USA 1997 Bag filter Bag filter
Gladstone QCL Australia 1997 Bag filter Bag filter
Hon Chong Morning Star Vietnam 1998 Bag filter Bag filter
Luga-it Alsons Philippines 1998 EP EP
Midlothian Holnam USA 2000 Bag filter Bag filter
Figure 15: New filters installed (not exhaustive)
Plant Group Country Year Filter Type

Compagny
Mississauga SLC Canada 1996 Bag filter Gravel to Pulse jet
Rockhampton QCL Australia 1997 Bag filter EP to Pulse jet
(lime)
Hver Alsen Germany 1998 Bag filter EP to Pulse jet
Cantagalo Holdercim Brasil 1998 Bag filter EP to Pulse jet
Eclpens HCB Switzerland 1998 Bag filter EP to Pulse jet
Lumbres COO French 2000 Bag filter EP to Pulse jet
Carboneras* Hisalba Spain 2001 Bag filter EP to Pulse jet
Gador* Hisalba Spain 2001 Hybrid filter EP to Hybrid filter
*: Not yet commissioned
2. PRESENT STATE OF CEMENT KILN EMISSION

According to the very much varying dust emission limits in countries where cement plants
are operated and the development in the dedusting technology, the average dust emission
from cement kilns varies between less than 10 up to 500 mg/Nm3 dry. Only a few exceptions
are above 1000 mg/Nm3.

3. PRESENT LEGAL SITUATION
Dust was one of the first stack emissions that were regulated and is today still the only
emission limit for some plants. The reduction efficiency required is much higher compared to
other emissions like SO2 or NOx. The dedusting efficiency of modern dust filters is about
99.999% compared to only 95% of very good SO2 wet scrubbers.
Dedusting efficiency
r
=1 1)
R
where: = dedusting efficiency
r = clean gas dust content
R = raw gas dust content
In Europe, emission limits are expressed as mass of particles per gas volume [mg/Nm3] and
usually the gas volume is calculated on dry base. In some countries the gas volume is
referred to a certain oxygen concentration, mostly 10% O2.
Emission Limits
Nm3dry= m3 at 273 K, 101300 Pa and 0% water

Europe: 20 - 50 - 500 mg/Nm3
Germany: 50 mg/Nm3 dry at 10% O2
In the United States of America not all plants do have a dust emission standard. But all of
them have an opacity limit, which is to some degree correlated to the dust emission.
For comparison reasons all emission limits/standards are indicated in mg/Nm3. The emission
standards in the USA are usually not using mg/Nm3 but ppm, lb/tdry feedgr/dscf, lb/1000 lbgas,
b/hr, etc. To convert them into mg/Nm3 certain assumptions were necessary.
USA (e.g.): 0.3 lb/tfeed, dry ca. 80 mg/Nm32%O2, wet

0.05 - 0.08 gr/dscf ca. 115 - 180 mg/Nm3dry
0.015 gr/acf ca. 60 mg/Nm32% O2, wet
5.5 lb/hr ca. 15 mg/Nm32% O2, wet

Some US plants also have to comply with PM10 limits. PM10 stands for particles smaller
than 10 m. Particles smaller 10 m are small enough to enter and mechanically damage
the lungs. To express the limits the same units as discussed above are used.
USA PM10 (e.g.): 0.016 gr/dscf ca. 37 mg/Nm32% O2, wet

4.7 lb/h ca. 13 mg/Nm32% O2, wet
0.015 gr/acf ca. 60 mg/Nm32% O2, wet t
US plants burning hazardous waste are regulated under BIF (Burners and Industrial
Furnaces). Other plants do have a state permit defining certain parameters like NOx, SO2
CO, particles (dust) and THC emission. The limits for these emissions are called emission
standards. These standards are individually defined for each plant and usually represent the
operating situation under certain conditions. Therefore, the US standards are different from
the emission limits in Europe where emission limits are valid for a whole state or country.
All the above explained emission limits do include definitions how and when the compliance
tests have to be carried out. It is, e.g., very important whether the emission has to be
measured continuously or not. Some dust filters like electrostatic precipitators (EP) are very
sensitive on process changes and can have an increased dust emission during transition
periods and may not be in compliance during that time.

4. DUST CHARATERISTICS
The characteristics of dust have a strong influence on the behavior and design of dust filters
and on the impact of the dust on its environment. Dust is characterized by
Size
Shape
Hardness
Chemical composition
Mineral structure
Electrical resistivity
Specific weight
Angle of repose
etc.
Figure 16: Comparison of sizes and physical characteristics of various dusts
The character of the dust is defined by its origin and the different treatments like grinding,
blending, classifying or burning. Dust from a preheater kiln is much finer than dust from a
clinker cooler and because of this more difficult to separate.
Figure 17: Particle size distribution of some dusts from cement kilns

Particlesize[m] 1000- 100 100- 10 10- 1 1- 0.1 0.1- 0.01
Examplesof Grid Dust Fume
different types
of particles Rain Mist Fog TobaccoSmoke
Drops
Foundery FlyAsh CarbonBlack
Sand
PulverisedCoal Aerosols
Pollens Bacteria Virus
TemporaryAtmosphericImpurities Permanent AtmosphericImpurities
Dust fromCement ProductionProcess
Particulatescollected SettlingChamber
bydifferent dust
collectors Cyclones
HighEfficiencyCyclones
Scrubbers
FabricFiltersandElectrostaticPrecipitators
Approximat Visability
Limit [mg/m3] 150 100 80 50 30 20 10
FreeFallingVelocity
for Spheres[m/s] 3 0.3 0.03 0.003 0.0003 0.00003 0.000003 Particulatesdonot settle
inStill Air at 25C, 1bar due tobrownian
(StokesCunningham) movement

Bag Filter
Arnaud De Luca
1. Categorizing Bag Filters ..............................................................................................101

1.1 Cleaning Method ..............................................................................................101
1.2 Filter-Media Type / Temperature Capabilities ..................................................101
2. Filter Media....................................................................................................................102
2.1 Selection Criteria..............................................................................................102
2.2 Properties of Various Media.............................................................................102
2.3 Hydrolytic Influences ........................................................................................106
2.4 Woven Fabrics and Needle Felts .....................................................................106
3. Cleaning Systems.........................................................................................................109
3.1 Overall View .....................................................................................................109
3.2 Pulse Jet Cleaning ...........................................................................................110
3.3 Reverse Gas Cleaning .....................................................................................115
3.4 Comparison between Pulse Jet and Reverse Gas Cleaning ...........................117

Bag Filter (BF)
The history of bag filters begins in 1886, when engineer Wilhelm Friedrich Ludwig Beth, of
Lbeck, was granted the Patent 38396 for a "suction tube filter with automatic cleaning
device". Since then, the bag filter technology was continuously improved and is today at a
level that makes it superior to any other dedusting system.
1. CATEGORIZING BAG FILTERS

To better understand one's bag filter, it is helpful to determine where it fits among the various
types of BF. When one tries to group bag filters into a number of categories, it soon
becomes obvious that the task is not simple. There appears to be an exception to each of
the rules. The creation of certain categories, though they are not rigid, is yet very helpful.
1.1 Cleaning Method

One such approach is to group BF designs by cleaning method. There are three major
cleaning methods: shakers, reverse gas and pulse jets. In addition to these three dominant
cleaning methods, there exist a large number of other cleaning methods which are less often
applied. Combinations of the three primary methods have been occasionally employed. For
example, reverse air and shake have been used in combination and reverse-air with a pulse
assist. Today, in cement plants mainly pure reverse gas or pulse jet BF are applied.
1.2 Filter-Media Type / Temperature Capabilities

The filter-media type and temperature capabilities provide two other ways to categorize and
view bag filters.
Media types categories: Woven vs. felted media
Operation Temperature:
High-temperature (> 200C)
Medium-temperature (150 to 200C)
Low-temperature applications (< 150C)
One should be aware of these distinctions and attempt to find where the collector in question
fits; thus, when considering operating or troubleshooting recommendations, one only applies
recommendations that are suited to the type of collector being used.

2. FILTER MEDIA
2.1 Selection Criteria

The filter medium is the all-important central feature of any dust collector operating on the
filtration principle. With the correct or incorrect choice of the filter material the whole
dedusting operation, however well received, will stand or fall in actual practice. Important
criteria are:
filter type, particularly cleaning principle
gas temperature (average and peaks)
composition and chemical properties of the gas
raw gas dust load
required dust load of the clean gas
physical and chemical properties of the dust
Furthermore, the filter medium must satisfy the following conditions:

high air permeability (low pressure losses)
good mechanical strength
good thermal stability at operational temperature
good dimensional stability at operational temperature
2.2 Properties of Various Media

Wool and cotton, the only two raw materials that were available to our grandfathers for
making filter media have been nowadays, to a great extent, superseded by synthetic fibres.
Even mineral and metal fibres are used for special applications. Table 1 summarizes the
main properties of various fibres. The most extensively used ones in the cement industry are
polyester, polyacrylnitrile, modified polyamide (aramide) and glass fibre.

Table 1: Properties of Various Fibres for Woven Fabrics and Needle Felts

Special treatment of the fabrics and needle felts can significantly improve the properties of
the bags.

Table 2: Special Treatment of the Bag Surface
Non-Fiberglass Finish Purpose Available For
Singe Recommended for improved cake Polyester, Polypropylene, Acrylic,

release Nomex, Ryton, P 84 (felts)
Glaze Provides short-term improvements Polyester, Polypropylene (felts)
for cake release (may impede
airflow)
Silicone Aids initial cake development and Polyester (felt and woven)
provides limited water repellany
Flame Retards combustibility (not flame- Polyester, Polypropylene (felt and
Retardant proof) woven)
Acrylic Coatings Improves filtration, efficiency and Polyester and Acrylic felts
(Latex base) cake release (may impede flow in
certain applications)
PTFE Surface For capture of fine particulate, Nomex, Polyester, Acrylic,
Treatments and improved filtration efficiency, cake Polypropylene (felt) (Laminates
Laminates release available in Polypropylene, Ryton and
Polyester only)
PTFE Improved water and oil repellency; Nomex (felt)
Penetrating limited cake release
Finishes
Acid Resistant Improved acid resistance and Nomex (felt)
water retardance
Fiberglass Finish Purpose Applications
Silicone, Protects glass yarns from For non-acidic conditions, primarily for
Graphite Teflon abrasion, adds lubricity cement and metal foundry applications
Acid Resistant Shields glass yarn from acid attach Coal-fired boilers, carbon black,
incinerators, cement, industrial and
small municipal boiler applications
Teflon B Provides enhanced abrasion Industrial and utility base load boilers
resistance and limited chemical under mild pH conditions
resistance
Blue Max- Provides improved acid resistance Coal-fired boilers (high and low sulfur)
CRF/70 and release properties, superior for peak load utilities, fluidized bed
abrasion resistance, resistant to boilers, carbon black, incinerators
alkaline attack, improved fiber
encapsulation

2.3 Hydrolytic Influences
One of the greatest enemies of the textile filter media is hydrolysis. By this is understood the
breakdown of the molecular chain of the polymerizate by the action of moisture. Hydrolysis is
intensified by the action of heat, acids and alkalis.
Polyester, for example, should not be used under conditions in which moisture and elevated
temperatures occur in combination. Also aromatic polyamides (Nomex) are subject to
hydrolytic influences at temperatures above 100C, especially if acids or alkaline agents
additionally act as a catalyst.
Fortunately, in recent years chemical modification processes have emerged which have
enabled these polyamides and also polyester to be substantially improved in this respect.
2.4 Woven Fabrics and Needle Felts

Filter media are available either in the form of woven fabric (cloth) or felt fabrics (more
particularly needle felt). The structure of these two types of filter media is shown as follows:
Figure 1 Woven fabric and needle felt with supporting structure
Characteristic of woven fabric is its system of warp and weft threads crossing one another.
Essentially the fabric pores, i.e. the holes between the warf and the weft threads, are
decisive with regard to the filter properties. Effective fabrics for dust collection purposes have
a free perforated area of about 40%. Therefore, they must have a permanent dust crust to
maintain their good filtering effect.
In contrast with woven fabrics, needle felts are "three-dimensional" filter media. Their active
filtering surface is located both on the surface and in the interior of the medium. The dust
collection process, beside the sieving effect as in woven fabrics, additionally takes place
through inertia and barrier effects. For reinforcement, needle felt can be provided internally
with a supporting woven fabric interlayer which is only of secondary importance as regards
its dust- collecting effect but which serves primarily to give tear resistance and dimensional
stability to the material. The pore volume of needle felts is 60 - 90%. This porous structure
allows higher admission velocities with lower pressure drops and higher dust collection
efficiencies.
The filtration process for both types of filter media is shown in Fig. 2:

Figure 2 Dust Separation with Needle Felt and Woven Fabric
For the cleaning of the woven fabric bags reverse gas is usually used. The needle felt bag
filters generally have a jet pulse system for cleaning.
The high cleaning air pressure of the jet pulse allows the use of denser filter media which in
turn achieve higher dust collection efficiencies.
Special finishing or application of membranes on the bag surface become more and more
important, especially for the jet pulse filters (see table 2). The purpose of those treatments is
to give the bags improved resistivity against chemical and mechanical attack as well as
optimum filtration efficiency and cake release (especially for fine dust particulates). Bags
with such a treatment may have a very much improved filtering efficiency and therefore do
not need any more cake formation on their surface to achieve a good dedusting efficiency.
Such bags can be operated with a much thinner dust crust and have therefore a reduced
flow resistivity.

Figure 3 Needle felts without and with membrane

3. CLEANING SYSTEMS
3.1 Overall View

Ten or twenty years ago, fabric filters used to be cleaned by rapping or shaking, sometimes
in combination with low-pressure air purging but in the early 1960's a new agent for filter
media cleaning was introduced: compressed air.
Fig. 4 gives a summary of the various cleaning principles used for bag filters.
Figure. 4 Cleaning Principles of Bag Filters
a, b: manual or mechanical, by rapping or shaking

c: mechanical, by vibrating
d: pneumatic, by reverse air flow (often combined with shaking or vibrating)
e: pneumatic, by compressed air (pulse jet)
Since most of the mechanically cleaned filters have been superseded by compressed air
ones, mainly in the cement industry, the mechanical cleaning devices will not be described
further here.

3.2 Pulse Jet Cleaning
Pulse jet cleaned bag filters normally employ felted fabrics of various types. The raw gas
enters the bags from the outside. The cleaned gas flows through the centre of the bag to the
clean gas plenum and from there to the stack. The dust is deposited on the outer surface of
the bag. To prevent bag collapse during filtering, metal cages are inserted inside each bag.
A periodic bag cleaning is required to remove excess residual dust cake. This is
accomplished by pulsing compressed air down into each filter bag. Bag cleaning can be
accomplished either with the compartment isolated or not isolated (off-line or on-line
cleaning, see Fig. 5).
The cleaning made is selected based on the particulates being filtered, process conditions
and bag quality.
Figure 5 Schematic of compartments in a pulse-jet cleaned baghouse filtering

flue gas, purging (or ventilating) prior to maintenance, cleaning bags
and filtering flue gas again
Most "jet" filters use injectors for the periodic purging of the individual filter elements with a
nozzle, usually disposed centrally.

Figure 6 Injector with central nozzle
Each row of bags or each individual bag is equipped with an injector which operates as
follows:
When the bag is in service, i.e. engaged in dust collecting, the clean gas flows from the
interior of the bag through the injector into the clean gas plenum.
When compressed air is released as a jet from the injector at a velocity which may be above
or below the velocity of sound (depending on its type and design), secondary air is entrained
from the clean gas chamber of the filter, and a purging air flow comprising the actual jet plus
this entrained secondary air is introduced into the filter bag. The ratio of secondary flow to jet
flow is called mass flow ratio.
Cleaning the bags involves three stages:
1) The normal filtration gas flow is briefly interrupted by the barrier effect of the purging air
flow in the opposite direction.
2) The purging air injected into the bag expands it to its original circular section (Fig. 7) and
removes the dust cake which falls down into the dust bunker.
3) The purging air then flows outward through the filter medium in the direction opposite to
that of the raw gas flow.
A compressed air pulse of only 0.1 - 0.2 s duration is sufficient to perform all three above
mentioned stages. The pulse is applied at intervals of 1 to 10 minutes. Thus, the duration of
the cleaning operation amounts to only 0.02 - 0.3% of the overall operating time of the filter.
Practically speaking, therefore, the whole filter surface area is always available for filtration,
and the net area is virtually equal to the gross area.

Figure 7 Jet pulse bag charged from the outside

3.2.1 Guide Values for Jet Pulse Filters
Although there exist many different forms of construction of jet pulse filters it is possible to
give some guide values.
For common jet pulse filters, with vertical bags and nozzle injectors, the following data can
be given:
Table 5: Typical data of jet pulse filters
Compressed air pressure (normal) bar 47

Low pressure bagfilter bar 0.8r
mass flow ratio - 3-6
Compressed air flow rate Nm3/m3h 0.05 - 0.10
spec. power consumption1) kW/m2 0.005 - 0.010
1) installed capacity
The main criteria that are defining the size of a jet pulse bag filter are:
maximum actual gas flow
maximum permissible air to cloth ratio (A/C)
Table 3: Recommended A/C ration m3/m2h for jet pulse filters
Kiln Exhaust Gas Clinker Cooler Cement Mill

Vent Air
with Cyclone without Cyclone
55 65 60 80 60 - 70 90
3.2.2 Bag Dimensions

Jet pulse bag filter bags are smaller than bags from reverse gas bag filters because of the
different cleaning principles. The diameters of the bags are usually between 130 and 150
mm. The length of the bag should be between 4 and 3 meters, if a conventional gas flow
pattern is applied, which means, that the gas enters through the hopper and then flows
upwards, along the bags. In newer applications, the gas is distributed partially through the
hopper and partially horizontally to the bags. Together with new low pressure cleaning, here
larger bags of up to 6 m are applicable.
Generally longer bags are more difficult to clean and have therefore an increased
compressed air consumption. It is also very difficult to remove a bag of more than 3 to 4
meters length if it is filled with dust because of a hole in the felt.

3.2.3 Bag Material
As mentioned above, most bags for jet pulse filters are made with needle felt. This material
allows a higher air to cloth ratio and resist the stress of the jet cleaning. For some exceptions
(mainly bypass bag filter) and other high temperature applications, fiber glass bags are used
which resist the high temperature and the aggressive environment.
Experience during the last few years showed that jet pulse bag filter (operated at lower
temperatures, < 130C) are more reliable and economical than others. At temperatures
below 130C low cost bags like polyester or polyacrylnitril bags can be applied and the
power consumption of the filter fan is reduced as well.
The still on-going fast development of fabrics significantly improved the performance of bag
filters. Due to this development it is possible to install very efficient bag filters at cost that are
in the same range or below the cost of EPs.
Special treatment of the bag fabric surface allows for more frequent cleaning without
increased dust emission. The smaller dust cake on the bags is reducing the pressure drop
and therefore saves fan power.
3.2.4 Maintenance of Jet Pulse BF (for Kiln and Clinker Cooler)

Except for the valves and dust discharge equipment jet pulse BF have no moving parts. The
most important aspect for maintenance are the bags. Holes in the bags can significantly
increase the dust emission. Defective bags in reverse gas BF are replaced on line by
isolating the respective compartment. Modern jet pulse BF do not need compartments like
the reverse gas BF but usually consist of about 4 to 6 modules (static stability) with manual
outlet and inlet valves for each module.

3.3 Reverse Gas Cleaning
The reverse air cleaned bag filters usually contain woven filter bags. The raw gas enters the
bags from the bottom. It flows from the bag center to the outside of the bag. The dust is
deposited on the inner surface of the bag. Removal efficiency is improved and maintained by
these particulate deposits (residual dust cake). With time excess particulates are deposited
on the bags and increase the system resistance to the gas flow (pressure loss).
To allow the filter fan to operate within the design parameters and to reduce the fan power
consumption, this dust cake must be partly removed. Bag cleaning methods must be
designed properly - not over cleaning or under cleaning.
The bag cleaning process is triggered either by a timer or, better, when the pressure drop
over the bag filter reaches some predetermined level. A reverse air bag filter consists of
several compartments, usually of more than ten. When the bag cleaning process is started,
the outlet valves of one of the compartments are closed (off-line cleaning). Then, an auxiliary
fan forces a relatively gentle flow of filtered gas backwards through the compartment and
bags to be cleaned. This causes the bags to partially collapse inward, dislodging the
dustcake. This falls through the bags, the thimble and the tubesheet into the hopper. Metal
anticollapse rings sewn into the bags along their length prevent complete bag collapse.
Fig. 5 is a schematic of compartments in a reverse-gas-cleaned BF showing operation
during filtering, cleaning and purging prior to shut sown or maintenance. To support the
cleaning of the bags low frequency, pneumatic horns ca be installed and used
simultaneously with the normal reverse gas flow.
Table 4: Typical data of reverse gas filter
Reverse gas pressure mbar 30 - 40

Reverse gas flow m3/m2h 2.0 - 3.5
Power consumption (installed capacity) kW/m2 0.0075
The new generation of jet-pulse filters have significantly reduced the pressure drop over the
filter and the maintenance cost compared to the reverse gas filters.

Figure 8: Schematic of a shake / deflate-cleaned baghouse filtering flue gas,
isolated prior to bag cleaning (null), with deflation gas entering the
compartment prior to bag shaking, bag shaking and purging (or
ventilation) prior to maintenance
The main criteria that are defining the size of a reverse gas by filter are:
maximum actual flow
maximum permissible air to cloth ratio (A/C)
number of compartments
Table 5: Recommended A/C m3/m2h for reverse gas filter (net, net)1)
Operating mode Compound Direct

Kiln exhaust gas 30 36
1)
A) Because one compartment is usually isolated for reverse gas cleaning only
the filtering area of n-1 compartments are used to calculate the A/C (net).
B) The of the reverse gas must be added to the filter inlet flow to calculate the
total gas flow passing the bags and the A/C (net).
C) A) + B) A/C (net, net)

3.4 Comparison between Pulse Jet and Reverse Gas Cleaning
The following table shows a comparison between pulse jet and reverse gas cleaning. Recent
experience shows that the dedusting efficiency of jet pulse filters is the same as or better
than the efficiency of reverse gas filters. Investment and operating cost of jet pulse filters are
significantly lower compared to reverse gas filters.
Table 5: Comparison of jet pulse filter with reverse gas filter for 4-stage
preheater:
Kiln Bag Filters
Jet Pulse Reverse Gas
Bag quality Polyacrylnitrile Fiber glass
Relative cost for one set of bags 1 3-4
Pressure drop [mbar] 8 12 10 - 20
Bag cleaning 2) [Wh/kg cli] 0.6 - 0.8 1.1 - 1.5
CT water pump [Wh/kg cli] 0.6 - 0.8 1) ---
Filter fan [kWh/kg cli] 2) 6.7 - 7.5 6.5 - 9.0

1) Only during direct operation 2) Installed capacity
Table 6: Comparison of Investment Cost
Value given for a 4-stages preheater, dust emission < 30 mg/Nm3, plant located in Europe

Figure 9 Injector Systems for Pulse Jet Baghouse Cleaning

Figure 10 3-D View on Jet Pulse Bag Filter with Low Clean Gas Plenum

Figure 11 Pulse jet Bag Filter designed for Off-line Cleaning with High Clean Gas
Plenum

Electrostatic Precipitator
Arnaud De Luca
1. How does an EP work? ................................................................................................122

2. How EP Efficiency is determined ................................................................................124
3. How EP Clean Gas Dust Content is determined........................................................124
3.1 Clean Gas Dust Content (r) in Function of the Raw Gas Dust Content (R) .....125
3.2 Clean Gas Dust Content (r) in Function of the Gas Flow (Q)...........................126
3.3 Clean Gas Dust Content (r) in Function of the Total Projected Collecting Area (A)
127
3.4 Clean Gas Dust Content (r) in Function of the Migration Velocity () ..............128
4. HT-Rectifier ...................................................................................................................145
4.1 Semi-Pulse Energization..................................................................................146
4.2 Pulse Energization ...........................................................................................147
4.3 Improvement of Voltage and Current Wave Form Shape ................................148
5. Voltage-Current Curves ...............................................................................................149
6. Minimisation of Electrostatic Precipitator Trips ........................................................152

Electrostatic Precipitator (EP)
Our companies, and thousands more throughout the world, faced with a substantial
environmental challenge over the next few years. Increased governmental regulation and
enforcement of clean air laws will require your air pollution control equipment to consistently
meet rigid performance standards. And with an electrostatic precipitator that is always a
challenge.
1. HOW DOES AN EP WORK?

Fig. 3 shows a schematic drawing of an electrostatic precipitator. An industrial precipitator
has a number of passages through which the gases pass at a velocity of about 1 m/s. The
passages are formed by two parallel rows of vertically mounted collecting plates and a
number of discharge electrodes vertically suspended between the collecting plates.
Normally the spacing of the discharge and collecting electrodes varies between 125 and 200
mm and the voltage applied between them is 35 to 110 kV negative DC according to
spacing, gas and dust conditions.
The high negative voltage applied to the electrically insulated discharge electrodes creates a
strong electrical field between the discharge electrodes and the earthed collecting plates.
The highest strength occurs near the discharge electrodes. As the voltage is raised,
electrical breakdown of the gas close to the electrode surface takes place. This breakdown,
called "corona", appears as a bluish glow extending into the gas a short distance beyond the
surface of the discharge electrode.
The corona produces large numbers of gas ions, the positive ions being immediately
attracted to the discharge electrodes while the negative ions migrate towards the collecting
plates.
Some of the moving ions attach themselves to dust particles suspended in the gas between
the electrodes. Dust particles are charged either by bombardment by the ions moving under
the influence of the electrical field, or by ion diffusion, both types of charging taking place
simultaneously. The particle size determines which type of charging is predominant, ion
diffusion being the prevailing mechanism for particle sizes below 1 micron.
Through the influence of the electrical field, the negatively charged particles migrate towards
the collecting plate to which they adhere while being electrically discharged.
These particles build up a layer of dust on the plate surface which is dislodged by rapping.
The dislodged particles fall by gravity towards the bottom of the precipitator, ending up in the
bottom hopper from where the dust is extracted by either mechanical conveyor (drag chain
or screw conveyor) or pneumatic type system.

Figure 1 How electrostatic precipitators work

2. HOW EP EFFICIENCY IS DETERMINED
There have been many attempts over the years to develop satisfactory equations based on
fundamental theories in order to enable the efficiency of a precipitator to be forecast. They
are contained in a large number of technical papers, which are conveniently summarized in
H.E. White's book entitled "Industrial Electrostatic Precipitation". Generally the performance
of EPs can be expressed by the following Deutsch formula:
L
( )
EP = 1e v s
2)
or
A
( )
EP = 1e Q
3)
where
= Efficiency of the electrical precipitator
= Particle migration velocity (m/s)
A = Total projected collecting area (m2)
Q = Gas flow (m3/s)
L = Field length (m)
v = Gas velocity (m/s)
s = Distance between collecting and discharge electrodes (m)
From equation 3 it follows that the dedusting efficiency of a precipitator depends on:
I the migration velocity w (m/s)
II the total projected collecting area A (m2)
III the gas flow Q (m3/s)
3. HOW EP CLEAN GAS DUST CONTENT IS DETERMINED

If equations 1 and 3 are combined then one can describe the clean gas dust content r in
function of
R the raw gas dust content (g/m3)
the migration velocity (cm/s)
A the total projected collecting area (m2)

Q the gas flow (m3/h)
A
( )
r = R e Q
[mg/m3] 4)

3.1 Clean Gas Dust Content (r) in Function of the Raw Gas Dust Content (R)
With respect to equation 4, one would expect that the raw gas dust content (R) is directly
proportional to the clean gas dust content (r). However, because the migration velocity is
increasing with R, the effect of R on r is much weaker than expected.
Figure 2 Example of correlation between raw gas dust content R and clean gas
dust content r
Cleangasdust content r [mg/Nm3dry]
45
40
35
designpoint
30
25
20
30 40 50 60 70
Rawgasdust content R[g/m3]

3.2 Clean Gas Dust Content (r) in Function of the Gas Flow (Q)
The equation 4 shows that r is an exponential function of the inverse gas flow (Q). However,
the migration velocity ( ) is also influenced by the gas flow. Therefore, the correlation of r
and Q is not exactly according to the equation 4 with = constant.
Figure 3 Example of a correlation between gas flow Q and clean gas dust content
r for a modern kiln EP during compound operation
100
Clean gas dust content r [mg/Nm3dry]
90
80
70
60
50
40
designpoint
30
20
10
0
80 85 90 95 100 105 110 115 120
RelativegasflowQ[%]
At relative gas flows above 100%, r is increasing exponentially because of the exponential
correlation of r and Q (equation 4) and the amplifying effect of turbulence and dust re-
entrainment from the collecting plates.
The latter is overlaid by other effects mainly based on physical and chemical changes of the
particulates caused by the lower clean gas dust content (r).
At this point we already realize that the calculation of r is very complex because,
unfortunately, migration velocity is not constant but a function of R, Q and other variables.

3.3 Clean Gas Dust Content (r) in Function of the Total Projected Collecting Area
(A)
The total projected collecting area (A) is
A =l h G F 2 [m2] 5)
where
l = Length of field (m)
h = Height of field (m)
G = Number of gaps of one field (-)
F = Number of fields (-)
The factor of 2 is required because both sides of the collecting electrodes are active during
the dust extraction process.
The correlation between A and r is about inverse to the correlation between Q and r (see
equation 4). It is important to notice that the required collecting area is increasing
exponentially with the reduction of the clean gas dust content.
Figure 4 Correlation between the projected collecting area A and the clean gas
dust content r

3.4 Clean Gas Dust Content (r) in Function of the Migration Velocity ()
Among the variables in equation 4 the migration velocity is the one with by far the most
practical significance. Not because it has a stronger impact on r than the others (see
equation 4) but because it represents the effect of all other variables besides Q, R and A on
the EP efficiency.
Figure 5 Clean gas dust content r in function of the migration velocity
The migration velocity can, somewhat simplified, be understood as the average velocity of
the dust particles in their migration from the discharge to the collecting electrode in the
electrostatic field.

The migration velocity () itself is a function of many other variables like
Nature of dust
Electrical resistivity
Size
Gas condition
Temperature
Volume
Humidity
Chemical composition
Dust load
False Air
Energization of electrical fields
Voltage
Current
EP design
Gas distribution
Electrode design
Electrode cleaning
and these are only the most important ones.

With this information we can rewrite equation 4 as follows:
A
( ( , , T, Q, , c, r, R, l, U, , ...))
r =R e Q
6)
For most of the mentioned variables there exist empirical graphs describing the correlation
between the variables and . Some of these graphs were published but others are the
secrets of the suppliers.
Various attempts to calculate theoretically were not successful.

3.4.1.1 Electrical Resistivity of Dust
The electrical resistivity of the dust particles plays a very important part in the precipitation
process and depends mainly on the type of the dust, the gas temperature and the gas
humidity.
Figure 6 Dust resistivity in function of temperature and dust source
Three ranges of electrical resistivity can be distinguished:

less than 104 cm
104 to 1011 cm
more than 1011 cm
For particles having a resistivity of less than 104 cm the electrical conductivity is so high
that although they are charged in the normal manner and move normally under the influence
of the electrical field, the attainable dedusting efficiency is poor. The reason thereof is that
as soon as they reach the collecting electrodes, the electric charge leaks away so rapidly
that the particles are repelled into the gas stream and most likely escape with the outlet
gases.

Dust types belonging to the range comprised between 104 and 1011 cm show a
favorable discharge behavior. This means that neither particle repulsion nor back-ionisation
occurs, i.e. the particles are nicely deposited on - and sufficiently attached to the collecting
electrode.
Cement industry dusts usually belong to these "easily" separating dusts. Dusts stemming
from long dry process kilns, suspension preheater kilns and grate clinker coolers, however,
may occasionally develop dedusting problems.
Particles having resistivities of more than 1011 - 1012 cm can form within a very short
period of time an electrically insulating layer on the collecting electrodes leading to the so-
called back-ionisation (back corona) effect. With "back-ionisation" already captured dust is
forced back into the gas flow and a reasonable dedusting efficiency of the precipitator
becomes impossible to obtain.
Figure 7 Back ionization of dust particles at high electrical resistivities
Dust resistivity at temperatures below 200C is primarily determined by the amount of

moisture present in the gas. Therefore, a wet kiln will have a much lower resistivity than a
standard long, dry or a preheater kiln. In fact that was the reason why a water spray /
conditioning tower was added to these kiln systems to tread the exhaust gas.

The variation of the resistivity as a function of the moisture content of the raw gas is due to
an extremely thin conditioning layer on the particle surface which modifies the resistivity of
the dust.
At higher temperatures (above 350C) the particles become increasingly conductive and the
gas composition ceases to have much effect as a such.
Figure 8 Dust resistivity in function of the temperature and the dew point
At middle-range temperatures of about 200 to 250C the resistivity curve of some dust
reaches a maximum.

Figure 9 Dust removal efficiency as a function of the EP operating temperature
3.4.1.2 Size of Dust Particulates

According to Stoke's law for particles larger than 1 m, the migration velocity is directly
proportional to the particulates diameter.

Figure 10: Dedusting efficiency in function of the particle size
A dust with a mass mean diameter of 10 microns would require a precipitator only one-third
the size of a system collecting dust with a mass mean diameter of two microns. As you can
see, goes down when it is dealing with particulate in the 0.5 micron range and then starts
to improve in efficiency when the particulate gets smaller (say 0.05 microns). That has to do
with the two principals of particle charging which predominate in a precipitator. Field
charging predominance for particulate greater than 1 micron in size and diffusion charging
predominates for particulate less than 1 micron in size. That range around 1 micron is kind of
a no-man's land where neither field charging nor diffusion charging has much effect. That is
why the efficiency drops dramatically and then improves once the particles get even smaller.
What are other consequences for the EP operation based on the correlation between and
particle size:
EPs are classifying the incoming dust. The coarse particles are found in the first fields
and the fine fraction in the last fields.
This classifying of the dust can be used to extract selectively a dust portion enriched with
condensibles like K2O, SO3 and heavy metals, thus avoiding generation of larger
quantities of "contaminated" dust or enrichment of certain compounds in the process.
Fine dust particulates and condensibles can be accumulated in the system and reduce
the EP efficiency if they are not extracted from the last field.
The particle diameter of the clean gas dust is generally below 10 m.
3.4.1.3 Gas Temperature T

Gas and particulate temperature are usually the same because the particulates are
suspended in the gas and the retention time of particulates in the gas is sufficient to reach a
temperature equilibration.

The influence of the gas temperature T is mainly:
Increased dust resistivity at higher temperatures below 250C (see para 1.3.4.2)
Decreased dust resistivity at higher temperatures above 250C (see para 1.3.4.2)
Increased actual gas flow Q at higher temperatures (see para 1.3.2)
3.4.1.4 Gas Humidity (dew point )

The water of the raw gas is originating from:
Combustion (4 CmHn + (4 m+n) O2 -> 4 m CO2 + 2n H2O)
Water in ambient air

Water in raw materials
Water injection for gas conditioning

The dew point can be calculated as follows:
336.48
= 179 [C] 7)
5.3362 17.045+ln(Vf Ptot )
where:
Vf = Volume fraction of water vapour in the wet gas (m3 H2O / m3wet gas)
Ptot = Total pressure (bar)
As described in para 1.3.4.1 the dew point is influencing the electrical resistivity of the dust
particulates at temperatures below 250C. This is responsible for the increased efficiency of
the EP at higher dew points
Figure 11 Example for clean gas dust content in function of the dew point at
temperatures below 250C
The figure above shows the strong effect of gas dew point on EP efficiency if no back
ionization occurs. With back ionization the clean gas dust content r would increase even
faster at lower dew points.
A typical example for the influence of the dew point are preheater kilns switching from
compound operation (mill on) to direct operation (mill off). When the raw mill is in service, the
moisture conditioning (11 % to 12 % at 110C) of the gas is optimum. When the raw mill
goes off line, the spray tower preceding both the raw mill and the EP cannot catch up quickly
enough to increase the volume of water to make up for the moisture content lost when the
raw mill goes down.

3.4.1.5 Gas Composition (not including water vapour)
The gas composition of clinker cooler vent air is fairly simple, however, the composition of
kiln exhaust gas is a complicated cocktail of many different compounds.
Some compounds like SO2 can enhance the EP operation by reducing the resistivity of the
particulate surface.
Others like organic compounds or condensible alkalis reduce the EP efficiency. It is
assumed that organic compounds attached to the particulate surface can increase their
resistivity. Condensible alkalis can occur as very fine particulates < 10 m significantly
reducing the average migration velocity.
Condensibles like chlorides can increase the stickiness of the deposited dust on the
electrodes which leads to thicker dust layers on the electrodes. This would increase the total
electrical resistivity of the dust layer and therefore reduce the EP efficiency.
3.4.1.6 Gas Dust Load S, r and R

An increased raw gas dust load R has a positive effect on the migration velocity but cannot
fully compensate the raise in the clean gas dust content according to equation 4.
An increase of the clean gas dust content r is also increasing the migration velocity.
According to the explanation under para 1.3.4.2 the lower the clean gas dust emission is the
lower is the diameter of the dust particulates and smaller dust particulates have a slower
migration velocity than larger ones.
Therefore, the required collecting area A is increasing exponentially with the reduced clean
gas dust content r (see Fig. 6).
3.4.1.7 Energization of the EP

The collection efficiency of a precipitator is directly related to the total power for all fields on
the precipitator. In general, the higher operating power levels that each field can achieve, the
higher collection efficiency for that field.

Figure 12 Example for clean gas dust content r in function of the relative power
input
Many people believe that a precipitator cannot work (achieve power levels) unless the gas is
loaded with dust. This question can be easily examined by energizing any field of an EP in
air. By that, it is meant that the kiln is not in operation, and that the temperatures have
settled to ambient conditions. Furthermore, the precipitator is not bottled up and dampers
are open, allowing for a natural stack draft through the precipitator. It is important to have
some air movement in order to obtain a good "air load".
When a precipitator is energized in air, the following results could be obtained:
Table 1: Example of energization of an electrical field of an EP under pure air

(without dust)
Precipitator Secondary Voltage (kV) Precipitator Secondary Current (mA)
0 0
1 0
5 0
10 0
15 0
16.5 1
24 100
28 200
30.6 300
33.5 400
34 500
35.2 600
36.3 700
36.8 750

Actual results are dependent on T/R set size, type of high voltage electrode, and the
electrical clearance between the electrodes.
The mA readings are synonymous with the actual current flowing in the precipitator. Current
flowing in a circuit is equivalent to the number of electrons that are moving past the point in
that circuit.
For current to flow in a precipitator, that means that electrons need to flow from the
discharge electrode to the collecting electrodes in the precipitator. That means that the air in
the precipitator must become a conductor. It is easy to think of the various conductors and
realize that an insulator is a very poor conductor, a piece of copper wire is an excellent
conductor, and an energized precipitator is somewhere in between. The air load
demonstrates that current does not start to flow in a precipitator until (in this case) a voltage
of 16.5 kV is achieved. That voltage is referred to as the corona onset voltage.
With moderate increases in voltage, a correspondingly increasing current results. If the
alignment is correct between the electrodes in the precipitator, then the air load test should
achieve either the primary or secondary current rating of the T/R set being energized. In the
above example, we ran out of secondary current (705 mA) first.

Therefore, in order to get corona discharge in a precipitator, dust particles are not required.
However, the concentration of particulate has a dramatic effect on the power levels in the
precipitator. The term "space charge" is used to indicate a precipitator field that is collecting
a significant number of fine particles or a heavy concentration of large particles. For our
example, we will examine the latter, which is a common occurrence in cement plant
precipitator applications.
Space charge - high dust concentrations
As we saw in the section on air load, since there are no particles (dust) in the inter electrode
space, there can be no space charge. However, with the influence of a large concentration
of large particles, see what affect it has on these two wet process cement kilns. Kiln No. 1
has a cyclone mechanical collector in series with the precipitator, whereas kiln No. 2 does
not. The automatic voltage controls for those two precipitators were found to be operating as
follows:
Table 2: Energization of two EPs with different dust loads Kiln No. 1: low dust
load, kiln No. 2: high dust load
Unit Amps Volts mA kV kW Sparks/Minute

Kiln No. 1-1 123 337 664 50.1 27 0
Kiln No. 1-2 142 247 758 36.2 23 0
Kiln No. 2-1 9 232 39 57.5 1a 20

Kiln No. 2-2 16 324 71 52.0 2 14
Kiln No. 2-3 115 465 940 48.0 38 3
Kiln No. 2-4 120 346 924 35.1 28 0
Because kiln No. 2 does not have the mechanical collector preceding it, the dust loading
(concentration) is significantly higher than kiln No. 1. The voltage control readings show the
affect of space charge. Space charge is indicated by high voltages, but more importantly, by
extremely low current. It is the absence of current flow that can be of significance.
When asked what is the more important parameter, precipitator voltage (kV) or precipitator
current (mA), often times people will say kV. They are partially correct in most cases, but not
in this case. Precipitator voltage is responsible for pushing the dust particles toward the
plates. Current is responsible for keeping them there. So although kiln No. 2, field 1 has a lot
of pushing forced, (57.5), it has no holding force. Most of the dust re-entrains onto the next
field.
The other important point to note is that sparking in a precipitator (an electrical breakdown of
the gas) is directly related to the precipitator voltage levels. That is why inlet fields have
sparking (because of the high kV) whereas outlet fields sometimes do not.
If one looks at the flow of current from the transformer / rectifier to set to the high voltage
electrodes through the dust laden gas, to the collecting plate and back to the T/R set
(through earth ground) as shown in Fig. 15 the effect of the ion mobility may become
apparent.

The air load demonstrated that in air without dust, the main current carriers are the free
electrons and the negative ions. These two characters can be compared to running backs on
a football team. They are very swift moving and seek the holes, and the mA meter counts a
lot of them during an air load.
With the introduction of dust into the precipitator, the ion mobility changes dramatically. The
charged particles, which move very slowly, establish a "particulate space charge" in the inter
electrode space. Fig. 15 gives an idea of their relative velocity.
The affects of high space charge can be both influential and detrimental. On the positive
side, high voltages created by space charge in turn create higher "electric fields". The
electric field is the pushing force against the dust particles, accelerating them towards the
collecting plates. Higher accelerations toward the collecting plates can result in increased
efficiencies.
However, as in our example in table 2, kiln No. 2 was operating with very low current levels.
Therefore, the space charge enhanced the particulate collecting field (high voltages), but
also contributed towards a suppression of the corona current. Corona current directly affects
particle charging. The higher the particle charging ensure that the dust loss due to particle
re-entrainment is diminished. If the corona is suppressed, this can promote re-entrainment.
That is the case on kiln No. 2.
Figure 13 Relative velocity (mobility) of current carriers

The peak value of the precipitator voltage is limited by the dielectric constant of the gas. The
arc-over voltage is the only value which determines the maximum possible precipitator
voltage. The total power input and therefore the EP efficiency are strongly influenced by
the applied voltage.
Up+Uv
Pc = lm 8)
2
where
Im = Mean secondary current
Up = Secondary peak voltage
Uv = Average secondary voltage
The factors determining the maximum possible precipitator voltage can change quickly.
Therefore, the efficiency of the automatic voltage control, that is adjusting the voltage to
operate at the maximum value, is directly correlated with the EP efficiency.
The functioning of HT-rectifiers and automatic voltage control is explained in para 1.4.
The electrical operating behaviour is also changing over the length of the field. Gas
turbulence and distribution, dust content and particulate size at the EP inlet are very different
from the ones at the EP exit. Therefore, to optimize the energization of the EP the electrodes
should be subdivided mechanically and electrically in the length direction.
Figure 14 Clean gas dust content r in function of the number of independent

electrical fields at constant collecting area A
3.4.1.8 EP Design
The equipment parts with the main influence on the migration velocity are:
Gas distribution screens
Electrode
Electrode cleaning systems

Gas distribution
In general terms the ducting leading to the precipitator and the inlet and outlet funnels should
be designed to ensure a proper gas velocity distribution in view of utilizing the whole
collecting area and avoiding negative velocity effects. From a practical viewpoint this implies
different requirements to the gas distribution in the different parts of the precipitator, and too
strictly formulated numerical rules for deviation from uniformity may not be justified.
The inlet gas distribution must be sufficiently uniform to secure a reasonable uniform current
distribution. This is especially important for precipitators for processes with high resistivity
dust and fine particles. A rule of thumb says that the standard deviation of the gas velocities
in the EP should be below 30%.
The velocity profile at the outlet should be specifically selected to reduce the risk of re-
entrainment in the bottom region.
Sneakage of dust laden gases around the electrically energized electrode system must be
kept at an absolute minimum, in particular at the bottom part of the precipitator. And large
eddies in the bottom hoppers caused by the velocity "slip" at the bottom of the electrode
system may aggravate the influence from sneakage because particles already picked up by
the hopper are swept into the main flow again.
High local velocities may scour away already precipitated dust from the collecting plates. In
this case a good gas distribution combined with high average velocity may not be superior to
a bad gas distribution combined with low average velocity.
The gas distribution may influence the dust space charge distribution and thereby the current
distribution in a separately energized precipitator field. In areas with low velocities or, in
extreme situations, areas with recirculating flow, the particle concentration will be much
lower than in corresponding areas with higher velocities. Consequently the power input will
be limited in the high velocity areas causing a reduction in overall efficiency. In particular
with high resistivity dust such uneven current distribution will cause back ionization and
frequent sparking, resulting in lower average voltage and current and increased dust re-
entrainment. Due to the turbulence the gas distribution in each separate duct will tend to
improve through the precipitator, thus smoothing the dust space charge. However, a skew
cross distribution at the field inlet will not be smoothed to the same extent, and so the
horizontal gas distribution should be fairly uniform in order to maintain a proper current
distribution.
Finally, high local gas velocity, combined with high dust content, can result in erosion of the
edges of the collecting plates and other internal parts of the EP. Low gas velocity can cause
dust build ups.
The cross section of an ideal EP should be designed to achieve an average gas velocity of
Kilns 0.8 - 1.0 m/s
Clinker coolers 0.7 - 0.9 m/s
The gas velocity should not drop below 0.5 m/s to maintain a suitable gas distribution.

Electrode design
The electrodes have two duties. First emission of electrons (discharge electrode) and
second the collection of the dust (collecting plate).
The energization of the fields or in other words the supply with voltage and current is
influenced by the discharge electrode design. Various electrode designs to achieve optimum
voltage or / and current are employed by the suppliers. An important factor is the corona
onset voltage which depends mainly on the radius of the electrode (plan strips) or the radius
of spike peaks. The corona onset voltage is increasing with the above mentioned radius.
Figure 15 Corona discharge voltage in function of the discharge electrode radius
In applications, where a high current is required (high dust load, low resistivity), the electrode
radius should be small. In situations, where current must be reduced and voltage increased
(high resistivity dust -> back corona) electrodes with larger radius (without peaks) can
improve the efficiency.
Since corona discharge is also greatly affected by dust settling, the discharge electrodes
need rapping, which means that their oscillation behaviour is of utmost importance. Best
results have been obtained with rigid frame-mounted electrodes or rigid electrodes.
For maximum collection efficiency, the collecting plates must be rigid to maintain the critical
spacing between the different electrodes and withstand bowing during operation. At the
same time, they must facilitate the efficient transfer of rapping energy for effective cleaning.
Not optimum cleaning can amplitude back corona effects and generally reduce the EP
efficiency.

4. HT-RECTIFIER
The High Voltage Rectifiers are responsible for optimum energization of the EP under
different operating conditions. Optimum energization means:
Clean gas dust content r below the target
Minimum energy consumption
The precipitator energization has a very strong influence on precipitator collection efficiency.
As a result of this recognition, the microprocessor-based controller for precipitator high
voltage power supplies have in recent years become the general standard. These
programmable, fast reacting, digital controllers can implement sophisticated control
strategies through their monitoring of secondary current and voltages, including
differentiation of reactions according to type of arc or spark in the precipitator, arc
quenching, fast voltage recovery after arcing without reignition of the arc, automatic current
limitation to the nominal current at overload or short circuit conditions and operation at a
precipitator current level just below the onset of "back corona". They continuously control
flash-over rate and power input to the precipitator for optimum performance.
Figure 16 Automatic voltage adjustment. Behavior of EP voltage at constant arc-

over limit
Most microprocessor-based transformer / rectifier controllers have or can easily be

supplemented with an option for semi-pulse energization, as described in the following:

4.1 Semi-Pulse Energization
An inexpensive method for reduction of precipitator power consumption, and in some
instances also improvement of precipitator performance, is also known under various trade
names such as semi-pulse intermittent energization and energy-control.
Semi-pulse energization is implemented at a conventional thyristor controlled full wave
transformer / rectifier simply by suppressing for instance two out of three, or four out of five
half waves. The ripple of the precipitator voltage hereby becomes more pronounced than
with conventional energization, resulting in a voltage wave form that resembles a DC base
voltage superimposed with long duration pulses.
The intermittent nature of the corona discharges gives this form of energization certain
properties resembling those of the later discussed pulse energization. Semi-pulse has, in
some cases, been able to improve the performance of precipitators operating with medium
to high resistivity dust, but as a rule not to the same degree as pulse energization. Its main
advantage is the resulting power savings. Power saving up to 90% and emission reduction
of up to 50% were reported.
Figure 17 Voltage wave form for semi-pulse energization

4.2 Pulse Energization
Advances in high-power switching technology in recent years have made it possible to
develop pulse energization systems with sufficient reliability and capacity to energize large
precipitators.
With pulse energization short duration, high voltage pulses are repetitively superimposed on
a DC base voltage. Some energy conserving pulse energization systems utilize pulses with
a duration in the order of 100 microseconds and pulse repetition frequencies up to 200
pulses per second.
Pulse energization makes it possible to attain more favorable electrical conditions for high
resistivity dust than is obtainable with conventional DC energization. Pulse energization,
therefore, can successfully be used to improve the performance of an existing precipitator
operating with high resistivity dust or to reduce the size of a new precipitator installation for a
high resistivity application, as for instance with the so-called "hot" precipitators for kilns.
Power saving of up to 90% and emission reduction of up to 60% were reported.
Figure 18 Voltage wave form for pulse energization

4.3 Improvement of Voltage and Current Wave Form Shape
Modern precipitator power supplies include silicon controlled rectifiers (SCR's) and current
limiting reactors (CLR's). SCR's and CLR's are designed to produce an optimum
energization of the EP at one specified process condition. Since it is known that this
condition can change very frequently, the current and voltage input have to be adjusted
continuously. Operation of CLR's at conditions which are not according to the design
specifications can produce a poor current wave form shape (poor form factor) which leads to
a reduced power input. This can be corrected with a variable inductance current limiting
reactor (VI-CLR).
Figure 19 Improving wave form shape with variable inductance current limiting
reactor
Another more expensive way to increase the power input is the utilization of a 3-phase
energization. The transformer is operated with a square wave voltage with a frequency of
500 Hz. This produces a very flat direct voltage that can under certain circumstances result
in a higher power input. Unfortunately, very little experience is available for this system.

Figure 20 Improving current wafe form shape with 3-phase energization
Intelligent EP control systems do limit the power input if additional power input does not
result in significantly reduced dust emission (see Fig. 14).
5. VOLTAGE-CURRENT CURVES
A voltage-current curve to a precipitator troubleshooter is like a stethoscope to a cardiologist.
When a precipitator is running, we cannot see what is happening inside that might affect its
performance. However, by a close examination of the relationship between the voltage and
current levels in the operating precipitator, one can predict what is affecting performance.
A V-I curve is run by taking the voltage controls to zero then slowly increasing the power
levels, recording both the kilovolts and milliamps at convenient intervals (usually 50 mA or
100 mA) until the voltage control sparks over. A curve can then be drawn from the points
collected utilizing the "X" axis for the kilovolts and the "Y" axis for the milliamps. Some typical
V-I curves for a dry process cement kiln are shown on Fig. 23. Note that the voltage and
current corresponding to each field reflects the voltage and current relationships as first
shown in Table 2 of our precipitator example.
When there are problems with the operation of the precipitator, Fig. 24, would be more
helpful for troubleshooting. For example, the high resistivity dust as indicated by low current
levels in the outlet fields may show up as the "moderately high" dust resistivity curve shown
on Fig. 24. This short, stubby curve shows corona onset voltage as normal (say around 18
kV), but current level only increases to a very low level as opposed to the way an outlet field
should, as shown on Fig. 23.
This contrasts with a misalignment of the electrodes (wire-to-plate spacing) in the
precipitator. Misalignment exhibits itself by a very low corona onset voltage (the electrical
clearance is decreased), and the spark over.
These curves can also be utilized to show if there is excessive dust buildup on the high
voltage electrodes. Excessive dust buildup exhibits itself almost as if the wire diameter of the
high voltage electrode has been increased. Dust buildup on the wire has the same effect of

increasing the corona onset voltage from the normal range of 15 - 20 kV on up to 25 to 35
kV. The problem with wire buildup is that you are not able to achieve as high a current as if
the wires were clean. Remember, a precipitator needs both high voltage and high current
levels.
Figure 21 Normal precipitator voltage-current (V-I) curves

Figure 22 Abnormal precipitator current-voltage curves
Table 3: Influence of some variables on EP's dedusting efficiency
Variables Variation Efficiency Dust

Emission
Raw gas dust content R
Gas flow Q
Collecting area A
Electrical resistivity of the dust / /
Temperature T1 (<200C)
Temperature T2 (>300C
Particle size (> 1 m)
Humidity ()

Organic emission c
Power input P
Standard deviation of gas distribution
Misalignment of electrodes
Speed of the controller for EP
energization
6. MINIMISATION OF ELECTROSTATIC PRECIPITATOR TRIPS

Electrostatic precipitators (EP) for dedusting of kiln/raw mill exhaust gases are normally
protected against gas explosions by installing a shut-off CO-level. If the CO concentration in
the exhaust gas rises above a defined level, the EP fields are shut off to suppress sparks in
the EP that could trigger an explosion. Grounding (shutting off) the EP results in dramatically
increased dust emissions, observed as a huge dust cloud hanging over the plant.
Therefore, HMC-TPT and ETPS have elaborated guidelines for a systematic approach,
which will allow any plant to reduce the number of EP trips due to CO peaks excluding those
produced by power failures to below 5 trips per year (objective). The approach consists of
three steps:
1. Assessment of the current situation with respect to all main factors influencing EP trips.
2. Comparison of the current situation with the desired, optimum situation (gap analysis)
and identification of the optimization potential and priorities
3. Optimization according to (2.), which may include:
Process optimization for minimization of the CO levels
Optimization of the CO analysis system to assure availability, reliability and speed of
information on CO levels
Optimization of the reaction to the measured CO levels, e.g. fact based CO limit or
optimized EP power management.

Figure 23: 3 Phases to minimise electrostatic precipitator trips
These 3 phases are developed and detailed in a handbook that can be directly used in the
plants in order to improve the situation.
Handbook for Minimization of Electrostatic Precipitator Trips, report TPT 01/21056/E
Figure 24 Longitudinal Section of a 2-Field EP (Lurgi)

Figure 25: 3D view on a 2-Field EP (ELEX)

Figure 26 Insulator Chamber FLS (Type C)

Figure 27 Electrode System (Lurgi)

Figure 28 Collecting Electrode Rapping System (Lurgi)

Figure 29 Suspension and Rapping of Discharge System (Lurgi)

Figure 30 Gas Distribution Screen (FLS)
Figure 31 Dust Removal System (FLS)

Dedusting Concepts for Cement Kilns
Arnaud De Luca
1. Concept 1 ......................................................................................................................162
1.1 Kiln Exhaust Gas..............................................................................................162
1.2 Clinker Cooler ..................................................................................................162
2. Concept 2 ......................................................................................................................163
2.1 Kiln Exhaust Gas..............................................................................................163
2.2 Clinker Cooler ..................................................................................................164
3. Concept 3 ......................................................................................................................165
3.1 Kiln Exhaust Gas..............................................................................................165
3.2 Clinker Cooler ..................................................................................................165
4. Concept 4 ......................................................................................................................166
4.1 Kiln Exhaust Gas..............................................................................................166
4.2 Clinker Cooler ..................................................................................................168
5. Concept 5 ......................................................................................................................168
5.1 Kiln Exhaust Gas and Clinker Cooler Vent Air .................................................168
6. Conclusion ....................................................................................................................170

Dedusting Concepts for Cement Kilns
Because of the changing needs and the different filter and cement production technologies,
the concepts for cement kiln dedusting are numerous. Each of them has certain advantages
and disadvantages.
The dedusting of preheater kilns with waste heat utilization (raw mill) and great cooler is the
most difficult case. Therefore, the following examples will be based on such a modern
cement kiln system.
Chart 1 shows the different possible dedusting concepts. Within the presented concepts
there is still some differentiation possible; e.g. the bag filter concepts can be further
differentiated by the fabric quality or the type of pulse jet that is applied.
Figure 1 Different dedusting concepts
It can be seen that the first decision which must be made is, if the kiln and clinker cooler
shall be dedusted separately or in one common filter. The standard solution is to use two
filters, one for the kiln and one for the cooler, but we will see later that the simultaneous
dedusting in one filter has some important advantages.

1. CONCEPT 1
Separate dedusting of kiln exhaust gas and clinker cooler vent air.
1.1 Kiln Exhaust Gas
1.1.1 Concept
Conditioning tower before ID fan to cool the exhaust gas to below 150C during direct
operation
3 fan system (separate raw mill fan and cyclones)
Electrostatic precipitator without pre-separation chamber
1.1.2 Advantages
Length of duct between preheater and conditioning tower very short because the
conditioning tower is placed ahead of the ID fan. Therefore, the investment costs are
reduced.
Improved raw mill control because of separate mill fan
Possibly reduced operating cost compared to bag filter
Less false air intake and danger of corrosion at the electrostatic precipitator because of
small negative static pressure (due to 3 fan system)
No precollection required at the filter (due to raw mill cyclones)
1.1.3 Disadvantages
Strong load changes on the ID fan because of different gas temperatures during direct
and compound operation (due to conditioning tower before ID fan)
More false air intake into conditioning tower and higher corrosion risk because of strong
negative static pressure (60 - 40 mbar) compared to conditioning tower positioned after
ID fan
Over all dedusting efficiency of electrostatic precipitator is lower compared to bag filter
1.2 Clinker Cooler
1.2.1 Concept
Electrostatic precipitator
1.2.2 Advantage
Possibly reduced operating and investment cost
1.2.3 Disadvantage
Efficiency of electrostatic precipitator is lower compared to bag filter

Figure 2 Concept 1
2. CONCEPT 2
2.1.1 Concept
Conditioning tower before ID fan to cool the exhaust gas to below 150C during direct
operation
2 fan system (no separate raw mill fan and cyclones)
Electrostatic precipitator
2.1.2 Advantages
reduced.
Slightly reduced pressure drop between ID fan and filter fan because of missing raw mill
cyclones (compare Concept 1)
Possibly reduced operating and investment cost compared to bag filters

2.1.3 Disadvantages
and compound operation.
Increased false air intake and corrosion at the filter because of very strong negative
static pressure
Electrostatic precipitator need separate precollector chamber
Efficiency of electrostatic precipitators is lower compared to bag filter
Gas recirculation to operate raw mill is increasing the required filter size (due to 2 fan
system)
2.2 Clinker Cooler

See concept 1
Figure 3 Concept 2

3. CONCEPT 3
3.1.1 Concept
Fresh air intake before and / or after the ID fan to cool the exhaust gas to below 240C
Reverse gas bag filter
3.1.2 Advantages
No cooling tower and water injection required
Higher overall dedusting efficiency than electrostatic precipitator
Simple filter inlet temperature control
Reduced corrosion in the filter because of high operating temperature and low dew point
3.1.3 Disadvantage
Higher operating cost than electrostatic precipitator and jet pulse bag filter
Very large filter
If the temperature control fails it is possible to burn the bags
With a bad design it is possible that the pressure drop over the filter is increasing to a
point where the capacity of the filter fan is not sufficient anymore to pull the gases
3.2 Clinker Cooler
3.2.1 Concept
Cooling of the vent air with air to air heat exchanger (designed for up set conditions)
Jet pulse bag filter; preferably equipped with polyester bags
3.2.2 Advantage
The air to air heat exchanger serves as a compensator in case of upset conditions and
allows therefore a very smooth operation of the filter.
Higher overall dedusting efficiency compared to EP
3.2.3 Disadvantage
If air to air heat exchanger and the jet pulse are designed properly, there are no real
disadvantages in this concept. With a bad design it is possible to damage the bags
during heat excursions or to increase the pressure drop over the filter up to the point
where the capacity of the filter fan is not anymore sufficient to pull the gases.

Figure 4 Concept 3
4. CONCEPT 4
Separate dedusting of the kiln exhaust gas and clinker cooler vent air.
4.1.1 Concept
Cooling tower before ID fan to reduce the temperature to below 150C during direct
operation
Fresh air intake to reduce the gas temperature to below 120C
Jet pulse bag filter; preferably with polyacrylnitrile bags.

4.1.2 Advantages
reduced.
Good mill control because of separate mill fan
Reduced investment and operating cost compared to reverse gas bag filter
Less false air intake and reduced corrosion risk at the filter because of low negative
static pressure (due to 3 fan system).
4.1.3 Disadvantages
and compound operation
More false air intake into cooling tower and higher corrosion risk because of strong
negative static pressure 860 - 40 mbar) compared to CT after ID fan
Slightly higher pressure drop between ID fan and filter fan because of the raw mill
cyclones
Very reliable temperature control required to protect the bags
Higher risk for corrosion at filter because of low operating temperature
With a bad design it is possible that the pressure drop over the filter is increasing to a
point where the capacity of the filter fan is not sufficient anymore to pull the gases.

4.2 Clinker Cooler
See Concept 3.
Figure 5 Concept 4
5. CONCEPT 5
Simultaneous dedusting of the kiln exhaust gas and clinker cooler vent air in one jet pulse
filter.
5.1 Kiln Exhaust Gas and Clinker Cooler Vent Air
5.1.1 Concept
Reduction of the clinker cooler vent air dust concentration in a cyclone
Direct operation: mixing of the clinker cooler vent air with kiln exhaust gas and reducing
the temperature to below 120C in an air to air heat exchanger
Compound operation: reducing the clinker cooler vent air temperature and mixing it with
the exhaust gas from the raw mill in a air to air heat exchanger
3 fan system (separate raw mill fan and cyclone)
Jet pulse bag filter preferably with polyacrylnitrile bags

5.1.2 Advantages
Only one filter
No cooling tower
No water injection
Simple temperature control
Reduced gas volume compared to all other solutions
Good mill control because of separate mill fan
Reduced investment and operating cost compared to the other bag filter concepts
Less false air intake and reduced corrosion risk at the filter because of low negative
static pressure.
Only one stack
Clinker cooler gas can be used to dry raw material in the raw mill
No visible plume at the stack because of the reduced dew point (no water injection and
"dry" clinker cooler vent air)
No load changes on the ID fan
5.1.3 Disadvantages
Because of raw mill cyclones slightly higher pressure drop between ID fan and filter fan
Kiln and clinker cooler not controlled by separate fan
Portion of the clinker cooler dust is mixed with the kiln dust

Figure 38 Concept 5
6. CONCLUSION
All concepts have certain advantages. Therefore, one cannot produce a ranking without
respecting the individual situation of the plants. Those that require a very reliable dedusting
without short time dust emission peaks should chose a bag filter concept. In case of water
shortage the concepts 3 and 5 without cooling tower are most suitable.
If alternative fuels are burnt it is possible that CO peaks are produced more frequently than
without, especially during the commissioning phase of the waste feed equipment. The plants
that are burning alternative fuels or those that are planning to do so are usually under more
intensive observation by the neighbours and the authorities. Frequent dust emission peaks
caused by EP CO-shutdown or changes of process conditions may be very embarrassing
when asking for a permission to burn alternative fuels or when applying for extension of the
permit.
Therefore, bag filters, especially the pulse jet type, will be for many plants the preferred
solution for the future.

Chapter 3
Maintenance
1. The Maintenance Elements ....................................................................................... 172
2. Condition Monitoring program ................................................................................. 197
3. Quality Inspections for Cement Plant Equipment................................................... 251

The Maintenance Elements
1. Maintenance Requirements Pyramid..........................................................................176

1.1 Description .......................................................................................................176
1.2 Purpose............................................................................................................176
1.3 Examples .........................................................................................................176
2. Critical Asset.................................................................................................................177
2.1 Description .......................................................................................................177
2.2 Purpose............................................................................................................177
2.3 Examples .........................................................................................................177
3. Works Order System ....................................................................................................177
3.1 Description .......................................................................................................177
3.2 Purpose............................................................................................................177
3.3 Examples .........................................................................................................177
4. Daily / Weekly Plan .......................................................................................................178
4.1 Maintenance Master Schedule.........................................................................178
4.2 Weekly Plan .....................................................................................................178
4.3 Daily Plan .........................................................................................................178
5. Maintenance KPIs........................................................................................................179
5.1 Description .......................................................................................................179
5.2 Purpose............................................................................................................179
5.3 Examples .........................................................................................................179
6. Daily Maintenance Report............................................................................................179
6.1 Description .......................................................................................................179
6.2 Purpose............................................................................................................179
6.3 Examples .........................................................................................................179
7. Maintenance Cost Structure ........................................................................................180
7.1 Description .......................................................................................................180
7.2 Purpose............................................................................................................180
7.3 Examples .........................................................................................................180
7.4 Description .......................................................................................................181
7.5 Purpose............................................................................................................181
7.6 Examples .........................................................................................................181

8. Production Plan ............................................................................................................182
8.1 Description .......................................................................................................182
8.2 Purpose............................................................................................................182
9. Spares Policy & Management......................................................................................183
9.1 Description .......................................................................................................183
9.2 Purpose............................................................................................................183
9.3 Examples .........................................................................................................183
10. Standards & Specifications.................................................................................183
10.1 Description .......................................................................................................183
10.2 Purpose............................................................................................................183
10.3 Examples .........................................................................................................183
11. Asset History System...........................................................................................184
11.1 Description .......................................................................................................184
11.2 Purpose............................................................................................................184
11.3 Examples .........................................................................................................184
12. Weekly Maintenance Report................................................................................184
12.1 Description .......................................................................................................184
12.2 Purpose............................................................................................................184
12.3 Examples .........................................................................................................184
13. Maintenance Cost Report ....................................................................................184
13.1 Description .......................................................................................................184
13.2 Purpose............................................................................................................184
13.3 Examples .........................................................................................................184
14. Resource Skills Matrix .........................................................................................185
14.1 Description .......................................................................................................185
14.2 Purpose............................................................................................................185
14.3 Examples .........................................................................................................185
15. Budget ...................................................................................................................185
15.1 Description .......................................................................................................185
15.2 Purpose............................................................................................................185
15.3 Examples .........................................................................................................185
16. Bill Of Materials (BOM) ........................................................................................185
16.1 Description .......................................................................................................185
16.2 Purpose............................................................................................................185
16.3 Examples .........................................................................................................185
17. Maintenance Master Schedule ............................................................................186
17.1 Description .......................................................................................................186

17.2 Purpose............................................................................................................186
17.3 Examples .........................................................................................................186
17.4 Overview ..........................................................................................................186
17.5 Description .......................................................................................................187
17.6 Purpose............................................................................................................187
17.7 Examples .........................................................................................................187
18. FMEA / RCM ..........................................................................................................187
18.1 Description .......................................................................................................187
18.2 Purpose............................................................................................................187
19. RCM Approach .....................................................................................................188
20. MTBF / MTTR / MTBCF .........................................................................................189
20.1 Description .......................................................................................................189
20.2 Purpose............................................................................................................189
20.3 Examples .........................................................................................................189
20.4 MTBF (Hours) ..................................................................................................189
20.5 MTTR (Hours) ..................................................................................................189
20.6 MTBCF (Hours)................................................................................................189
21. Predictive Routines / Condition Based Monitoring...........................................190
21.1 Description .......................................................................................................190
21.2 Purpose............................................................................................................190
21.3 Examples .........................................................................................................190
22. Plant Master Plan .................................................................................................190
22.1 Description .......................................................................................................190
22.2 Purpose............................................................................................................190
23. Multi-Skilling .........................................................................................................191
23.1 Description .......................................................................................................191
23.2 Purpose............................................................................................................191
23.3 Examples .........................................................................................................191
23.4 Description .......................................................................................................191
23.5 Purpose............................................................................................................191
23.6 Examples .........................................................................................................191
24. Computerized Maintenance Systems .................................................................192
24.1 Description .......................................................................................................192
24.2 Purpose............................................................................................................192
24.3 Examples .........................................................................................................192
25. Autonomous Maintenance ..................................................................................192
25.1 Description .......................................................................................................192

25.2 Purpose............................................................................................................192
25.3 Examples .........................................................................................................192
26. Business Plan.......................................................................................................193
26.1 Description .......................................................................................................193
26.2 Purpose............................................................................................................193
27. Area Work Teams .................................................................................................193
27.1 Description .......................................................................................................193
27.2 Purpose............................................................................................................193
28. Process Maintenance Teams ..............................................................................194
28.1 Description .......................................................................................................194
28.2 Purpose............................................................................................................194
28.3 Examples .........................................................................................................194
28.4 Description .......................................................................................................195
28.5 Purpose............................................................................................................195
28.6 Examples .........................................................................................................195
28.7 Examples .........................................................................................................195
29. Integrated Process / Maintenance System.........................................................195
29.1 Description .......................................................................................................195
29.2 Purpose............................................................................................................195
29.3 Examples .........................................................................................................195
30. Glossary Of Terms ...............................................................................................196

1. MAINTENANCE REQUIREMENTS PYRAMID
Maintenance Requirements Pyramid

World Class
Maintenance
Integrated
Area Process Risk based
Process/
Work Maintenance Maintenance
Maintenance
Teams Teams (RbM)
System
Shutdown
Computerized
Multi Cycle Time Autonomous Business
Maintenance
Skilling Compression Maintenance Plan
Systems
(CTC)
Predictive
B.O.M. Maintenance Planned MTBF Plant
FMEA/ Routines/
Parts Master Maintenance MTTR Master
RCM Cond. based
List Schedule Routines MTBCF Plan
Monitoring
Spares Standards Asset Weekly Maintenance Resource

Weekly
Policy & and History Maintenance Cost Skills Budget
Plan
Management Specifications System Report Report Matrix
Asset Short
Works Daily Maintenance
Numbering Critical Daily Maintenance Interval Production
Order Maintenance Cost
System Assets Plan KPIs Control Plan
System Report Structure
(HAC/PNS) (SIC)
Fully installed Process initiated To be addressed later Input Elements
Asset Numbering System (HAC/PNS)
1.1 Description
Unique asset numbering system describing :
All assets (to the lowest discrete maintainable level)
Its physical location
The equipment numbering system should be consistent for a plant and ideally, but not
necessarily, across the whole business
1.2 Purpose
Allow tracking of reliability, activity and costs against each item of maintainable asset
It is a requirement for basic history reporting
1.3 Examples
HAC, PNS Code

2. CRITICAL ASSET
2.1 Description
Failure of the asset (for more than n hours) interrupts production of the finished product
Failure of the asset may result in a failure to meet legislative, safety or environmental
requirements
Failure to repair the asset immediately will result in significant damage to that or another
item of equipment
No other back-up equipment is available
The equipment requires special or external attention
2.2 Purpose
To focus and prioritize maintenance effort for maximum gain
2.3 Examples
Kiln Girth Gear, Mill Drive, Cooler Grates
3. WORKS ORDER SYSTEM
3.1 Description
Information and control system providing :
- Instruction to perform a task
- The priority of the task
- Task description
- Feedback of what was done, lost time and parts used
3.2 Purpose
Controls and monitors maintenance activity
Provides an auditable trail for all jobs
Provides feed to other information systems
3.3 Examples
Mapcon / SAP / Marcam / etc. Works Order System

4. DAILY / WEEKLY PLAN
LEVELS OF MAINTENANCE PLANNING
4.1 Maintenance Master Schedule

Capacity and Resource Planning
Medium Term (13 weeks)
What-If Modeling
4.2 Weekly Plan

Weekly Scheduling
Task Prioritization
Planning of skills, resource and equipment requirements
4.3 Daily Plan

Allocation of tasks to individuals
Flexing of the plan on a daily basis
DAILY / WEEKLY PLAN

LEVELS OF MAINTENANCE PLANNING
Maintenance n Capacity and Resource Planning

Master n Medium Term (13 weeks)
Schedule n What-If Modeling
Weekly n Weekly Scheduling

Plan n Task Prioritization
n Planning of skills, resource and
equipment requirements
Daily n Allocation of tasks to individuals

Plan n Flexing of the plan on a daily basis

5. MAINTENANCE KPIS
5.1 Description
The key operational performance measurements which can be used to manage
Maintenance or an area within Maintenance
A KPI should have a base, plan and target
For a KPI to be useful it should be timely and capable of being influenced by the person
using it
5.2 Purpose
Allows performance to be measured and reviewed
Enables good fact based business decisions
5.3 Examples
Management Report, Daily/Weekly Operating Report (DOOR)
and Short Interval Controls
OEE, Availability, Performance, etc.
6. DAILY MAINTENANCE REPORT
6.1 Description
Timely reporting of plan attainment of maintenance activities and equipment performance
including cause of variation
6.2 Purpose
To allow structured review and to assign corrective actions to improve towards the
agreed target levels
6.3 Examples
SIC Log sheet

7. MAINTENANCE COST STRUCTURE
7.1 Description
Definition of the level of detail for cost reporting
7.2 Purpose
Systematic cost roll up to enable analysis and reporting on all required asset levels
7.3 Examples
MAINTENANCE COST STRUCTURE

l Description
Definition of the level of detail for cost reporting
l Purpose
Systematic cost roll up to enable analysis and reporting on all
required asset levels
l Examples
Current
Major
Maintenance
Company Cost Maintenance
Wages Cost
Cost Plant Salaries
Roll Cost Center Overheads
up Contract Labor
Group /Area Parts, Supplies
Other
Equipment
Cost Types

SHORT INTERVAL CONTROL (SIC)
7.4 Description
Regular monitoring and control of a process or activity
The frequency of SIC should reflect the span of control that an individual has to influence
the process or activity
7.5 Purpose
To identify problems early and prevent them of becoming
bigger ones
7.6 Examples
Maintenance activities
Operation performance [t/h]
Maintenance costs
SHORT INTERVAL CONTROL (SIC)
IG ONES
com ing B
be
l ems f rom
b
all pro
e v e nts s m
p r
n trol
g & co
r mo
nitorin
la
Regu
TIME
SP
C ONT AN OF
RO L

8. PRODUCTION PLAN
[Input Element to Maintenance System]
8.1 Description
A Weekly Plan indicating production requirements for the week should be broken down
into Daily Plans indicating target production levels
The Production Plan should be linked to the Maintenance Plan to identify the agreed
equipment availability for both maintenance activities and production needs
8.2 Purpose
To support the co-ordination of activities and requirements between Production and
Maintenance
Note: The production plan is not a development of MAC. However the production plan is
an important input to the maintenance system as described above. The maintenance system
as developed by MAC will feed the production plan with more accurate information and
therefor help to improve its content.

9. SPARES POLICY & MANAGEMENT
9.1 Description
Spares policy and management takes into account:
Spares criticality
Lead-Time of critical spares
Economic Order and Stocking Quantities
Parts Availability and Quality
Inter-plant parts sharing agreements
Systematical planning and control (reporting) of spare parts in order to maximize
availability and minimize cost
9.2 Purpose
To maximize critical equipment availability at minimum cost
A stocking policy is a pre-requisite for maintaining a spares management system
9.3 Examples
Decision whether or not to store a kiln tyre
10. STANDARDS & SPECIFICATIONS
10.1 Description
Standard short Description of a planned or routine maintenance activity like :
What needs to be done
How it should be performed
The optimum time to complete it
How many people are required
What skill or trade is required
What materials & tooling
Quality and safety requirements
Standard short description of failure cause
Standard short description of lost time causes
10.2 Purpose
To enable the identification of lost time and to provide clear and consistent instruction of
the best way to perform a task.
Clear identification of failure causes in order to make statistical analysis
10.3 Examples
Instructions for a routine inspection or tensioning of a belt

11. ASSET HISTORY SYSTEM
11.1 Description
A performance history record for each item of asset including:
Downtime and number of failure
Descriptions of major failures
Causes for those failures
Maintenance activities performed
Maintenance costs
11.2 Purpose
Allows the simple analysis showing basic history and performance of all assets.
Supports strategic or capital decisions.
11.3 Examples
Equipment performance log book
Work Order History
12. WEEKLY MAINTENANCE REPORT
12.1 Description
Timely reporting of maintenance KPIs to allow the review and analysis of maintenance
activity and equipment performance.
12.2 Purpose
To allow structured review and to assign corrective actions to improve towards the
agreed target levels
12.3 Examples
Maintenance KPI Report.
Management Report
13. MAINTENANCE COST REPORT
13.1 Description
A life-cycle cost analysis system to determine the true costs of operating and maintaining
an item of asset.
13.2 Purpose
To enable improved decision making including repair, replacement or re-engineer
decisions.
13.3 Examples
Cost Report out of machine history (top ten spendings)

14. RESOURCE SKILLS MATRIX
14.1 Description
Matrix identifying the skills needed for all people who perform maintenance activities.
14.2 Purpose
To identify the skills base, needs /gaps and training requirements for an individual or
group of people in order to optimize skills flexibility.
14.3 Examples
Human Resource Training records.
15. BUDGET
15.1 Description
The Budget should explicitly identify projected maintenance costs by period including:
Labor
Materials
Spare parts
Major Project
Contractors
The Budget should be linked to the planned KPIs levels.
15.2 Purpose
To identify and plan maintenance costs and to set operating targets.
15.3 Examples
Annual maintenance budget
16. BILL OF MATERIALS (BOM)
16.1 Description
Explosion of all of the parts and consumables, to the level of each purchasable item,
required to maintain an item of asset.
16.2 Purpose
To identify the parts required to perform all maintenance activities.
Maintenance Planning and Stock Management.
Maintainability improvement.
16.3 Examples
Asset - Spare Part identification/relation, (HAC-PNS)

17. MAINTENANCE MASTER SCHEDULE
17.1 Description
A long term plan (3 month) indicating all maintenance activities and the resources
required to complete them, considering :
Labor Availability
Labor Productivity
Planned / Predictive Maintenance Routines
Task Priorities
Materials and Spare Parts Availability
17.2 Purpose
To identify maintenance resource requirements and to optimize maintenance efforts.
17.3 Examples
13 week Master Schedule for the cement grinding area
17.4 Overview
MAINTENANCE MASTER SCHEDULE

OVERVIEW
Frequency Crew & Duration Labor Availability
Productivity /
Maintenance Maintenance Labor
Reactive -
Routines Standards Schedule
Allowance
Packing & Dispatch Contractor

Finish Mill Process Master
Raw Mill Mechanical Schedule
Quarry Electrical Summary
Kiln
13 Week Tradesmen
Maintenance Summary
Schedule
Output Output Output
n Preventive Maintenance Routines n Earned Standard Labor Hours n Re-schedule jobs.

n Planned Corrective Jobs (E.S.L.H.) by process area. n Cancel Jobs.
n Capital Improvement Projects n Actual Labor Hours inclusive of n Re-allocate jobs.
breakdown & productivity factor. n Increase / Reduce Overtime.
n Adjust Holiday / Training plans.
n Increase / Reduce Crewing.
n Modify Shift Patterns.

PLANNED MAINTENANCE ROUTINES (PMRs)
17.5 Description
Routine activities designed to minimize the risk of unplanned failures, including:
Routine overhauls
Fixed frequency replacement of parts or equipment
17.6 Purpose
To minimize unnecessary downtime and increase predictability by reducing the level of
unplanned maintenance.
17.7 Examples
Replacement of Cement Mill liners every 20000 hrs. of production.
18. FMEA / RCM
18.1 Description
Failure Mode Effect Analysis (FMEA) is an analytical tool to systematically establish the
failure mode and effect of a failure.
Reliability Centered Maintenance (RCM) is a process, utilizing FMEA, for determining
what maintenance, if any, should be performed in order to respond to the demands for :
Safe Operation
Environmental Protection
Production Quality
Plant Availability
18.2 Purpose
To proactively identify the optimum maintenance activity.

19. RCM APPROACH
A continuous process for determining the optimum preventive

maintenance plan for each item of plant in its operating context.
RCM APPROACH
CRITICAL EQUIPMENT
MONITORING & REVIEW FUNCTION
Select by greatest need /
Monitor performance of KPIs most benefit.
Define function and
against targets. performance level
IMPLEMENTATION FUNCTIONAL FAILURES

- Maintenance Tasks How the function may fail to
- Areas of re-design perform to the desired or
- Operating procedures. specified level of performance.
A continuous process for

FINAL ACTIONS
determining the optimum FAILURE MODES
- Modifications to equipment
- Changing operating procedures preventive maintenance How and reason-root cause.
- Other.
plan for each item of plant
in its operating context.
TASK FREQUENCIES FAILURE EFFECTS
Effects of failure, what needs to
Dependent upon failure be done, time required and total
characteristics of the equipment. lost-time.
PREVENTIVE ACTIONS CONSEQUENCES

Check action that can and should OF FAILURE
be taken to prevent failure : PROBABILITY OF
- Cleaning & lubrication Safety, Environmental,
- On condition tasks FAILURE
Operational.
- Fixed interval overhauls
- Fixed interval replacements Likelihood of failure occurring.
(Pareto of causes)

20. MTBF / MTTR / MTBCF
20.1 Description
Indicators to measure maintenance effectiveness
Mean Time Between Failure (MTBF)
Indication of the average time between failure for an item of asset.
Mean Time To Repair (MTTR)
Indication of the average downtime duration for an item of asset.
Mean Time Between Cause & Failure (MTBCF)
Indication of the MTBF by specific cause.
20.2 Purpose
To measure and focus on the correct alignment of maintenance activity.
20.3 Examples
Focus Maintenance activities in the Bottle Neck area.
20.4 MTBF (Hours)

Total Time Controlled - Duration of Breakdowns
________________________________________
Number of Breakdowns
20.5 MTTR (Hours)

Cumulative Downtime
_____________________
Number of Breakdowns
20.6 MTBCF (Hours)

Total Time Controlled - Duration of Breakdowns by Cause
_________________________________________________
Number of Breakdowns due to that cause

21. PREDICTIVE ROUTINES / CONDITION BASED MONITORING
21.1 Description
Predictive inspection routines, condition based monitoring and condition based
maintenance to asses the condition of equipment to predict failure and perform planned
maintenance activities.
21.2 Purpose
To minimize the level of intrusive maintenance.
To optimize the use of asset lifetime.
21.3 Examples
Oil analysis to determine change.
Vibration measurement to determine the optimal time to replace a rolling bearing.
22. PLANT MASTER PLAN

22.1 Description
A long term plan (1 to 3 year) indicating :
Production Requirements
Labor Availability
Training Plan
Major Maintenance Activities
Investments
22.2 Purpose
To identify planned major maintenance activities and investments in order to optimize
resource requirements (capital, people, training).
Note: The Plant Master Plan is not a development of MAC. However the Plant Master
Plan is an important input to the maintenance system as described above. The maintenance
system as developed by MAC will make a sound base for updating the Plant Master Plan.

23. MULTI-SKILLING
23.1 Description
Training of all operational employees in the core maintenance skills either between
Process Operators and Maintenance Technicians or between different trades.
Note : Multi-Skilling does not mean that everyone should be expected to do everything.
23.2 Purpose
To maximize labor utilization.
Basic requirement for Autonomous Maintenance
23.3 Examples
Quarry truck drivers doing their own oil changes.
Production doing daily inspection.
SHUTDOWN CYCLE TIME COMPRESSION (CTC)
23.4 Description
Approach for planning, controlling and reviewing activities during a shutdown period in
order to minimize downtime:
Critical path planning to identify and manage parallel or critical activities
Value Added (VA) / Non-Value Added (NVA) analysis to identify and remove non-
essential activities.
23.5 Purpose
To minimize equipment downtime by identifying activities which can be performed
outside of a shutdown or in parallel
To minimize equipment downtime by improving maintainability.
23.6 Examples
Kiln shut down
Mill shut down

24. COMPUTERIZED MAINTENANCE SYSTEMS
24.1 Description
Integrated, comprehensive maintenance management system linked to all of the other
relevant business systems including purchasing, stock control, engineering & finance.
24.2 Purpose
To automate the development and management of maintenance information.
Note : It is only appropriate to fully automate the maintenance system once it has been
developed, tested and utilized in a live application.
24.3 Examples
Mapcon / SAP / Marcam / JDE / etc.
25. AUTONOMOUS MAINTENANCE
25.1 Description
High frequency maintenance tasks which can be performed routinely by the immediate
operator, outside the control of the planning system. Often utilizing check sheets or
where the operator does not need to be told to do a task.
Dependent upon behavioral change and true ownership of the process.
25.2 Purpose
To devolve and simplify maintenance tasks.
25.3 Examples
Simple lubrication, cleaning to identify contamination, gauge marking.

26. BUSINESS PLAN
26.1 Description
A long term plan (5 year) indicating the company's strategies and activities, and the
resources required to complete them, considering :
Market Development
Business Focus
Labor Availability
Training Plan
Investments
KPIs Targets
Mission
26.2 Purpose
To manage direction of business.
To identify operation resource requirements and to optimize operation efforts.
Note: The Business Plan is not a development of MAC. However the Business Plan is an
important input to the maintenance system as described above. The maintenance system as
developed by MAC will make a sound base for updating the Business Plan.
27. AREA WORK TEAMS
27.1 Description
Individuals from the Maintenance and Process functions aligned to an area responsible
for the operation, maintenance and improvement of all assets in that area, supported by
specialists (assessor) when required.
27.2 Purpose
To create ownership of the process and to ensure continuous improvements.

28. PROCESS MAINTENANCE TEAMS
28.1 Description
A Multi-Functional Team replacing discrete or Cross-Functional Maintenance and
Process Operator teams.
Everyone is a Maintainer Operator.
28.2 Purpose
To minimize maintenance cost whilst maximizing flexibility.
28.3 Examples
PROCESS MAINTENANCE TEAMS
l Description
A Multi-Functional Team replacing discrete or Cross-Functional
Maintenance and Process Operator teams.
Everyone is a Maintainer Operator.
l Purpose
To minimize maintenance cost whilst maximizing flexibility.
l Examples
Process
Maintenance Process Maintenance
Team

RISK BASED MAINTENANCE (RBM)
28.4 Description
Extension of the RCM process to determine the optimum maintenance approach based
upon cost and reliability requirements within a changing environment.
28.5 Purpose
To achieve the optimum operational cost for a business.
28.6 Examples
When there are no sales and the silo is full, why incur the cost of fixing an item of asset?
28.7 Examples
Risk Profile
29. INTEGRATED PROCESS / MAINTENANCE SYSTEM
29.1 Description
Systems developed and utilized for process control used to take maintenance decisions.
These would include expert systems, rate loss and downtime accounting systems.
29.2 Purpose
To return rate to optimum or to predict the deterioration of plant condition in order to
identify maintenance requirements.
29.3 Examples
Computerized maintenance system linked with a fully developed TIS

30. GLOSSARY OF TERMS
MAC MAintenance Cement
KPI Key Performance Indicator
ESLH Earned Standard Labor Hours
W/Os Works Orders
BOM Bill of Materials
CTC Cycle Time Compression
SIC Short Interval Control
PM Preventive Maintenance
PDM Predictive Maintenance
CBM Condition Based Monitoring
FMEA Failure Mode Effect Analysis
RCM Reliability Centered Maintenance
OEE Overall Equipment Efficiency
MTBF Mean Time Between Failure
MTTR Mean Time To Repair
MTBCF Mean Time Between Cause and Failure
RbM Risk based Maintenance
HAC Holcim Asset Code
PNS Part Numbering System
TIS Technical Information System
DWOR Daily / Weekly Operating Report
PMR Planned Maintenance Routines

Condition Monitoring program
TABLE OF CONTENTS
1. Introduction...................................................................................................................202
2. Objectives .....................................................................................................................204
3. What is Condition Monitoring......................................................................................204
3.1 Potential failures and the P-F-Curve ......................................................................206
3.1.1 P-F Curves and inspection interval timing ......................................................207
4. Implementation of a CM program................................................................................209
4.1 Phase 1- Current Situation Assessment ................................................................209
4.2 Phase 2- Gap analysis and definition.....................................................................210
4.3 Phase 3- Implementation .......................................................................................213
4.4 Phase 4- Commissioning .......................................................................................215
4.5 Phase 5 - Sustainability review and handover .......................................................216
5. Training .........................................................................................................................216
6. Tracking Metrics ...........................................................................................................217
7. Program Performance Control ....................................................................................218
8 Condition Monitoring Technics...................................................................................218
8.1 Visual Observations, Listening, Touching ..............................................................218
8.2 Machinery vibration analysis ..................................................................................219
8.2.1 Measuring points ............................................................................................221
8.2.2 Vibration severity ............................................................................................223
8.2.3 Analysis frequency .........................................................................................224
8.2.4 Advantages / limitations..................................................................................225
9 Bearing vibration analysis ...........................................................................................225
9.1. Shock Pulse Method ..............................................................................................225
9.2 Limitations/advantages...........................................................................................226
10. Motor Current Signature Analysis ......................................................................226
10.1. Analysis frequency .............................................................................................227
10.2 Advantages/limitations ...........................................................................................228
11. Oil Analysis...........................................................................................................228
11.1. Oil sampling procedure ......................................................................................228
11.2 Analysis frequency .................................................................................................229
11.4 Contamination Analysis..........................................................................................231

11.5 Wear Particle Analysis ...........................................................................................233
12. Performance Trending .........................................................................................233
13. Thermographic Analysis .....................................................................................234
14. Kiln ovality and creep ..........................................................................................238
15. Elimination of failures..........................................................................................238
15.1 Pareto Analysis ......................................................................................................238
15.2 Failure Modes Effects & (Criticality) Analysis (FME(C)A) ......................................239
15.2.1 Benefits...........................................................................................................239
15.3 Cause and Effect Diagrams (Fish Bone or Ishikawa Diagrams) ............................239
15.4 Fault Tree Analysis (FTA) ......................................................................................240
16. References ............................................................................................................242
Annex 1: Summary of Typical Monitoring Techniques
Annex 2: Choosing the best spot for Vibration measurement
Annex 3: Holcim Definitions of the Measurement Points
Annex 4: Holcim proposal of "Scope of deliverables"
Annex 5: Training - Level definitions

CONDITION MONITORING PROGRAM
HOLCIM LTD
Handbook / Working Document

Maintenance Department
ondition onitoring
l
visu a
ph y io
r e Thermo
g ra ud
hie
P M
atu
ap
lis ts R
a
e r Fibr C h e ck
gr
e Lo
rro
e ll
s
op
In
sm
p
ue
fra
Fe
Te m
rq
Re
tion
To
ibra tou c
m ic
d
AVc h
le
na ou
ic
y st n
ic s io
rt
rro
D
Pa o
r, C tion
ea etra
W ye Pen
a t i o n MD
g ra d ag
ne
e
En
tic
d Pa
t r ic a l
a ly s i s
e rtic
os
le
c
ac
e f i c i e nc i
op
a l D es S o nic
y
rf
rn U ltra-
e
Su
t p e r t ie s X-
le c
l An
In ro Ra
P Ha y
E
rd n
Meggres ess
ica
De f or
nce m
is ta a
m
Res
he
ti
o
C
n
en t
Sh e
lltes
s
t
Ca
ak
em
Le
lib
rat
n
Pressure i
oni
g
Al
1 2 0

Lase
s
Cement Manufacturing Course Version 2005 r Volume 5 - Page 199

2 4 0

0
3 6 0

ire
dW
Lea
CONDITION MONITORING PROGRAM
for the Cement Industry
The purpose of this manual is to document in details the whole process leading to the
successful implementation, development and usage of the Condition Monitoring program
recommended by Holcim Group Support - Maintenance Department
Solomon Baumgartner
Holcim Group Support Ltd
CTS - Maintenance Department
Im Schachen
CH-5113 Holderbank
Phone +41 58 858 52 34
Fax +41 58 858 52 30
April 2004

Executive Summary
In today's business environment ever-higher requirements are placed with regard to the
reliability, availability and economical performance of plants and equipment. Loss of
production due to machine damages, in particular if these occur unexpectedly, diminish the
economic success of an enterprise and must be prevented. In recent years the emphasis in
maintenance has shifted to using emerging technology for measuring machinery condition
and predicting maintenance requirements. Key words like preventive, predictive or condition-
based maintenance describe this tendency. If plant people can always make an accurate
statement about the condition of their machines and plant assets, then they are in the
position to plan and introduce appropriate maintenance measures promptly, before larger
and, serious damages and associated unplanned machine downtimes occur. The bottom
line benefits of such an approach, however, can also consist of gaining experiences and
learn relevant mechanism or correlations to better control the own production processes.
The key component in this approach is thus the strive to achieve an accurate failure
prediction, and therefore an effective method of monitoring the status of an item or system is
essential. Clearly the ideal technique is one in which the condition of the equipment is
known at all times and which would provide an accurate prediction of any potential failure on
demand. Attempting to fulfill this requirement to a greater or lesser degree are the
techniques embodied within C ondition M onitoring (CM).
Well, most plants are using one or the other CM technology in order to predict their plant
condition. However, what makes a successful Condition Monitoring Program? Successful
programs implemented nowadays have something in common; they are approached in a
systematical manner and have focused efforts on the following issues.
Ease of Use (minimum learning curve, implemented at any level, quick results)
Management Supported (dedicated human resources, dedicated capital resources,
trusting in the information)
Integrating Maintenance Technologies (best in class, open database, the right
information to make the call)
Support (training (program advancement), results assurance, high end analysis)The
purpose of this report is to provide a detailed outline of Holcim Group Support's
understanding on how to start and execute a Condition Monitoring program at any Holcim
cement plant. The proposal will explore on the theoretical aspect necessary to understand if
dealing with condition monitoring and the general concepts for the implementation that were
developed during the two pilot projects at Portland (Holcim US) and Untervaz (Holcim
Schweiz) in co-operation with Rockwell Automation / Integrated Condition Monitoring, USA,
as well as Prftechnik, Germany. The document suggests a clearly outlined five-phases
process, which intends to assure that the Condition Monitoring Program reaches its
maximum effectiveness.
The five-phase process consists of:
1. Initial assessment
2. Gap definition and project outline
3. Implementation
4. Commissioning
5. Compliance Review

It is designed to assume sustainability and longterm benefits, especially through
incorporated training and periodic checks.
31. INTRODUCTION
High machine availability and low maintenance costs are important conditions for the
economical operations of production plants. However, the number of obstructing problems in
cement production has drastically increased in the past.
Common problems are:
Series of major breakdowns of critical equipment

Aging equipment
Preventive/ predictive measures vary from plant to plant and may be questioned
Missing systematic approaches to risk reduction
High turn over of maintenance personnel
Many maintenance organisations face an excessive level of reactive or breakdown

maintenance. This is expensive in terms of both, maintenance cost and downtime
consequences. Excessive reactive maintenance also contributes to a negative
performance spiral that at best counters other valuable improvement initiatives and,
in the worst case, leads to an almost completely reactive environment.
Standards Drop Backlog grows
Morale Declines PMR is missed
Head / Budget Band-aid

Reduction Maintenance
Resources taken More repeat

by Breakdowns work
More preventable
failures

Figure #1; The Vicious Circle of Reactive Maintenance (Turner, 2001)
A challenge for many companies is to move out of a reactive maintenance cycle.

Intuitively, the answer lies mostly in the ability of an organization to obtain more labor
resources or gain better access to the assets for maintenance. Most commonly, the
option of increasing labor resources is not available and asset maintenance periods
are constrained by factors external to the maintenance department. The only option
therefore is to increase the productivity or effectiveness of existing personnel and
make better use of the time allocated to maintenance. In other words an organization
has to get hold of highly qualified personnel and state of the art technologies.
The truism that "unless things are done differently, the past will continue" has never
had a greater place than in organisations caught in a troublesome maintenance spiral.
In recent years the emphasis in maintenance has shifted to using emerging technology for
measuring machinery condition and predicting maintenance requirement.
4th Generation
3rd Generation Design for
reliability
Preventive
Maintenance
2nd Generation Maintenance as
Routines
a source of
Relationship benefit
between failure Multi-tasking
1st Generation probability and and
Teamwork Risk Analysis
age
Cause / effect Commitment of
Break down Planned
analysis all departments
corrective
Maintenance
maintenance
Involvement of Maintenance
Fix it when it other process
breaks System for
departments
planning and Learning
(TPM)
controlling organization
Figure #2; Evolution of Maintenance in Industry
Key words like preventive or condition-based maintenance describe these tendency. If plant
people can always make an accurate statement about the condition of their machines and
plant assets, then they are in the position to plan and introduce appropriate maintenance
measures promptly, before larger damage, serious damages and associated unplanned
machine downtimes occurr. Thus this will break the Vicious Circle of Reactive Maintenance.
This report is expanding on the general concepts for the implementation of a condition
monitoring program that was developed during the two pilot projects at Portland (Holcim US)

and Untervaz (Holcim Schweiz) in co-operation with Rockwell Automation / Integrated
Condition Monitoring, USA, as well as Prftechnik, Germany.
32. OBJECTIVES
The objective of the condition monitoring program is to obtain real, measurable and practical
benefits in the performance of monitored equipment along with reliability thereof at lowest
sustainable costs. Further to that, it is also aims to set a base for comparable and
measurable efforts supporting Holcim drive towards performance excellence.
In other words:
Optimize the plant overall availability
Cost effective maintenance by targeting identified problems
Reduction of catastrophical failures
Root cause analysis and weak point elimination
Having specialized knowledge and understanding the characteristics of cement machines

are key requirements to achieve this objectives.
33. WHAT IS CONDITION MONITORING
Maintenance
Due to the randomness
of many failure modes,
planned maintenance
(based on time elapsed,
Planned Breakdown
Maintenance Maintenance machine running time,
machine cycles, etc.)
will have limited effect
Preventive Planned on improving the
Maintenance Corrective performance machinery.
Routine Work
Often it can actually
make it worse. In many
Condition
situations it can prove
Service Repair Monitoring Predictive cost effective to
Analysis
measure the condition
of the equipment and
Daily plan maintenance as
Visual Vibration Oil Wear Thermo. On-Line
inspections Monitoring Sampling Measurement Measurement
Inspections
(Checklist) cond. monit. required. This is known
as condition based
maintenance (CBM).
Figure # 3
This approach to preventive maintenance attempts to overcome the "crystal ball" data
analysis/prediction requirement inherent in scheduled maintenance. It does so by seeking to
determine the actual condition of a system or component at any point in time. The
establishment of the item's correct status, or condition at the required time, is obviously

critical to the viability of the approach and therefore various monitoring techniques are used
within CBM.
Formally condition-based maintenance is defined as (BS 3811:1984):
The preventive maintenance initiated as a result of knowledge of the condition of an
item from routine or continuous monitoring.
Or more informally as: Components are periodically inspected by a manual or an automatic
systems in order that their condition may be assessed (to identify their degradation rates.) A
decision is then taken regarding replacement and this is based upon an analysis of the
monitored data (Seddon, 1984).
This concept, of undertaking maintenance only when it is required and based on the actual
condition of a system or component is obviously very attractive. Some of the perceived
advantages are (Davies, 1990):
An early knowledge of impending failure, through the use of trend analysis
techniques and/or qualitative inspection. Preventive maintenance put in hand as a
consequence of this information reduces the system/component failure rate and therefore
increases equipment reliability.
Actual condition knowledge and failure prediction. This allows sufficient flexibility for the
plant to be shutdown for repairs at a point in time convenient to the manufacturing schedule
thereby improving system availability and minimizing random failure.
Diagnostic potential. This is present in a condition-based maintenance approach and can
pinpoint impending component failure together with the cause of that failure. The use of
diagnostic routines also permits system modification, to remove identified failure modes and
thereby improve inherent design reliability. In addition, through the rapid identification of an
impending failure, system maintainability is improved together with workload planning and
spare part stockholding.
Flexibility. If used in conjunction with and adaptive control system, the monitoring aspect of
condition-based maintenance can extend the useful life of plant and equipment, by limiting
the effect of a dynamic failure mechanism within the system. This reduces the likelihood of
early failure and thus improves system reliability.
Hence the ability to forecast failure at a time sufficiently in advance of the event for action to
be taken is the fundamental premise upon which the approach of condition-based
maintenance is built. The key component in this approach is thus the need to achieve an
accurate failure prediction, and therefore an effective method of monitoring the status of an
item or system is essential. Clearly the ideal technique would be one in which the condition
of the equipment would be known at all times and which would provide an accurate
prediction of any potential failure on demand. Attempting to fulfil this requirement to a greater
or lesser degree are the techniques embodied within condition monitoring (CM).
Formally, condition monitoring is defined as follows (BS 3811:1984):
The continuous or periodic measurement and interpretation of data to indicate the
condition of an item to determine the need for maintenance.
The standard also notes that condition monitoring is normally carried out with the item in
operation, in an operable state or removed but not subjects to major stripdown. More
colloquially, condition monitoring has been described as the assessment of the current
condition of plant and equipment by the use of techniques which can range from
sophisticated computer-driven instrumentation to human sensing, in order to predict failure
and to economically perform maintenance only when a potential failure is identified and at a
time convenient to the production schedule (Davies, 1990).

Consequently, condition monitoring is the means by which condition-based maintenance can
be carried out, utilising as it does the various techniques available to monitor and identify
impending failure. The use of condition monitoring thus ensures that all necessary
maintenance actions can be undertaken at a time appropriate to the predicted failure, and
therefore allows the maximisation of equipment availability through the timely and systematic
organisation of maintenance. In essence therefore the implementation of condition
monitoring can be described as, the performance of periodic or continuous comparative
measurement on parameters which are suspected of reflecting the condition of a
component, sub-assembly or system with the object that on analysis, the measurements
may indicate the item's current condition and the future trend of its possible deterioration
(Davies, 1990).
Thus through its contribution towards the planning of future maintenance action, condition
monitoring ensures that all decisions are made on substantive and corroborated diagnostic
information, thereby providing a basis for cost-effective and logical decision making
(Collacot, 1980). The integration of condition monitoring techniques within the wider
approach of CBM can yield many advantages.
3.1 Potential failures and the P-F-Curve

Condition Monitoring rely on the fact that failures do not occur instantaneously, but actually
develop over a period of time. At the initial point of failure it is not possible to detect the signs
of an impending failure since the deviation to the normal condition is too small. As the failure
develops, there comes a point at which it can be measured. At this stage the failure has not
yet reached a stage in which the equipment is unable to operate effectively. The
maintenance department has now the opportunity to correct the fault, before it becomes
more serious. The point in the failing process at which it is possible to detect that the failure
is occurring or is about to occur is known as a potential failure (potential because a serious
failure has not yet occurred). If nothing is done this condition will further deteriorate to the
point where the equipment fails.
Examples include:
hot spots on bearing housings and in electrical panels.
vibrations indicating imminent bearing failure
particles in gearbox oil showing imminent gear failure
visible leaks and wear.
Show the optimal point of intervention
Point
where
failure
starts to Changes in Wear
occur vibration particle in
oil
Audible
noise
Condition
Heat
generation
Functional
failure

Figure #4: The P-F curve
The condition being measured can take a variety of forms. Any condition that shows a
change, as the "health" of the spare equipment deteriorates, can be used. the warning must
however be early enough to allow actions to be taken., otherwise it is not possible to use
CM.
Equipment condition being measured could include:

reducing pressure supplied by a pump indicating impeller wear, slip ring wear, etc.
increased surface temperature on the outside of an insulating surface indicating
insulation deteriorating
increased vibration from rotating equipment; these vibrations must be analysed further to
identify possible causes.
3.1.1 P-F Curves and inspection interval timing

The life of a component and the P-F curve are often confused. On-condition task frequencies
are often based on the real or imagined "life" of the item. If it exists at all, this life is usually
many times greater than the P-F interval, so the task achieves little or nothing. The
component life is measured forward from the moment it enters service. On the other hand
the P-F interval is measured back from the functional failure to some point that a warning of
the forthcoming function failure can be detected (potential failure).
The two concepts are often unrelated.

Potential failure Inspections
detected at least done at
Failures occur on 2 m onths before 2 m onthly P-F Interval
a random basis functional failure intervals = 4 m onths
A ge (years) 6 m on ths 12 m onths 18 m onths
Figure #5:
In the above example (Figure 5), the batch of components has a random failure pattern with
two failures in the first year. Let us assume that over a large batch these components have
an average life of 3 years. From condition monitoring it has been observed that these
components have a P-F interval of about 4 months. Because they started to fail in the first
year of service the condition monitoring tasks must commence immediately after installation
on a 2 monthly basis. The timing of the inspections has nothing to do with the age or life of
the component.
P-F curves can have considerable variation in length from minutes to months. A P-F curve of
4 months is desired because:
fewer on-condition inspections are required

there is more time to organise the people and materials needed to correct the potential
failure
it is easier to plan to correct the potential failure without disrupting operations or other
maintenance activities
it is possible to do whatever is necessary to avoid the consequences of the failure in a
more considered and hence more controlled fashion.
As a guideline, the frequency of a condition monitoring task has to be half of the (shortest) P-
F-Interval of the failure. Therefore the frequency for the task depends mainly on two things:
the characteristic's in which a failure occurs
the deviation needed to detect a potential failure
The characteristics of the failure is predetermined mainly by design and equipment
operation. This fact does not give maintenance the possibility to act directly on them. The
only way to influence the frequency is to recognize a potential failure earlier in time. This
means the deviation from the normal condition will be less and therefore the method of
detecting the deviation has to be more sophisticated.
In the case of a detected abnormal condition (potential failure), the frequency can be
adjusted if the repair task will not be carried out immediately and the future development of
the failure is not known well (lack of experience, inconsistencies of the P-F-Interval). The
idea is to have the equipment, or part, remain in service for as long as possible.
However, here are also some limits to the usage of the P-F Curve with respect to the
application of condition monitoring.
The failure occurs without warning or too fast to undertake any action (P-F-Interval close
to zero); there is no condition monitoring task to prevent it from occurring.
The deviations are too small to be detected or if it is impossible to establish limits for the
condition to be monitored.
The P-F-Interval is so inconsistent, that no meaningful task interval can be established.
The ultimate limit is given by the cost for the task in comparison to the money saved.

34. IMPLEMENTATION OF A CM PROGRAM
Most of successful programs implemented nowadays have something in common; they are
approached in a systematical manner. The following are the essential steps, or phases, to
take in order to install an effective condition monitoring program. As shown in the chart
(Figure #6), the assignment, incorporating all basic condition monitoring techniques,
methods and corresponding workshops, has a proposed timeline of approx. 12 months.
However, it must be remarked that the last phase, Training, is an ongoing commitment, in
the sense that the CM technicians' abilities will improve with the acquisition of experience
and with the knowledge and usage of the latest technological breakthroughs.
Training
Phase 1 Phase 2 Phase 3 Phase 4 Phase 5

Situations- Definition / Implementation Compliance Commissioning Sustainability
Assessment Gap Analysis Review review &
Handover
Figure #6; the methodology
4.1 Phase 1- Current Situation Assessment

A strong focus has to be given on the initial assessment process. The main objective of this
phase is to establish the current needs of the plant. To achieve that, the first step calls for
getting a common understanding of the condition monitoring concept and benefits with the
involved personnel. It is of outmost importance that all parties fully agree on the scope of
the project.
Financial Assessment
The first stage in optimizing the performance of the installation is to conduct a study
leading to a pro forma assessment of the benefits If the pro forma indicates a return
on investment, the next stage is conducting a formal motivation study. In partnership
with plant personnel, the typical project team will carry out an in-depth study of the
organization to understand the motivation level for possible changes, which may
come up with the implementation of a condition monitoring program. This will be
achieved by asking specific and well-tested questions directed towards plant
personnel on all levels.
Improvement Opportunity Assessment
Improvement opportunities are to be derived from the answers of the study and the
questionnaire. In addition, formulas such as the KPI's installed trough MAC will be used to
help focus the assessment. The formulas outlined below provide a quantitative analysis to
reinforce the qualitative analysis that is performed using the questionnaire/interview process.
Preventive Maintenance Routine to Corrective Maintenance ratio - the number of

scheduled PMR work to the number of unscheduled corrective maintenance.
Over Time ratio this is the ratio of the number of regular hours worked to the total
number of hours worked

Emergency Work Order ratio this is the ratio of the number of regularly scheduled work
orders to the total number of work orders generated. War schon oben
MTBF and number of repetitive jobs (downtime Pareto)
The purpose is to include measures, which provide indication of the effectiveness of the
current maintenance program, and facilitates benchmarking of similar facilities. More
importantly, it provides indication of the effectiveness of a current condition monitoring
program or the need for implementing one.
Skill Sets of Plant Maintenance Personnel
During the interview process the technical skill sets of available work force will be assessed.
Performed in conjunction with the plant staff responsible for maintenance this will allow to
identify what additional training or resources would be required to both begin and operate an
effective CM program. A matrix is developed using a spreadsheet to score the level of
maintenance that is being performed relative to industry standards for similar equipment. A
quantifiable improvement potential is thus determined.
Example of a specific technology skills matrix:
Level I Level II Level III Equipment Available

Basics Advanced Professional
Vibration
Technology
Ultrasonic
Testing
Oil Analysis
Motor Current
Analysis
And more
4.2 Phase 2- Gap analysis and definition

The initial assessment was performed in order to form a basis for a qualitative analysis of
improvement opportunities. The questionnaire provided the statistics necessary to quantify
the improvement opportunities factor. Based upon these results, key area for improvement
opportunities were identified. It is now necessary to investigate the maintenance practices
performed on the equipment.
Equipment review
Two main equipment will be selected to perform an evaluation of the level of maintenance
being performed on them. In addition to this a system visual inspection is executed in an
effort to identify the strengths and weakness of operations and maintenance. Task includes
an on-site physical assessment of the equipments, to determine if they are in a sound
enough condition before being integrated within the program. This qualitative analysis can
include items such as spare parts and material storage, plant and system design with
respect to maintainability, overall plant health, etc. Also the plant equipment criticality
analysis will be consulted to get in-depth knowledge. However, should the plant not already
have performed their own equipment criticality assessment, also this would be executed
based on best experience available. Other to this, the existing data is being reviewed and
proposals elaborated.

It is now time to find the best strategy, in order to gain the full benefits of the project.
According to criticality rating, Pareto and other analysis, the machines to be included in the
condition monitoring program will be chosen. In a cement plant this normally implies the
clinker cooler(s), as well as kiln(s) and mills (vertical or ball ones), along with their
corresponding feeding and dedusting systems. However, this will be decided on an
Process flow diagram

Determine Critical Units
Personnel
Assess Availability
Maintenance History
Reliability Database
Activities / History / Modes of Failure
Failure Analysis
Select optimum technique to
Match each mode of failure CM Techniques
Design-out Condition-based Time-Based On failure Opportunity
Assemble a monitoring and activity

plan for the maintenance of each unit Production Windows
Maintenance Resources
Assemble a monitoring and activity
plan for a Plant
individual basis.
Figure #7; Assessment process
Once the equipment has been selected, the measuring and inspection points will be
determined. On this stage, the efforts will concentrate on walk-by inspections, vibration
monitoring, online temperature tracking and oil sampling analysis. These techniques
represent the core of the condition monitoring, are widely known and easy to install and
develop (in most of the cases, it is expected that cement plants are already acquainted with
them). Along the project, continuously other techniques and methods will be implemented
and included to the routines (thermography, motor circuit evaluation, etc). Of course, this
approach may vary if an assessment of the existing routines proves the need for
implementing further complementary techniques in an earlier stage. On the other hand,
systems where a simple walk around inspection would suffice the operational and
maintenance requirements will not be further elaborated. An important part of the Holcim
condition monitoring approach is the cost/benefit ratio, which means choosing the proper
technique according to the need and not always the most expensive one.
Figure #8 shows the
strategy to be pursued,
Critical
Surveillance
$$$
Systems in terms of the cost of
the potential failure and
Impact of Failure
On line the probable warning

Walk-
Walk- Around time (see also P-F
Essential
Data Collectors diagram). For example,

$$
and Analyzers on line monitoring is
Off line recommended when the
Walk-
Walk- Around machine under
but frequent
Inspection routines surveillance is a vital

Low
$
Walk by
Gradual Degradation Sudden Failure
Time to Failure
piece, whose malfunction would prove critical, not to say catastrophic, for the production. At
the same time, this failure would appear in short time notice, demanding a constant
monitoring. However, such a method entails a significant investment and for this reason, it
has to be carefully pondered if it is the best profitable solution.
Cost Justification/ROI
Depending upon the amount and complexity of the equipment to be installed, the capital
outlays for a complete condition monitoring program amount approximately in the order of
the magnitude of costs, which would result from an three-days loss of production. Now, one
of the most frequently asked questions by companies who are investigating the
establishment or enhancement of a CM program, and one of the most difficult questions to
answer is What is the expected ROI? The question is difficult to answer because there are
multiple metrics that are used to measure the success of a program. Furthermore, investing
in hardware, software and manpower does not guarantee that a program will necessarily
provide a positive return. A program will only be as successful as the understanding and
commitment that is given by all involved parties.
The only accurate way to determine cost justification is to review historical performance and
cost data. Then that data should be compared to industry benchmarks. Once a discrepancy
is found, a reasonable estimate of improvements potential can be calculated. There is no
canned formula for cost justification because each plant is performing differently. Once the
opportunity is identified, specific actions must take place to realize the improvement.
For example:
Incorporate aquired knowledge of equipment condition into the shoutdown planning
Increase productivitiy of workforce through intenitfied training needs
Identified redundant equipment must be removed from the maintenancee program
Holcim Group Support Ltd has developed first templates according to the HARP
structure for tracking, and comparison reasons. Thus the team has to work closely
together to determine on a plant-by-plant basis the costs, goals and methodologies
for calculating returns for identified measures.
Measures that may yield the greatest return for Holcim cement plants would be:
Net plant availability improvement

MTBF increase
Costs of maintenance avoidance
Unplanned shutdowns reduction
Planned shutdowns optimization
Reduction / elimination of impact of major failures

There is no question that investment in a CM program will yield a net positive result
as others in the cement industry have proved this to be the case. For example,
Adelaide Brighton (Australia) has published the following returns, which justified their
program.
15% (US$5M) improvement in plant availability

Avoid 6-10 unplanned kiln shutdowns per year
Reduced the cost of outside maintenance consultants by US$30,000 per year
Saved $125,000 per year by reducing time-base maintenance schedules
Reduced spare parts inventory resulting in savings of approximately $130,000 per year
Also Rohrbach Zement, plant Dotternhausen Germany, presented considerable impacts of

their condition monitoring initiative. They stated having reduced the maintenance cost by 7.8
million DM over a period of 9 years.
Perceived key success factors were as following:
Installation of a reliability improvement program (Conditon Monitoring)

Consequent applying of root cause analysis
Usage of high quality material as of the beginning
Conversion of the new knowledge acquirements
Highly qualified, motivated and dedicated people
Reference and agreement report

Phase 2 will come to its end when the software/hardware and training requirements are
determined. At this moment a reference report will be established, which includes project
targets, expected costs and savings, as well as a jointly developed implementation plan
outline, a roadmap to move forward. This report will serve as orientation for the future project
controlling.
4.3 Phase 3- Implementation

This endeavor will demand a strong involvement of the selected supplier for
instrumentation, in terms of the assistance needed to setup both hardware and
software. Holcim Group Support will closely cooperate with the plant people to
implement the plan that was agreed upon in phase two, adhering to the original
scope, schedule and budget, and being open to improvements along the way.
Establish the bounds: so far, the different variables to be measured have already
been clearly defined. However, their bounds need to be established. To accomplish
this, existing norms should be consulted (ISO 2372, DIN 45665), as well as other
sources. There is no sense in monitoring, if the obtained values can not be
compared to a functional standard, which will allow to rate the machines' condition.

Set up data collecting route: according to the physical layout of the plant, the routes
for off line monitoring (walk around) have to be optimally designed in terms of
distance between the machines and the measuring sequence. A good knowledge of
the plant is essential to avoid unnecessary back and forth trips during the data
collection.
Installation of sensors: this includes online sensors as well as the identification of the
off line points. For the first, all essential modifications or allowances have to be
performed (drilling on reducers' housings, bearings' housings, etc), along with the
proper wiring. For the second task, the method may vary, depending of the system
chosen. There are some so sophisticated, where the data collector recognizes
automatically the point and rules out the chance of a mistaken reading, to the
simplest one of marking the point with painting. The important thing is that the
measuring points be clearly identified, so that the readings can be always performed
on the very same spot.
Interface to CMMS and TIS: the software must be correctly configured, so that it
might be fully able to share information with the already existing CMMS and TIS
systems. The data from the on line sensors has to be also integrated. It should be
remembered that this software must fulfill the role of a Decision Platform.
Set up software package: this installation should be tailored to the plant needs, to
become an user friendly tool. One of the most important aspects is the proper
configuration of the Database. All of the information gathered in the previous phase
must be properly organized: equipment clearly identified, besides their measurement
points set up with the predefined variables (vibration, bearing condition). It will also
be essential that the alarm levels system might allow for the different operating
conditions, because this constitutes the "thermometer", that will easily allow to
identify any troublesome machine.
Other points to settle are the already defined collecting data routes and the design of
the initial reporting system. The information included in the reports has to be clear
and concise, to allow taking proper decisions afterwards (change the frequency of
the measurements, write a notification for a planned shutdown of the machine, etc.).
Trail run on the system: once set up, it must be ascertained if the system is
completely functional and if there is need to amend any pending point. That is, if the
data from the online sensors is reliable, if all the collecting data routes have been
correctly set up, if the machine's measuring points have been properly identified
Another concern would be if the involved personnel has access to the databases and
knows how to carry out their assigned tasks. The training received by the technicians
who will perform the actual data collection has also to be tested, in case of any
shortcomings. This is valid as well for the oil analysis: If the personnel has been
properly trained to gather the oil samples, if the schedule for these tasks has been
programmed in the system.
Baseline signature analysis: a preliminary step to the commissioning phase foresees
the taking of baseline measurements, in order to establish the current condition of the
machine or process. The results will then be compared to expected performance.
(Actual performance will vary from machine to machine within like makes and sizes).
Expected performance criteria can be obtained from the vendor, an industry standard
(ISO 2372, DIN 45665), or the historical performance history of the machine at the
location. It is strongly recommended to consult as well the HGRS document
'Vibration severity assessment criteria for gear reducers, girth gears and fans", by
Benno Saxer (it will shortly be available on Holspace, ICM and in the Maintenance
Community). All suspected variances to accepted bounds will be investigated, faults

diagnosed and problems corrected before including the machine or process in the
program. Another benefit from this measure is the possibility to compare the ensuing
measurements against this baseline, which will help to point out where the deviations
from the normal condition are appearing. A guideline will be thus obtained to
facilitate the analysis.
Up to this stage, the sequence of events, and how each part fit into the whole, can be
pictured as shown in Figure #9.
FME(C)A, Pareto, Establish the

existing MAC need
analysis
Select
Equipment to
audit
Taking into
consideration Select possible
the differed audit method
operating and devices
conditions
Initial Cost- Suppliers
Benefit analysis information
Decide on
monitoring level
Periodic
Spare parts Walk by Inspec. Off-line On-line
condition
measurement
Install the
system
Data recording
Establish the
bounds
Data Collection
Machine baseline
measurement
Condition
Fault correction Trend analysis
analysis
Figure #9: Siquence of implementation siquenci
4.4 Phase 4- Commissioning

After setting up the whole system, data must be fed to it. Through periodic monitoring, the
data will be collected (online, off line, walk by); when there is enough information, a trend
can be established. Its behavior will then be evaluated, to ascertain if there is a fault, which
demands corrective action. The modification of the machine baseline cannot be completely
ruled out, whereas a change of the operating conditions may demand it.
Only through sustained use a fine-tuning of the system will be achieved. The modifications
can be of several types - data collection routines setup, alarm levels, monitoring frequencies
- just to mention a couple examples. A good communication within the Project Team (Client
and HGRS) will be essential to carry out these adjustments; the whole program will be
reviewed with this team. Any concerns and questions from the plant personnel will be given
due consideration.
On the other hand, the reporting procedure will be also scrutinized, to determine whether it is
actually fulfilling the expectations. By this moment, the quality and quantity of the information
contained can be accurately judged. The modifications that may be relevant will be done.
The tracking metrics defined should be registered by now. These data must be carefully
stored and continuously updated, in order to start trending the results. A thorough analysis
will be undertaken, in order to find the root cause of any negative deviations and to decide

upon the best course of action to eliminate them. If the deviations are of a positive nature,
the reasons behind them will also be targeted, in order to register them as good practices,
which could be implemented in future projects.
4.5 Phase 5 - Sustainability review and handover

The condition monitoring program is a dynamic process, which seeks for permanent
adoption to internal (equipment, frequencies of measurement) as well as to external
(monitoring technology) changes. Through phases 1 to 4, the CM process has
established a framework of rational and cost effective maintenance. In phase 5 the CM
program is consolidated and the project team will mutually establish the results
achieved and handover the program to the local person in charge, performing a final
sustainability review. It is the intention to create an organisation that constantly seeks
to improve its methods by continued appraisal of every task it undertakes and every
unplanned failure that occurs. This requires good problem solving skills along a
process of integration and iteration of technology to be in place.
5. TRAINING
Although it has been depicted just now, the training is an ongoing process, from the very
beginning of the project. From the understanding of the CM concept to the correct
interpretation of the vibration spectra along the knowledge of how to analyse root cause and
fix the problem right, the personnel must be always kept well-informed of the latest
technological and methodological developments in this field. This training can be
accomplished through courses and on the job practices.
A series of maintenance training workshops is proposed, which would be appropriated as a

means of increasing awareness and stimulating action among the attendees. Everyone in
the plant needs to be aware of what is tried to be accomplished, and how it is going about to
accomplish it. These workshops are designed to keep people informed and integrated. The
workshops should be open to the general plant personnel, business and maintenance
leaders, however, also others in the company.
Six major themes of the workshop session are proposed:

1. Overall Project Introduction What are we doing? Why are we doing it
2. What is Condition Monitoring?
3. Vibration Analysis Level I and II
a. Introduction and basics of Condition Monitoring
b. Basics of single transformation and analysis (FFT, DFT..)
c. Registration of measured values
d. Error recognition on machine elements (Frequency analysis, Trouble
Shooting Chart)
e. Set up of intervals, routines and routes
f. Case studies and practical exercises
4. Lube Oil courses, Levels I and II
a. Oil Analysis Level I and II
5. Root cause failure analysis - Cause and Effect Diagrams, Fault Tree Analysis

6. Vibration analysis of drive systems such as planetary reducers, girth gear and pinion
drive systems
6. TRACKING METRICS
In order to justify such investments, the program must clearly show its benefits. This proof of
success should be furnished over tracking metrics such as MTBF, MTTR, Availability and
Maintenance cost avoidance. Whenever possible, benchmarks comparable within similar
Holcim Plants should be set.
The actual target values to be reached will be determined according to each plant situation
and individual assessment. It is understood that not only the condition monitoring program
does and can affect these costs. There are other activities, which have just a large effect as
the CM program (e.g. planning, annual shutdown, technical competence etc.).
Below are additional metrics listed, which are recommended to be tracked:
Percentage of PMRs completed
Percentage of PMRs deferred (in hours)
Track number of hours associated with PMRs deferred based on CM direction.
Track the cost avoidance for the deferred PMRs
Note: The deferred PMRs should result in an improvement in equipment availability as well
as a reduction in maintenance costs. Since the maintenance spending cannot be controlled
and the deferred costs may be applied to other maintenance needs, the tracking of cost
avoidance is recommended for this program. Assigning a unique project number to work
orders generated throughout this program will collect this data. Standardised downtime costs
will be used to calculate cost avoidance. If this cost avoidance results in an overall
maintenance cost reduction, fine, if not, at least the program effectiveness can be tracked by
this number.
A review is needed of the current PMR procedures to see which ones can be impacted by
the CM activities. Those procedures will be a target for the deferral program. It would not be
the intention to impact PMRs that do not affect equipment availability or operating
conditions.
Percentage of Corrective Maintenances completed
Number of unplanned work hours
Plan Attainment
OEE

7. PROGRAM PERFORMANCE CONTROL
Experience has shown that condition monitoring programs often fail because of
decrease in motivation, missing system fine tuning, insufficient training, incorrect set
points setting and wrong expectations. Therefore, regular meetings and interim
measurements (after six month a HGRS compliance review) will take place during the
project to assure the continued development and to determine compliance with the
original targets.
After evaluating the general condition of the plant, studying the existing databases and
interviewing the responsible personnel, a precise idea of the overall performance will be
obtained. The main aspects to be examined are:
Updated database: data collected as per schedule and repairs or other interventions in
the machines easily traceable.
Clear alarm levels and easy identification of the troublesome spots.
Distil, validated and make available generated knowledge and successes.
Number of equipment failures due to variables already monitored.
Tracking metrics and deviation from their target figures.
Knowledge and commitment of the involved personnel.
The main benefit from this procedure resides in the compliance to the Holcim guideline and
therefore being comparable to other programs within the Holcim Group. The tasks to be
carried out to improve and sustain the program are established and converted in a
comprehensive action log..
8 CONDITION MONITORING TECHNICS

How to monitor equipment or component best requires knowledge which measurements
to take and where and how to take them. Commonly measured operating parameters
include:
8.1 Visual Observations, Listening, Touching

Visual observation, listening and touching are perhaps the oldest and most common
predictive maintenance techniques used in industry today. They are certainly the least
expensive and they can be surprisingly effective on virtually any machine found in an
industrial plant. For these reasons, it is important that any discussion of machine condition
monitoring techniques include these basic human senses.

Visual Inspections
Belt Conveyor (unsatisfactory) Belt Conveyor (good)
In many cases human observation helps to identify a problem that was undetected by
other predictive techniques or maintenance inspections. This can include looseness,
visibly worn or broken parts, oil leaks, chattering gears or hot bearing housings.
Typical applications include less critical plant machines that are not included in a
formal predictive data collection route. The value of observations is not limited to
unmonitored equipment. However, since sensory information may also be extremely
valuable as a supplement to predictive data and analysis. It is recommended that both
operations and maintenance personnel are e trained as observers, since that will
provide the most complete and knowledgeable coverage of plant machinery.
8.2 Machinery vibration analysis

Vibration analysis can be used to determine the operating condition of rotating or
reciproking equipment, This can include deteriorating or defective bearings,
mechanical looseness and worn or broken gears. Vibration analysis can also detect
misalignment or unbalance before these conditions result in bearing or shaft
deterioration. Evaluation of long term vibration analysis trends can identify poor
maintenance practices such as improper bearing installation and replacement,
inaccurate shaft alignment or imprecise rotor balancing.
These techniques are based on the fact that all rotating machinery produces
vibrations that are a function of the alignment and balance of the rotating parts.
Measuring the intensity of vibration at specific frequencies can provide valuable
information about the preciseness of shaft alignment and balance, the condition of
bearings or gears and the effect on the machine of resonance from housings, piping
and other structures. It is an effective, non-intrusive method to monitor machine
condition during start-up, shutdown or in a normal operation.
Vibration analysis is used primarily on rotating equipment such as: steam and gas
turbines, pumps, motors, compressors, paper machines, rolling mills, machine tools
and gearboxes. Recent advances in the technology now allow limited analysis of
reciprocating equipment such as large diesel engines and reciprocating compressors.

Because different types of problems often occur at different frequencies it is very useful to
analyse measured vibration signals with respect to their frequencies. The Fast Fourier
Transformation (FFT) can mathematically breakdown the overall vibration signal and arrange
it according to frequency. FFT analysis can be very useful when analysing the causes of
vibration. E.g. Vibration associated with a bent shaft or impeller problem will occur with the
frequency of rotation of the shaft. Problems associated with the roller bearings will occur at
higher frequencies that are a function of the number of rollers in a bearing.
A vibration analysis system usually consists of five basic parts; signal pickup(s), a
signal recording device, a signal analyzer, analysis software and a computer for data
analysis and storage. These basic parts can be configured as a continuous on-line
system, a periodic analysis system using portable measurement and diagnostic
equipment or a multiplexed system that samples a series of points every few minutes.
Hard wired and multiplexed systems are more expensive per measurement point. The
determination of which configuration is more practical and economical will depend on
the critical nature of the equipment and the value of continuous or semi-continuous
measurement data for that particular application.

8.2.1 Measuring points
With the selection of the measuring points it is important to know how the machines are built
as well as to know where,
which signals can be Foundation: Torsional system: Motor shaft
18 Hz, 82 Hz, 100 Hz 17 Hz, 31 Hz, 80 Hz 16 Hz
measured. An overview is 109 Hz, 117 Hz 102 Hz, 114 Hz
depicted in the following
diagram
Pinion mesh Shaft bending Pinion shaft 1 st stage: 590 Hz

67 Hz system: 2.0 Hz 2nd stage: 210 Hz
108 Hz, 253 Hz,
Figure #10a: Sources of the
oscillation and noise
components

Figure #10b: Sources of the oscillation and noise components
B i
Misalignment
Field error
Transmis
sion error
A selection of common detectable failure modes are:

Imbalance: a peak at shaft speed.
Misalignment: typically 1x, 2x & 3x shaft speed
Looseness: often at 1x or 2x shaft speed
Bearing Damage: higher frequency peaks typically between 2 KHz and 5 KHz
depending on shaft speed and transducer resonance.
Electrical Problems: synchronous frequency and side bands.
Gear Damage: gear mesh frequency depending on shaft speed and number of
teeth and side bands.
Oil Whirl: approximately half shaft speed.
Blade Damage: number of blades by shaft speed.
Cracked Shaft: typically 2x, 3x shaft speed.
To receive the desired information it is of utmost importance to place signal pickup(s) at the
right place and in the right direction. Detailed discussion on this can be found at ANNEX 2.

8.2.2 Vibration severity
It is common engineering practice that the vibration severity of industrial gear drives are
assessed in accordance with the ISO 2372 standard that is more or less identical with DIN
45665 or the VDI 2056 guide-line.
However, the standards and guidelines take into account the operating condition of electric
motors, turbines and generators but not gear drives, therefore it should be used somewhat
careful and flexible for the rotating machinery in the cement industry, especially since some
of the frequencies are so low that they cannot be detected with normal techniques. For open
girth gear drives the above mentioned guidelines are definitely not applicable.
The vibration severity guide-lines, as recommended in the following for fans, gear reducers
and open girth gear drives are based on the ISO 2372, for open girth gear drives the severity
guideline is based on own experience and digress from the standards.
Vibration severity for gear reducers and fans
Large machines, operating Large machines, operating speed

speed below the natural above the natural frequency of the
frequency of the foundation foundation
Good < 2.8 mm/s RMS < 1.8 mm/s RMS
Allowable 2.8 7.0 mm/s RMS 1.8 4.5 mm/s RMS
Just tolerable 7.0 18 mm/s RMS 4.5 11.0 mm/s RMS
Not permissible > 18.0 mm/s RMS > 11.0 mm/s RMS
It is recommended in the relevant standards that for fans the residual unbalance is also
assessed, the balance quality grade be outlined in the standard ISO 1940.
Guidance values for the permissible residual unbalance

Figure #11
Quality grade G 6.3

is generally used for
big fans in the
cement industry

Vibration severity for girth gear drives
Pinion with friction bearings Pinion with roller bearings

Very good < 5 mm/s RMS < 2.5 mm/s RMS
Fair 5 10 mm/s RMS 2.5 5 mm/s RMS
Usable 10 15 mm/s RMS 5 10 mm/s RMS
Just tolerable 15 - 20 mm/s RMS 10 - 15 mm/s RMS
Not permissible > 20 mm/s RMS > 15 mm/s RMS
In case that the vibration severity in a gear drive system was found to be not permissible
this should be the reason for more detailed investigations. Possible causes are mechanical
defects (forced vibrations) or resonance vicinities (free vibrations) that requires an extended
investigation procedure by experts including analytical studies, field measurements and an
analysis in the time and frequency domain.
8.2.3 Analysis frequency

Condition Based Maintenance utilising vibration analysis involves setting up "tours"
(a list of condition based maintenance tasks to be carried out). Spot checking
(monitoring a measurement point once a year or less) is used primarily when
maintenance or operations personnel detect unusual noises or vibrations and want to
determine if a serious problem actually exists. If a problem is detected, additional
spectral analysis can be made to define the problem and estimate how long the
machine can continue to operate before a serious failure occurs.
Spot checking can be cost effective for less critical equipment of the plant,
particularly when budgets or manpower are limited. Its effectiveness relies heavily, ,
on someone detecting unusual noises or vibration; a practice which may not be
reliable on large or complex machines or in noisy parts of a plant. Also, by the time
the problem is serious enough to cause noticeable noise or vibration, the amount of
deterioration or damage may already be considerable.
Another application for spot checking is as an acceptance test to verify that a

machine repair has been done properly. This analysis can verify proper bearing or
gear installation and alignment or balancing to the required tolerances. Additional
information can be obtained by monitoring machinery on a periodic basis, for
example, once a month or quarterlyr. Periodic analysis and trending can provide a
more subtle indication of bearing or gear wear, allowing personnel to project machine
condition into the foreseeable future. This advance notice means that equipment can
be repaired during normal machine shutdowns, rather than after a machine failure has
caused unscheduled downtime.
Long term vibration analysis can identify improper maintenance or repair practices.
These can include improper bearing installation and replacement, inaccurate shaft
alignment or imprecise rotor balancing. Since more than 80 % of common rotating
equipment problems are related to misalignment and unbalance, vibration analysis
can be important in reducing or eliminating recurring machine problems.

Long term trends can also identify improper production practices, such as operating
equipment beyond design specifications (i.e., at higher temperatures, speeds or
loads.) The trends can be used to compare similar equipment from different
manufacturers to determine if design benefits or flaws that are being reflected in
increased or decreased service life.
Ultimately, long term analysis can be used as part of an overall program to

significantly improve equipment reliability. This can include more precise alignment
and balancing, higher quality installations and repairs and continuously lowering the
average vibration level of equipment in the plant.
8.2.4 Advantages / limitations

Though the costs have been reduced and the ease of use improved significantly over
the past five years, vibration analysis is still one of the more complex and expensive
predictive techniques. The complexity stems in large part from the relatively
subjective nature of interpreting vibration spectra and the difficulty in setting effective
alarm limits for a wide variety of rotating machinery configurations.
The relatively high cost per measurement point is a result of the need for
sophisticated electronic instruments to collect, analyze and store the data, the cost of
personnel to collect the data and the cost of personnel and training to interpret the
data. The technique is also limited in application to rotating machinery, though there
is recent experience with reciprocating equipment such as diesel engines and
compressors.
For those companies that are willing to make a commitment of manpower and
resources, the payback can be considerable. Some companies report being able to
accurately identify specific gears within a gearbox that are failing, substantially
reducing the amount of downtime required for troubleshooting and repair. Others
have been able to identify and solve complex resonance problems that were causing
damage to shafts, bearings and couplings.
In spite of the higher cost and complexity, the investment in vibration analysis equipment
and manpower is often paid back within the two years. For companies with limited budgets,
there is a variety of service companies that will perform vibration analysis on a contract
basis.
9 BEARING VIBRATION ANALYSIS
9.1. Shock Pulse Method

The shock pulse method detects the mechanical shock waves caused by the impact
of two masses. Used on anti-friction bearings this is a measurement of the
mechanical impacts caused during rotation by contact between the surfaces of the
ball or roller and the raceway. The magnitude of these pulses are dependent upon the
surface condition and the peripheral velocity of the bearing (rpm and size). As bearing
condition deteriorates from good to imminent failure, shock pulse measurements
taken on the bearing housing can increase up to 1000 times.

The shock pulse measurement instrument uses a piezoelectric accelerometer to
measure the mechanical impact of shock pulses without being significantly
influenced by factors such as background vibration and noise. The transducer is
tuned mechanically and electrically to a resonant frequency of 32 kHz. The shock
pulse caused by a mechanical impact sets up a dampened oscillation in the
transducer at its resonant frequency. The peak amplitude of this oscillation is directly
proportional to the impact velocity.
Compared with other vibrations, shock pulses from anti-friction bearings are very
weak signals that hardly affect the general vibration spectrum of a machine. This
means that even strong shock pulses (which indicate significant bearing
deterioration) may not be easily differentiated by conventional vibration analysis from
other sources of machine vibration. If detected by vibration analysis, shock pulses
may be difficult to analyze or interpret accurately. As a result, the shock pulse method
can be used to identify subtle changes in bearing condition or lubrication, prior to
substantial bearing deterioration or failure.
9.2 Limitations/advantages
The shock pulse method is moderately complex and expensive predictive
techniques. The complexity stems in large part from the need to have accurate
bearing size and speed information available prior to taking a measurement. The
expense is moderate due to the cost of the handheld instrument(s), the cost of
training and the need for data archiving. The technique is also limited to rolling
element bearings. The advantages of the technique are that the instrument is portable,
easy to use and it can be used on virtually any rolling element bearing. This makes
the shock pulse method particularly effective on equipment that historically had
bearing or lubrication problems or on less critical plant equipment. The technique
can also be used as a troubleshooting tool prior to planned equipment shutdown or
repair; to identify, prioritize, plan and schedule important bearing maintenance
activities.
10. MOTOR CURRENT SIGNATURE ANALYSIS

Motor current signature analysis (MCSA) provides a nonintrusive method for detecting
mechanical and electrical problems in motor driven rotating equipment. The system is the
development of Oak Ridge National Laboratory, as part of a study on the effects of aging
and service degradation of nuclear power plant components. The basis for MCSA is the
recognition that an electric motor driving a mechanical load acts as an efficient, continuously
available transducer (the motor can be either AC or DC.) The motor senses mechanical load
variations and converts them into electric current variations that are transmitted along the
motor power cables. These current variations, though very small in relation to the average
current drawn by the motor, can be monitored and recorded at a convenient location away
from the operating equipment. Analysis of these variations can provide an indication of
machine condition, which may be trended over time to provide an early warning of machine
deterioration or process alteration.
Although MCSA was developed for the specific task of determining the effects of aging and
service wear on motor operated valves used in nuclear power plants, it has found application
within a much broader range of plant machinery. Motor current signals can be obtained
remotely, typically at a motor control center which may be located several hundred feet from
the equipment being monitored. The signals are obtained nonintrusively with a single split-
jaw current probe placed on one of the power leads. Because no electrical connections need

to be made or broken, the hazard of electrical shock is minimized. The resulting raw signal is
amplified, filtered and further processed to obtain a measurement of the instantaneous load
variations within the drivetrain and ultimate load.
A comparison of motor current and mechanical vibration signatures obtained simultaneously

from a motor operated valve provide similarities and distinctions. Both spectra contain
frequency peaks corresponding to the motor speed, worm gear tooth meshing, and its
harmonics; though the amplitude relationships are different.
One distinction is that there is a strong spectral component in the motor current signature
that is defined as the slip frequency. This signal is a general characteristic of AC induction
motors and reflects the rate at which the spinning armature continually falls behind the
rotating electrical field generated by the motor's field windings (no such peak appear on DC
motor signatures.) Since this motor slip frequency component is electrical rather than
mechanical in origin, it has no vibration counterpart and it is not present in the vibration
spectrum.
Tests on motor operated valves indicate MCSA is capable of detecting and tracking the
progress of: stem packing degradation, incorrect torque switch settings and/or varying switch
trip points, degraded stem or gear case lubrication, gear tooth wear, restricted valve stem
travel, obstructions in the valve seat area, and disengagement of the motor pinion gear.
10.1. Analysis frequency

Spot checking a point less frequently than once a year is of significant value only after a
plant has developed its own baseline of application data.Once historical files have been
developed, spot checking can be cost effective for critical or expensive machines, as well as
less critical or balance of plant equipment. It could also be used effectively as a
troubleshooting tool.
Information can also be obtained by monitoring machinery once a month or n quarterly.

Periodic motor current signature analysis can provide a more subtle indication of bearing,
packing, coupling or gear wear; allowing personnel to project acceptable machine
performance into the future. This advance notice of developing problems means that they
can be repaired during normal machine shutdowns, rather than allowing a serious machine
failure to cause unscheduled down time. Since problems are detected when they are
relatively minor, they are usually less expensive to repair.
Long term motor current signature analysis can be used to identify improper maintenance or
repair practices. These can include improper seal/packing installation, improper bearing or
gear installation, inaccurate shaft alignment or imprecise rotor balancing. This information
can be of particular importance in reducing recurring machine problems.
Long term trends can also be used to identify improper operating conditions, such as
continually running equipment beyond design specifications (i.e. at higher temperatures,
speeds, or loads.) They can be used to compare similar equipment from different
manufacturers to determine if there ear any inherent design flaws of benefits that can be
reflected in increased service life.

10.2 Advantages/limitations
Motor current signature analysis is one of the moderately complex but expensive predictive
techniques. The complexity stems in large part from the relatively subjective nature of
interpreting the spectra, and the limited number of industry-wide historical or comparative
spectra available for specific applications.
In the last several years the technique has been simplified by several vibration data
collector/analyzer manufacturers. This has improved the technique from a data collection
and analysis standpoint, and significantly increased the amount of practical field experience.
The technique is particularly useful due to its non-intrusive nature. Measurements can be
taken without the need to make or break electrical connections, and shut down or open up
machinery. This eliminates equipment downtime for inspection, and improves personnel
safety. In addition, since readings can be taken remotely, the technique can be more
conveniently and safely performed on large, high speed, or otherwise hazardous machines.
For those companies willing to make a commitment of resources into this developing
technique, the payback appears as attractive as for vibration analysis. And, for companies
with limited budgets, there are several service companies that will perform motor current
signature analysis on a contract basis.
11. OIL ANALYSIS

Oil analysis can be used to identify the condition of fluids and lubricants, and
determine if they are suitable for continued use or should be changed. It can also
indicate the condition of internal, oil-wetted components, identifying excessive wear.
Finally, oil analysis can be used to identify the presence of contamination which can
lead to premature failure.
For the most accurate results, samples should be taken from an active, low pressure
line, ahead of any filtration devices. For consistent results and accurate trending,
samples should be taken from the same place in the system each time (using a
permanently installed sample valve is highly recommended.)
11.1. Oil sampling procedure

Strictly avoid any pollution from the surroundings to enter the sample
Seal the bottle right after the sample has been taken
Provide the bottle with following data:
Company name
Sample reference number
Date and time of sample
System reference e.g. HAC
Fluid type, ISO VG and supplier
Sampling location
Service life of oil.

Often the oil samples are analyzed by the supplier, sometimes it is good to change to an
independent body. As the analysis is an evaluation of a rather costly product, it can be
recommended to asks the supplier for analysis data for a new oil as well as the rejection limits.
Fore more details regarding extraction of fluid samples from lines of operating systems see ISO
4021.
Most independent labs supply sample containers, labels and mailing cartons. If the oil
analysis is to be done by a lab, all that is required is to take the sample, fill in
information, the machine number, machine type and sample date; and sent it to the
lab. Results are normally available within 24 hours of receipt of the sample. If the
analysis is to be done on-site, analytical equipment must be purchased, installed, and
standardized. Sample containers must be purchased, and a sample information form
created and printed. Once a sample is taken, information about it is entered on a form,
it is taken to the on-site lab, and the results are available a soon as the analysis can
be scheduled. However, for starting a oil analysis program and also to minimize costs,
it is recommended to have an on site tool kit which is basicly alowing to measure
viscosity, wear paricle, water content.
Oil condition is most easily determined by measuring viscosity, acid number and
base number. Additional tests can determine the presence and/or effectiveness of oil
additives such as antiwear additives, antioxidants, corrosion inhibitors and anti-foam
agents. Component wear can be determined by measuring the amount of wear metals
such as iron, copper chromium, aluminium, lead, tin, and nickel. Increases in specific
wear metals can mean a particular part is wearing, or wear is taking place in a
particular part of the machine. Contamination is determined by measuring water
Recommended Analysis
Gear Oil Hydraulic Oil
R - Infrared Spectrum IR Infrared Spectrum
Appearance / Color Appearance / Color
Viscosity cSt Viscosity cSt
TAN Total Acid Number TAN Total Acid Number
Water Content Water Content
Wear Particle (weight %) Wear Particle (weight %)
Trace metal analysis (ppm and size) - optional Foam
Trace metal analysis (ppm and size) - optional
Particle counting - ISO 4406 (optional)
content, specific gravity, and the level of silicon. Often changes in specific gravity
mean that the fluid or lubricant has been contaminated with another type of oil of fuel.
The presence of silicon (usually from sand) is an indication of contamination from
dirt.
11.2 Analysis frequency

Spot checking (sampling a system once a year or less) is used primarily to determine
whether the fluid or lubricant should be changed, or to confirm if a suspected problem
actually exists. For example, if a machine is experiencing noticeable vibration or
noise, an oil sample may be taken to confirm if there is bearing damage or excessive
wear.

Tips and tricks
The lubricant supplier or For failure investigation: in addition to oil analysis, cutting open the
equipment manufacturer filter element to examine trapped debris can lead to further
should provide the plant with information.
the percent change in
The industry has no guideline for new oil cleanliness. Pre-filtering of
lubricant viscosity or acidity new oil before reintroduction to the system (10 - 25 m filter rating)
that can be tolerated before is recommended to reduce contamination.
requiring an oil change (i.e.
+25% or -10% change in Once you are familiar with the test methods being used and has
viscosity, +2.00 mgKOH/gm sufficient historical data, it is best to start setting your own limits
since this limits can be set to reflect the real life conditions and
change in acidity.) Once that
reliability goals in place.
value is reached, the oil
should be changes as soon as Laboratory analysis do not guarantee repeatable results between
possible. Continuing to different laboratories. One reason is that the exact same test
operate equipment after the method may not have been used by the two labs. Also suppliers of
guidelines have been test dust do not guarantee the number of particles of a given size in
their samples. Therefore, carefully selection of a proper laboratory
exceeded can cause
is required.
excessive component wear
and failure.
Spot checking can also detect high levels of contaminants such as water or dirt,
which will require an immediate oil change to prevent excessive wear. If unusually
high wear metal levels are detected, additional oil samples should be taken and the
results trended over a short period of time to determine if excessive wear is
occurring.
Sampling machinery on a periodic basis (once a month or once a quarter) can provide
a more subtle indication of lubricant or machine deterioration, or the slow
introduction of contamination. Most bearing or gear failures occur after their
condition has deteriorated slowly and steadily for a period of moths or even years.
Contamination may be introduced when oil is added to the system, and periodic
monitoring will indicate this. Early warning of contamination means that repairs can
be planned for a scheduled shutdown when they can be made more efficiently.
Long term monitoring of oil condition over six or eight sample periods can identify
improper maintenance or repair practices. These can include the failure to properly
flush out a system after repairs, improper fluid or lubricant handling procedures,
which introduce water or dirt contamination, or improper filter handling or
replacement techniques.
Unusually rapid oil degradation can indicate that the oil is not suitable for the
equipment or application. For example, a rust and oxidation inhibited oil, rather than a
straight mineral oil, may be required where there is the possibility of high
temperatures or water contamination. Rapid oil degradation may also indicate that the
equipment is being operated beyond its original design capacity, creating excessive
temperatures or bearing/gear surface loading.

For details in specification, please consult the training material "Oil analysis " in the
Maintenance Community at Holspace.
Oil analysis is one of the simplest predictive techniques to use and certainly one of
the least expensive. Independent labs can help select machines and frequencies,
suggest which tests to run, supply sample bottles and mailers, interpret the results
and archive data. The maintenance departments of most companies have some
experience with oil analysis if only on a limited basis. In spite of its low cost and
simplicity, oil analysis can be an extremely effective technique; particularly when the
data is trended over and extended period of time (12-24 months). Trended data can
identify poor maintenance and operating practices which, if corrected, can result in
substantial maintenance and operating savings.
One company found that contamination levels increased significantly each time oil
was added to a gear reducer on a coal handling system, and the contamination
resulted in bearing and gear failures. On examination they found that removing the
cover plate to add oil allowed coal dust to fall into the sump. They installed a covered
oil reservoir and piped it to all the gear boxes. Now clean oil can be added by opening
a valve and the incidence of bearing and gear failures has been significantly reduced.
Oil analysis can be used only on equipment that contains a circulating oil system. In
most cased it can indicate that a problem exists (i.e. there is excessive wear)
however, it may not be able to identify the specific problem (e.g. what is causing the
wear and which of similar or identical parts are wearing.)
Oil analysis is only as good as the timeliness and consistency of the sample. The
longer a sample sits before it is shipped and analyzed, the less significant the data;
and the value of trended information diminishes quickly if samples are not taken from
the same place on the machine each time.
11.4 Contamination Analysis

Contamination analysis can be used to identify the presence of particals, water, or
gases in a fluid or lubricant; and determine if the oil is suitable for continued use or
should be changed. The presence of contamination can lead to excessive wear and
premature component failure, and in some cases it can represent a safety hazard. It
can also indicate the deterioration of seals or packings, or poor fluid or lubricant
storage and dispensing practices. It can be used on machines that have a circulating
oil system, including: steam or gas turbines, generators, hydraulic systems, diesel or
gasoline engines, gearboxes, paper mills, dryers, stamping presses, machine tools an
other metal-working equipment.
For the most accurate results, samples should be taken from an active, low pressure line,
ahead of any filtration device (s). For consistent results and accurate trending, samples
should be taken from the same place in the system each time (using a permanently installed
sample valve/port is highly recommended.) Most independent labs supply sample
containers, labels and mailing cartons. If the contamination analysis is to be done by a lab,
all that is required is to take the sample, fill in information about the machine number,

machine type, and sample date; and send it to the lab. Results are normally available within
24 hours of receipt of the sample at the lab.
If the analysis is to be done on-site, analytical equipment must be purchased, installed, and
standardized. Once a sample is taken, information about it is entered on a form, it is taken to
the on-site lab and the results are available as soon as the analysis can be scheduled.
If an in-line measurement device is used, usually a permanent sample port must be installed.
Analysis is performed by connecting the instrument to the sample port; and the analysis
results are often immediately available. Some instruments allow data to be collected from
several sample ports and downloaded into a PC for trending and further analysis.
Contamination analysis can be used to determine the number, size and composition of
particles; the amount of water or other liquid contaminants (such as fuel); and the presence
of gaseous contaminants.
Visual or electronic particle counters provide information on the number and size of particles.
The presence of large particles can indicate severe wear or goss contamination from an
outside source; both of which need immediate corrective action. The presence of large
quantities of smaller particles may be critical if there are close tolerance valves whose
surfaces can become eroded. Visual examination can also be helpful in determining the
composition of particles; for example, identifying small pieces of material from a severely
worn or damaged seal.
Determing the presence and amount of water can identify leaks in coolers, steam seals, or
condensation in high humidity areas. Water can significantly increase component wear, or
cause corrosion of internal metal surfaces; and its presence requires immediate corrective
action. Contamination such as fuel dilution should be corrected immediately. Measuring gas
contamination is usually important for gas compressors or electrical generators, where
contamination from hydrogen or other gases can be a safety hazard.
Purchasing and installing analytical equipment requires additional expense and manpower
(including training). This investment can often be paid for within twelve months by a
maintenance department that takes several hundred samples a year, or that monitors critical
/ expensive equipment. This analysis technique is limited to equipment that has a circulating
oil system. In most cases it can indicate that a problem exists (i.e. there is excessive
contamination) however, it may not be able to identify the specific source of the problem (i.e.
where the contamination is coming from.) The value of contaminant analysis is also limited
by the consistency and timeless of each sample. The value of trended information
diminished quickly if samples are not taken from the same place on the machine each time;
and the longer a sample "sits" before it is analyzed, the less significant the data will be.
These concerns can be overcome by using an on-site lab or an in-line diagnostic system
with a permanent sample port. For critical or expensive equipment, the added hardware and
training cost can be a cost-effective investment.

11.5 Wear Particle Analysis
Wear particle analysis can be
used to determine the type,
location and severity of
component wear occurring
within the lubrication or
hydraulic system of a machine.
What is wear causing This information can be used to
troubleshoot suspected machine
problems and project an estimated time to failure. The technique allows equipment
wear problems to be identified prior to serious deterioration or failure so that repairs
can be made on a planned basis during scheduled downtime.
The presence of large ferrous wear particles can indicate severe wear and the
presence of large quantities of small wear particles may be important if there are
close tolerance valves whose surfaces can become eroded. One analysis technique
calculates the ratio of small to large particles, noting that when this ratio changes
significantly, a severe wear process has begun. Another calculates the total weight of
wear particles present in the sample and trends that value as an indicator of the
severity of the wear process. A third technique measures the concentration of ferrous
wear particles, noting that an increase in concentration is often indicative of the start
of a severe wear process.
Visual examination can also be helpful. Under a microscope, wear particles formed
from different wear processes often have different shapes (i.e. cutting wear produces
long, curly particles similar to metal cuttings, while rubbing wear produces thin, plate
like particles.) Visual examination can also identify non-ferrous materials such as seal
or gasket material, which can be helpful in defining or locating the problem. One
technique allows samples to be heat treated, with different ferrous alloys
distinguishable by their different colors.
Wear particle analysis is a moderately sophisticated predictive technique and it is

fairly expensive. Independent labs can help to select machines and frequencies,
determine which wear particle tests to run, supply sample bottles and mailers,
interpret the results and archive data. This technique can be used only on equipment
that contains a circulating oil system. In most cases, it can provide an accurate
diagnosis of the specific source, severity and location of the problem. With data
trending, it can also provide a good indication of how long a machine will run until it
fails. The same limitations apply as for the contamination analysis.
12. PERFORMANCE TRENDING

Performance trending can be used on virtually any machine found in an industrial
plant. This can include critical machines with permanently installed instruments
which measure pressure, temperature, flow, rpm or electrical power consumption. It
can also include readings taken by portable instruments on less critical equipment of
plant

In many plants performance readings are already being taken by operations personnel
as part of a routine process or quality control program. Performance readings may
also be taken as part of a routine data collection program (vibration, thermographic or
ultrasonic), since many of the portable data collectors allow manual entry of process
statistics through a keypad. This data can then be downloaded into a PC and trended
concurrently with other predictive data, trends and information. The real value of
performance trend data is to confirm a problem identified by other predictive
techniques and provide further information on the location or seriousness of the
problem.

Spot checking (monitoring a machine once a year or less) is a very limited value
unless the reading is so extreme as to indicate significant performance deterioration.
In a limited manner, spot checking can also confirm the seriousness of a problem
already detected by other means.
Monitoring a machine at least once a month and trending the data can provide a more
subtle indication of gradual performance deterioration; for example, the slow
deterioration of a pump impeller. This information can be used to confirm the
conclusions made from other predictive measurements and provide and indication of
when performance will deteriorate to an unacceptable level.
Monitoring performance trends over longer periods of time can provide an indication
of improper maintenance or operation practices. For example, several cases of rapid
deterioration in pump or compressor performance may indicate poor maintenance or
overhaul procedures, poor quality replacement parts, excessive load or speed, or
poor equipment design.
There are two primary advantages of performance trending. The first is that the data is
usually readily available and relatively inexpensive to collect (in some cases the data
may already be part of a periodic collection process and in others it can be easily
added to another predictive data collection route). The second is that, unlike other
predictive techniques, performance trending monitors variables that often have a
direct relationship to production output. This means that performance trending can
help to indicate the direct impact of a problem on production.
This analysis technique is limited to equipment that has permanently installed

instrumentation or that can be monitored by portable instruments. In reality that is
often not much of a limitation but it may be unsuitable to take measurements on
hazardous or difficult to reach machines. Performance trending can also be of limited
use if the data has to be manually recorded and plotted since this can significantly
increase the time and cost of data analysis and the usefulness of the data when
comparing it to other predictive trends. Since many companies have already made a
commitment to other predictive techniques, it is usually very cost effective to expand
a program to include performance trending as well.
13. THERMOGRAPHIC ANALYSIS

*>34.6C Diff. 1 Thermographic analysis provides a high-
7.8
resolution, non-contact means of monitoring the
condition of electrical and electromechanical
30.0
equipment, roofing and wall insulation and kiln
25.0
refractories. Infrared scanners, similar in
20.0 appearance to video cameras, detect differences
15.0 in surface temperatures and highlight those
10.0 differences in B/W or color images that are
Bearing 2 Bearing 1 displayed on a television screen. These images
25.6
*<8.3C 33.4 can be photographed with conventional films or
recorded on videotape or picture files and the
Thermograph images, called thermograms, are used to analyze
patterns of heat gain or loss.
Thermographic analysis is an effective

predictive maintenance tool because
mechanical or electrical breakdowns are often
preceded or accompanied by changes in
operating temperatures. This information can
be particularly important in electrical
machinery where circuits and connections
may show no visible signs of deterioration
until moments before a complete failure.
Thermographic analysis can also detect
cracks or deterioration in roof or wall
Reference
insulation and kiln refractories, which can
increase heat loss or reduce the efficiency of production
processes. Infrared scanning is non-destructive and it can be performed at a distance
for machinery that is difficult or awkward to reach. Since surveys are best done while
the equipment is in operation, there is no need for machine downtime and lost
production. In addition, to helping avoid costly or even catastrophic equipment
failures, thermographic analysis can be used to help prioritize repairs prior to planned
maintenance, to evaluate completed repair work and to check new installations prior
to start-up.

Spot checking (monitoring a measurement point once each year or less) is used
primarily when maintenance or operations personnel suspect that a problem exists or
prior to planned maintenance to identify developing problems and prioritize the work.
One advantage of thermographic analysis of electrical equipment is that the scan

provides an immediate and quantifiable indication of the temperature difference
between a properly functioning circuit or connection and one that is overheating. This
means that the technician can often determine the seriousness of the electrical
problem without having a review historical patterns or past analysis. This quantitative
analysis can also be helpful in determining how long the equipment can continue to
operate before there is a catastrophic failure.
Information can be obtained by performing thermographic analysis on a periodic

basis, for example once a year or after each major overhaul or plant turn-around.

Periodic thermographic analysis can provide a more subtle indication of electrical
deterioration, bearing or coupling wear, roof deterioration or leakage and refractory
cracking. This can allow personnel to project acceptable performance for the coming
six months or one year. Advance notice of developing problems means that they can
be resolved or repaired during normal shutdowns, rather than allowing a catastrophic
failure to occur. Since problems are detected when they are relatively minor, they are
usually much less expensive to repair. Long-term thermographic analysis over a
period of several years can be used to identify improper maintenance or repair
practices. These can include improper roof installation or repair; improper bearing or
coupling installation or lubrication; poor quality electrical installation or repair and
improper refractory installation or repair. This information can be particularly useful in
reducing or eliminating recurring problems and improving personnel safety. Long
term trends can also be used to identify improper operating conditions, such as
operating equipment beyond design specifications (i.e. at higher temperatures,
speeds or loads). The trends can be used to compare similar equipment from different
manufacturers to determine if there are design benefits or flaws that can result in
increasing service life, improved operation or reduced maintenance costs.
Dirk V. Kreunen
Regional Mgr- Midwest Region
SKF Reliability Systems
In general, we tend to use thermography for 1) electrical system inspections, 2) facility

(roofing) inspections, 3) process monitoring. The electrical systems (MCC cabinets and
switch gear) can be covered on a bi-annual or even annual basis. Given the amounts of
electricity consumed
by a cement plant, as well as safety considerations, bi-annual is probably warranted. We
normally look at facility roofing on an annual basis. We have had good success identifying
whether roofing needs to be repaired/replaced and limiting the repairs to the areas of the
roof which require the work. This type of work, while a somewhat less known use of the
technology, has created savings well into six figures for a single customer. Finally, we
monitor process equipment where thermography can have an impact. Typically, this
involves a look at electric motors, especially those recently replaced (looking for bad
connections). We also look at couplings and PTO systems where vibration may not provide
quick or positive identification of alignment problems (jackshafts for example). There is often
application on the process itself, for example, we monitor the felt in paper machines for
deterioration
Thermographic analysis has found particular application in scanning large, distant or
hazardous surfaces for temperature variations. This includes roofs and walls, kiln or
high temperature refractory linings, overhead power lines or cables, high voltage
transformers or electrical connections.
There has been limited application in mechanical troubleshooting (for example,

identifying hot bearings or couplings). Plants usually have separate groups for
thermographic and mechanical troubleshooting and historically they have had limited
interaction.

Infrared scanners can only measure the temperature of visible, radiating surfaces (i.e.
they cannot take readings through glass or metal housings or covers unless the metal
covering is thin enough for secondary heat patterns to appear). Readings are also
affected by the reflective nature of the surface being measured and this reflectivity
value must be taken into account when scanning different objects or surface
materials.
The equipment cost is moderate to expensive, particularly when many of the more
sophisticated options are included. Infrared scanners vary in their accuracy,
sensitivity, temperature range, resolution and portability. Options include a variety of
graphic displays, video/audio capabilities, zoom, interchangeable lenses, battery
operation, datalogger/printer, image storage and digital outputs. Due to the variety of
industrial applications, the cost of infrared scanning equipment can often be
recovered in a year or less. Scanner operation is relatively easy; however,
interpretation of the scanned images requires a moderate amount of
training/experience. Training and certification can be obtained from either the scanner
manufacturer or one of several independent certification companies. For companies
with limited budgets, there are service companies that will perform thermographic
analysis on a contract basis.

14. KILN OVALITY AND CREEP
In order to prevent kiln shell plastic deformations in the tire
sections, the creep (or relative movement) must be
monitored, either manually or with an automatic device.
The shell ovality must also be determined by using the
"Shelltest" instrument. "Shelltest" records in the form polar
diagrams, taken close to the supporting planes reflect the
mechanical condition of the kiln tube. Since the elastic
deformations are proportional to the bending moments to
which the kiln shell is subject during one rotation, the peak-
to-peak stresses can also be estimated from the "Shelltest"
diagrams. This option is of particular interest if cracks in the kiln shell

Kiln Shelltest are found and the cause of them has to be analyzed. The radial
deformation behavior reflects quite well the mechanical "health" of a rotary kiln. Therefore
"Shelltest" measurements should be taken as a reference data source on newly installed
kilns, at least twice a year or before and after kiln shell repair and overhaul work and
whenever refractory lining problems occur. This measurement will also help to corroborate if
the creep values presently used are the most proper ones. For further details, it is
recommended to consult the "Shelltest" Manual.
There are other as importand inspection methods to be applied in the kiln area. Methodes
such as the lead wire test, girth gear inspections, etc. To have a better understanding of this
methodes, consult the kiln training papers in the maintenance community / Holspace.
15. ELIMINATION OF FAILURES
15.1 Pareto Analysis

For maintenance management purposes it may be desirable to graphically represent
historical maintenance data. This type of data can be represented on histograms, pie-
charts, etc. Another way of presenting data is to carry out a Pareto Analysis. This tool
can be applied to a wide range
of maintenance control
problems, for example, to
highlight the relative
importance of particular key
plant registers to the total
maintenance effort on a line.
The following histogram
shows the maintenance man-
hours recorded against units
numbers 1 to 30 in a 12 month
period.
Fi #12 M M H d d i t h it b
The Pareto analysis shows
that unit numbers 23, 1, 15,
23, 18, 14, 28, 6, 13, 17, 2 & 4
have incurred the majority of
the maintenance man-hours.

The maximum benefit would be gained if these 12 units were targeted to reduce their
maintenance requirement.
Fig re #13 Pareto Anal sis of Maintenance Man Ho rs recorded b nits

15.2 Failure Modes Effects
& (Criticality) Analysis (FME(C)A)
FMEA & FMECA are engineering analysis that, if performed properly, can be of great value
in assisting the decision making process of the engineer during the design stage or
maintenance review to analysis possible and existing modes of failures of a piece of
equipment. The analysis is often called a 'bottom up' analysis as it examines equipment at
the spare level, and considers the system failures that result from their different failure
modes. In an FMECA analysis a relative criticality is calculated for each failure mode.
15.2.1 Benefits
It provides designers and maintenance engineers with an understanding of the
structure of the system, and the factors, which influence reliability.
It helps to identify items that are reliability sensitive or of high risk, and so gives a
means of deciding priorities for corrective action.
It identifies where special effort is needed during manufacture, assembly or

maintenance.
It establishes if there are any operational constraints resulting from the design.
It gives assurance to management and/or customers that reliability is being or has

been properly addressed in the project.
A good FME(C) A will present in tabular form for each failure mode of all the components:
The effect on sub-assemble, assembly, sub-system etc. to system level;

Likelihood of occurrence;
Severity and criticality (for a criticality analysis).
The report may also include further information, such as recommendations for: manufacture
and assembly (inspection, test, quality of components etc.); maintenance (inspection, test,
replacement periods etc.); detectability (user, maintainer, etc.).
15.3 Cause and Effect Diagrams (Fish Bone or Ishikawa Diagrams)

Fish bone diagram is a graphical method of listing causes of an event. "The Top Event" is
shown at the end of a horizontal line on the right of the diagram. The major categories of
possible causes are arranged as branches slanting off above and below this line. Typical
causes of the top event include the four "M's" - man, methods, materials, and machines.
Specific possible causes become the detailed twigs as shown in Figure below.
Maintenance Practices & Procedures

Poor workmanship (man) can result in some of the errors such as misalignment. Errors such
as impeller balance may result from material (materials) build up on the impeller.

Equipment Design
The basic design of the machine (machine) may be deficient such as a seal design will not
protect the lubricating oil from water contamination.
Operating Practices
The system design may cause errors such as cavation. Inappropriate operating procedures
(methods) such as frequent stop starts may result in early machine failure. The fish bone
diagram helps to focus one's mind on the possible causes of equipment failure. Maintenance
should look to cure the cause and not the symptoms of the problem. For this reason causes
should be pushed as far back as possible. Each possible cause must then be analysed
using techniques such as FMECA to analyse the exact cause of failure.
Maintenance Operating
Equipment
Practice and Practice and
Design
Procedures Procedures
Excess Water in
Cavitation
Vibration Lubricant
Impeller Bal. Pressure

Coupling Fit Flow
Shaft Runout Level
Coupling House
Misalignment Keeping
Type
Spills
Base Skills
Wash-up Practice
Thermal Growth
Method of
Pipe Stress Impeller Pump Packing
Adjustement Procedures
Alignment Pract.
Lubrication
Procedures
Shop
assembly Start and Stop
Procedures Lubrication
Pump Failure
Wear
Humidity
Impeller
Castings
Bearings
Faulty Parts
Stores
Quality Control
etc... Environmental Materials

Figure #14
15.4 Fault Tree Analysis (FTA)

Fault Tree Analysis can be a very detailed process and is used in the analysis of complex
systems. It complements FMEA in that it is a "top-down" analysis, starting with a system fault
(the top event) and analysing this fault in terms of sub-system faults. For example in the
nuclear industry the top event could be a leak of radio-active water from a heat exchanger.
The heat exchanger is then analysed to identify every possible way it could leak such as a
leaking flange or cracked pipe. Each sub event is then analysed to identify how each one
could occur, such as incorrect flange packing fitted, incorrect flange pressure, excessive

water pressure, etc. One of the next levels would then be to identify all possible reasons for
a pressure build up inside the heat exchanger pipes. The process is complete when every
possible root cause of a water leak form the heat exchanger has been identified. Each event
can have a probability associated to it. The same techniques as used in the chapter on
Reliability Block Diagrams can then be used to estimate the probability of a radio-active leak
form the heat exchanger.
Example
In Figure 15 the top event is no liquid display on a calculator. This can be caused either
because the LCD has failed, or because there is no power reaching the display. The latter
fault is analysed further, and may be caused by the faults shown, either switch failed at off,
or internal wiring failure, or no power to the calculator.
In this example the secondary basic events of:

Multiple failure of LCD's
Battery uncharged
No battery
will not be investigated because it does not really matter why the battery is uncharged or
why the battery is missing.
No Liquid Crystal Display

The secondary basic events of:
OR
Gate Internal wiring or joint failure in
the calculator
Insufficent or no
Multiple
Power to LCDs
Failure of LCDs
Wiring or joint failure in transformer

OR
Gate could be investigated if desired.
However it may be easier to replace
On / Off
Power
Supply Failure
the transformer or calculator than try
Internal Wiring
Switch Failed
at Off
or Joint Failure to fix them.
AND
Gate
Fault trees utilise many different

Low or no
Power From
Low or no
Power From
symbols, some similar to digital
Charger Battery Pack electronics e.g. AND gates (in which
all of the inputs must occur together
OR OR
Gate Gate for the event to operate) & OR Gates
(in which only one of the events must
occur for the event to operate).
Faulty Faulty Plug No Faulty
Transformer or Fuse Power on Battery
Mains
Internal Wiring
or Joint Failure Battery No
Urcharged Battery
Internal Wiring Faulty These problems

or Joint Failure Transformer will not be investigated
further
Figure #15:
Bas-035.dsf / Kma 14.3.02

16. REFERENCES
B.K.N. Rao, Handbook of condition monitoring, Frist Edition, Elsevier Advanced Technology
J.H. Williams and P.R. Drake(1994); Condition based maintenance and machine diagnostics
J. Moubray (1997), Reliability-centred Maintenance, Second Edition
R. Anderson and L. Neri (1990), Reliability-centred Maintenance and management and
engineering methods
J. Kolerus (2000), Zustandsberwachung von Maschinen, Third Edition
W. Meng (2001), Maintenance techniques and analysis
S. Turner (2001); A Tool for Improving Operations and Maintenance in the 21st Century
B. Lyons (2001); "Straight Talk" on Maintenance
J. Turnbull (2001); "CbM Rollout Proposal"
W. Nigg (2001), Zustandorientierte Instandhaltung in Untervaz
R. Burkhalter (1995), Condition Monitoring Report VA 95/4277/E
S. Baumgartner and J. Pucillo (2001); Holcim Group Support Report MTC/01/9053/E

Annexes
Annex 1: Summary of Typical Monitoring Techniques
Typical Monitoring Techniques

Short Term
Monitoring Equipment Long Term Benefits
Benefits
Technique Monitored
Contamination analysis Lubrication / hydraulic Identifies excessive gas, Identifies improper
systems: liquid or particle motor installation or
contamination
compressors, repair, excessive
turbines, large pumps, operating conditions,
diesel/gasoline design differences
engines between motor
manufacturers
Electrical surge Induction or Identifies turn-to-turn Identifies improper
comparison synchronous motors, and phase-to-phase motor installation or
dc armatures, shorting, short repair, excessive
synchronous field circuits, reversed or operating conditions,
poles open coils design differences
between motor
manufacturers
Motor current Motors and Identifies worn or Identifies improper
signature analysis generators, damaged bearing or gear
gearboxes, pumps, air gears/bearings, installation practices,
conditioning seal/packing wear, improper
equipment, blowers misalignment, seal/packing
imprecise balancing replacement
practices, poor
alignment or
balancing practices
Oil analysis Lubrication/hydraulic Identifies excessive Identifies improper
systems: wear of lubricated fluid or lubricant
compressors, parts, deterioration selection, improper
turbines, large pumps, of fluid or lubricant oil handling or
diesel/gasoline condition, the dispensing practices,
engines presence of excessive operating
contamination conditions
Performance Heat exchangers, Identifies losses in Identifies improper
trending pumps, motors, pumping efficiency, maintenance or
compressors, flow rates, and heat repair practices, poor
refrigeration transfer due to machine design or
equipment, deteriorating installation,
diesel/gasoline components excessive equipment
engines loading or operation

Shock pulse method Rotating machinery Identifies Identifies improper
with anti-friction deteriorating or bearing installation
bearings, motors, defective bearings, or repair practices,
large pumps, turbines, excessive wear, lack improper oil handling
compressors of lubrication or dispensing
practices, improper
machine operation
Typical Monitoring Techniques

Short Term
Monitoring Equipment Long Term
Benefits
Technique Monitored Benefits
Thermo graphic Roofs, boilers, furnaces, Identifies roof and Identifies poor electrical,
analysis steam systems, electrical insulation leaks, poor roofing, steam system or
switching, motor control insulation repair
centers, high voltage lines electrical contacts or procedures, excessive
connections, leaking operation, poor
steam traps, equipment design
refractory
deterioration
Ultrasonic analysis Steam systems, plant Identifies Identifies improper
air or pneumatic compressed air, seal/packing
systems, hydraulic vacuum or steam installation
systems, vacuum seal leaks; hydraulic practices, poor
systems leaks, bearing wear quality replacement
parts, poor
equipment operation
or design
Vibration analysis Rotating machinery: Identifies Identifies improper
pumps, compressors, deteriorating or alignment or
turbines, gearboxes, defective bearings balancing
diesel/gasoline and gears, procedures,
engines looseness, rubs, improper bearing or
misalignment, gear installation
imprecise balancing practices, excessive
operating conditions
Visual observations, Virtually any machine Identifies loose or Identifies improper
listening, touching shown above visibly worn parts, oil maintenance or
leaks, excessive repair practices,
noise/vibration, hot poor quality parts,
bearings/couplings excessive operation
Wear particle Lubrication/hydraulic Identifies excessive Identifies improper
analysis systems: wear of lubricated bearing or gear
compressors, parts, the presence installation
turbines, large pumps, of contamination practices, improper
diesel/gasoline oil handling and
engines dispensing
practices, excessive
operating conditions

Annex 4: Scope of deliverables
Recommendation/proposal
In order to ask for a ROP (request of proposal) from your suppliers, one should
consider following items and services as integral part of a CM program. This
suggestion is based on an average cement plant with normal complexity. It includes
the first phase of any CM program, starting to monitor only A-criticality equipment
such as drives systems, ventilators etc. Typically we are talking of about 500
vibrations measuring points, 100 oil sampling stations, thermography, kiln ovality
measurements and motor evaluations. However, not all technologies and tool need
to be purchased as services are existing and the frequency of measures to be taken
may led only to once a year (e.g. kiln ovality).
Qty Item Deliverables

1 Initial Audit and Criticality Detailed proposal on required equipment and strategy
Assessment for the first twelve months
Required training activities
ROI calculation and progress tracking system
Integration and iteration process on equipment and
technology
Base documentation to refer to => including mid- and
long term goals
Implementation plan outline
1 Condition Monitoring All state of the art analysis tools and techniques
Software (multi-usage Open architecture => most allow users to
abilities) communicate with existing data collectors and
software e.g. SAP PM, Knowledge Manager Systems,
Hand held etc.
Ability to have a main DB at a cooperate level (like
SAP)
All kind of signals, be they temperature, flow,
corrosion, speed, vibration, pressure or others need to
be feed-able in to the system (as long as we can
measure them) => Hardwired and/or manual entry
Collection of relevant process and machinery condition
data => online and/or offline, locally and/or remotely
Communication of this information to the right people,
at the right time and format. This must occur well
before reaching any alarm levels
Handy and easy manageable SW and DB
User defined alarm levels correlating diagnose
frequencies
Alarms must allow easy recognition of troublesome
spots

Intelligent alarm setting possibilities
Potential into the future:
Correlation of various data to understand the
relationship between machinery stress, malfunctions
and operating conditions
Ability to combine with other data or variables
statistically, mathematically manipulate
Potential to let the user generate their own rules for
automated analysis
Generation of automated work orders trough the
CMMS
1 Data collector 24 Mb of on-board memory
PCM/CIA memory cards usable
Dynamic and static data collection ability e.g. manual
readings
5000 and more measuring points should be collectable
Battery must last of 9 hours at worst-case power
consumption
Battery changeable and rechargeable => without data
loos age
Provide power for transducers => typically 24 Vdc
Data viewable on the equipment
Balancing features optional
Dynamic measurements on the time domain of min. 30
sec.
Analysing tools => FFT
1 (2) Data collectors (1in the 24 Mb of on-board memory
first year and a second Dynamic and static data collection ability e.g. manual
the following year) readings
5000 and more measuring points should be collectable
Battery must last of 9 hours at worst-case power
consumption
Battery changeable and rechargeable => without data
loosage
Provide power for transducers => typically 24 Vdc
Data viewable on the equipment
Ca. Permanent Probes For points of difficult accessibility and/or future online
20 "optional" (Accels, Prox, measurements
Velocity)

Ca. 480 Hand held fast connectors / Assure that measures are take always at the
couplings => not magnetically same position and best possible transmission
of data is guaranteed
1 Onsite Oil analyser Ability to analyse
Water content
Viscosity
Total sum of alien particle
Service Thermography services 1 to 2 time a year => according to plant
requirement
16 Online Surveillance Systems Installation and commissioning included
points (Critical Machinery)
Service Hardware and Software Start- 5 days to 10 day event (should be part of the
Up Assistance and Training HW/SW package)
Service Results Assurance Program 3 days per months according specific scope
(One Year Support)
Service Training as integral part of CbM program introduction to all
program and to a certain What is Condition Monitoring (AP)
extend of the assurance Basics in vibration measurements and
program (AP)
analysis (AP)
Basics in oil sampling and analysis (AP)
Vibration analysis Level I/II
Oil analysis Level I
RCFA
Vibration analysis of drive systems
Not integral part of the "CM
technologies and services"
suppliers deliverables but to
be take into consideration
Service Sheltest measurement 1 per year => only through a experienced
company
Service Motor evaluation Motor with a performance above 100 KW

Annex 5: Training requirements
Training should be directed according the pre-defined knowledge base of the different levels.
Vibration
Level 1 (Basics) Level 2 (Advanced) Level 3 (Professional)
Understand the Covering everything of LEVEL 1 Covering

fundamental everything of
Understand how vibration
information necessary LEVEL 1 & 2
detection and analysis can be
to implement and
used in a cost effective program Expert
operate successful
to identify machinery problems Knowledge of
predictive
and schedule repairs to avoid vibration analysis
maintenance program
costly machine downtime. technologies and
Understand basic methods
Understand advanced alarm
vibration measures (Envelope, SPM,
setting techniques based on
including amplitude, etc)
statistical analysis (both
frequency and phase.
overall and band alarms). Expert knowledge
Able to recognizing of the most
Usage and correctly
common mechanical important
application of VIBRATION
problems from equipments in the
DIAGNOSTIC CHART
vibration plant (drives,
measurements. Understand advanced reducers) and
measurement techniques being able to
Understand the
including natural frequency analyse root
relationship between
testing and high frequency causes.
vibration and
envelope signal analysis.
machinery condition Able to train Level
Able to develop strategies for 2 people
combining this range of
diagnostic techniques and
alarm methods for the
maximum effectiveness of the
condition-monitoring program
Able to explain the benefits of
properly balanced rotating
machinery.
Understand various methods
of solving unbalance
problems including single
plane; two plane and multi-
plane
techniques
Able to train level 1 people

Oil analysis
Level 1 (Basics) Level 2 (Advanced) Level 3 (Professional)
Understand the Covering everything of Covering
fundamentals of LEVEL 1 everything of
lubricants, including LEVEL 1 & 2
Understand the following
a basic introduction
Expert level of
to lubricant Oil Sampling and Frequency
understanding
composition and an Recommended Analysis for
understanding of Train level 1
Gear and Hydraulic Oil
additives. and 2 people
Test, Warning and
o From Crude Condemning Limits
Oil Distillation
to Raffinate Frequent Damages due to
Contamination
o Mineral and
Synthetic Oil Wear Causes
o Classification Mechanical Pollution
of Oil (ISO Impact of Water and
VG, SAE etc) Catalysation to Service Life
o Operating Handling of Oil
Viscosity
Application Cost Reduction with Oil
Care
o Additives and
Test Methods Selection Criteria for Filter
Elements
o Greases
Comparison of Different
Understand how to Filter Design
correctly take
samples Contamination Rating
according to NAS, ISO, Mil,
SAE
Mobil Filtration Units
Element Performance and
Cost Comparison
Tank Breather Filter
Train level 1 people
Infrared Thermography Level 1
have a thorough understanding of the fundamental concepts and theories of this

technology
have the ability to develop a method of thinking required to accurately evaluate
temperature-related problems for virtually any application.
make accurate temperature and emissivity measurements
perform an inspection
account for small object size,
collect, document and analyze thermal data
write reports

Quality Inspections for Cement Plant Equipment
W. H. Brgi, R. Burkhalter, T. Elvermann

PT 97/14233/E (Revision 2, February 1999)
1. Introduction...................................................................................................................253
2. Objective of quality inspections..................................................................................253
3. Examples of damages..................................................................................................253
4. Reasons for defects .....................................................................................................254
5. Inspection methods / tests ..........................................................................................255
6. Suppliers / manufacturer's inspections ....................................................................256
7. Steps of quality assurance ..........................................................................................257
8. Quality inspection program .........................................................................................257
9. HGRS as an inspection company Benefits .............................................................258
10. Concluding Remarks ............................................................................................259
11. Annexes .................................................................................................................260

Summary:
Quality is not an invention of the twentieth century. Conscientious manufacturers have
always been endeavored to assure the good quality of their products. But through the
demand for substantial cost reduction, manufacturers reduced or outsourced their internal
quality control.
Suppliers also place their orders to manufacturers in low cost countries where quality control
is not yet sufficiently observed.
At the same time cement machines change to always larger dimensions and capacities. With
this trend new problems appear as well.
Our statistic shows that during the last few years only 60% of the items inspected are without
any shortcomings. Out of the 40% rejected due to non-conformances, 2% could not be
repaired and had to be discarded and re-fabricated. If not detected during the inspection, this
would have caused substantial loss of production due to the down time, which means a loss
of money and in the worst case even loss of customers.
The Vision for our Quality Control Equipment was therefore clearly defined.
No break downs in the Holcim group due to components failures on purchases of new
equipment
The objective of quality services is always the profit of the purchaser. The investment into
quality will be paid back in lower production costs as the items are more reliable (less down
time), lower maintenance costs (breakdown maintenance) and a longer live time (reduction
of NOA).
One of the most important points from the very beginning is to include adequate quality
requirements (specifications) in the purchase order/contract to assure a trouble free roll-out.
Following this procedure the quality requests have to be included in the contract and the
supplier has to follow it. Each deviations from the specifications must be communicated with
the inspectors and needs an agreement. After a seriously review of operating requirements it
can lead into longer guarantee time, price reduction or manufacturing of a new part.
Finally, from HGRS experience we can say: Quality depends very much on the suppliers
care and regular quality inspections remind them for this responsibility.

1. INTRODUCTION
For the purchaser of a production plant, new machine or component the mayor problem is
normally the loss of production due to equipment break down and less the cost of a
replacement. The failed item is sometimes still under guarantee. Due to production loss,
substantial loss of sales or even loss of customers can occur.
To avoid unexpected down time due to manufacturing defects of items, companies started to
contract independent inspectors to supervise the quality control of the manufacturer. The
history of the quality inspection service at Holcim shows typically this evolution.
Recalculating the main items as girth gears, gear boxes and kiln tyres to influence its design
was the first step. Anyhow, damages still occurred and it was found that some of the reasons
were manufacturing defects. The experience with quality inspections showed that still
nowadays it is necessary to supervise the fabrication through an independent body. This not
only because the suppliers move their production into low price countries, but also because
sometimes there is poor quality produced even in well-known companies with normally good
products.
Consequently Holcim's objectives were set:
Achieve high reliability of the equipment at adequate cost
Maximize effective life time of assets
Form the suppliers to deliver the expected quality
Train our inspectors in supervising efficiently and precisely the quality inspections
Full responsibility for the products remains with the supplier and the quality inspections have
to be cost effective. Therefore our inspections are limited to important items as mentioned in
Annex 1.
2. OBJECTIVE OF QUALITY INSPECTIONS

The objective of quality inspection services is always the profit of the purchaser. The
investment into quality will be paid back in lower production costs as the items are more
reliable (less down time), lower maintenance costs (breakdown maintenance) and have a
longer life time (reduction of NOA). Further valuable information with regard to applied
materials and procedures, which normally are not supplied by the manufacturer, are handed
out.
3. EXAMPLES OF DAMAGES
In a cement plant the catalogue of possible examples of damages is quite long, e.g.
Cracked / fractured girth gears and pinions
Cracked / fractured kiln tyres, rollers, shafts
Cracked kiln or mill shells
Cracked mill end-castings and flanges
Damaged roller mill bodies
Damaged roller press rollers etc.

4. REASONS FOR DEFECTS
Reasons for defects are many fold. In some cases it is lack of knowledge, unfavorable
design or a price that was too low for a good quality product. It may be added that the
production of large castings not only demands excellent knowledge and modern facilities but
also involves a great deal of luck. Errors, such as insufficient supervision or checks, a
carefree attitude, fatalism and taking calculated risks up to deliberate attempts to cheat have
all been observed.
The life cycle of an equipment/component explains very well that the quality of a part is
made at different stages, starting with the design (specification) followed by the
manufacturing (inspections) and finally at the installation.
The quality inspections focus on the manufacturing of the equipment/component, which is a

very important step in its life. Defects can originate in every manufacturing step: casting,
forging, welding, machining or assembling. The most common defects are:
Material inhomogeneities
Cracks
Surface deficiencies
Dimensional/Geometrical errors

5. INSPECTION METHODS / TESTS
During the course of fabrication of components / machines quite a series of quality tests are
carried out and different inspection methods are applied.
Destructive Testing (Mechanical tests and Chemical analysis)
Non-destructive testing
Ultrasonic test
Magnetic particle test
Dye penetration test
Hardness test
Dimensional check
Verification of procedures / documents
Destructive Testing and Chemical Analysis
Materials samples taken from the component in question are tested for the ultimate strength,
yield strength, elongation, reduction of area and impact energy. Their chemical composition
is analysed as well. The values received are compared with current standards and
specifications.
Visual Testing
Visual testing includes checking of general appearance, surface quality, dimensions,
conformity with drawing, functions and is the most often employed method and the one
which has proved its worth for acceptance inspections of cement equipment.
Ultrasonic Testing
The most difficult of the above mentioned test methods might be the check for internal
defects with ultrasonic sound waves. The microstructure of the material, the geometry and
surface quality of the test piece has an important influence on the sound waves. A broad and
permanent experience is needed to interpret the respective signals on the screen.
A wide variety of ultrasonic probes is available, each of them has its special features. For the
inspection of cement equipment it is not possible to take advantage of all technical
possibilities (as for nuclear equipment) owing to the high cost. Therefore, a certain risk of
failure during operation can not be excluded.
Magnetic Particle Testing
A test piece is magnetized and then sprayed with a solution containing very fine iron
particles over the area to be inspected. If a crack exists, the iron particles will be attracted to
the magnetic flux leaking from the area caused by the discontinuity and giving an indication.
These leakage fields act as local magnets. The indication is then interpreted and evaluated.
Fluorescent magnetic particle sprays provide greater sensitivity; the inspection should be
carried out under ultra-violet light in a darkened room.
Dye Penetration Testing
The penetrant liquid is applied to the test surface and sufficient time is permitted for it to
penetrate the surface discontinuity. The excess surface penetrant is removed. A developer is
applied which draws the penetrant from the discontinuity to the test surface where it is
interpreted and evaluated. The liquid penetrants are categorized according to dye type
(visible dye, fluorescent or penetrates of double sensity) and the required procedure to
eliminate them from the test surface (washable with water, post emulsified or solvent
removed).

Hardness Testing
A test body is accelerated and collides with the test surface. The resistance of the material
against penetration of the body into the surface is an indication of the hardness. Two types
of evaluations are used depending on the method applied: A typical dimension (i.e.
diameter) of the trace of the test body on the surface is measured or the energy difference of
the test body before and after the collision is measured. For both methods the value is
converted in hardness and/or strength using tables.
Verification of Procedures / Documents
No quality inspection is complete without the verification of the welding and heat treatment
procedures/parameters (if applicable) and the complete documentation. This documentation
includes all forms completed during the manufacturing and the quality inspection. It has to
come along with the part for further reference at the plant in case of a failure during the
operation phase.
Inspectors Qualifications
The inspectors carrying out quality checks on components and machines during fabrication
must have knowledge, sufficient training and practical experience in the different testing
methods and management of such jobs.
Documents and Arrangements necessary for Quality Inspections
In order to be able to carry out tests, certain specific information and documents about each
component must be available. Also the necessary arrangements for the inspectors must be
provided by the supplier. These needs are listed in Annex 2 which should become part of
the purchasing contract in case HGRS is entrusted with the quality control.
6. SUPPLIERS / MANUFACTURER'S INSPECTIONS

In most cases the supplier is not or only up to a certain degree the manufacturer. He plans
his order to a sub-supplier, e.g. a foundry, forge, gear maker etc. The manufacturer (sub-
supplier) must perform quality inspections himself at various stages of manufacturing
according to contractually agreed specifications, standards and procedures. He is, together
with the supplier, fully responsible for ensuring that his products meet the specified qualities.
In practice however we experience too often that the manufacturer does not bear sufficiently
the responsibility for the quality of his products. Statistical data collected by Holcim show that
only about 60% of the items inspected were without any shortcomings.
Out of 40% rejected due to non-conformances, 2% could not be repaired and had to be
discarded and re-fabricated (see Annex 3). If not detected during inspection, many of these
cases would have later caused substantial loss of production due to downtime of the
installation.

7. STEPS OF QUALITY ASSURANCE
In order to avoid the above-mentioned shortcomings and non-conform behavior of
manufacturers/suppliers it is recommended to engage an independent quality inspection
company. This should, whenever possible, take place before signing the purchasing
contract, specially when working with a new supplier.
The consultant can help to define and verify the expected quality level of the products as
well as the inspection methods / procedures to be applied. The various steps of the
proposed proceedings are shown in Annex 4.1 and 4.2.
If the inspection company is contracted late, the purchaser will not get the full profit of their
experience. Also the work for the company becomes much more difficult, since the
manufacturer will not accept any directions from them towards better quality. Especially if it
would cause additional work for fabrication or testing and if such activities will add cost
(reduce his margin).
In the worst cases experienced, at the time of involvement of the inspection company,
fabrication had progressed so far that only a final inspection was possible. If there are any
mayor deficiencies detected then, they can not be corrected anymore or only with a
substantial loss of time and money. Both endanger the delivery and installation schedule of
the item.
8. QUALITY INSPECTION PROGRAM

Not all components and machines are in the same way critical for failures or for the time
when failures occur. HGRS established a list of the most important items regarding
Highest stress
Influence on the operation reliability
History of similar items
Long delivery
High cost items
The list is presented in Annex 1. For each component mentioned its priority for inspection is
also indicated (1= high priority).
Depending on its type of material, way of fabrication and complexity every component has
an individual inspection program. While some are only checked once at the end of
fabrication, others are checked two or three times at different stages. In Annex 5 a typical
inspection program with three inspections is shown:
After casting and pre-machining (before any welding repairs are carried out)
After construction/repair welding and heat treatment
Final inspection after completion
Certain tests are performed on a random basis others are full-scale tests. The extent of the
non-destructive test amounts generally to 100%, but can be reduced according to
requirements.

9. HMC AS AN INSPECTION COMPANY BENEFITS
HMC members of the inspection team can look back on a 20 years experience in machine
techniques, maintenance and trouble shooting. Operational problems are therefore not new
for them.
HMC is now on the way of establishing their own inspection specifications for the key
components. These can then be used as a "bar" for new suppliers specifications or directly
for placing an order.
Whether HMC sends their own or assigned inspector, all of them are highly trained and
experienced inspectors. Training and knowledge exchange is a permanent issue also in the
field of quality control.
Once a customer entrusts HMC with the quality assurance and inspection of their purchases,
an independent party will watch his interests as if they were their own.
Some specific benefits are:
The risk of unexpected sudden and premature failure is generally largely reduced. Such a failure
could incur a downtime for carrying out of a (temporary) repair up to 3 weeks, resulting in a great
loss of production with corresponding high monetary loss.
If the quality inspection reveals shortcomings or deficiencies, action can be taken in time to avoid
or shorten a delay in the delivery schedule of the item.
Information that is not included in the documentation from the supplier will also be provided to the
purchaser.
An item unacceptable from a quality point of view will be disclosed to the client, the supplier will
have to provide a replacement of acceptable quality.
Due to the long time we have been in this business, we know most of the respective suppliers
and their products very well; all the particularities and specialties, but, most of all, also their weak
points. This allows for more effective inspections.
Due to the well-structured documentation and the valuable hints given, lower maintenance cost
can be expected and the lifetime of the parts may be increased.
Based on their knowledge, our inspectors can decide on the spot about the impact of any
such problem. Furthermore, they have the full back-up support of our department.
Additionally, we are always able to call in specialists within our organization for process
matters. This makes the inspections very efficient and effective, i.e. less time (and money )
is needed.
All this will help that erection, commissioning and operation will be more efficient and
economic.

10. CONCLUDING REMARKS
Quality inspections are not only a matter of costs but also an investment for the lifetime of
new products and will be paid back by lower operation costs (reliability, lifetime). Besides the
assurance of greater reliability in operation they also provide valuable information on the
product. Therefore purchasers should look at it in the same manner as at an investment for a
new line or an upgrade of a plant. However, no quality can completely exclude all risks.
The quality requirements have to be determined together with the supplier and set down in
the acceptance specification before signing the contract. They should actually form an
integral part of the purchase contract or order.
In the long run the collaboration between the customer (plant) and the manufacturer/supplier
has to become closer, so that only good quality items are delivered, even if there is no third
party carrying out additional inspections. Unfortunately we are still far away from that point.

11. ANNEXES
Annex 1
Recommended item to be inspected No. of Inspections Inspection Priority
Kiln
Kiln Roller 2 1
Kiln Roller Axle 1 1
Thrust Roller 1 2
Tyre 2-3 1
Tyre Shell Section 1 1
Shell Section (10 pieces) 1 2
Planetary Carrier 1 2
Tube Mill
Mill End 2-3 1
Mill Flange 1-2 1
Mill Body 2 1
Mill Shell Section 1 1
Mill Shell with Tyre 2-3 1
Tube Mill Tyre (Bandage) 2 1
Slide Shoe 1 2
Vertical Mill
Grinding Table 2 1
Roller Yoke (Polysius) 2 1
Roller Carrier (FLS, Pfeiffer, etc.) 1 1
Rocker Arm (Loesche, Onoda) 1 2
Fork (Loesche, Onoda) 1 2
Roller Body [if GGG] 1-2 2
Roller Shaft 1 2
Table Liner (Segment) 1 2
Roller Wear Segment/Tyre 1 2
Roller press
Roller Body (Studded/ Welded Design) 1 1
Roller Segments 2 1
Roller Shaft 1 1
Bearing Block 1 2
Roller Press Frame 1 2
Drives
Girth Gear (Cast) 3 1
Girth Gear (Welded) 2 1
Pinion 1 1
Pinion Shaft 1 1
Mill Gear Drive 2 1
Mill Gear Planetary Carrier 1 1
Kiln Gear Drive (Mechanic Standard) 1 2
Kiln Gear Drive (Mechanic Direct) 1 2
Kiln Gear Drive (Hydrostatic) 1 2
Roller Press Drive 1 1
Various
Hammer Crusher Shaft 1 1
Main Process Fan 1 2
Reclaimer Chain 1 2

Annex 2
Documents and Arrangements necessary for Quality Inspections

The following documents are to be provided to Holcim at least one month before
inspections:
Fabrication program
Inspection program of supplier
Manufacturer for each item to be inspected (company address, telex, fax, telephone-No.,
responsible person for quality inspection, place of fabrication)
Acceptance specifications for each item to be inspected (acceptance criteria, state of
item when inspections take place, scope, etc.)
Applied standards for non-destructive testing and other inspections
Drawings (fabrication and assembly if available)
Guarantee clauses of contract for items to be inspected
Documents and certificates to be provided to Holcim during inspections or latest at
the final inspection (two sets):
Non-destructive testing
Surface quality
Hardness test
Materials (chemical composition, mechanical properties)
Record of dimensions
Heat treatment (indicating temperatures, heating, holding and cooling time)
Welding (indicating properties of electrodes, preheating and post heating procedures,
type of welding, welder legitimation)
Map of defects of repair welds (indicating place and size of defects/repair welds)
All the above mentioned documents have to be delivered in the language specified in the
contract. This also applies to documents from sub-suppliers.
Arrangements of Inspection
The HMC's inspectors and additional inspectors from the client must be given access to
manufacture at any time (time to be co-ordinated in advance).
The HMC's inspector must be given due notice (at least 5 working days in advance) of the
individual inspection stages by the supplier. Provision should be made by the supplier to
allow the examination of several items at the same time (if possible) to minimize traveling
expenses.
All inspections will be carried out together with the supplier whereby all tests and
measurements are to be performed by the manufacturers' personal using their own
instruments and equipment.
In case of suspect results of inspections the HMC inspector has the right to demand
additional suitable inspections.

Annex 3

Annex 4.1
Quality Control of New Equipment
PT-QCE
PT-QCE Checks specifications and
Discussion on possible scope of procedures
machines in project Proposes adjustments when
Cost assumption for Quality Control necessary
for the CE offer to the customer Prepares HMC inspection
programs for all items to be
checked
HES/HEC/Customer Discussion with supplier
Send tender documents to supplier Customer, HES/HEC, Supplier

for offering
Quality requirements are included in Sign purchasing order
tender
Customer/HES/HEC
Order inspection services from PT-
Suppliers QCE
Send tender to HES/HEC/Customer
Supplier
Sends the following documents /
HES/HEC/Customer/PT-QCE
information to customer/HMC:
Evaluate tender modified acceptance specifications
Negotiate with preferable suppliers inspection programs of supplier
Discussion on quality / workshop information on sub-supplier, if
inspection, specifications applicable
fabrication schedule
drawing
Supplier
PT-QCE
Adjust their offer
Studies these documents
Discussion with supplier
Customer, HES/HEC, Supplier Agreement on proceedings
Contract negotiations (agreed quality Follow up of project
inspection requirements specified)
PT-QCE (+Customer)
HES/HEC/Customer Supervise inspections
Initiate corrective measures if
Sign letter of intent with selected
necessary
supplier
Inform customer with interim reports
Supplier
Send contract specifications and PT-QCE
inspections procedures to HMC Elaborates final report
Invoicing (periodically and at the end
of project)
Final work

Annex 4.2
Quality Control of Individual Spare Parts
Customer
Requests offer from HMC-PT for
workshop inspections Supplier
Send the following documents /
information to customer / PT-QCE
confirmation of order
PT-QCE
modified acceptance specification
Offers services inspection program of supplier
Request information / documents for information of sub-supplier, if
inspections applicable
fabrication schedule
drawing
Customer
Accepts offer Supplier
Procures acceptance specification /
information from supplier
PT-QCE
Follow-up of information and project
PT-QCE
Opens project
Checks specification / information
Recommends adjustments of Supplier
specifications if necessary
Informs on inspection program Invite PT-QCE (+Customer) for
inspections
PT-QCE (+Customer)
Customer Supervise inspections
Initiate corrective measures if
Orders spares on base of adjusted necessary
specification Inform customer with interim reports
Informs supplier on QCE inspection
program
Requests for detail information
PT-QCE
Elaborates final report
Invoicing
Final work

Annex 5
GIRTH GEAR (CAST)
INSPECTION PROGRAM
For the quality assessment of a cast girth gear three inspections are to be carried out at
different stages of the manufacturing.
First Inspection
After casting and pre-machining (before any welding repairs are carried out)
Visual inspection
Ultrasonic test of the internal material homogeneity (100%)
Magnetic particle test of the surface in general and the areas of casting defects (100%)
Dimensions
Review of repair welding and heat treatment procedure
Second Inspection
After repair welding and heat treatment
Visual inspection
Magnetic particle test of repaired areas
Ultrasonic test of the repaired areas
Witness of material strength test (tensile and yield strength, elongation and reduction of
area, impact work)
Hardness test across repair welds and at gear rim
Dimensions
The second inspection can be dispensed if only minor repair welding is necessary. In this
case the material strength test would be carried out either during the first or the last
inspection.
Third Inspection
Final inspection after completion. The two gear halves must be assembled.
Visual inspection
Magnetic particle test of tooth (100%) and remaining areas
Hardness test of tooth flanks
Surface roughness test of teeth
Tooth rectilinearity (if straight teeth)
Tooth contact of pinion with girth gear (if possible)
Dimensions in general and of toothing in particular
Review of documents and certificates
The extent of the Non Destructive Tests amounts generally to 100%. It can be reduced on
the HMC inspector's decision.

Chapter 4
Gas Analysing Systems

PRACTICAL GAS ANALYSIS IN CEMENT WORKS
H. Nyffenegger
PT 98/14340/E (substitute for 93/4080/D)
1. Introduction...................................................................................................................275
2. Purpose .........................................................................................................................275
3. Definitions .....................................................................................................................275
4. Measuring Methods ......................................................................................................277
4.1 In-situ measurement ........................................................................................277
4.2 Extractive measurement ..................................................................................278
4.3 Measuring points..............................................................................................278
5. Gas Sampling................................................................................................................282
5.1 Kiln inlet (measuring point A) ...........................................................................282
5.1.1 Preheater kilns (cyclone preheater and grate preheater kilns).....................282
5.1.2 Wet kilns and long dry kilns ..........................................................................285
5.1.3 Special instructions for gas sampling at the kiln inlet ...................................285
5.2 Other measuring points (B to G) ......................................................................289
5.2.1 Kiln inlet chamber and riser duct (measuring point B)..................................289
5.2.2 Lower cyclone stages (measuring point C) ..................................................289
5.2.3 Upper cyclone stages (measuring point D) ..................................................289
5.2.4 Downstream of cyclone or grate preheaters (measuring point E) ................290
5.2.5 Downstream of electrostatic filter (measuring points F, G)...........................290
5.2.6 Coal pulverising mill......................................................................................291
6. Gas Sampling Probes ..................................................................................................293
6.1 Sampling probes in the temperature range 900 to 1500C..............................293
6.1.1 Siemens type FLK ........................................................................................293
6.1.2 Harman & Braun type 60S (formerly type 13)...............................................294
6.1.3 Hartmann & Braun type H (Holcim probe) ....................................................295
6.2 Sampling probes in the temperature range 500 to 900C................................296
6.3 Sampling probes in the temperature range below 500C ................................296
7. Sample Gas Preparation ..............................................................................................298
7.1 Sample gas pipe ..............................................................................................298
7.2 Sample gas coolers..........................................................................................299
7.2.1 Specifications of a suitable sample gas cooler.............................................299
7.3 Sample gas pump ............................................................................................300

7.4 Flow control and distribution ............................................................................301
7.4.1 Simple gas control ........................................................................................302
7.4.2 Electronic gas volume control.......................................................................304
7.5 Sample gas filters.............................................................................................305
7.5.1 Coarse filters ................................................................................................305
7.5.2 Fine filters .....................................................................................................305
7.5.3 Valves...........................................................................................................305
7.5.4 Pressure control valve ..................................................................................305
7.6 Adjustment .......................................................................................................306
7.7 Further information on sample gas preparation ...............................................306
7.7.1 Sample gas discharge ..................................................................................306
7.7.2 Tightness test ...............................................................................................306
7.7.3 Filter condition monitoring ............................................................................306
7.7.4 Analyser room ..............................................................................................307
7.8 Space requirement...........................................................................................309
7.9 Location............................................................................................................309
7.10 Climatic conditions ...........................................................................................309
8. Analysers ......................................................................................................................310
8.1 Infrared absorption ...........................................................................................311
8.1.1 Space case, sulphur dioxide (SO2)...............................................................311
8.2 Ultra-violet absorption ......................................................................................312
8.3 Paramagnetism ................................................................................................312
8.3.1 Oxygen analysers.........................................................................................312
8.4 Flame ionisation (FID)......................................................................................312
8.5 Solid-state electrolytic systems ........................................................................313
8.5.1 Zirconium dioxide .........................................................................................313
8.5.2 Other electrochemical measuring methods ..................................................313
8.6 Multi-component measuring systems...............................................................313
8.6.1 Sick GM 31 ...................................................................................................314
8.6.2 OPSIS AR 620; ER 650................................................................................314
8.6.3 LDS 3000 diode laser of AltOptronic (Sweden)............................................315
8.6.4 Advance Cemas FTRI of Hartmann & Braun ...............................................316
9. Recording and Evaluation ...........................................................................................317
9.1 Trend curves, recorders ...................................................................................317
9.2 Averaging computers .......................................................................................317
9.3 Data logging .....................................................................................................324
9.4 Emission computers.........................................................................................324

10. Maintenance and Quality Assurance...................................................................324
10.1 Visual checks ...................................................................................................326
10.2 Adjustment checks, cleaning............................................................................326
10.2.1 Tightness checks..........................................................................................326
10.3 Replacement of wearing parts .........................................................................326
10.3.1 Sample gas pump.........................................................................................326
10.3.2 Cleaning sintered metal and ceramic filters..................................................326
10.3.3 Fine filters .....................................................................................................326
10.3.4 Gas analysers...............................................................................................327
10.4 Function test.....................................................................................................327
10.5 Test gases........................................................................................................327
10.5.1 Mixing test gases..........................................................................................328
10.6 Fault signals .....................................................................................................329
10.7 Automated maintenance equipment ................................................................329
11. Measuring Errors...................................................................................................330
11.1 Sample gas sampling.......................................................................................330
11.2 Sample gas preparation ...................................................................................330
11.2.1 Sorption and chemical reaction ....................................................................330
11.2.2 CO reduction ................................................................................................330
11.2.3 NO2 formation in the probe and sample gas preparation system .................330
11.2.4 SO2 reduction in the probe and sample gas pipe .........................................330
11.2.5 SO2 reduction of the filter dust of the sampling probe ..................................330
11.2.6 Interaction with a liquid .................................................................................331
11.3 Volumetric errors due to the solubility of accompanying components .............331
11.4 Gas analysers ..................................................................................................331
12. Special Functions of Gas Analysis......................................................................332
12.1 CO monitoring for protection of electrostatic filter ............................................332
12.1.1 Basic concept of filter shutdown ...................................................................332
12.1.2 Optimisation of CO measurement ................................................................332
12.1.3 Specifications ...............................................................................................333
12.1.4 Shutdown procedure ....................................................................................333
12.1.5 Special devices.............................................................................................334
12.1.6 Laser analysers ............................................................................................334
12.2 Monitoring of coal pulverising mills ..................................................................335
12.2.1 Limiting values..............................................................................................335
13. Conversion Factors...............................................................................................336
14. Literature................................................................................................................337

SUMMARY
The measurement of individual process gas components in connection with emissions and
for the purpose of process optimisation is increasingly becoming one of the most important
procedures in cement works. Measuring techniques have advanced to the extent that there
is now virtually no such thing as a problem that cannot be solved technically with the
necessary specialised knowledge and means.
In principle, distinction is to be made between analysers that measure (in-situ) directly in the
measuring channel (process gas flow) and extractive measuring instruments. Extractive
measuring instruments measure in a partial gas flow that is prepared prior to entering the
analysers according to specifications (sampling). Extractive measuring instruments are
mainly used in process measurement, as the conditions with respect to temperature and
dust loading does not permit the use of the in-situ technique.
The point at which measurements are carried out is mainly a question of the measuring
problem involved. The extent of sampling increases with increasing process temperature at
the sampling point. For gas sampling at the kiln inlet and up to gas temperatures in excess
of 500C, cooling and an emergency cooling or withdrawal system is required. Withdrawal
devices with compressed-air drives are more reliable for emergency use, so that preference
is given to electric drives.
The separation of dust from the sample gas preferably takes place as close as possible to
the entrance of the sampling probe. Sintered metal filters (alloyed steel) clog less quickly
than the usual ceramic filters. Compared to the filters outside the sampling probe, internal
filters require less maintenance and are smaller in volume. Wet sampling probes are
considered out-dated.
Three systems for gas sampling at the kiln inlet are described here in brief: Siemens type
FLK, Hartmann & Braun type 60 and type H (Holcim probe). The latter has proved highly
reliable in use worldwide. For gas sampling below 500C, in addition to standard products,
proposals are submitted for the specific construction of sampling probes.
The purpose of sample gas preparation is to supply the analysers with qualitatively and
quantitatively specified sample gas. The apparatus used for this purpose must be
generously dimensioned and equipped. This requirement applies in particular to the sample
gas pump, sample gas cooler and sample gas distribution. The temperature in the
measuring cabinet should be between 15 and 25C. A monitoring system must inform the
control system of any irregularities such as insufficient sample gas, system being
adjusted, etc.
The analyser room must provide the ambient conditions for the perfect functioning of the
instruments, i.e. free from vibrations, as close to the sampling point as possible, free from
dust and between 15 and 25C. As it also serves as a workroom for maintenance
personnel, it must offer sufficient space, good ventilation and a low noise level as possible.
The analysers used for the determination of individual components in the sample gas are
based on different physical effects such as light absorption. Paramagnetism, flame ionisation
and solid-state electrolysis. The instrument functions and theoretical relationships will be
dealt with only insofar as they are of significance for practical application. When choosing
analysers, the manufacturing company is not as important as the specifications and local
services offered.
The majority of gas components (CO, NO, SO2, CO2) can be measured with NDIR (non-
dispersive infrared analysers). They are simple, sturdy, robust, durable and maintenance-
friendly. Disturbing cross-sensitivities to CO2 and water vapour are appropriately corrected.
When measuring SO2 with NDIR analysers by the classic method, undefinable quantities of
SO2 are absorbed in the sample gas cooler. Corrections are possible by acidifying the

condensate in the sample gas cooler or using hot NDIR analysers. Preference is given to the
use of in-situ analysers.
Oxygen cannot be measured with light-sensitive analysers. The most reliable analysers are
still considered to be those that exploit the magnetic properties of oxygen. The life of solid-
state electrolytic systems such as zirconium dioxide probes is limited and depends on the
gas composition. Improved electrochemical measuring cells have been more widely for the
past few years and have a useful life of three years. They are small, lightweight, require a
minimum degree of evaluation and are therefore used more and more in extractive multi-
component analysers.
The flame ionisation detector is used for measurement of the total hydrocarbons (VOC =
Volatile Organic Compounds). The increasing requirements for continuous measurement of
the VOC - in connection with alternative fuels - are maintenance-intensive.
Microprocessor technology makes it possible to use well-known and new measuring
methods and therefore to measure several components simultaneously. For the installation
or replacement of measuring instruments, only the use of multi-component analysers is
possible.
The GM 31 in-situ analyser of Messrs Sick enables three gas components to be measured
simultaneously. Compared with the earlier GM 21 (SO2 and dust) and GM 30 (additionally
NO) series, the GM 31 does not offer dust measurement.
The AR 620 and AR 650 series of Opsis are modern emission measuring instruments.
Depending on the type and number of components to be measured, infrared or ultra-violet
light used to the two together. The analyser or analysers are connected to the receiver at the
measuring channel via glass-fibre cable (light conductor). Control, monitoring and evaluation
is carried out by a standard PC. The analysers have been used several times with good
experience.
The new gas analysers with lasers are expected to have a promising future. These are in-
situ analysers that offer the benefits of high sensitivity, short response times and high
stability. They could provide the ideal preconditions for CO monitoring, but are (still)
relatively expensive.
Of the extractive measuring instruments, good experiences have been made with the
CEMAS FTIR analyser of Hartmann and Braun. The analyser can be programmed with all IR
selectable components. In order not to alter the originality of the sample gas, all gas
conducting parts are heated.
Recording the measured values with strip chart recorders should still suffice for a general
overview. However, in the majority of countries, digital evaluation of emissions is required
according to specific criteria. In the field of process optimisation, the measured values are
also only useful when available in the required form, e.g. as mean values, spreadsheet for
further evaluation, etc. At the kiln inlet, averaging or damping for lump fuel (tyres) feeding in
precalcining is absolutely necessary. In modern plants, the measured values can be partly or
wholly processed by the process control system.

The reliable functioning of each measuring instrument depends largely on the maintenance
provided. Visual checks and function tests must be carried out according to the age of the
measuring instruments and the specifications of the supplier.
Personnel must be suitably trained. Critical spare parts such as the sample gas pump, filters
valves, seals, etc., must be carried in stock.
The gas analysers must be checked according to the manufacturers instructions and
emission measurements carried out in compliance with the local requirements and
corrections made as necessary. If test gas should (still) be required, this must be obtained in
good time. Any fault signalling devices must be checked for proper functioning.
Theoretically, maintenance can be reduced by appropriate automation. Adjustment via
internal analyser reference devices has proved highly disadvantageous in this respect.
However, if solenoid valves should be required for the application of test gas, their
usefulness is negligible or negative. Solenoid valves in sample gas are highly susceptible to
faults. For this reason, sample gas changeover from two or more measuring points to the
same analyser group is not considered worthwhile.
The accuracy of gas analysis as a whole is often overestimated, whereby the analysers are
least to blame in this respect. Significant errors occur when sampling (representative) due to
sorptions and reactions in sample gas preparation.
Gas analysis also has special monitoring functions for protection against explosions in the
electrostatic filter and coal pulverising mills. Simple monitoring of the CO content is usually
not sufficient for the protection of the electrostatic filter. The consequences are unnecessary
shutdowns of the filter. In addition, reference is made to a new analyser of Messrs Sick,
which has been designed specially for this task.
Finally, a table is provided showing the various conversion factors from ppm to mg/m3 for the
relevant gas components.

1. INTRODUCTION
Gas analysis is among the most important, but also the most sophisticated methods of
measurement in cement works. This is due to the constantly growing demands in the field of
gaseous emissions and related requirements. Gas analysis fulfils an equally important
function in the field of automation, where the reliable and exact measurement of certain gas
components plays an essential role.
Advances in the field of gas analysis have taken place parallel to the trend of demands, so
that one can hardly speak of technically unsolvable problems any longer. If problems
should nevertheless occur, they are mainly due to the following
Incorrect choice of equipment
Lack of specialised knowledge
Underestimation of expenditure involved (investment and/or maintenance expenditure)
The aim of this report is to prevent problems of this kind occurring when renewing and
adding gas analysing apparatus.
2. PURPOSE
The continuous measurement of the constituents of certain components in the gas flows
(gas analysis) in cement works consists of three areas of activity:
a) Emission measurements (pollutants, e.g. NO, SO2, CO, etc.).
b) Safety measuring systems (e.g. CO monitoring in filtering plants, monitoring of coal dust
silos, etc.).
c) Process measuring systems (e.g. O2, CO, NO) for optimisation of the combustion
process.
Important for all these activities is a high availability of the measuring equipment and
sufficient accuracy of the measured values. The report therefore attaches particular
importance to sample gas sampling and preparation, maintenance and quality assurance as
well as possible measuring errors. Finally, hints and practical advice are given, based on
many years of experience.
3. DEFINITIONS
Indication delay (90% time) of an analyser or a measuring instrument: The time that
elapses from a sudden change in the gas concentration at the entrance of the analyser or
probe up to indication of 90% of this change.
Downtime: The time during which faulty conditions of the measuring instrument and their
correction unforeseeably make the generation of measuring or adjusting signals impossible.
Accompanying gas: Sample gas without the gas components to be measured.
Extractive sampling: The sample gas is extracted as a partial gas flow from the process
gas prior to analysis and conditioned.
Gradient monitoring: Monitoring the gas concentration change gradient (e.g. control of CO
cut-off).
In-situ: The measuring instrument is located in or on the process gas duct; analysis is
carried out in the process gas (e.g. optical dust content measuring instrument or zirconium
oxygen probe).

Adjustment: Setting or adjustment of the gas analyser with test or zero gas with the aim of
adjusting the indication of the gas analyser as close as possible to the gas concentration to
be measured in the sample gas. With newer analysers, this can also take place with built-in
gas-filled glass cuvettes.
Calibration: For a given measuring instrument, determination of a valid relationship
between the indicated measured value or the value of the output signal and a reference
value that comes closest to the true measured quantity. Determination of the scale divisions
of a measuring instrument with the aid of a reference measuring method (generally required
only for officially specified continuous emission measurements).
Life zero: A specific quantity (normally mA) prevails in the electric output signal when the
analyser physically measures zero. The advantage of this system is that the output signal
is electrically set to zero only if the electrical measuring circuit fails.
Sample gas: That part of the process gas extracted from the process gas and passed to the
analyser for measurement.
Sample gas preparation: The entire equipment used for the qualitative and quantitative
preparation of the sample gas to ensure that the same conforms to the specifications of the
analysers.
Measuring equipment: All devices and instruments required for the measurement of gas
component concentrations.
Measured quantity: Physical quantity, the numerical value of which is to be determined by
measurement.
Measured component: The gas component in sample gas, the concentration of which is to
be measured.
Zero gas: Test gas used for the adjustment or readjustment of the zero of a gas analyser.
On-Line: The measured quantity is coupled directly with the process.
Test gas: Gas with a known composition for adjustment or readjustment of gas analysers. It
consists of the measured component and one or several accompanying gases (e.g. nitrogen/
carbonic acid + measuring component).
Cross-sensitivity: Influence of the concentration of a disturbing component on the
measuring result of another gas component under identical measuring conditions.
Representativity: The correspondence of the measured signal with the definition of the
measured quantity, e.g. the composition of the gas at the kiln inlet measured at a point in
relation to the average gas composition over the entire cross-section.
Disturbing component: Component in the sample gas that falsifies the measured value,
e.g. if the gas analyser has a cross-sensitivity or chemical reactions distort measurements.
Dead time of an analyser or measuring instrument: The time that elapses from a sudden
change in the gas concentration at the entrance of the analyser or probe up to the start of
the indication of this change.
Availability: The time during which the measuring instrument generates usable signals.
Supply pressure: The sample gas pressure necessary for supplying the analysers.
VOC Volatile Organic Compounds: Volatile organic hydrocarbons.
Maintenance time: Time required for maintenance of the measuring instrument (planned
preventative maintenance).

4. MEASURING METHODS
4.1 In-situ measurement

In-situ measuring methods mainly offer the following advantages:
No gas sampling
Sampling forms a path in the measuring channel (as opposed to a point in extractive
methods)
Undelayed indication
Less maintenance
Good long-term stability
In-situ measuring methods are used increasingly in modern measuring techniques.
Two different measuring principles are in the foreground:
Solid-state electrolytic and
Optical systems
Stabilised zirconium dioxide is a typical solid-state electrolyte with whose properties oxygen
can be measured directly in the process (similar to thermocouples) (see chapter 8.5.1). This
measuring method has proved reliable in the cement industry. In individual cases, the useful
life of the zirconium sensors was unsatisfactory, as their activity was destroyed by certain
accompanying gases.
Optical instruments are flanged directly on to the gas channel to be measured (e.g.
chimney). These consist of a transmitter (normally light) and a receiver or reflector arranged
opposite. While passing through the channel, the light beam emitted from the transmitter
(infrared, ultraviolet or laser) is changed by certain gas components. After appropriate
processing, this change results in the measured quantity (see chapter 8.6).
Optical measuring instruments (8.6.1, 8.6.2, 8.6.3) are used mainly for the measurement of
pollutant emissions, e.g. dust content, SO2, NO, etc. The range of instruments available for
measuring further gas components is increasing in line with technological advances. The
application of optical measuring in-situ instruments is restricted by two factors; the dust
content and the given length of the measuring path, which is in direct relation to the
sensitivity. For use in the cement works, these restrictions mean that they only function in
pure gas and their application is limited solely to the measurement of emissions, although
good experience has been made in this respect. Thanks to improved representativity, in-situ
measuring methods are technically superior to extractive methods.
Investments in-situ instruments are generally higher compared with extractive measuring
instruments. As will be shown later, additional investments in in-situ measuring methods are
justified under certain conditions. The level of maintenance is normally less, but partly more
sophisticated than extractive measuring instruments, from a technical point of view.
The measured values of in-situ instruments are always based on gas in an operational
condition. However, values in a standard condition are usually required, .i.e. at 0C,
1013mbar, dry. For conversion to the standard condition, the parameters pressure,
temperature and moisture must be known or additionally measured.

4.2 Extractive measurement
The measuring procedure is subdivided into different stages:
Sample gas sampling
Sample gas supply
Sample gas preparation
Analysers
With the exception of the dust content, the majority of components can be measured by the
extractive method. In-situ measuring instruments offer the following advantages:
Several analysers can be connected to the same sample gas preparation apparatus.
The positioning of dust and heat-sensitive analysers is more flexible.
Adjustment and calibration is simpler (with built-in reference cuvettes or test gas).
Drying the gases during gas preparation by means of cooling to a dew point of about 3C
can be disturbing for certain components. These are logically water vapour (H2O) sulphur
dioxide (SO2) and certain hydrocarbon compounds (VOC) (see chapter 8.4).
To exclude disturbing influences during sample gas drying, there are instruments with
operating temperatures that are far above the dew point (up to 250C). The components of
these instruments are subject to increased load when exposed to these temperatures, which
inevitably leads to an increased susceptibility to faults. Experiences made with an instrument
of this type are described in chapter 8.6.4.
The multiple advantages of extractive measuring instruments, compared with in-situ
measuring methods, are confronted with a slightly higher level of maintenance, although the
maintenance necessary for the extractive measuring method is less demanding than that
required for in-situ measuring instruments.
4.3 Measuring points

Figures 1 to 3 show the typical measuring points at various kiln systems (cyclone
preheaters, grate preheaters and wet kilns). The associated tables show the usually
measured components as well as the purpose of their measurement. The choice of
measuring point at the kiln system at which measurements are to be made, depends
primarily on the purpose of measurement.

Figure 1: Measuring Points at Cyclone Preheater Kilns
G
E
F
Purpose Components A B C D E F G Remarks
Process Control O2 CO NO CO2 Without Precalcination
Process Control O2 CO NO CO2 With Precalcination
CO-Monitoring O2 CO NO Without Precalcination
CO-Monitoring O2 CO NO With Precalcination
Emissions O2 CO NO Without Precalcination
Emissions O2 CO NO CO2 With Precalcination
Emissions SO2 NO HC etc All Kiln Systems
Suitability: Good Measuring Points: A Kiln Inlet (Probe in rotating Part)

B Inlet Chamber, Riser Duct
Satisfactory (Probe not in rotating Part)
C Lower Cyclon stages (Calcinator)
Bad D Upper Cyclon stages
E After Cyclon preheater
Standby (Reserve) F After the Filter before Fan
G Stack (After Filter and Fan)
Unsuitable
Table 1: Cyclon-preheater Kilns

Figure 2: Measuring points at the grate preheater kilns
Process control O2 CO NO CO2
CO-monitoring O2 CO NO
Emissions O2 CO NO
Emissions O2 CO NO CO2
Emissions SO2 NO HC etc
Suitability: Good Measuring points: A Kiln inlet (probe in rotating part)

B Kiln inlet (probe not in rotating part)
Satisfactory E After the grate preheater
F After the filter, before the fan
Bad G Stack, after the filter and the fan
Unsuitable
Table 2 Grate preheater kilns

Figure 3: Measuring points at wet kilns and long dry kilns
Process optimization O2 CO NO CO2
Emissions O2 CO NO CO2
Emissions SO2 NO HC etc
Suitability: Good Measuring points: A Kiln inlet (probe in rotating part)

E before the filter
Satisfactory F After the filter, before the fan
G Stack, after the filter and the fan
Bad
Unsuitable
Table 3 Wet and long dry kilns

5. GAS SAMPLING
5.1 Kiln inlet (measuring point A)

At the rotary kiln itself, only the measuring point at the feed end of the kiln is realisable with
reasonable effort. The components normally measured there are needed as reference
quantities for an optimal combustion process (usually O2, CO, NO).
To prevent distortion of the measured signals due to false air from the kiln inlet seal, the
probe tip must be located distinctly within the rotating part. The sampling conditions at the
kiln inlet vary considerably, depending on the type of kiln (preheater kilns, long dry kilns and
wet kilns).
5.1.1 Preheater kilns (cyclone preheater and grate preheater kilns)
Process conditions at the sampling point

Gas temperature 1000 to 1300C
Dust content >100g/m3 (cyclone preheater)
Dust content >5/g/m3 (grate preheater)
Dew point temperature 35 to 40C
The process gas contains large proportions of alkalis, chlorides and sulphates that are
subject to high local and time variations. Depending on the process and raw material
situation, the alkali compounds combine with the dust to form strongly adherent deposits. To
counteract their effects to make gas sampling possible at all, elaborate measures are
necessary from a measuring point of view.
Owing to the high temperatures, the provision of external cooling is necessary for all types of
sampling probes in the area of the kiln inlet.
Mounting sampling probes
The principle mounting arrangement is shown in Figure 4. Owing to the wide range of kiln
systems available, it is impossible to define an exact mounting point for all types of kilns in
advance. The following criteria play a dominant role:
Available space
Accessibility
Direction of rotation of the kiln
Internals in the kiln inlet chamber
According to experience, the gas composition at the kiln inlet is not homogeneous. In
consequence, the sample is only representative of the average gas composition to a limited
degree, irrespective of where the probe is positioned. The measuring point or sampling point
in the rotating part of the kiln, as will be shown later, cannot be determined primarily on the
basis of measuring criteria. An individually adapted interpretation of the measured values
(averaging), correlation with disturbance variables, etc.) is far more important here than the
position of the probe (see chapter 9).

The mounting position shown in Figure 4 is for reference only. Positions and dimensions can
vary considerably from plant to plant. The undisturbed, reliable function of the probe and the
warranty that no false air will be drawn in from the kiln inlet seal always has the highest
priority.
The probe must be easily accessible. If necessary, special platforms must be erected. A
lifting device placed above the probe will considerably facilitate maintenance. The minimum
distance between the platform railing and probe axis should be about 1.3m (at least on one
side).
As the probe consists of relatively long sections, the access way must not be too narrow.
The mounting position should be accessible via steps not ladders.
Pokeholes, which must be operated regularly, are located in the immediate vicinity of the
probe. It must be ensured that the sampling probe does not interfere with the working space
provided for maintenance personnel.
In principle, the probe must be mounted on the side where the direction of rotation of the kiln
shell is directed downwards. If conditions do not permit, the other side can also be selected,
provided the probe can be positioned at a sufficient distance to the material bed.
Meal inlet pipes, pokeholes, measuring connections, etc., mostly restrict free selection of the
mounting position, the main aim is to minimise hindrance of maintenance personnel.
The inevitably cooled probe surfaces encourage the formation of deposits in the kiln inlet
chamber. These deposits not only hinder subsequent removal of a probe, they can
negatively influence the geometry of the kiln inlet chamber and also cause pressure loss. To
prevent this happening, it is recommended to regularly blow-out the gap between the pipe
and probe by means of a mounted air cannon. The pipe must be dimensioned, so that a gap
of about 25mm is produced between the probe and inner pipe wall.

Figure 4: Mounting of the gas sampling probe at kiln inlet
Retraction device
Kiln inlet seal

B X A
E
E
ca L+1.5m L
A = 0....300 mm = 30....45
B = Min. 500 mm L = Nominal length of the probe
C = B + wall thickness N = Inclination from horizontal axis
E = Min 200 mm X = As short as possible
Fig. 4 Mounting of the gas sampling probe at kiln inlet

5.1.2 Wet kilns and long dry kilns
Process conditions at the sampling point:

Temperature wet kilns 150 to 200C
Temperature long dry kilns 250 to 500C
Dust content wet kilns about 40g/Nm3
Dust content long dry kilns about 300g/Nm3
Dew point temperature long dry kilns 35 to 40C
Under these conditions, gas sampling is much simpler than for kilns with preheaters. No
cooling of the probe is necessary and the measured signal has improved representativity, as
better intermixture of the gases takes place due to the internals (chains, crosses). With
regard to positioning, mounting conditions, etc. roughly the same conditions apply as for
kilns with preheaters.
5.1.3 Special instructions for gas sampling at the kiln inlet

Cyclone and grate preheater kilns
The above process conditions show that gas sampling from this difficult environment
requires particular measures:
Owing to the high temperature, the sampling probe must be cooled.
A filter system must ensure that the large amounts of diversely structured dust is
removed prior to sample gas preparation
Very often, deposits build up at the sampling system and clog the probe entrance, so
that measures must be taken for their prevention.
Probe cooling
Liquid cooling is used for all known probes. The cooling medium, which is usually water,
transfers the heat in a closed system to water-air or water-water heat exchanger to prevent
calcareous deposits. The energy produced varies considerably. Depending on the condition
of the probe surface, a large amount (clean surface) or small amount (dust or deposits on
the probe) must be removed. Based on experience, an energy flow of maximum 30kW per
probe meter must be removed within the kiln (probe with 10cm outside diameter). In extreme
cases, the energy flow varies between 1kW and 30kW per probe meter.
According to experience, the risk of deposit formation increases with reducing probe or
cooling circuit temperature. As a countermeasure, some probes are operated at higher
temperatures. This takes place by controlling the cooling circuit or by internally heating the
probe.
If a water-air heat exchanger is used, it must be taken into account in dimensioning that the
maximum cooling capacity is still sufficient even with heavy clogging of the heat exchanger
surfaces. It is also recommended to add an anti-freeze to the water in cold regions.
In order to better control incrustations and deposit build-up in and on the probes, cooling
systems were developed which enable the probes to be operated at higher temperatures. An
example is the steam jacket probe. This probe functions as a heat conducting tube, whereby
the heat of evaporation of water is used as a heat carrier to the directly flanged-on heat
exchanger (condenser).

The probe temperature is 120 to 150C, the internal pressure is 1.5 to 5 bar. The probe only
functions properly if it can be mounted almost vertically (less than 30 from the vertical axis).
This requirement virtually excludes the mounting method shown in Figure 4 in an almost
horizontal position, so that its possible applications are very restricted.
If a synthetic heat carrying liquid is used instead of water, the probes can be operated at
even higher temperatures (up to 200C) (see chapter 6.1.1).
Safety and maintenance equipment
In the event of failure of the cooling system, the probe would be damaged within a few
minutes. Countermeasures are essential. For emergency cooling purposes, water supplied
from storage tanks can be used and feed directly into the probe circuit if required.
Emergency cooling systems of this kind must also function in the event of a total power
failure. Pneumatically operated withdrawal devices have proved more reliable than
emergency cooling systems (see Figure 5). Since compressed air can be stored without
problems, pneumatic operation has the advantage that the probe can be withdrawn from the
danger, zone even in the event of a total power failure, within a reasonable time. The
automatic insertion and withdrawal device also has further significant advantages:
Regular insertion and withdrawal prevents the formation of deposits on the probe and
their build-up.
Maintenance is considerably facilitated.
All connections to the probe (hoses, cables, etc.) must be flexible.
Dust filtration in the probes
With dry filtration, the sample gas is withdrawn from the process by suction in an unpurified
condition. Dust separation takes place inside the probe or at its outer end. The produced
dust must be removed from the filter from time to time. Whether this takes place manually by
replacing the filter element or automatically by blowing back into the sampling space, the
sample gas flow is always interrupted. This interruption has a negative effect on the
availability of the measuring instrument. To prevent condensation of any kind, the filter must
have a minimum operating temperature of 150C.

Figure 5: Probe extraction device
Rail with hose guide
Suspension point
Suspension
Drive chain
Drive
Pneumatical Drive
Pneumatical
Gas
Gas tube
Shutter flap
Shutter
Mounting flange
Mounting
With wet filtration, the sample gas flows through a water curtain at the probe tip. This
produces a slurry of water and dust, which together with the sample gas discharges through
the wet probe, which in contrast to the dry probe is inclined outwards. The slurry discharges,
the gas is dried and passed to the analysers.
Wet filtration has the disadvantage that various gas components are also dissolved in water.
SO2 is completely washed out and the CO2 concentration reduces to an uncontrollable
degree. As a result of this change in the total composition, the relative contents of the other
gas components increase, so that the composition of the sample gas is no longer correct. If
the injector water circulates in a closed system, the circulated water becomes acidic and
behaves virtually inert towards CO2. However, susceptibility to corrosion of all parts coming
into contact with the water must be observed.
The level of maintenance necessary with wet filtration is slightly less and above all
technically less demanding than with dry filtration. However, in view of the distortion of
measured values and the reliable dry systems that are currently available, its application is
no longer recommended.
Filter cleaning
The dust separated in the dry filter must be removed from time to time. The degree of
clogging of the filter, including the gas sampling probe, can be determined by means of a
vacuum gauge arranged between the filter and sample gas pump. The maximum
permissible value for cleaning differs from plant to plant and must be individually determined.
Compressed air is normally used for cleaning and removing the dust from the filter. If the
filter is arranged at the outer end of the probe, cleaning should take place in two stages.

1) Loosening the dust via the sample gas pipe.
2) Blowing out the dust via a separate compressed-air pipe.
The compressed air must be free from oil and water. Oil in the purging air produces incorrect
CO in the filter and excessive moisture encourages encrustation of the filter.
With regard to gas sampling, the conditions for the wet kiln and long dry kiln are
considerably simpler than for preheater kilns. The lower temperatures usually obviate the
need of probe cooling. The lower dust load enables the use of simpler filter systems without
special cleaning, as used following the heat exchanger in the case of dry kilns (see chapter
6.3).

5.2 Other measuring points (B to G)
5.2.1 Kiln inlet chamber and riser duct (measuring point B)

Dust content 200 to 1200g/m3 (N.tr.)
In the kiln inlet chamber and riser duct, as at the rotary kiln inlet, local and time-related
concentration differences can be expected. Above the meal inlet of the lowest cyclone, there
is no longer any danger of a deposit build-up through circulated materials. However, the gas
composition no longer corresponds at this point with the gas composition in the rotary kiln
inlet. It is distorted by the reduction of CO in the riser duct and through dilution with false air
and CO2 from deacidification of the burned product.
When burning secondary fuel, the measuring point must be arranged below the secondary
fuel inlet.
With regard to gas sampling, virtually the same conditions apply as at the kiln inlet. This
applies similarly to a large extent to sampling from the calcining chambers of grate preheater
kilns.
5.2.2 Lower cyclone stages (measuring point C)

From a measurement point of view, gas analysis in this area is virtually pointless, particularly
in plants with pre-calcining. However, if measurements are still carried on a frequent basis,
these are not for measurement reasons, but the more favourable conditions with regard to
sampling (deposit formation and temperature) than further below, e.g. in the rotary kiln inlet.
A cooling system is necessary in any event.
5.2.3 Upper cyclone stages (measuring point D)

Protection of the electrostatic filter against explosions due to impermissible CO

concentrations, presupposes a rapid reaction of the gas analysis (see chapter 12.1). For this
task, gas sampling below the uppermost cyclone stage is preferable, as the residence time
of the main gas flow in the uppermost cyclone stage prolongs the reaction time of the
monitoring device. Two-section preheater systems should always be provided with two
complete measuring instruments (one for each section).
The sampling conditions are relatively simple. No cooling is necessary. In order to achieve
rapid reaction times, the maximum amount of gas must be drawn in (60 to 300 l/h). However,
as the amount of gas increases, so too does the dust content in the sampling system, so
that, depending on the particular system, it must be separated by the filter and blown back
again. More suitable for this purpose are probes whose filters are located directly in the gas
flow (internal filter probes) than those with external filters (see chapter 6.3).
5.2.4 Downstream of cyclone or grate preheaters (measuring point E)

Gas temperature for rotary kilns with cyclone preheaters 300 to 400C
Gas temperature for rotary kilns with grate preheaters 100 to 150C
Dust content for rotary kilns with cyclone preheaters 20 to 70gm/3 (N.tr.)
Dust content for rotary kilns with grate preheaters 2 to 10g/m3 (N.tr.)
Dew point temperature for rotary kilns with cyclone preheaters 35 to 45C
Dew point temperature for rotary kilns with grate preheaters 50 to 65C
At the point, similar conditions prevail for gas sampling as in the riser duct (measuring point
D). In the case of grate preheaters, this point approximately corresponds with the conditions
downstream of the grate, but less dust loading. To be taken into account for CO monitoring
in grate preheater kilns is the fact that no cooling tower is usually available to delay the
residence time of the gases in a positive sense. For this, the gas quantity can be increased
on account of the low dust loading.
5.2.5 Downstream of electrostatic filter (measuring points F, G)

Dust content >50mg/m3 (N.tr.)
In the majority of cases, the components relevant for emission are measured. Preference
must be given to the point downstream of the fan, as the gas composition is no longer
subject to change before the exhaust gas discharges from the stack. Owing to the
turbulence of the process gas in the plant sections and fans, it can generally be assumed
that the process gas is well intermixed.
Owing to the probability of filter shutdowns, gas sampling requires a dust filter as provided at
the measuring points in the raw gas flow. The high dew point makes it necessary for the
probe to be heated at the critical points. No differences exist with regard to the type of kiln
(cyclones, grate preheaters or wet kilns).

5.2.6 Coal pulverising mill
Gas analysis in the area of the coal pulverising mill serves solely for the purpose of quickly
detecting smouldering fires and/or potentially explosive gas concentrations. The possible
measuring points are shown in Figure 6.
Figure 6: Measuring points at the coal grinding plant
Fig. 6 Measuring points at the coal grinding plant
Upstream of coal pulverising mill (measuring point H)

In filter (measuring point I)

The measuring point in the filter serves especially for monitoring glow nests when the coal
pulverising mill is stationary. Owing to the large filter volume, glow nests can only be
detected quickly enough with sufficient circulation of the process gas. For this reason, 200-
300m3/h process gas should be drawn out by suction via a separate gas pipe and fed back
again at the filter inlet when the coal pulverising mill is stationary. The gas sampling probe is
then arranged in the gas pipe. When the coal pulverising mill is in operation, the gas pipe is
closed via a valve.
Downstream of coal pulverising mill fan (measuring point E)

Coal dust silo (measuring point L)

Gas temperature >60C
Dew point temperature >35C
The gas sample is preferably positioned in the silo roof.

6. GAS SAMPLING PROBES
In principle, gas sampling probes can be divided into three categories according to
temperature range:
Sampling probes in the temperature range 900 to 1500C
Sampling probes in the temperature range 500 to 900C
Sampling probes in the temperature range up to 500C
The following assessment of various gas sampling probes is based on the manufacturers
documents as well as our own experience and information from the works, if available.
6.1 Sampling probes in the temperature range 900 to 1500C

This category includes gas sampling at the inlet of preheater kilns with and without pre-
calcining.
6.1.1 Siemens type FLK

Brief description
Compact gas sampling system with special cooling fluid permitting probe operating
temperatures up to 230C. The condensation of alkali vapours and in turn the danger of
incrustations is drastically reduced. The suction opening is arranged on the side of the probe
tube and should be located in the flow slipstream. This arrangement offers selective dust
separation. An electrically or pneumatically operated insertion / withdrawal device is an
integral part of the probe.
Design: Extremely compact and professional design. Oval stainless
steel sampling tube.
Suction opening: Lateral, about 30mm in diameter
Mounting tube: 325mm diameter (relatively large)
Dust removal filter: External, heated, blow-back with compressed air, large volume
Maximum probe length: 3200mm
Comment
Despite effective measures, stubborn blockages can occur, particularly when kiln operation
is loaded by high alkali circuits. Owing to the lateral arrangement of the suction opening, the
gases within the probe tube flow through an elbow. The removal of incrusted deposits from
this elbow is far more difficult due to poor access than if the gases were drawn out by
suction at the end of the stem and through a straight tube. In view of the turbulent flow
conditions at the sampling point, it is doubtful whether lateral suction results in selective dust
separation.
The large overall volume of the dust filter permits a relatively large gas throughput. The
resulting delay and damping of the indication is of little significance from a measurement
point of view.
Available for the drive of the withdrawal device is either an electric motor or compressed-air
drive. As compressed air can be stored without problems, the compressed-air drive offers
more reliability than an electric motor in the event of a power failure. Emergency operation
with crank handle is also provided.
With difficult kiln operation, alkali condensation can occur and block the entrance of the
probe, as already mentioned, despite increased operating temperature.

6.1.2 Harman & Braun type 60S (formerly type 13)
Brief description
Gas sampling probe with closed cooling system (cooling medium water) and temperature
control. Integrated emergency cooling system with fresh water. Lateral arrangement of two
suction openings (Figure 7).
Comment
Widely used. Functions well in non-extreme alkali conditions. The same remarks apply here
with regard to the lateral arrangement of the suction openings as for the Siemens system.
With large amounts of dust, the filter must be blown-out at brief intervals, The resulting loss
of signal can have a critical effect on the availability of the measuring signal.
The operating temperature is too low to completely prevent alkali condensation. An
automatic insertion and withdrawal device is optionally available.
Figure 7:
1 2 10 11 3 7 5 8
9
12
12
6 5
1 Gas inlet holes 7 Ceramic fliter
2 Gas tube 8 Protection case
3 Filtering unit 9 Connecting for air blaster
4 Gas outlet 10 Built in tube
5 Compressed air inlet 11 Mounting flange
6 Ring heating element 12 Cooling water in/outlet
Fig. 7

6.1.3 Hartmann & Braun type H (Holcim probe)
Brief description
The main features of this probe are the automatically functioning mechanical removal of any
incrustations at the probe inlet as well as a new type of filter system (see Figure 8). Cooling
takes place in a closed system (cooling medium water) with temperature controller.
About 20 probes of this type are in use worldwide and are functioning efficiently. Marketing
takes place via Messrs ELSAG BAILEY Hartmann & Braun in Frankfurt under the
designation type H.
Design: Probe body of alloyed steel

Suction opening: At end of stem, about 40mm in diameter
Probe temperature: External, 30 to 85C, internal up to 250C
Mounting tube: 150mm in diameter
Dust removal filter: Internal, heated filter of sintered metal, cleaning with
compressed air
Maximum probe length: 3000mm
Thanks to its modular concept, the probe can be provided with the same insertion and
withdrawal device as type 60.
Figure 8:
1 2 3 4 7 10 11
9
12
12
5 8 6
1 Measuring gas inlet 7 Heated gas tube

2 Piston 8 Protection case
3 Filter 9 Anschluss fr Luftkanone
4 Tube 10 Montagestutzen
5 Expansion seal (Stainless steel) 11 Montageflansch
6 Air cylinder 12 Khlwasseranschlsse
Fig. 8

Comment
This sampling system has proved highly reliable in practical use, even under difficult
conditions. The availability satisfies the high requirements of high-level control systems.
The maintenance costs are low thanks to the simple and maintenance-friendly design of the
system.
6.2 Sampling probes in the temperature range 500 to 900C

With regard to the sampling probes, similar conditions prevail in this temperature as at
higher temperatures (e.g. at the kiln inlet of preheater kilns).
However, the danger of incrustations is lower. Cooling of the probes is essential down to
500C, so that in principle, the same sampling probes can be used as in the higher
temperature range.
Temperatures between 500 and 900C are typical for cyclone preheater kilns in the area of
the lower cyclone stages (measuring point C). As already noted, measurements there are
almost pointless, particularly in plants with precalcining. If measurements are still to be
carried out despite this, steam jack probes are considered ideal for this purpose when
mounted vertically in the cyclone roof.
6.3 Sampling probes in the temperature range below 500C

Typical for the application of such probes are the measuring points downstream of the heat
exchanger or grate preheater, upstream or downstream of the electrostatic filter as well as at
the kiln inlet of wet or long dry kilns. Cooling is unnecessary. With regard to the place of use
or purpose of measurement, there are certain differences in the design of the probes (e.g.
mounting length). Decisive for the quality of these probes is the dust filtration. The discovery
of fine-pore filter tubes of sintered stainless steel brought about great advances. The level of
maintenance is low, a block-back device is unnecessary. The sample gas is clean and
available for weeks to months without interruption. Probes with such filters are not available
from established suppliers. They still use fine ceramic filters that enable a comparatively long
useful life. However, it must be ensured that the gas conducting tube is heated at the
penetration point (Figure 9).

The optimal porosity of sintered metal filters is only 0.5m. As the main proportion of the dust
particles is larger, the pores clog at a correspondingly slow rate. According to experience,
they have a useful life of four weeks to several months, depending on the properties of the
dust. If the pressure drop becomes excessive due to clogging, the filter tubes can be
cleaned with acid.
With wet kilns and long dry kilns, the probes must be lengthened, depending on the size of
the smoke chamber. In order to ensure the necessary mechanical strength, supporting tube
of a larger diameter must be mounted, into which the probe is inserted.
A further important element is the heating in the penetration zone from the measuring
channel to the outside. In order to prevent any kind of condensation, the temperature must
not fall below 100C at any point.
For all gas conducting parts, stainless steel, e.g. No. 1.4541 or 1.4571 must be used.
Figure 9: Gas sampling probe, up to 500C
Thread e.g. 25 x 1.5
430 Length as required max 1500
10
ca 40
25 (1")
1 2 3 4 5 6 7 8
1 Stainless steel closing

2 Filter but: Typ 1400.1.000-.18-0.5 threaded on both sides e.g. M 25
Manufacturer:
Mott Metallurgical Corp.
Farmington Industrial Park
Farmington CT 06032
Fax No (USA) 203 674
3 Gas condcuting tube, Outside diam. 10 mm, inner diam. 6 mm, stainless
4 Centering discs approx. every 500
5 Carrier tube. Outside diam. 40 mm, wall thichness 2 mm, stainless
6 Heating coil Heated lenght approx. 300 mm ca. 200
7 Mounting flame
8 Sample gas connection

7. SAMPLE GAS PREPARATION
The purpose of sample gas preparation is to supply the downstream gas analysers with
sample gas, so that it conforms qualitatively and quantitatively with the specifications of the
analysers. Such an arrangement normally consists of the following elements:
Supply > Sample gas pipe
Drying > Sample gas cooler
Delivery > Sample gas pump
Dosage > Controller
Distribution > Pressure control valve
Calibrating option > Manual or automatic
Experience has shown that the availability of a measuring instrument depends essentially on
a properly functioning sample gas sampling and preparation. Insufficient sample gas
preparation leads to fouling in the gas analysers and not infrequently to expensive damage.
7.1 Sample gas pipe

The probe itself as well as the sample gas pipe between the sampling probe and sample gas
cooler should be heated to between 100 and 150C (for measurements with FID devices up
to 200C). This heating has the following purpose:
To prevent condensation and freezing of the condensate at low temperatures, because:
Condensation in the sample gas pipe can distort certain measured values.
Incrustations and blockage of the sample gas pipe are accelerated by the formation
of condensate.
To delay the conversion of NO into NO2 during emission measurement.
Heated sample gas pipes are available in various qualities:
Low price pipes with a self-regulating heating band and
Pipes with resistance heating and separate temperature controller
For selection purposes, it should be noted that with heating band controlled pipes, the
temperature is limited to a maximum of 105C. This temperature does not normally suffice
for process measurements but for emission measurements. The following specifications
must be observed:
Inner tube material: Teflon
Inner tube dimensions: Outside diameter 6mm
Inside diameter 4mm
Connection ends: Clamping ring tube fittings
Control: Thermocouple with controller
Protective sleeve: PVC or metal

Heated pipes are relatively expensive. When planning a measuring system, the sample gas
pipes must therefore be as short as possible. The following points must be observed:
a) If only O2, CO or NO is measured, heating can be dispensed with if the sample gas pipe
(several metres) is short and can be laid with uninterrupted gradient from the sampling
point to the ample gas cooler (no siphons).
b) Possible installation of the sample gas cooler in the vicinity of the sampling point.
For automatic compressed-air cleaning via the sample gas pipe, it must be ensured that this
has sufficient pressure resistance.
For connections, high-quality stainless steel clamping ring fittings, e.g. Swagelok, Serto or
similar, must generally be used.
7.2 Sample gas coolers

Sample gas coolers serve to remove most of the water vapour from the sample gas at
temperatures of about 3C and keep the residual content at a constant value. The latter is
important particularly for emission measurement. The temperature indication of the sample
gas cooler normally shows the operating temperature of the cooling medium. If the sample
gas cooler is subject to excessive load or ambient temperature, the sample gas temperature
at the cooler outlet can be considerably higher than the indication. The dew point
temperature then no longer corresponds with the cooling medium indication.
The sample gas cooler is usually accommodated in a cabinet together with the analysers
and the remaining sample gas preparation apparatus. Instead of using long heated sample
gas pipes, it may be advantageous to install the sample gas cooler immediately downstream
of the sample gas probe in a separate equipment cabinet.
Condensate collecting tanks have the disadvantage that indication of the measured values is
greatly delayed due to the additional volume. The produced condensate should therefore be
pumped out using automatic hose pumps or diaphragm liquid pumps.
7.2.1 Specifications of a suitable sample gas cooler
Dew point at outlet +5C (1C)

Dew point stability 0.25C
Temperature at sample gas inlet max. 180C
Dew point at sample gas inlet max. 80C
Gas flow max 100l/h
Ambient temperature +5 to +45C
Cooling capacity 860kJ/h, 25C
Material of gas conducting parts Teflon, PVDF
Permissible gas pressure min. 3 bar
Volume max. 100cm3
Sample gas connections 6mm or 1/4
Time until ready for measurement max. 30 minutes
Also available for the majority of sample gas coolers are moisture monitoring devices at the
sample gas outlet. Such devices are urgently recommended.

7.3 Sample gas pump
Mainly diaphragm and compressor pumps are used. The sample gas volume necessary for
the gas analysers is between 30 and 100l/h. In order to keep indication delays to a minimum,
a larger sample gas volume is often extracted from the process gas. The excess sample gas
can be diverted via a bypass shortly upstream of the analyser. If possible, a large part of the
measuring instrument should be operated under pressure, as false air can infiltrate during
suction phase.
Sample gas pumps installed upstream of the sample gas cooler must be heated. This
arrangement has the advantage that no false air is sucked in the event of leaks in the
sample gas cooler. However, according to experience, heated pumps are more susceptible
to faults than cold operated pumps, so that arrangement downstream of the sample gas
cooler is still more advantageous.
The pump capacity should be generously dimensioned in respect of the suction pressure.
The flow rate must be dimensioned, so that the available gas flow about 1.5 to 2 times
higher than that required by the analysers according to specifications. Throttling of the flow
preferably takes place on the low-pressure side.
With throttling on the pressure side, a pressure control valve must ensure that the pressure
downstream of the pump does not increase to an unnecessary degree. Significant pressure
drops in the sample gas encourages the formation of powdery salts (white powder). The
following diagram shows the characteristics of a suitable sample gas pump.
Vacuum
Excess pressure
Gas flow [l/min] at 20C
Figure 10:

7.4 Flow control and distribution
The arrangement for control and distribution of the sample gas depends on the number and
specifications of the analysers to be supplied. Multi-component analysers with only
connection solve the problem of gas distribution internally. However, as soon as two or more
analysers have to be supplied with gas, the inlet pressure must be taken into account in
distribution. At a low supply pressure, the analysers can be connected in series. If analysers
have to be connected with varying supply pressure, they must be connected in parallel via
appropriate gas distribution.

7.4.1 Simple gas control
Figure 11:
1 3
7 13
8 9
12
2
6
5
4 11
10
1 Sample gas inlet 8 Analyzer 1

2 Sample gas cooler 9 Analyzer 2
3 Fine filter combined with wetness detector 10 Condensate pump
4 Sample gas pump 11 Calibration gas input
5 Changeover valve measurement / Calibration 12 Overpressure valve
6 Throttle valve, Gasflow setting 13 Min. flow contact
7 Floating body flow meter
A B C
The flowmeter monitors and controls the set sample and test gas volume. Rotameters fitted
with variable limit transmitters are normally used for setting the flow rate. The signal
transmitters should signal especially when the necessary flow is not reached. This is
particularly important for gas analysers with safety functions.
As already mentioned, not all gas analysers have the same conditions with respect to supply
pressures. Optical analysers normally require less pressure than oxygen analysers, for
example, which use the paramagnetic measuring principle. However, special attention must
be given to the supply pressure of the analysers when dimensioning the sample gas
preparation apparatus. As soon as several analysers have to be operated, parallel
connection is recommended, despite the additional work involved. For this purpose, the
sample gas flow must be divided appropriate to the number of gas analysers to be supplied.
The following flow diagram shows how this division can take place. The sample gas pump
supplies a common supply pipe, which is under increased sample gas pressure. A spring-
loaded pressure control valve controls the pressure in this pipe. The excess sample gas is
diverted. A further, slightly more elaborate option, is to use a separate sample gas pump for
each analyser.
Parallel connection with pressure control valve
Figure 12:
A B C
Parallel switching with 3 pumps

7.4.2 Electronic gas volume control
The available electronic flowmeters are reliable and should be used to an increasing extent
for automatic gas volume control purposes.
These are flowmeters whose measuring principle is based on a thermal measuring bridge,
whereby the transported heat of a flowing substance is used as an indicator. The calibrated
reference quantity is normally air. For other gas compositions, appropriate correction values
(e.g. CO2) must be used for conversion.
Figure 13:
Setpoint
S Power
Supply
Flow Controller
Fig. 11 Electronic Gas flow Control
Tests with automatic control have confirmed that the reliability of the gas supply can be
considerably improved with this system. It offers decisive functional advantages compared
with rotameters and is less susceptible to faults. A constant gas flow positively affects overall
gas analysis in every respect. Since the controllers have a throttling effect on the gas flow,
they are preferably arranged on the suction side of the pump. For flow indication purposes, a
volume-proportional signal is available. This enables ideal monitoring via the control system.
In addition, a rotameter (without needle valve) can still be used for monitoring locally.

7.5 Sample gas filters
7.5.1 Coarse filters

Coarse filters are normally installed inside or downstream of the probe, so that a large
proportion of the dust is already separated before the sample gas enters the sample gas
preparation apparatus. If sintered metal filters with a porosity of <0.5m are already used in
the probe, the sample would be sufficiently clean. Nonetheless, fine afterfiltration directly
upstream of the analysers is recommended, as impurities can also infiltrate the system
downstream of the probe. Ceramic filters of silicon carbide, mineral fibres or filter casings of
borosilicate glass fibres are also frequently used. Silicone carbide composites have a high
separation efficiency of about 99.9% at an average grain diameter of 1.2m and are
temperature resistant up to 660C.
The disadvantage of all these filters compared to sintered metal filters, apart from their lower
separation efficiency is that they have to be cleaned and blown out or replaced at shorter
intervals
7.5.2 Fine filters

For separation of the dust particles and sublimated salts left behind in the sample gas,
membrane filter, e.g. of glass fibres or PTFE should be provided.
Recommended is a combination of membrane filter and condensate monitor. The
condensate monitor measures a change in the electrical conductivity between two
electrodes. If a limiting value is exceeded, the sample gas pump is stopped and the fault is
indicated in the control room.
7.5.3 Valves
Changeover valves
If automatic blowing out of the filter and/or probes is required, the gas path must be able to
be diverted automatically. For this purpose, stainless steel solenoid valves or mechanically/
pneumatically operated ball sliding valves of Teflon (housing) and stainless steel (ball) are
used.
Valves installed in moist sample gas (upstream of the sample gas cooler) must be mounted
on a temperature-controlled valve plate in order to prevent corrosion as a consequence of
the dew point not being reached.
Valves, particularly solenoid valves, have the tendency to leak, even when slightly fouled.
They enable a high degree of automation of the measuring instrument, e.g. automatic
calibration, however, experience has shown that their susceptibility to faults almost cancels
out the convenience they are expected to provide (see chapter 11.7).
7.5.4 Pressure control valve

The purpose of the pressure control valve is to ensure that the supply pressure of the
analysers remains fairly constant. The excess gas delivered by the pump discharges via a
spring-loaded body. The nominal pressure of the valve must be slightly higher than the
highest specified supply pressure of all parallel connected analysers.

7.6 Adjustment
The necessity of regular adjustments depends on the purpose of the analysis and the
stability of the analysers. For emission measurements, a rotation of one to two weeks is
sufficient, provided the regulations in the country concerned do not contain any tighter
requirements. During adjustment, a preferably manually-operated two-way directional control
valve switches the gas supply to a bottle containing test gas. Newer gas analysers are
normally provided with an automatic adjusting device. By means of an internal or external
command a glass cuvette filled with the respective gas is inserted into the sensor and the
analyser is automatically readjusted. During this time, the last measured value remains
indicated, so that the subsequent signal processing system is not disturbed by the
procedure. With older analysers that do not have this type of automatic device, operating
personnel or the process control system must be informed of the adjustment. With CO
monitoring systems, activation of the automatic filter shutdown system must be prevented.
Analysers with so-called calibrating cuvettes have the big advantage that no more test gas is
necessary. Manufacturers and most nationally customary regulations accept this type of
adjustment over a period of 2...3 years. After this time, the analysers must be overhauled
and inspected and readjusted by an appropriately licensed institute.
7.7 Further information on sample gas preparation
7.7.1 Sample gas discharge

No sample gas must discharge into the analyser room. Every gas outlet must be connected
to a pipe and discharged to the atmosphere or returned to the main gas flow. Pipes with
large cross-sections must be used in order to prevent any back pressure.
7.7.2 Tightness test

An option should be provided to easily connect the sample gas pipe gas-tight to a nitrogen
bottle directly downstream of the sampling probe.
The complete measuring apparatus is purged with nitrogen with the gas preparation
apparatus operating normally. The oxygen analyser shows zero indication soon afterwards.
After several minutes, the nitrogen supply is interrupted, so that a vacuum is produced in the
system. Caution! The regulating valves on the test bottles are often not vacuum-tight. If the
measured value of the oxygen analyser should increase again within a space of about 20
seconds, this means that there is a leak in the system. An increase of several per cent within
several minutes is considered normal.
7.7.3 Filter condition monitoring

The suction pressure is a good indicator of the condition of the filter or filters between the
probe and sample gas pump. By measuring this pressure with a pressure transmitter or
contact pressure gauge, the automatic cleaning process can be initiated when a certain level
is reached. As a result of this optimising measure, the measured signal is not interrupted
unnecessarily and the filters are cleaned or blown back before being irreversibly clogged.

7.7.4 Analyser room
The accuracy and reliability of gas analysis essentially depends on temperature variations,
vibrations, draughts, dust, moisture and electrical interference fields being avoided. For this
reason, the gas analysers including sample gas preparation apparatus should be installed in
a closed room or equipment cabinet that can be locked to prevent unauthorised access. In
the analyser room, the individual instruments and apparatus must be arranged with an
emphasis on clarity and easy accessibility, so that rational operation and maintenance of the
measuring equipment is ensured.

Figure 14:
13 9
A
10
14
1 18
8
Calibration Operation
16
3 4 12
2
7
6 17
15
Blow-out Operation
1 Gas sampling probe

2 If necessary: Blow-out device for outside of coarse filter)
3 Heated sample gas pipe
4 If necessary: Heated magnetic valve for blow-out device inside the filter
5 Sample gas cooler
6 Condensate pump
7 Suction pressure measurement (Manometer or pressure transmitter)
8 Sample gas pump
9 Fine filter with humidity monitoring
10 Changeover valve Operation/ Calibration (Manual actuation)
12 Control valves with floating body flow monitoring
13 Gas analyzers
14 Exit of superfluous sample gas
15 Condensate discharge
16 Calibrating gas
17 Cabinet air conditioning with suction filter
18 Cabinet
Fig. 12 Complete gas preparation

"HOLDERBANK"

7.8 Space requirement
For the analysers and sample gas preparation apparatus, a double 19 equipment cabinet
with one field respectively for the gas preparation apparatus and analysers is normally
sufficient. However, other equipment is often installed in this clean, air-conditioned analyser
room. In such cases, it must be ensured that sufficient room is provided to enable
maintenance personnel to move freely and that there is adequate room for material (e.g. test
gas bottles, tools, etc.) and additional space.
7.9 Location
In principle, the analyser room should be located as close as possible to the measuring
point. This aspect is particularly important for CO monitoring, where fast indication is
necessary. Unheated sample gas pipes must be laid descending to the sample gas cooler. It
is therefore advantageous when the analyser room is located one floor below the sampling
point.
If space is available at a suitable location that is not exposed to excessive dust, heat
radiation and noise, a switchgear cabinet is sufficient for accommodating the measuring
equipment. A separate, closed room is unnecessary under such conditions. The volume to
be air-conditioned can be considerably reduced as a result.
7.10 Climatic conditions

The temperature in the analyser room or cabinet should be able to be regulated between 15
and 25C. For this purpose, heating and ventilation is necessary; a cooling unit must be
provided to counteract heat radiation. The ventilation fans must be provided with efficient
dust filters. All filters produced during gas analysis must be discharged to the atmosphere.

8. ANALYSERS
The determination of individual components in the sample gas is based on various physical
basic principles, such as:
Light absorption
Paramagnetism
Flame ionisation and
Solid-state electrolytic systems
All newer analysers process the analogue signals output by the sensors digitally.
Digitalisation has provided the following advantages:
Increased operating convenience (menu guidance)
Multi-channel technology
Automatic adjustment
More compact design
Increased stability
Remote monitoring and control via modem and/or data bus
The internal functions of the analysers will not be dealt with here. Sufficient information in
this respect is provided in the manufacturers documents and technical literature. Important
for practical application is knowing how to use the various measuring methods effectively.
The manufacturer plays a secondary role in analyser selection from a qualitative point of
view. Measuring equipment faults are rarely attributed to faulty analysers. It goes without
saying that only types designed for industrial purposes are suitable for use in cement works
and certainly no laboratory equipment. Suitability for industrial purposes means, for example:
Solid, totally enclosed housing (degree of protection IP 65 or IP 54)
Insensitivity to vibrations
Minimum temperature application range 15 to 35C
Insensitivity to system disturbances
Isolated (floating) output signals
The majority of established manufacturers are in a position to meet these specifications.
Service and spare parts availability are therefore more important than the purchase price. If
measuring equipment is already available and the above criteria are met, there is no reason
to change manufacturers when purchasing new equipment or additions.

8.1 Infrared absorption
With the exception of oxygen, the majority of interesting gas components can be measured
by infrared absorption. Available are analysers for extractive and in-situ application. In the
main, infrared is used for the extractive method. The following advantages are offered by
these NDIR (non-dispersive infrared) analysers:
Broad application spectrum
Relatively simple method, favourable price
Robustness
Low wear, long useful life
Good stability
The not particularly high sensitivity of the NDIR analysers does not normally pose a problem
in the cement industry. The smallest measuring range, e.g. for CO is about 100ppm, for NO
about 500ppm and for CO2 about 20 %. Cross-sensitivities to water vapour and CO2 are
present. Water vapour as a disturbing component plays a secondary role when the gases in
the sample gas cooler are dried with sufficient stability. The cross-sensitivity of CO2 is
usually within the tolerance when compensated in the analyser. The influence can be further
reduced when an amount of CO2 corresponding to the average value of the sample gas is
mixed with the test gas. As the CO2 content in the cement process is extremely high,
unestablished analysers should be tested in this respect prior to their use.
The most well known manufacturers today produce multi-channel NDIR analysers, enabling
the prices per measured component to be distinctly reduced. The majority of manufacturers
also offer analysers in different quality classes (low-cost analysers). As a rule, the low-cost
analysers are not temperature-stabilised, but temperature-compensated and are therefore
not as stable as the analysers at the top end of the price scale. However, if external
fluctuations are kept to a minimum, the temperature-compensated analysers meet the
requirements in the majority of cases.
8.1.1 Space case, sulphur dioxide (SO2)

In principle, NDIR analysers are suitable for the measurement of SO2. However, tests on kiln
systems with cyclone preheaters have shown that insufficient SO2 is measured with
extractive sample gas sampling and preparation (in accordance with DIN 2462, page 4). The
magnitude of the deficiency can vary considerably and can therefore not be calibrated.
These measuring errors are attributed to chemical reactions of the SO2 with NH3 (ammonia)
in the sample gas cooler (see chapter 11.2). The following options are available to eliminate
these errors:
Acidification of the condensate by addition of phosphoric acid in the sample gas cooler
(TV) tested method developed by and available from Hartmann & Braun).
Use of a hot analyser (measurement with moist sample gas).
In-situ measuring method
With the addition of about 8 to 12ml phosphoric acid (5%) in the sample gas cooler, the pH
of the condensate reduces to a value below 1.5. The reaction of the SO2 with ammonia is
accordingly prevented.
It is obvious that hot analysers, whose entire measuring system must function at
temperatures between 150C and 250C, are more susceptible than cold analysers.
According to experience, the heated sample gas pump causes the greatest problems.

With in-situ measuring instruments, no distortion through chemical reaction is noticeable. As
part of comparative tests, they showed to be more reliable than measuring instruments with
extractive sampling and sample gas cooler. However, absolute measured values in mg/Nm3
are initially available after calibration using a reference measuring procedure. In addition, if
dry standard values are required, the correction must be corrected with the moisture and
temperature of the sample gas.
8.2 Ultra-violet absorption

The difference of the NDUV analysers compared with NDIR is mainly that they are less
cross-sensitive to water vapour and CO2 and higher sensitivities can be achieved. In the
cement industry, NDUV analysers were initially used for the measurement of NO
components. Meanwhile, the less expensive NDIR analysers are being more widely used for
NO and have proved reliable.
8.3 Paramagnetism
Oxygen does not have a usable spectrum either in IR or UV light. In future, laser analysers
/chapter 8.6.3) will be the first to be able to measure oxygen with light in addition to other
components. The paramagnetic properties of oxygen (oxygen molecules are strongly
attracted in a magnetic field) are still used for oxygen measurement, however with a
declining trend. In practice, two methods are widely used:
The magnetic torsion balance
Thermomagnetic oxygen measurement
8.3.1 Oxygen analysers

In recent years, analysers operating on the torsion balance or also dumb bell principle have
become increasingly popular.
The measuring effect is based solely on the magnetic forces of the oxygen molecule and is
therefore comparatively less cross-sensitive to other gas components. However, the
measuring chamber reacts sensitively to dirt and condensate.
The paramagnetism of oxygen reduces with increasing temperature. This thermomagnetic
effect is used in so-called annual chamber or hot wire analysers for O2 measurement. The
advantage of these analysers is that the measuring chambers contain no sensitive internals,
so that they are less sensitive to dirt and are also easier to clean than analysers that operate
on the torsion balance principle. However, the measuring method is cross-sensitive to other
gases and only functions accurately with a known or more constant accompanying gas
composition.
8.4 Flame ionisation (FID)

This measuring method is used for measurement of the volatile content of unburnt
hydrocarbons (VOC) (Volatile Organic Compounds). Organic carbon compounds contained
in the sample gas are ionised in a hydrogen flame. The ion quantity is almost proportional to
the number of carbon atoms involved in combustion. The ionic voltage is detected with the
aid of an electrode and brought to indication via a high-resistance amplifier. Propane gas in
nitrogen is normally used for calibration purposes.
Continuous measurements with the flame ionisation detector (FID) are increasingly
prescribed, as soon as alternative fuels are burned. FID analysers are unfortunately
susceptible to faults and require considerable maintenance by suitably trained personnel.

The entire sample gas preparation system must not fall at any point below the sample gas
temperature, i.e. the sampler, sample gas pipe and pump must be heated. The temperature
in the ionisation chamber of the FID analyser is normally about 200C.
Modern multi-component systems such as OPSIS are capable of measuring the majority of
interesting VOC compounds as individual components. Programs for direct measurement of
the volatile hydrocarbons (VOC) have also been developed. Initial tests have shown almost
one hundred per cent correlation with FID values. This means that in countries where non-
explicit FID measurements are required, an OPSIS or perhaps also some other type of multi-
component analyser will preferably be used at a later time.
8.5 Solid-state electrolytic systems
8.5.1 Zirconium dioxide

Zirconium dioxide (ZrO2) has the characteristic of building up a differential voltage at varying
partial pressures and temperatures between 500 and 1000C due to the flow of oxygen ions.
This effect is used for the purpose of oxygen measurement by exposing both sides of a ZrO2
membrane heated to about 800C to varying oxygen concentrations (sample gas and
reference gas). The supplied voltage is tapped and measured on the two surfaces of the
membrane with thin, porous platinum electrodes. The measured potential difference
increases exponentially with the oxygen concentration difference. To be noted is that the
indicated measured values are based on gas in a moist condition.
In addition to the advantageous zirconium dioxide based in-situ measuring instruments,
there are also extractive types that are hardly worth mentioning.
8.5.2 Other electrochemical measuring methods

Electrochemical measuring cells (EC cells) have the characteristic of supplying a voltage in
the presence of certain components proportional to the concentration. As they are small and
lightweight and require only a small number of peripheral devices, they are used mainly for
portable measuring equipment (e.g. flue gas controls), although they had proved unsuitable
for use in the cement industry in the past. However, in recent years, suitable EC cells have
been developed that offer advantages mainly for oxygen measurement. The oxygen sensor
may not be used when the accompanying gas contains H2S, chlorine or fluorine containing
compounds, as well as heavy metals and aerosols. The guaranteed useful life is three years.
The many years of good experience prompted the majority of manufacturers to complement
the NDIR analysers with EC cells for oxygen measurement.
8.6 Multi-component measuring systems

Development in the field of analysis has witnessed significant advances. Thanks to modern
microprocessor technology, large computer capacities and high speeds, measuring methods
are employed that enable disturbing secondary effects to be compensated by calculation
and new techniques to be used, for example:
Gas filter correlation technique (GFC)
Fastfourier transformation technique (FTIR, FTUV) Differential Optical Absorption
Spectroscopy (DOAS)
Laser Diode Spectrometer (LDS), etc.

Without going into details about the individual techniques, the following points are of
importance:
Multi-component analysers are less expensive for simple measuring tasks with more
than one component than individual analysers in all respects.
When using alternative fuels, the necessity may arise that new, particularly critical
pollutant components have to be measured. This aspect justifies the use of flexible,
programmable measuring instruments.
The suitability of multi-component measuring instruments in the cement industry has been
demonstrated by many years of reliable use. In the following, three instruments will be
presented that have proved reliable and are state of the art. These are both extractive and
in-situ measuring instruments.
8.6.1 Sick GM 31
Sick, Waldkirch (D), is a pioneer of the design and construction of in-situ measuring
systems. The instruments GM 21 for dust, and later, the GM 30 for NO SO2 and dust, have
been used successfully for more than ten years. The successor instrument, GM 31, differs
considerably from its predecessors in two ways:
Dust can no longer be measured with this instrument.
The flanged design as a so-called lance instrument does not measure the whole channel
cross-section.
The GM 31 is designed for the components SO2, NO, NH3 and NO. A maximum of three of
these components can also be measured simultaneously. Despite the lack of the benefit of
dust measurement, the GM 31 is a low-cost emission measuring instrument, which
complemented with a dust measuring instrument for systems not subject to particular
requirements, meets all specific requirements.
8.6.2 OPSIS AR 620; ER 650

Within the Opsis measuring system, a light beam is projected on to a receiver and then
passed through a glass fibre cable to the Opsis analyser.
In emission measurement, the light beam or measuring path continues through the interior of
a stack or exhaust gas duct. Every gas absorbs light in known spectral regions of the total
light spectrum, in a manner that is characteristic of the respective gas. This enables the
analyser to detect and measure gases defined by the user. The results can be indicated in
real-time or used for statistical and further processing purposes.
Opsis uses a scientifically researched principle for the identification and measurement of
concentrations of various gases: Differential optical absorption spectroscopy (DOAS) which
is based on the Lambert-Beer Law. It describes the correlation between the absorbed
quantity of light and the number of molecules in the light path..
As every gas has its own unmistakable absorption spectrum, the so-called finger print, the
concentrations of several different gases can be determined simultaneously in the light path.
With the DOAS method, a light beam is sent from a special light source - a Xenon high-
pressure lamp - over a specific path; by means of complex, computer-aided calculations, the
light losses through molecular absorption along the path are evaluated and analysed. The
light of the Xenon lamp contains both the visible spectrum as well as the ultra-violet and
infrared wavelengths. The light is detected by a receiver and passed on via a fibre optic
cable to the analyser. This fibre optic cable makes it possible to install the analyser at a
sufficient distance from any harmful environmental influences at the measuring point.

The analyser consists, among other things, of a high-performance spectrometer, a computer
and the associated control unit. The spectrometer breaks down the light with the aid of an
optical grid in narrow wavelength bands. This optical grid can be adjusted for examining an
optimal wavelength region.
The light is converted into electrical signals. A narrow slot moves at high speed across the
detector; this combines a large number of instantaneous values that provide an image of the
spectrum in the relevant wavelength region This scanning procedure is repeated a hundred
times per second; the recorded spectrums are added up in the multi-channel memory of the
evaluation unit until their evaluation.
Evaluation is carried out individually for each wavelength region and is based on the
comparison of absorption curves. The respectively last recorded absorption spectrum is
compared with a computer calculated spectrum. The calculated spectrum consists of a
summation of the reference spectrums for the respective evaluation. The computer alters the
size factors for each reference spectrum until optimal correspondence is achieved, so that
the various gas concentrations can be calculated with high accuracy.
In order to be able to measure a diverse range of gas components as possible, the light
spectrums must be divided into IR and UV. Measurement is similarly divided into a UV
spectrometer and an IR interferometer. Depending on the components to be measured, both
or one of the two instruments is used. Both instruments are usually necessary for
comprehensive measurement of cement kiln exhaust gases.
The system is approved as a recognised emission measuring instrument worldwide.
For calibration: In principle, each component must be calibrated at last once (as part of the
commissioning procedure). This takes place either via a convention method or by inserting
cuvettes circulated by test gas in the light path. In this simple way, the work involved in the
normally usual convention method is avoided. In two works of the ChB, these measuring
instruments have proved highly reliable and may be referred to as standard-setting
technology.
8.6.3 LDS 3000 diode laser of AltOptronic (Sweden)

The design is basically comparable with OPSIS. However, as a light source, a
semiconductor laser is used. The LDS 3000 enables measurements to also be carried out in
aggressive atmospheres with a high dust content; this takes place in a way that the special
features of the diode laser are utilised in conjunction with a patented signal evaluation
method. The instrument can distinguish whether a change in transmission is due to dust or a
change in the gases to be measured.
The LDS 300 can be used for measurements in atmospheres with a varying dust content up
to 50g/m3, depending on the size of the dust particles.
The light spectrum of the semiconductor laser can be adjusted to the absorption lines of the
gas component to be measured via temperature and current. This ensures that
measurement can only take place on a selected absorption line of the respective gas. The
absorption lines of the remaining gas components are on other wavelengths and do not
influence measurement.
In the three years in which the LDS has been used, continuous tests were carried out in
order to determine the stability of the system. It emerged that no recalibration of either the
zero or measuring range were necessary.
The diode laser is located in the main unit, from where the laser beam is passed via an
optical fibre to the sensors at the measuring points. The distance between the instrument
and the measuring point can be up to 1000m. In the main unit, the laser beam is scattered in

an optical distributor and so enables simultaneous measurements on three different process
measuring levels.
The short measuring time of the LDS 3000 provides direct information (t90<1 sec). This is
made possible by in-situ measurement, the high efficiency of the diode laser and appropriate
signal evaluation. The short resonance time depends on the number of measuring points, as
the laser light is scattered in a fibre-optic distributor and passed on to all measuring points
simultaneously. Each measuring point can therefore function independently of the other
measuring channels.
Thanks to its unusually short reaction time, this measuring system was to be able to
considerably improve the CO monitoring of electrostatic filters. Measurement directly
following WT would perhaps be possible, since the manufacturer mentions that up to 50g/m3
dust has no influence. Unfortunately, the measuring distance this value applies to is
unavailable.
The following gases can be measured at the present time:
Gas ppm, at 1m mg/m3, at 1m
NH3 0.6 0.4
O2 200 260
H2O/NH3 200/1 150/0.7
H2O 0.3 0.2
Hcl 0.3 0.4
HF 0.3 0.4
Also other gas components can be measured on request.
The LDS 300 is one of the first laser analysers to appear on the market. At present, the price
is (still) relatively high. Development is absolutely in line with the trend and could also be
used advantageously in the cement industry in the near future.
8.6.4 Advance Cemas FTRI of Hartmann & Braun

Development has also continued with extractive measuring instruments. Hartmann and
Braun builds in addition to the maximum four-component measuring NDIR systems a
modern FTIR (Fastfourier Transformation Infrared) spectrometer, which has identical
functions to those of the OPSIS, but is designed as an extractive measuring instrument. It
offers high selectivity as well as easy upgrading to additional infrared components. Based on
the measuring principle and automatic zero correction, calibration is only necessary twice a
year to maintain the smallest measuring ranges. All parts coming into contact with the
sample gas: Sampling system, sample gas preparation system and measuring cuvette are
heated to 100C. The instrument features a self-diagnosis system and, as with all modern
instruments of this type, can be monitored by H & B Service and faults diagnosed via an
integrated modem.
A German cement works has had good experience over a two-year operating period. The
instrument is overhauled twice a year as part of a maintenance contract. The weak point, the
hot gas pump, is changed each time or the diaphragms replaced as a precautionary
measure.

9. RECORDING AND EVALUATION
Following the measured values on the basis of recording strip charts is advantageous for the
general assessment of a process sequence. In the area of emissions, the majority of
countries require that measured values are prepared according to certain, adapted criteria;
hourly and daily average value, criteria relating to the observance or non-observance of
limiting values, etc. In such cases, digital processing of the measured values is unavoidable.
9.1 Trend curves, recorders

For the purpose of clarity, not too many components should be recorded simultaneously.
Strongly varying signals must be appropriately dampened. Monitors or line recorders are
better suited for this purpose than dotted-line recorders. Scaling should correspond with the
actual physical values. This applies particularly to life-zero signals. For strip chart recorders,
a paper feed rate of 20mm/h is standard.
9.2 Averaging computers

When lump fuel (e.g. tyres) is fed into the precalcining zone, the measured values at the kiln
inlet are subject to extreme fluctuations. Without damping of averaging, the signals are
difficult to interpret. It has emerged that sliding, linear averaging is easier to interpret than
logarithmic damping, as individual values are less important in linear averaging. The more
favourable intervals for optimal averaging are between 10 and 15 minutes. Figures 15 and
16 show the behaviour of the signals of a 1500-7/d kiln with tyre feed without averaging and
with various averaging times.

Figure 15:
O2 [%] O2 Direct recording

O2 Floating mean value 10 minutes

O2 Floating mean value 15 minutes

Figure 16:
CO [ppm] Direct recording (every 5 sec)
7000
6000
5000
4000
3000
2000
1000

CO Floating mean value 10 minutes
5000
4500
4000
3500
3000
2500
2000
1500
1000
500
0

CO Floating mean value 15 minutes
5000
4500
4000
3500
3000
2500
2000
1500
1000
500
0

9.3 Data logging
The interpretation of measured signals on the basis of trend curves is too rudimentary for
sophisticated tasks. Visual comparisons of such curves easily lead to misinterpretations. In
connection with optimisations or in the search for faults in the process, the signals from gas
analysis must be processed with other measured quantities from the system in the PC. A
precondition for this is that the data can be digitally acquired, logged and input into the PC.
In plants with modern process control systems, this precondition is usually met. Where this is
not the case or as part of temporary measurements, independent data logging systems
should be used for data recording purposes.
9.4 Emission computers

The form of documentation of the measured emissions is prescribed in more and more
countries. In Europe, there are plants whose emission computers are connected via a data
line to the authorities concerned. The emission computers are normally designed for
documentation of the following values:
Half-hourly average values
Hourly average values
Daily average values
Exceeded limiting values
Emission computers primarily meet the officially prescribed evaluation and documentation of
emission values. As the emission data also contains valuable process-relevant information,
they are also useful for the operator. Depending on the regulations of the country concerned,
the valuation of emission values can be taken over from internal process control systems.
10. MAINTENANCE AND QUALITY ASSURANCE

Reliable continuous operation of continuous gas analysis is dependent not only on the
efficient design of the measuring equipment, but also on systematic maintenance and repair.
The maintenance work to be carried out can be subdivided into the following categories.
Condition check
Adjustment check, cleaning
Replacement of wearing parts
Function test
Table 4 shows the maintenance intervals of a typical gas analyser. Maintenance must be
carried out daily (D), weekly (W), monthly (M), quarter yearly (QY), half yearly (HY) or yearly
(Y). The time cycle of the individual working procedures depends on the system-specific
conditions and essentially on the age of the measuring equipment. The work undertaken is
noted in a logbook for practical purposes.

Table 4: Maintenance Intervals
Function Test
Replacement of worn parts
Adjustments, checks, cleaning works
Visual checks
Sample gas preparation:
Flow measurement (floating body instrument) D W

Suction pressure (Manometer) D M
Humidity in the sample gas pipes D
Function check of the wetness monitor
Heated sample gas pipe temperature HY HY
Sample gas cooler condensate pump D HY
Sample gas pump check and replace membrane Y
Y
Tightness check QY
Cabinet air conditioning D Y
Magnetic valves, function and tightness check HY Y
Cabinet cleaning QY Y
Probes:
Tightness check QY
Cooling D
Filter change HY
Cleaning HY
QY
Analyzers:
Indication plausibility D
Adjustement W
Filter change QY
Total overhaul 3Y
The operating conditions and type of construction of the equipment used varies
considerably. Instructions for maintenance and repairs must be adapted to local conditions
and defined according to individual experience.

10.1 Visual checks
The term visual check includes visual inspection of the equipment, indicating instruments,
signal states and checking heated parts for a hot or cold condition by trained and
experienced personnel.
10.2 Adjustment checks, cleaning

This category includes adjustments, readjustments, etc., of conditions which can change for
different reasons.
10.2.1 Tightness checks

Vacuum check
The hose connections at the probe are sealed and the sample gas pump is started. The
maximum vacuum which the pump can produce must be noted when commissioning the
measuring equipment. At maximum vacuum, the part of the pump is closed and the pump
stopped. The sample gas preparation system up to the sample gas pump (suction area) is
considered tight when the vacuum does not change significantly for a period of 5 minutes
following temperature equalisation.
Nitrogen check
See Chapter 7.7
10.3 Replacement of wearing parts

When replacing parts, it must be ensured that no leaks occur. According to experience,
weak points in this respect are filter housings and connecting points in the sample gas pipe.
Suitable sealing material must be available at all times, so that any type of seal, union, etc.,
can be replaced if there is the slightest suspicion of a leak.
10.3.1 Sample gas pump

The sample gas pumps are usually diaphragm pumps, whose diaphragm and valve plates
should be replaced at least once a year.
The ball bearings of the motor and eccentric should be replaced every two years in
continuous operation. Standby sample gas pumps should be available.
10.3.2 Cleaning sintered metal and ceramic filters

Ceramic filters can be cleaned with diluted hydrochloric acid (1 part concentrated
hydrochloric acid to 10 parts water). The filters must be placed in the acid for about 2 hours.
The parts must subsequently be flushed with water pressure from the inside to the outside,
blown out with compressed air and dried in a drying cabinet at about 100C.
10.3.3 Fine filters

The filter element should be replaced at regular intervals, e.g. every three months. Fine
filters with condensate monitor are tested by moistening the filter element. The instrument
must then respond.

10.3.4 Gas analysers
The gas analysers are tested with test gases or by means of the built-in calibrating cuvettes
and readjusted if necessary. For testing with gases, it must be ensured that the flow rate
roughly corresponds with the operating values. When selecting the test gases, the specifi-
cations of the instrument manufacturer must be observed. The gas analysers should be
adjusted at regular intervals, at least every two weeks. Adjustments must be noted in the
maintenance report. If significant adjustments are required, or the adjustment range limits
(greater than 80% or less than 20%) are reached, the instrument must be tested and
inspected if necessary.
Wearing parts are, e.g. emitter, receiver, diaphragm motor.
For zero readjustment, nitrogen is normally used. With the exception of oxygen
measurement, air can also be used as a zero gas if the analysers are not too sensitive.
For sensitivity adjustment, a test gas is used that normally consists of the measuring
components and nitrogen. For compensation of the residual moisture, it is advantageous to
feed the test gas upstream of the sample gas cooler. The test gas concentration for
sensitivity adjustment should be 80 to 90% of the respective measuring range end value.
10.4 Function test

This term means that certain functions gas analysis are extensively tested, e.g. to ensure
that the detector in the fine filter responds if moisture is present and the sample gas pump is
stopped.
10.5 Test gases

As initially mentioned, new gas analysers no longer require any test gases. The internal
calibrating and control options ensure that the accuracy is maintained over a long period
(normally two years). After this time, the analysers should be tested by the manufacturers
service department and in particular the internal control devices (calibrating cuvettes)
checked. Test gases are available in various accuracy classes according to particular
directives, e.g. VDI 3490, sheet 2. Depending on the manufacturing and analysis accuracy,
distinction is made between three classes. The test gases are delivered with analysis
certificates that contain all essential data, such as:
Measuring component and accompanying gas
Test gas production method
Measuring component concentration
Relative error of this concentration
Pressure of container filling
Minimum application pressure
Minimum and maximum storage temperature and maximum test gas storage time (limit
date)

The concentration specified on the analysis certificate should be written in large and easily
readable letters on the test gas bottle.
The test gases are available in bottles of varying size. A 10 litre bottle (about 1100mm high,
preferably of aluminium) is normally sufficient for a two-year supply. The bottle pressure is
between 100 and 200 bar, depending on the gas composition. The main valve must be
closed after each use. Test gas deliveries max take time in certain regions; this should be
taken into account when the bottle pressure starts to get low.
10.5.1 Mixing test gases

Certain gas components can be combined as a mixture in a bottle. This possibility
considerably facilitates maintenance (fewer bottle transports and savings on gas bottle hire
charges). Caution! New multi-component gas analysers must only be tested with gas
mixtures but not calibrated. Only individual components in nitrogen are permitted for
calibration.
With one exception, all standard components can be mixed together. This exception applies
to oxygen O2.
With oxygen analysers, the measuring range should be such that air can be used for
calibration. Nitrogen must always be used as a residual gas.
Example of a test gas mixture:
Component Chemical Analyser measuring Concentration

symbol range in test gas
Sulphur dioxide SO2 3000ppm 2800ppm
Carbon monoxide CO 5% 0.45%
Nitrogen monoxide NO 2000ppm 1900ppm
Carbon dioxide CO2 as accompanying gas 25%
Nitrogen N2 as residual component 74.08%
The accuracy of the test gases reduces the more components are mixed. The number of
mixing components is thus limited.
Further, general information on test gases is provided in the annexed instructions.

10.6 Fault signals
Faults in the measuring instrument should be indicated by status signals, e.g.:
Insufficient pressure in the suction pipe (blockage of the probe or gas path)
Excessive pressure in the suction pipe (possible failure or leakage of a solenoid valve)
Status signals of the analysers (failure of the analyser or electrical fault)
Operating temperature of the sample gas cooler or failure of the sample gas cooler
Operation of the condensate monitor (can indicate failure of the sample gas cooler or
faulty water-cooled probe
Adjustment/operation changeover valve should automatically inform the control room that
the measuring equipment is being serviced and is therefore not ready for operation
Minimum contact of flowmeter (insufficient sample gas)
Status signals for measuring equipment faults are checked for their plausibility by simulated
operating deviations.
10.7 Automated maintenance equipment

Theoretically, numerous maintenance procedures can be automated, e.g. automatic zero
and limit range adjustment, sample gas flow rate control, blowing out of the filter, etc.
However, sample gases are media that can give rise to various difficulties.
They have a corrosive effect, contain dust, must not be adulterated, can be very moist, etc.,
to mention but a few of the unpleasant characteristics that make automation difficult.
Practice shows time and again that solenoid valves are the weak point within measuring
equipment, so that only valves of the best quality must be used for evaluation. The number
of necessary solenoid valves increases with an increasing degree of automation; due to their
susceptibility to faults, the failure rate also increases, so that the savings expected from
automation are accordingly lost.
The same applies when a group of analysers is switched to two or more measuring points
with the intention of saving on expensive analysers. Apart from the operating difficulties,
experience has shown time and again that the saved investment costs are quickly absorbed
by the corresponding maintenance costs.

11. MEASURING ERRORS
The accuracy of gas analysers is often overestimated. The reason for gas analyses being
relatively inaccurate is not because of the analysers, as is often assumed. The following
examples explain the most important influences responsible for this.
11.1 Sample gas sampling

Significant measuring errors can occur as a result of non-representative sampling. In rotary
kiln systems, these occur particularly at the rotary kiln inlet and kiln inlet chamber measuring
points. These measuring errors occur mainly at the filter measuring points in coal
pulverising mills.
11.2 Sample gas preparation
11.2.1 Sorption and chemical reaction

In the probe and sample gas preparation system, sample gas components can react with
each other. These reactions are reinforced by catalytic effects. Furthermore, reactions with
dust or gas conducting components are possible.
11.2.2 CO reduction
At high temperatures (above 600C), CO is reduced to a large degree depending on the
residence time. This effect is the reason why only cooled probes may be used at
temperatures above 500C.
11.2.3 NO2 formation in the probe and sample gas preparation system
At temperatures of about 100C and a concentration of 900ppm NO, the theoretical
formation rate of NO2 is about 0.5ppm per second.
At lower temperatures and high gas residence times, this formation takes place more rapidly.
The indication delay of the NO measuring equipment should therefore be as short as
possible.
11.2.4 SO2 reduction in the probe and sample gas pipe

Rust, dust deposits, non-ferrous metal parts (especially copper) in the gas path of the
sample gas can reduce SO2 up to 100%. For the sample gas pipes and all other gas
conducting parts, Teflon, high-alloyed steel or steel with special coating (e.g. PTFE) must
therefore be used.
11.2.5 SO2 reduction of the filter dust of the sampling probe

SO2 reacts with the CaCO3 and CaO contained in the filter dust of the sampling probe with
the formation of CaSO4. With high dust contents in the process gas, measuring errors of up
to 100% can occur. SO2 can therefore only be measured fairly accurately process gas with a
low dust content. This applies in particular to SO2 measurements in the rotary kiln inlet.
Tests have shown that SO2 is initially indicated only a short time after cleaning the filter.
Within several minutes, the indication falls to zero again, because the SO2 is absorbed in the
dust building up in the filter.

11.2.6 Interaction with a liquid
Errors due to the solubility of the measuring component
The solubility of certain measuring components in water must be taken into account for wet
sampling probes. If the injector water of the wet sampling probe is not fed into the circuit,
CO2, NO2 and SO2 can go into solution, depending on the probe design.
Condensation of the water vapour which is separated as water in the sample gas cooler by
the extractive method, can also distort gas analysis. This is of particular significance for SO2
emission measurement when low concentrations are present. The theoretical solubility of
SO2 in water is low and could be compensated by appropriate calibration. However, tests on
cement kiln systems in a direct operating mode have shown that measuring errors for SO2
are far above the theoretical distribution equilibrium when SO2 reacts with other gas
components such as NH3. An SO2 measuring device can be tested for these errors by the
sample gas being transferred from the sampling probe directly into concentrated sulphuric
acid (H2SO4) as the drying medium and not via the sample gas cooler. If distinct differences
in the measured value indication occur, a chemical reaction is most probably taking place in
the sample gas cooler.
Apart from SO2, NO2 also dissolves in the condensate of the sample gas cooler. However,
since the proportion of NO2 in the total NO concentration can be disregarded with proper
sampling, the resulting measuring error can normally be disregarded.
11.3 Volumetric errors due to the solubility of accompanying components

If accompanying components, e.g. CO2 go into solution, the concentration of the other
measuring components increases. The partial absorption of CO2 in an injector probe,
increases, e.g. the O2 concentration.
11.4 Gas analysers

The accuracy of the gas analysers, provided they are in a perfect condition, is not an issue.
If a gas analyser is viewed as a whole, the analyser is the most accurate link in the
measuring chain, with a few exceptions. The analyser manufacturers specifications relating
to cross-sensitivities, disturbing components as well as temperature and pressure influences
must be observed.

12. SPECIAL FUNCTIONS OF GAS ANALYSIS
12.1 CO monitoring for protection of electrostatic filter

If, for any reason, combustible gases should develop (ring fractures, faults in fuel dosage,
etc.), in the kiln system, there is a risk of a potentially explosive mixture forming in the
electrostatic filter. High-voltage discharges inside the filter cause sparks, which can lead to
explosions in a gas atmosphere of appropriate composition.
As a safety measure, the CO component of the gas discharging from the kiln system is
measured. Once a certain level is reached, the high-voltage should be switched off before
the gases reach the electrostatic filter. Measurement can initially take place in precalcining
plants following complete combustion in the auxiliary firing equipment, i.e. in the uppermost
cyclone stage of the heat exchanger at the earliest. The residence time of the gases from
this point to the electrostatic filter is relatively short (several seconds) and varies
considerably from plant to plant (with or without cooling tower, combined/direct operation,
etc.).
Filter monitors, whose effect alone depends on an adjustable maximum value, reassure
responsible personnel, however, they are not optimal solutions. On the one hand,
unnecessary filter shutdowns are initiated, on the other hand, it is not ensured that they
would operate promptly in an emergency. A tried and tested possibility of optimising this
process is offered by sequential filter shutdown.
12.1.1 Basic concept of filter shutdown

Compared with the usual method, filter shutdown is optimised in two essential steps. The
first step consists of shutting down the filter chambers step by step. The second step
consists of taking into account a reducing or increasing CO trend.
The residence time of the gases in the electrostatic filter itself is used in order to further
delay shutdown. The filter chambers are shutdown at approximately five second intervals,
depending on the size and gas flow rate. However, as soon as CO measurement indicates a
reducing trend, the shutdown cycle is stopped.
12.1.2 Optimisation of CO measurement

The following measures must be taken in order to make CO measurement as efficient as
possible:
Short distance between sampling point and analyser
Large sample gas volume (up to 300l/h setting;short delay time)
Small volumes in gas path, particularly in sample gas cooler
Select analyser with fast response time
Use double measuring system for two-section WT systems
Formation of logical signals CO increasing and CO reducing

12.1.3 Specifications
Selection of measuring point according to Figure 1
Gas preparation system appropriately adapted according to Figure 12
NDIR gas analyser, measuring range 0 to 5%, CO with 3 adjustable maximum limiting
values
Arrangement of a shutdown procedure
With regard to the CO analyser, it should be noted that the same must have a high
measuring range for CO monitoring. It cannot be used simultaneously for continuous CO
observation, as the normal CO content is below the detection limit of this analyser. If CO is
to be continuously measured, a second analyser with a much lower measuring range of
5000ppm (0.5%) must be used.
12.1.4 Shutdown procedure

The procedure for shutdown of the individual filter chambers is shown on the basis of two
examples in Tables 5 and 6. The specified numerical values of the shutdown sequences and
the CO level are only guide values and must be adjusted to the respective plants.
For CO values below 10%, a risk of explosion is virtually excluded. However, the threshold
values must be set far lower, as the measurement reacts with delays despite optimal
preconditions (dead time + indication delay).
Table 5: Switching procedure, for a simple heat exchanger kiln with 3-changer
filter

Table 6: Switch-off pocedure, for a two strings heat exchanger kiln with 3-
chamber filter
Time- CO - Analyzer 1 CO - Analyzer 1 Trend Alarm- Filter chamber Fuel

point switch. threshold switch. threshold Stage
Sec. 0,8% 1,5% 4,8% 0,8% 1,5% 4,8% 1 2 1 2 3
X 0 0 0 0 0 0 +- +- 0 1 1 1 1
X 1 0 0 0 0 0 +- +- 1 1 1 1 1
X 1 0 0 1 0 0 +- +- 1 1 1 1 1
X 1 1 0 1 0 0 +- +- 1 1 1 1 1
5 1 1 0 1 1 0 + + 2 0 1 1 1
5 1 1 0 1 1 0 - - 1 1 1 1 1
5 1 1 1 1 1 1 + - 2 0 1 1 1
5 1 1 1 1 1 0 + + 3 0 0 1 1
10 1 1 0 1 1 0 - - 1 1 1 1 1
10 1 1 0 1 1 0 + + 3 0 0 1 1
10 1 1 1 1 1 1 + + 5 0 0 0 0
10 1 1 0 1 1 0 - - 1 1 1 1 1*
15 1 1 0 1 1 0 + + 4 0 0 0 1
15 1 1 0 1 1 0 - - 1 1 1 1 1
15 1 1 1 1 1 1 + + 5 0 0 0 0
1 = "On" * Check the purging time befor switching

0 = "Off" the filter on again
12.1.5 Special devices

Sick GM 950
Via an in-situ sensor (zirconium dioxide) CO is measured at a high response rate and
processed to a shutdown command if required. As zirconium is fairly instable as a CO
indicator, extractive supplied semiconductor sensors correct any deviations.
The device is new and still in the testing phase. The currently available test results have
been positive up to now. The measuring method is relatively elaborate and the device gives
the impression of being somewhat complex. During the test, no functional disturbance
occurred.
12.1.6 Laser analysers

In-situ laser analysers offer promising prospects as already mentioned in chapter 8.6.3. It is
conceivable that low-cost options will soon be available, as several measuring points can be
served with the same analyser, e.g. filter monitoring (upstream and downstream of the filter)
and various points in the coal pulverising mill.

12.2 Monitoring of coal pulverising mills
In the safety concept of coal pulverising mills, the monitoring of critical gas components
covers only a small part of the necessary equipment and measures. In addition, coal
pulverising mills vary considerably in their design, so that the application and function of gas
analysis is always determined as part of the entire plant.
12.2.1 Limiting values

In 1984, the following limiting values were listed by a working group of the Association of
German Cement Works (VDZ):
O2 content in moist exhaust gas downstream of the filter 14%
Temperature at mill inlet (coal moisture 12%) 400C
Gas temperature downstream of mill upstream of filter 120C
Gas temperature downstream of filter 120C
Gas temperature upstream of filter minimum 30C above dew point about 75C
CO concentrate during filter shutdown 50ppm
Operation with hot gas generator 70ppm
Operation with kiln exhaust gase not practical
Pressure upstream of mill -1.5bar
Coal dust temperature 110C
Air temperature of pneumatic transport 80C
The limiting value for the oxygen concentration applies only to plants with inert gas
operation. All other limiting values are the same for the safe operation of inert and non-inert
gas operated plants. For coal pulverising mills heated with kiln exhaust gas, CO monitoring
is not practical during operation, because the gases from kiln operation contain up to 10
times more CO than the limiting value of 70ppm permits.
CO shutdowns of the electrostatic filter for the kiln, must be coupled with the filter of the coal
pulverising mill.
This information on the safety equipment of coal pulverising mills only contains the most
important information and is intended to draw attention to the complexity and importance of
such plants.

13. CONVERSION FACTORS
Conversion of the gas components from volume to units of weight and vice versa, is often
the cause of errors in adjustment or interpretation. Such errors occur most frequently in
connection with NO and NO2. The problem with these NO components is that the
instruments measure NO, but the emission must be converted to NO2.
Furthermore, emissions are considered in weight-related units (mg/m3), but the instruments
and the test gases are often specified in volume-related units (e.g. ppm or %)
Conversion from [ppm] to mg/m3 Factor f = M

22.4
Whereby M = Mol mass (kg/kmol) of the gas components
Table 7
Components Converted to Mol mass [M] Factor (f) 1/f

CO CO 28.01 1.25 0.80
NO NO 30.01 1.34 0.75
NO NO2 46.01 2.85 0.49
NH3 NH3 17.03 0.76 1.32
Example
Given:
Test gas with 8.75ppm NO
NO analyser, measuring range 0-2000mg/m3NO
Required:
Analyser indication at 875ppm
How much NO2 in mg/m3 corresponds to the adjusted value 875ppmNO?
Solution:
Analyser indication = 875 x 1.34 = 1172mg/m3 NO
Emission value NO2= 875 x 2.05 = 1794mg/m3 NO2
or from mg/m3 NO to mg/m3 NO2 1172 x 1.53 = 1794mg/m3 NO2

14. LITERATURE
1) Verein Deutscher Zementwerke (VDZ) (Association of German Cement Works):
2) Continuous Gas Analysis in Cement Works, Notice VT9, June 1990
3) B. Thier: Safe Operation of Coal Pulverising Mills. ZKG 4/1984, page 163
4) M. Ascherfeld and W. Fabinski Multi-component Analyser for Oxygen and Infrared Active
Gases. tm Technisches Messen 59 (1991) Volume 5
5) H.G. Loos, Erlangen: A New Operational Compact Gas Sampling Device for Cement
Rotary Kilns. ZKG NO. 6/1987
6) M. Birrer, H. Nyffenegger: Gas Sampling at the Kiln Inlet with Improved Probe.
7) Holderbank NEWS 6/90
8) K. Utzinger: Reduction of Dust Emissions on Startup and CO Shutdowns of the
Electrostatic Filter at the Rekingen Works. Holderbank NEWS 7/8 (1986)
9) Bronkhurst High-Tech B.V., Ruurlo, Holland: Operating Manual for Mass Flowmeters and
Controllers for Gases
10) KNF Neuberger (Switzerland) AG, CH 8362 Balterswil: Prospectus and Data Sheet for
Diaphragm Compressors and Vacuum Pumps
11) Annette Schrck: Preparation of Decision-making Criteria for Plant Operators for
Purchasing an Emission Measuring Instrument. Thesis SS 1992 (available from Messrs
Sick)

Chapter 5
High Level Control

1. HIGH LEVEL CONTROL SYSTEMS........................................................................... 339
2. Field Preparation for High Level Control - Process Preview ................................. 355
3. Field Preparation for High Level Control ................................................................. 362

HIGH LEVEL CONTROL SYSTEMS
U. Haberstich
PT 98/14350/E (Revision of PT 94/4191/E)
1. Introduction...................................................................................................................341
2. Evolution of HLC Systems...........................................................................................341
3. Specific Requirements of a HLC System ...................................................................342
4. Process Optimization with HLC ..................................................................................343
5. Principles of Operation ................................................................................................344
6. Benefits and Keys of Success.....................................................................................347
7. How to justify an investment in HLC ..........................................................................348
8. Conclusion ....................................................................................................................348
9. References ....................................................................................................................349
10. Annex .....................................................................................................................350

Summary
For more than 30 years, the cement industry has been exploring computer-based techniques
to control and optimize the operation of cement kilns. The major reasons behind these
endeavours are clinker uniformity, savings in energy consumption, increase in production,
savings in refractory consumption and NOx reduction. Basically, cement kilns are difficult to
control because of their non-linear, multivariable, behaviour and the poor quality of the
available process signals. After several trials to describe the burning process with
mathematical models a new approach in cement kiln control was investigated in 1980.
Operator control strategies were studied, and a detailed record of the operators behaviour
was made while controlling the kiln. These fuzzy rules can imitate multivariable control
actions and can combine information from variables to identify the kiln conditions. Within the
Holcim group, specific requirements of such a High Level Control (HLC) system was defined
to compare the several suppliers. The system from ABB LINKman is the most powerful real-
time expert system with very advanced features and a high user-friendliness. A standard
implementation plan was made to reach a successful application of the system. The
preparatory work in the plant must be carried out according to the recommendations made
during the pre-project study. The identified process problems have to be solved. To achieve
the Best Operator Performance the detailed and rigorous monitoring of the HLC
performance is obviously essential at this stage. The optimization of the process with respect
to product quality and related process factors, on a long-term basis, is the key to the ultimate
level of savings arising from the HLC application. It is this optimization that can give the plant
additional benefits over and above those arising from the average operator to Best Operator
Performance. The experience shows that substantial benefits like higher kiln output, lower
heat consumption, longer refractory life, lower NOx emissions and lower standard deviation
of the key variables can be expected from HLC systems, if they are properly implemented
and used. The experience shows that the payback of 1 to 2 years is realistic, when
considering only the investment costs in the HLC system.

1. INTRODUCTION
For more than 30 years, the cement industry has been exploring computer-based techniques
to control and optimize the operation of cement kilns. The major reasons behind these
endavors are clinker uniformity, savings in energy consumption, increase in production,
savings in refractory consumption, NOx reduction (Fig. 1 + 2, Annexes 1 + 2). The still high
energy consumption of the cement manufacturing process, the stringent requirements on
cement quality and the environmental aspects which are leading the governments to apply
severe legislation regarding the emissions.
The classical process automation approach, consisting in defining a mathematical model of
the process, led to only a very few successful kiln control applications. The improvement in
the performance of hardware equipment, combined with the advent of artificial intelligence, is
leading to a major step toward kiln control and optimization.
2. EVOLUTION OF HLC SYSTEMS

Basically cement kilns are difficult to control because of their non-linear, multivariable,
behaviour and the poor quality of the available process signals. The control is then usually
limited to a few secondary measurement loops, whereas the control of the primary
parameters and the operating conditions are the responsibility of the kiln operators.
The first applied techniques were based on empirical or mathematical models. Although
successful in simulating the kiln operation, these techniques were generally based on too
many assumptions and used very complex theoretical models. For this reason they were not
applicable and could not be extended to a broad selection of kilns. Other approaches such
as the "hill climbing" techniques (Young et al, 1971) or statistical identification combined with
optimal controller design by the state space method (Otomo et al, 1972) was also used but
did not achieve any significant success.
Since the mid-seventies, a new approach based on the analysis of the human decision
making in cement kiln control has been investigated (Umbers and King, 1980). Operator
control strategies were studied, and a detailed record of the operator's behaviour was made
while controlling the kiln. Basically this approach rests upon the concept of fuzzy logic
introduced by Prof. L. Zadeh in 1965. The basic operator control rules were already
prescribed by Peray and Waddell (1972). These "fuzzy" rules can imitate multivariable
control actions and can combine information from variables, they work by identifying the kiln
conditions and prescribing suitable corrective actions.
FL Smidth supplied the first commercially available kiln control system based on fuzzy logic
in 1980. The concept of High Level Control was introduced at that time and is used to refer
to systems, which provide not only supervisory control but also optimising control. Since
then, many systems have been developed and are marketed. Some are using the concept of
fuzzy logic and are called rule-based systems or are based on expert system shells. Others
are more conventional and apply PID control or adaptive-predictive controllers. The
penetration of HLC in the cement industry has been very intensive over the last decade,
about 300 applications have been reported in kiln control applications.

3. SPECIFIC REQUIREMENTS OF A HLC SYSTEM
The following basic requirements were specified:
1) The provision of a high degree of user-friendliness; this aspect is extremely important,
since kiln control strategy needs to be adapted when process conditions change. The
maintenance of the application control strategy must be easy for the plant engineers to
carry out.
2) The concept of autopilot as used in the previous version of the suppliers HLC system
must be included to make it possible for the operator to switch the system on-line or off-
line at any time without disturbing the process.
3) The use of a toolkit based on the G2 expert system shell which provides advanced
features such as real-time facility, graphical interface and object-oriented programming;
the toolkit, which is a software layer between G2 and the HLC applications, is the
support for developing and implementing control strategies without programming skills.
4) The possibility for having multiple applications on the same system, typically one kiln,
one cooler, mills and the kiln simulator.
5) The provision of a system incorporating tools and facilities allowing for consistent
process optimization.
6) Very helpful for the introduction is the inclusion of a standard interface between the HLC
system and the tailor-made kiln simulator in order to provide a training platform for the
operators; this platform would allow for the simulation of kiln upsets and disturbances.
The plant engineer would thus be able to develop and test new strategies before real
implementation in the actual plant application. Figure 3 (Annex 3) shows a typical
configuration.
7) The standardisation of the control strategies, in order to shorten commissioning time and
make exchange of experience easier between different users. It is important to have a
standardised way of both configuring and maintaining the control strategies.
Within the Holcim Group LINKman from ABB LINKman Systems Ltd, London, is applied.
LINKman Graphic, the new version of this HLC system, is based on G2, a very powerful real-
time expert system shell from Gensym Corp. LINKman Graphic is particularly user-friendly
and offers very advanced features [4].

4. PROCESS OPTIMIZATION WITH HLC
As mentioned above, process optimization is the major target to be achieved. There are
typically three phases associated with our concept of successful application. See also [3].
These phases are outlined below.
Phase 1: Plant Preparation
The preparatory work in the plant must be carried out according to the recommendations
made during the pre-project study.
Phase 2: Achievement of Best Operator Performance
Very often bottlenecks are detected, process problems are identified and experience is
gained in the period immediately following HLC implementation. Detailed and rigorous
monitoring of the HLC performance is obviously essential at this stage. The identified
problems then have to be addressed during the secondary commissioning. Attempts must
be made to achieve the highest possible run time of the HLC system, typically 90 % or more,
using an adequate and consistent control strategy. At this stage, the HLC is expected to
operate with the same performance as the best operator with respect to production output,
product quality, heat consumption, etc. The associated benefits depend on the size of the
plant and the performance previously achieved in manual operation. This phase, based on
the best operator know-how, can be considered as the foundation of the whole HLC project.
Phase 3: Process Optimization
The optimization of the process with respect to product quality and related process factors,
on a long-term basis, is the key to the ultimate level of savings arising from the HLC
application. It is this optimization that can give the plant additional benefits over and above
those arising from the average operator to Best Operator Performance. Process
optimization mainly involves the plant technical staff. It requires an evaluation of the process
performance, an estimation of the potential savings, which can be achieved, and an
assessment of the control strategy performance. Since plant conditions change in respect to
raw material quality, availability of alternative fuels, product quality requirements, etc.,
process optimization must be considered as a permanent task.

5. PRINCIPLES OF OPERATION
Basically, in the case of a conventional preheater kiln, HLC manages the control of the
following parameters:
the kiln feed rate
the kiln speed
the IDF speed (or damper position)
the fuel to the kiln burner
In case of a precalciner installation, the precalciner fuel and the position of the tertiary air
damper (AS system) have also to be controlled automatically.
The HLC system needs to access the process relevant data, such as preheater
temperatures and pressures, the gas composition, the kiln amperage, the burning zone
temperature, etc.
The principles of kiln control operation depend on the HLC system used. The principles used
in LINKman are hereafter presented.
LINKman
LINKman works in two basic modes:
Normal Actions and Interrupt Actions.
The approach used in LINKman is sequential. In a first phase, a specific logic checks
whether the kiln is in normal operation or in upset situation.

NORMAL ACTIONS INTERRUPT ACTIONS
BZT, OX, BET Manual Override
Rule Block Selection CO Action
Ring Action
NORMAL ACTIONS
Fuel Override
Scalling Liter Weight Action
Post-Ruleblock Break Action

Normal Actions
If the kiln is in normal condition, then the parameters BZT, OX and BET are calculated.
BZT (Burning Zone Temperature) represents the clinker burning degree, calculated out
of NOx, Kiln Amps, and zone temperature.
OX, the draught index, based on the gas composition at kiln inlet.
BET (Bet End Temperature): the degree of preparation of the material when entering the
kiln.
LINKman selects then one of 4 ruleblocks to determine the required setpoint changes,
depending on:
the deviation between the actual feed rate and the desired target,
the spare draught capacity,
The process conditions.
Rule Block Actions on
General Feed, Fuel, Fan
Top-feed Fuel, Fan
Top-damp Feed, Fuel
Stable Fuel
The changes calculated using the ruleblocks are scaled to physical changes. However these
changes are merely based on the present kiln conditions. Specific actions are then carried
out, if required, to take into account the previous actions and the process dynamics, they are
included in the "Post-Rule Block Processing" module.
Interrupt Actions
If a kiln-upset condition is detected, a dedicated control action is applied. As these actions
have a higher priority than the normal actions, they are called interrupt actions.
Examples:
Break-Action if the kiln is in unstable conditions for a longer period
Ring-Action if a ring fall is detected
Hot-Action if the kiln gets very hot
CO-Action if a high amount of CO is detected

6. BENEFITS AND KEYS OF SUCCESS
The experience shows that substantial benefits can be expected from HLC systems if they
are properly implemented and used. The following table shows typical ranges of quantifiable
benefits:
Item Typical Range

Kiln output + [0 - 5 %]
Heat Consumption - [0 - 5 %]
Refractory Life + [0 - 30 %]
Long-term clinker strength + [0 - 5 %]
Electrical energy for clinker grinding - [0 - 10 % ]
NOx emissions - [0 - 30 %]
Number of kiln stops - [0 - 30 %]
Standard deviation of key variables - [0 - 50 %]
In addition, the following qualitative benefits have to be mentioned:

The working conditions of the operators are greatly enhanced. As the computer handles
the routine tasks, the operator can concentrate on more important matters.
Process analysis and optimization are made more easily since opportunities for testing
new control strategies and new ideas are available, assuming that the HLC is user-
friendly enough.
The use of HLC imposes to keep the instrumentation in a good operating state.
The keys of success with HLC systems are:
An adequate and reliable instrumentation (sensors, actuators, PID controllers, etc.)
A stable and uniform raw mix chemistry
An optimized combustion
A highly motivated personnel
A follow-up of the performance of the system

7. HOW TO JUSTIFY AN INVESTMENT IN HLC
In order to justify an investment in HLC, it is important to estimate the potential cumulative
savings over the lifetime of the system. Figure 4 (Annex 4) gives an illustration of these
savings over a period of ten years.
In this figure, curves 1, 1a and 1b indicate the cumulative savings with HLC when achieving
Best Operator Performance. If the HLC has a robust strategy and is permanently adapted
to the evolution of the burning process, then the cumulative savings will increase
consistently according to curve 1.
However, if the system is not properly maintained or not adjusted to changing process
conditions, sooner or later it will stop functioning and the savings will obviously stop. These
situations can take place very shortly after installation (curve 1a) or later on (curve 1b).
Curve 2 illustrates the cumulative savings, which can be realized in cases where the HLC is
used as a tool for permanent process optimization. These extra savings are on top of the
savings indicated on curve 1. The benefits resulting from this optimization can be achieved
after reaching an HLC run time of more than 90 %. The savings through optimization can
increase even more if the original HLC application is extended to other plant areas such as
raw milling, stack gas flow control or integration of an efficient raw mix control strategy.
To summarize, simply by achieving Best Operator Performance, the payback is generally
less than two years (considering only the investment in HLC). Furthermore, if the HLC is
used efficiently as a permanent optimization tool, then the cumulative savings will increase
even more.
The experience shows that a payback of 1 to 2 years is realistic, when considering only the
investment costs in the HLC system.
8. CONCLUSION
Over the last decade, HLC systems have penetrated the cement industry. About 300
applications have been reported which represents roughly 15 % of the cement
manufacturing installations.
There is no doubt that the proper implementation and use of a HLC system, although
requiring relatively low investment costs, provide significant enhancements in terms of
productivity.
However it is of the utmost importance to select the adequate system, that means a system
which has proven to be efficient and performant in the long term.
It must be remembered that the success of any HLC system depends upon:
the quality of the instrumentation
the raw meal preparation
the quality of the combustion system
the motivation of the works personnel and the acceptance of the system by the
operators

9. REFERENCES
1) Bertrand I Automatic control of kilns and mills by expert systems.
World Cement, July 1991.
2) Bauer C., G. Jger, J. Patzer and K.H. Walen
Cost-effective clinker production with the PYROEXPERT optimization system.
Zement, Kalk, Gips Nr. 4/1993.
3) Hasler R. and E.A. Dekkiche
Experience in kiln optimization with LINKman Graphic.
IEEE Cement Industry Technical Conference, Seattle 1994.
4) Haspel D. and W. Henderson
A new generation of process optimization systems.
International Cement Review, June 1993.
5) Maynard B. Hall
Kiln stabilization and control - A COMDALE/C expert system approach.
IEEE Cement Industry Technical Conference, Toronto 1993.
6) Mende N.
Erfahrungen mit dem Ofenfhrungssystem KCS an einer grossen Drehofenanlage in
Sdkorea.
Zement-Kalk-Gips, Nr. 2/1993.
7) Ostergaard J.J.
FUZZY II - The new generation of high level control
Zement-Kalk-Gips, Nr. 11/1990.
8) Otomo T., T. Nakagawa and H. Akaike
Statistical approach to computer control of cement rotary kilns.
Automatica, Vol. 8, 1972.
9) Ruiz Navarro M., J.M. Martn-Snchez and C. Corzo Carreno
Minimizing energy consumption in kilns by the SCAP system.
World Cement, March 1993.
10) Umbers I.G., P.J. King
An analysis of human decision-making in cement kiln control and the implications for
automation.
Int. J. Man-Machine Studies, Nr. 12, 1980.
11) Young S.C.K., K.L. Todd and K.H. Lau
On-line optimization of rotary cement kilns.
3rd IFAC/IFIP Intern. Conf. on Digital Computer Appl. to Process Control. Helsinki 1971.

10. ANNEX
Figure 1: Example of Improved Clinker Uniformity
Manual Operation
40
35
30
% Frequency
25
20
15
10
0
0-0.8 0.8-1.1 1.1-1.4 1.4-1.7 1.7-2.0 2.0-2.3
C linker Free Lime [%]

HLC Operation
40
35
30
% Frequency
25
20
15
10
0
0-0.8 0.8-1.1 1.1-1.4 1.4-1.7 1.7-2.0 2.0-2.3
C linker Free Lime [%]

Figure 2: Example of Improved Heat Consumption
3460 100%
90%
3440
80%
3420
70%
3400
60%
3380 50%
40%
3360
30%
3340
k J/k g Clk 20%
3320 HLC Operation 10%
3300 0%
J F M A M J J A S O N D J F M A M J J A S O N D

Figure 3: Operator Training
Engineer- and Training Room Control Room
Engineer Operator
Station Station
Fine Tuning Autopilot
HLC
Simu- Real
Simula- lator Kiln
tor
I/O
Plant
Training

Figure 4: Lifetime Benefits of Process Optimization with HLC
1: B est Operator Performance
1a: S ystem not properly maintained
1b: S ystem not properly maintained

C umulative Savings
2: HLC used as Optimisation Tool
Time

Field Preparation for High Level Control - Process Preview
U. Haberstich
PT 99/14495/E
1. Introduction...................................................................................................................357
2. Detailed Scope of work ................................................................................................357
2.1 Definition of responsibilities and meetings .......................................................357
2.2 Data collection..................................................................................................358
2.3 Combustion check............................................................................................359
2.4 Definition of the strategy ..................................................................................360
2.5 Control loops ....................................................................................................360
2.6 Instrumentation check ......................................................................................360
3. Appendix 1 ....................................................................................................................361
3.1 KILN MANUFACTURING DATA ......................................................................361

Summary and conclusion:
The introduction of a High-Level Control System requires a proper preparation of the plant to
ensure an efficient and successful implementation and commissioning. For that purpose a
so-called preproject from HMC is proposed in order to disclose eventual difficulties or even
disability of the plant to implement such a system. A team of the Corporate Technical
Services and Holcim Corporate Engineering will help the plant to achieve the best
preparation.
HES will check the possibilities to connect the LINKman system to the installed plant control
system and the sensitivity of measuring and control devices.
HMC/PT will check the quality aspects and the kiln behavior in order to disclose eventual
process problems.
The main items of the LINKman preproject are:
Kick-off meeting with explanation of the LINKman system and the procedure to
implement it.
Collection of all available plant data (Flow-sheets, technical descriptions of the
installation).
Collection of all available process data (Log-sheets, quality data, flow-sheets and
trends).
Combustion check.
Description of the used strategy and rules for manual kiln and cooler control (which
parameters were used) and a description of the most common kiln and cooler problems
(dusty clinker, coating falls, etc.).
Control of the adjustments of the most important control loops.
Instrumentation check (type, position, sensitivity).
Definition of the preparatory work to be done before the LINKman implementation.
Conclusion meeting with the definition of the implementation schedule (ordering and
commissioning). Definition of the project leader and of the responsible for the LINKman
Hardware and Software (Strategy).

1. INTRODUCTION
The introduction of a kiln High-Level Control System like LINKman requires several
preparatory works by the plant to ensure an efficient and successful implementation. The
capability of the kiln to be driven by an automatic system has to be ensured by a pre-project.
As in every project, the organization has to be determined and responsibles have to be
nominated.
The LINKman pre-project has to cover mainly the following capability checks of the kiln
system:
Disclosing eventual process problems
Defining an adequate kiln and cooler control strategy
Checking the instrumentation
Checking the sensitivity of the measuring and control devices (HES)
Defining the connection of the LINKman system to the installed plant control system
(HES).
2. DETAILED SCOPE OF WORK
2.1 Definition of responsibilities and meetings

A kick-off meeting has to be organized at the beginning of the pre-project. Experience has
shown that a more detailed explication of the LINKman system avoids confusions. The
implementation schedule (ordering and commissioning) has to be defined as well as the
project leader and the responsibles for the LINKman Hardware and Software (Strategy).
The responsible for the Hardware issues is usually an electrical engineer with knowledge of
the plant control system and the automatic control loops.
The responsible of the Software (strategy) needs detailed knowledge about the burning
process and kiln control. Therefore, a process engineer is strongly recommended. He will be
the future LINKman Champion and doing all further modifications of the strategy after the
commissioning.
Before the implementation also the exact position of the Hardware has to be defined
(location of the operator- and engineer-station in the control room and location of the
LINKman cabinet).
After finishing the study, all requirements will be discussed during the conclusion meeting at
the end of the visit. The preparatory work has to be terminated until the first commissioning
of ABB. If stronger process problems were detected, it may be recommended to postpone
the order of the system.

2.2 Data collection
To prepare the system before delivering, ABB needs sufficient data about the plant.
Therefore, a collection of all available plant data is required. This contains mainly:
Plant descriptions:
Flow-sheet of the kiln system
Flow-sheet of the plant control system (PLC-System)
Kiln manufacturing data (see also Appendix 1)
Layout of the control room (to place the Hardware and pre-configure the cabling)
List of available control loops
Instrumentation list.
Process data
(During 1 week of representative production)
Quality data of Raw meal, Hot meal and Clinker
Operator log-sheets
Statistical distribution of the freelime.
For 24 hours
1-day-trend of all important process values of that week (burning zone temperature,
NOx, Amps, Calciner temperature, Preheater exit temperature, etc.)
Additional information
Printout of all available kiln process screens (only 1 momentary printout)
Short description of the used strategy and rules for manual kiln and cooler control (which
parameters were used). Is there a correlation between the clinker quality and some kiln
control parameters (NOx, kiln amps, burning zone temperature)?

2.3 Combustion check
Process problems strongly influence the kiln behavior. Incomplete combustion leads to
enhanced Alkali/Sulfur cycles within the kiln system. If the molar Alkali/Sulfur ratio of the total
input of Alkalis and Sulfur is within the desirable range (0.8 to 1.5), minor encrustation and
Sulfur ring formation take place.
To avoid a bad or wrong implemented strategy and bad availability of the LINKman system,
a combustion check is required.
A combustion check contains:
Description and analysis of all fuel at the main firing
Gas analysis data at kiln inlet (O2, CO, NOx)
Assessment of sintering zone, coating and the burner position
Alkali/Sulfur balance. For this, sufficient samples of the raw meal, hot meal and clinker
have to be taken and analyzed.
Burner check (Primary air amount, fuel oil pressure, exit velocities, pressure fluctuations,
etc.)
Description and analysis of all fuels at the secondary firing
Gas analysis data at preheater exit (O2, CO, NOx)

2.4 Definition of the strategy
LINKman offers standard strategies for almost every kiln type. One of those strategies will be
pre-configured before delivery and modified on site during the commissioning.
To ensure a proper predefinition of the kiln and cooler control strategy, a description of the
current used strategy and the rules for manual kiln and cooler control is required.
Kiln Control:
Define the most important parameters estimating the burning zone temperature, the kiln inlet
temperature and the oxygen level.
A description of the most common kiln problems (dusty clinker, coating falls, etc.) will help
defining required special actions.
Cooler Control:
Define the most important parameters to ensure proper clinker cooling and high efficiency.
A description of the most common cooler problems (kiln rushes, red rivers, hot plates, etc.)
will help defining required special actions.
2.5 Control loops

For a proper working High Level Control System, the in the PLC installed control loops have
to work properly and smoothly. Therefore, the most important loops (kiln hood pressure
control, under grate pressure control, etc.) have to be revised and adjusted, if required. A
further possibility is the complete removing of a loop from the PLC in order to add it into the
LINKman strategy or to switch it off during LINKman control.
2.6 Instrumentation check

The instrumentation is the most important point for an automatic kiln control. The signals
have to be convenient and reliable. Therefore, the type and position of the most important
sensors have to be revised and corrected if required. Especially the gas analyzers need
periodic calibrations and maintenance inspections.
For a list with required process signals and its position see Report HES 98/6347/E.

3. APPENDIX 1
3.1 KILN MANUFACTURING DATA
Plant
Kiln No. :
Supplier:
Nominal Capacity: (t/d)
Max. Capacity: (t/d)
Type of fuels :

Heat Consumption: (kJ/kg)
Length: (m)
Diameter: (m)
Slope: (%)
Preheater: Number of strings:

Number of stages:
Precalciner: Type:
Type of fuels:

% Fuel PC / Kiln:
Burner Type:
% Primary air:
Cooler Type:
Number of grates:
Number of fans:
Width x length:

Field Preparation for High Level Control
F. Fehr
HES 98/6347/E
1. Input Signals For High Level Control .........................................................................363

1.1 Checklist: Assignment of Input Signals ............................................................365
1.2 Example of a Checklist.....................................................................................366
2. Output Signals For High Level Control ......................................................................367
2.1 Primary Control Loops, Actuators ....................................................................367
2.2 Example of a check list 1 .................................................................................371
2.3 Example of a check list 2 .................................................................................372
3. Interface To Process Control System.........................................................................373
3.1 Data Exchange.................................................................................................376
3.2 Security ............................................................................................................377
4. Examples Of Most Important Inputs In Respect Of The Location And Calibration Of
The Transmitter ............................................................................................................378
4.1 General ............................................................................................................378
4.1.1 Definition:......................................................................................................378
4.1.2 Location of the Sensor..................................................................................379
4.1.3 Calibration of the Transmitter .......................................................................379
4.2 Examples .........................................................................................................380
4.2.1 Temperature .................................................................................................380
4.2.2 Pressure .......................................................................................................385
4.2.3 Speed ...........................................................................................................388
4.2.4 Current..........................................................................................................388
4.2.5 Power/Energy ...............................................................................................389

1. INPUT SIGNALS FOR HIGH LEVEL CONTROL
The Input Signals of the High Level Control system (HLC) are the measured variables of the
controlled systems (kiln, cooler, mill). In the following a selection of the normally used input
signals is indicated.
Selection of Input Signals
Wet kiln
Sintering zone temperature (Pyrometer)
Secondary air temperature (if available)
Amps of kiln drive (Torque)
O2/CO/NOx at kiln inlet
Temperature at chain zone
Backend temperature
Backend pressure
Kiln hood pressure
Any other significant process variable used by the kiln operators
Preheater kiln
NOx at kiln inlet or at preheater exit
O2/CO at kiln inlet
Kiln inlet temperature
Kiln inlet pressure
Kiln hood pressure
Preheater exit temperature or second stage from top
Preheater exit pressure
O2/CO at preheater exit

Precalciner kiln
Basically the same signals as for preheater kilns, but additionally:
NOx at kiln inlet
Temperature exit lowest cyclone stage
Tertiary air temperature
Lepol kiln
O2/CO/NOx at kiln inlet
Temperature in hot chamber of Lepol grate (Pyrometer or Thermometer)
CO/O2 after intermediate fan
Temperature intermediate fan
Pressure in hot chamber above grate
Pressure in hot chamber underneath grate
Kiln hood pressure
Clinker cooler
Air rates of the individual fresh air fans
Pressures of chambers 1, 2, 3
Exhaust air temperature
Temperature of cooler plates
Middle air temperature (if any)
Cement mill
KW of mill motor
KW of bucket elevator
Rate of separator returns (t/h)
Noise level by electronic ear
Temperature/pressure at mill inlet and outlet
Production rate, e.g. belt weigher, pressure of pneumatic transport, etc.

1.1 Checklist: Assignment of Input Signals
INPUT SIGNALS Min. Max. Target

Name of Signal, Value
Sensor:
Location:
Significance:
Stability:
Comment:
Sensor:
Location:
Significance:
Stability:
Comment:
Sensor:
Location:
Significance:
Stability:
Comment:
Sensor:
Location:
Significance:
Stability:
Comment:
Sensor:
Location:
Significance:
Stability:
Comment:

1.2 Example of a Checklist
INPUT SIGNALS Min. Max. Target

1) Secondary air tempe rature 650C 1040F 815C
1200F 1900F 1500F
Sensor: o.k.
Location: o.k.
Significance: Indication is sensitive. Tendency is o.k.
Stability: o.k.
Comment: Measuring equipment is adequate. Useful
signal for LINKman II. Upper/lower limits are
exceeded, therefore the limits have to be
adjusted.
2) Cooler exhaust air temperature 95C 230C 150-175C
200F 450F 300-350F
Sensor: o.k.
Location: o.k.
Significance: not looked at
Stability not stable due to unstable cooler operation.
Comment: Measuring equipment is adequate.
3) Clinker temperature - -
Sensor: o.k.
Location: Existing location gives no representative
signal.
Significance: not looked at
Stability: not looked at
Comment: As a better location is not available, this
temperature should not be considered for auto-
matic control.
4) Grate speeds of cooler - -
Sensor: - -
Location: -
Significance: -
Stability: -
Comment: unproblematic signal.

2. OUTPUT SIGNALS FOR HIGH LEVEL CONTROL
2.1 Primary Control Loops, Actuators

Basically all High Level Control output signals go as setpoints to primary control loops which
then drive actuators.
Example:
Figure 1:
Setpoint Open
0-100% Posit.
PID M Flap
Close
Actual Value
Position 0-100%
Example of a Primary Loop
The behaviour of the variable controlled by the primary loop is influenced by the following
factors:
repeatability of the sensor (specifications)
absolute accuracy of sensor (calibration)
type of actuator: continuous, step-wise
tuning of PID controller
deadband to protect actuator
disturbances from outside (e.g. flushing material)

For HLC the primary loops have to fulfil the following criterias:
1) Tolerable deviation from setpoint
Figure 2:
Tolerable Deviation from Setpoint
x: max. tolerable deviation during a time > t

x and t depend upon what is controlled by the primary loop (e.g. coal, slurry to a wet kiln)

2) Sensitivity of setpoint change
What is the minimum applicable setpoint change that causes a reaction of the controlled
variable?
Figure 2:
Tolerable Deviation from Setpoint
The criterias which have to be fulfilled are given in the following list for every type of primary
loop. Most of the loops are not critical so that no criterias for the tolerable deviation are
given:

Table 1
Tolerable Deviation from Sensitivity of Setpoint

Setpoint Change
x[% of Span] t [min] SPmin [% of Span]
Cooler
Volume rate fresh air fans
- speed - - 0.5%
- damper position - - 1.0%
Kiln hood pressure
- speed - - 0.5%
Grate speed ratio - - 1.0%
Under grate pressure - - 1.0%
Kiln
Kiln fan
- fan speed - - 0.5%
Kiln drive - - 0.5%
Fuel rates 3% 0.5 0.5%
Feed: wet 3% 10.0 0.5%
preheater 3% 2.0 0.5%
Dust insufflat. rate 10% 0.5 1.0%
Tertiary air dampers - - 1.0%
Intermed. fan (Lepol) - - 0.5%
Lepol grate speed - - 1.0%
Water to granulator - - 1.0%
Mill
Mill feed 3% 0.5 1.0%
Separator speed - - 0.5%
Cooling air fan - -
- fan speed - - 0.5%
Water injection - - 1.0%

2.2 Example of a check list 1
PID Loop: Kiln drive
HAC 0% to 100% Unit
Code
Setpoint: ..... ..... to ..... [ ]
Actual Value: ..... ..... to ..... [ ]
Manipulated Variable: ..... ..... to ..... [ ]
(actuator)
Visual check of actuator:
Tuned:
Parameters: KP = .... Tl = .... [ ] TD = .... [ ]
Max. deviation: Delta x = .... [ % of span ]
during: Delta t = .... [ min ]
Sensitivity: .... [ % of span ]
SPmin =
Remarks:
PID Loop: Precalciner fuel rate

HAC 0% to 100% Unit
Code
Setpoint: ..... ..... to ..... [ ]
Actual Value: ..... ..... to ..... [ ]
(actuator)
Tuned:
PID Parameters: Kp = .... Tl = .... [ ] TD = .... [ ]
SPmin =
Remarks:

2.3 Example of a check list 2
PID Loop: Main burner fuel rate
HAC 0% to 100% Unit
Code
Setpoint: ..... ..... to ..... [ ]
Actual Value: ..... ..... to ..... [ ]
(actuator)
Tuned:
Parameters: Kp = .... Tl = .... [ ] TD = .... [ ]
SPmin =
Remarks:
PID Loop: Kiln feed

HAC 0% to 100% Unit
Code
Setpoint: ..... ..... to ..... [ ]
Actual Value: ..... ..... to ..... [ ]
(actuator)
Tuned:
PID Parameters: Kp = .... Tl = .... [ ] TD = .... [ ]
SPmin =
Remarks:

3. INTERFACE TO PROCESS CONTROL SYSTEM
For any process control application, the High Level Control System needs to read data from
the process (inputs from sensors) and to write data to the process (outputs to primary control
loops). Moreover, for kiln/mill control, the High Level Control System requires also operator
inputs such as:
operating targets (NOx, O2, etc.)
operator setpoints (kiln feed, fuel feed, etc.)
laboratory data (clinker factor, fuel heat value, % free lime etc.)
These inputs can be entered either from the LINKman II or from the Process Control System
(PCS) The data communication between the PCS and the HLC is mostly based on a RS
232/422 serial link with the PCS brand specific communication protocol. The data set for
exchange has to be pre-processed and stored in a specific memory location to be available
for the data communication software.
For a kiln application for example, process data are accessed every 10 seconds and outputs
are sent every 5 minutes, or with higher frequency if required, to update setpoints. The kiln
strategy, for example runs every minute.
To connect the LINKman II to the PCS, two possible configurations are proposed:
If the PCS has a bus connecting the whole plant (plant loop), then the LINKman II is
interfaced by using a serial link to the interface box of the loop. An example is presented
on the next page where the LINKman II is interfaced through a CIU (Computer Interface
Unit) to a Bailey Network 90.
If the PCS is structured according to the plant departments, the LINKman II has access
to the individual departments through individual serial lines. The second picture shows a
typical interface in a application where an Allen Bradley Programmable Logic Controller
is used in connection with the Holcim ODH system as Men Machine Interface (MMI).

Figure 4:
Integration of HLC System

PCS: Bailey
MMI
LINKman Graphic
CIU
SUPERLOOP
PS
Q:\sek -pbu\Kurse\zem-kurs\Folien\020596 Ff/ip

Figure 5:
Integration
Integration of HLC
HLC System
Plant: UV PCS: ODH, AB
MMI
LINKman Graphic
RS232
KF2
DH+
PLC5
Q:\sek -pbu\Kurse\zem-kurs\Folien\020596 Ff/ip

3.1 Data Exchange
The PCS does not only act as data acquisition system for the LINKman II, it serves also as
operator interface (MMI, setpoint change, manual mode control), it hosts the basic loop
controller, is responsible for automatic/manual switching including fail safe procedure if the
LINKman II or communication breaks and does the alarm handling for the whole process
(also HLC alarms). Because of this, the communication data set consists not only of analog
inputs and outputs, but also of digital signals:
Digital control bits from the PCS (digital inputs)
Digital alarm and status bit from the HLC (digital outputs)
DIGITAL INPUTS
The PCS system has to provide different status bits from the process to indicate specific
conditions which the LINKman II needs to work for proper operation.
Example:
LINKman II on/off
failure of instrumentation
group ready/running
direct/indirect operation of raw mill
select status (type of cement etc.)
DIGITAL OUTPUTS
In a similar way, the LINKman II has to send status bits to the PCS to inform the operator
about its status.
Example:
LINKman II available
on/off line
normal condition
upset condition
alarms
etc.

3.2 Security
The communication and proper execution of the program have to be guarantied. Therefore,
a watch-dog function on the PCS has to be realised:
Example:
Communication/HLC watch-dog
To check the communication as well as the LINKman II operating status (HLC on or off), a
watch-dog toggle-bit has to be programmed.
Figure 6:
LINKman
NOT
Toggle-bit Toggle-bit
OUT IN
Timer on
OR LINKman off
PCS
NOT Timer on
If the LINKman and the communication are on, the toggle-bit oscillates
The two on-delay timers control the on- and the off-time of the toggle-bit. If the time exceeds
the pre-set delay time of the timers (typical 30 s), the toggle-bit has not been inverted and
this means, that the LINKman II or the communication is off.
The PCS is responsible to switch the HLC on- or off-line (on operator's request) and to
monitor the HLC's on/off-state feedback and to switch the setpoint signals accordingly
(setpoint from operator MMI or from HLC). If the HLC goes off-line and the watch-dog
detects a problem, the setpoint will be set to the operator control and an alarm has to be
evoked.
Note: The setpoint switching has to be made bumpless. (see PID subject)

4. EXAMPLES OF MOST IMPORTANT INPUTS IN RESPECT OF THE LOCATION
AND CALIBRATION OF THE TRANSMITTER
4.1 General
4.1.1 Definition:
Range = Max. - Min. of the sensor
Example: 0C - 1000C
Span = Used range for the electrical signal
f or monitoring the process
Figure 7:

4.1.2 Location of the Sensor
The location of the sensor is the most important thing in the instrumentation. There are a few
locations which are really false and a lot which are good. For maintenance of the sensors it
is quite frequently necessary (especially in the kiln area), that the location of the sensor
should be above all easy accessible.
4.1.3 Calibration of the Transmitter

The calibration includes first of all the checking of the transmitter. There are different
methods how to check:
1) Reference measurement: The sensor will be tested on a well-defined media (boiling oil,
known weight, etc.)
2) Comparison measurement: Measure the process value by an other measuring device
(hand-held thermometer/multimeter, pressure U-tube, etc.)
3) Find out the actual physical measurement by measuring the electrical signal coming out
of the sensor (only if the characteristics are well-known,should not be used on
thermocouple due to the cold reference junction).
4) Simulating the process with an alternative source by entering a calibrated signal on the
primary side.
While applying method 1 or 2 as described above, the process value on the highest level
(screen of the supervisor or high level control system) - if already installed - has to be
monitored. If o.k., the procedure of course is quickly and successfully finished. If not, the
output of the transmitter has to be measured by an Amp-Meter (4...20mA) and the
transmitter has to be calibrated according its manual. In very seldom cases, the problem is in
the scaling of the PLC or the supervisory system.
Note: Two hints on calibration

Do a Plausibility Check. This means that the signal should be checked of its
plausibility. Is it on its expected value? If not, there is something wrong. Do not
accept everything you measure. Repeat the measurement if it is not plausible or
try to make the calibration by an other way.
Accuracy. If the sensor is not accurate enough or the transmitter has a too large
range, try to calibrate it within the operating range.

4.2 Examples
4.2.1 Temperature
EXAMPLE 1: KILN INLET TEMPERATURE
LOCATION:
Figure 8:
Easy accessible for maintenance

Cleaned by pulling out (daily)
Tip of sensor should be in the air-stream but not affected too much by incrustation
Sensor should not be damaged by falling material
The final position has to be evaluated

CALIBRATION:
Figure 9:
Inject a voltage in mV without the sensor into the transmitter, while measuring in the
4...20mA line. Take readings at the display for 0 and 100%.
SPAN: minimum: 0% = 1000C = 4.0mA
maximum: 100% = 1250C = 20.0mA
unit: 1% = 2,5C = 0.16mA
typical value: 80% = 1200C = 16.8mA (16:100x80+4)

EXAMPLE 2: PYROMETER (at kiln hood for clinker temperature)
LOCATION:
Figure 10:
Aim spot 30 cm (1 foot)

Do not point into flame (radiation)
Aim the pyrometer below the flame into the clinker just before the clinker flows out of the
kiln
Dust on the lens of the pyrometer or between lens and clinker affects the measurement,
therefore, choose a short measuring distance
CALIBRATION:
Calibration according to the manufacturer. (Does not have to be calibrated under normal
circumstances)
maximum: 100% = 1600C = 20.0mA
unit: 1% = 5C = 0.16mA
typical value: depending on the kiln
Note: The absolute temperature is not so important, changes have to be monitored.

EXAMPLE 3: PRECALCINER LOWEST CYCLONES
LOCATION:
Figure 11:
1 Dead zone
2 Good position for material temperature
Find the 2 right locations out of 4
Parallel probes recommended
CALIBRATION:

maximum: 100% = 1000C = 20.0mA
unit: 1% = 4C = 0.16mA

EXAMPLE 4: PREHEATER EXIT TEMPERATURE
LOCATION:
Figure 12:
Good for minimizing of exhaust gas temperature

Equilibrium of tower streams
If the temperature of the second stage is more indicative, it is advantageous to use this
value
CALIBRATION:

maximum: 100% = 500C = 20.0mA
unit: 1% = 5C = 0.16mA

4.2.2 Pressure
EXAMPLE 1: KILN HOOD
LOCATION:
Figure 13:
1 Measuring points (hood pressure)

2 ambient pressure (heated up air)
3 transmitter
4 U- Tube for calibration purpose
CALIBRATION:
Measure the pressure by an U-Tube (differential measurement)
grate cooler: SPAN: minimum: 0% = 0.1mbar = 4.0mA
maximum: 100% = -0.3mbar = 20.0mA
unit: 1% = 0.004mbar = 0.16mA
typical value: 37.5% = -0,05mbar = 10.0mA (16:100x37.5+4)
planetary cooler: SPAN: minimum: 0% = -2mbar = 4.0mA
maximum: 100% = -3mbar = 20.0mA

EXAMPLE 2: KILN INLET PRESSURE
LOCATION:
Figure 14:
Daily cleaning required

Indicates ring formation, limits max. production
CALIBRATION:
grate cooler: SPAN: minimum: 0% = 0mbar = 4.0mA

maximum: 100% = 10.0mbar = 20.0mA
unit: 1% = 0.1mbar = 0.16mA
typical value: = 2-4mbar
planetary cooler: SPAN: minimum: 0% = 0mbar = 4.0mA
maximum: 100% = 10mbar = 20.0mA

EXAMPLE 3: PREHEATER EXIST PRESSURE
LOCATION:
Figure 15:
Please note, that this pressure may be different from the ID-fan inlet pressure because of
the resistance in the tubes.
Typical value: 5mbar
i.e. 5mbar pressure difference from preheater top to
ID-fan (at bottom, difference in height : 100 m)
CALIBRATION:
grate cooler: SPAN: minimum: 0% = 0mbar = 4.0mA

maximum: 100% = 10.0mbar = 20.0mA
unit: 1% = 1mbar = 0.16mA

4.2.3 Speed
EXAMPLE: DC-DRIVE KILN
LOCATION:
DC-Drive Panel. Usually the transmitter is in the panel as well.
CALIBRATION:
By a hand-tachometer (analogous or digital) on the kiln motor or motor coupling.
SPAN: minimum: 0% = 0rpm = 4.0mA

maximum: 100% = 1500rpm = 20.0mA
unit: 1% = 15rpm = 0.16mA
typical value: 85% = 1275rpm = 17.6mA (16:100x85+4)
4.2.4 Current
EXAMPLE: KILN DRIVE CURRENT
LOCATION:
DC-Drive Panel. Usually the transmitter is in the panel as well.
CALIBRATION:
By the Amp-meter on the DC-drive panel while measuring in the 4...20mA line.
SPAN: minimum: 0% = 0A = 4.0mA

maximum: 100% = 1000A = 20.0mA
unit: 1% = 10A = 0.16mA
typical value: 30% = 300A = 4.8mA (16:100x30+4)

IMPORTANT:
The kiln amps are used to indicate the torque (and with the torque the:
coating falling
hot or cold clinker)
If the field current is not kept constant by a thyristor controlled unit, the kiln drive
amps have to be multiplied by the field current.
The value has to be filtered in case of a planetary cooler, but in a way that ring breaks
still can be registered
Span must be 100% nominal motor current
4.2.5 Power/Energy
EXAMPLE: POWER OF MILL DRIVE
LOCATION:
Medium-voltage Switchgear.
CALIBRATION:
Calibration is usually not necessary. For checking the value, it can be calculated: voltage
(phase to phase) x current x 1.732 (square root of 3) x power factor (cos, see motor
data, typical 0.97).
RANGE: minimum: 0% = 0kW = 4.0mA

maximum: 100% = 2000kW = 20.0mA
unit: 1% = 20kW = 0.16mA
typical value: 98% = 1960kW = 19.68mA (16:100x98+4)

Chapter 6
Practical Work
1. Control Technique - Practical Work ......................................................................... 391
2. PRACTICAL EXERCISE - Fuzzy-Control .................................................................. 415
3. Gas Flow Measurement - Practical Work................................................................. 419

Control Technique - Practical Work
1. The Control Loop..........................................................................................................392

2. Dynamic Behavior Of The Process .............................................................................395
2.1 Proportional Factor of the Process...................................................................395
2.2 Dead Time Element .........................................................................................395
2.3 Delay Element..................................................................................................396
2.4 Combinations ...................................................................................................396
3. Controllers ....................................................................................................................397
3.1 Different Types of Controllers ..........................................................................397
3.2 PID - Controller ................................................................................................397
3.3 Analog and Digital Controller ...........................................................................405
3.4 Terminology .....................................................................................................406
4. How To Turn The Knobs - Optimal Settings Of The Controller................................407
4.1 Stabilization Point.............................................................................................407
4.2 Adjustments of the PID Settings ......................................................................408
4.3 Optimal Settings of the Controller ....................................................................410

SUMMARY
Lack of manpower, quality requirements and complicated processes have increased the
importance of control to industry.
This lecture gives an introduction into the basic control theory and its use. Because of the
variety of todays control techniques, applications are purposely omitted.
1. THE CONTROL LOOP

The functioning and the components of a control loop are explained by the example of
cooling tower control loop.
Figure 1 Cooling tower control loop
1) Process: The Cooling Tower

The process consist of those machines and equipment which are controlled to produce a
desired output.
2) Process Inputs and Outputs
For example:
Process inputs: flowrate hot gas
temperature hot gas
flowrate cooling water
Process outputs: flowrate cooled gas
temperature cooled gas
3) Controlled Output (c): Outlet Gas Temperature
The purpose of the cooling tower control is to achieve a constant outlet gas temperature.
Variations of the other process outputs are of no importance.
4) Disturbances: Variations in Flowrate and Temperature of the hot Gas
Variations of the process inputs cause disturbances to the process and therefore also to
the process outputs, especially to the controlled output.

5) Manipulated Variable (m): Cooling Water Flowrate
The cooling of the gas stream is achieved by spraying water into it. The degree of the
cooling depends on the amount of water sprayed. Therefore, the outlet gas temperature
can be controlled by manipulating the cooling water flowrate.
6) Set point (s): Desired Outlet Gas Temperature
The level at which the controlled output shall be kept has to be defined. This reference
input or setpoint can either come from an operator or from an other control system.
7) Feedback Elements (h): Thermocouple
The process output you want to control has to be measured because you cannot control
what you dont know. The feedback elements are always measuring devices.
8) Actual Value (x): Temperature Signal
The signal which is delivered by the feedback elements
9) Error (e): Deviation between measured Gas Temperature and the desired
Temperature
The actual value is compared with the setpoint (desired value). The controller will react
according to the deviation of these two signals.
Depending on the process it has to be selected between forward and reverse action:
Forward Action:
If the actual value is below the setpoint an increase of the controller output is required
and vice versa.
Example: If the temperature in a kiln is too low the fuel has to be increased to heat up
the kiln.
Reverse Action
If the actual value is below the setpoint a decrease of the controller output is required
and vice versa.
Example: If the temperature after a cooling tower is too low, the flowrate of the cooling
water has to be reduced to increase the temperature.
The switching between the two actions can simply be done by calculating the error either
by subtracting the actual value (x) form the setpoint (s) (forward action) or by subtracting
the setpoint (s) from the actual value (reverse action).
10) Control Elements (g1)
According to the error, the control elements act upon the manipulated variable. Two main
elements:
Controller: The controller is a device which generates the appropriate control signal
according to the error (e) (see also the chapter Controller).
Actuator: The actuator transforms the controller output (y) into a change of the
manipulated variable. In our example: A control value which acts upon the cooling
water flow.

11) The commonly used expressions introduced above are summarized in the figure
below:
Figure 2 Cooling tower control loop
12)Block Diagram
Every closed loop control system can be represented with the same diagram.
Figure 3 Block disgram of control loop

2. DYNAMIC BEHAVIOR OF THE PROCESS
Depending on the nature of the process the output response to an input change or a
disturbance varies. The most important tool to characterize the dynamic behavior of a
process is the transient responses: the reaction of the process to a step-change of the
input.
2.1 Proportional Factor of the Process

The proportional factor describes how big the change of the output is caused by a one-unit
change of the process input:
Proportional factors of the process = B / A
Figure 4 Proportional factor of the process
2.2 Dead Time Element

Most processes require a finite response time, often with a magnitude of minutes or hours.
The dead time is defined as the difference in time between an input change and the resulting
output change. The dead time slows down controller response because the controller cannot
act until an error in output is detected.
An example for a typical dead time element is a belt conveyor. The dead time is given by the
length and the speed of the belt.
Figure 5 Dead time

The cooling tower as an example contains multiple dead time elements which can be added
together:
reaction time of the control value
time the change in the water flow rate needs to reach the spraying nozzles
time the gases need to pass the height of the cooling tower
reaction time of the thermo-couple
2.3 Delay Element

The reaction of a process to an input change is depending on the physical and chemical
reactions controlling this process: heat transfer, mass transfer, chemical reactions. As all
these reactions do not just happen within no time but follow certain time laws, the whole
process shows certain time delays. These time delays can be classified in different types.
Example:
First order element, second order element. The most frequent is the second order delay.
Figure 6 Delay time
The delay in the cooling tower system is the time between the first noticeable change in the
outlet gas temperature and the moment when the new stable temperature is reached.
2.4 Combinations
In most of the cases the process is a combination of a dead time and a delay element
(second order).
Figure 7 Dead time plus delay time

3. CONTROLLERS
As mentioned before the controller is a device which generates according to the deviation
between desired and actual value an appropriate control signal to bring this deviation to
zero.
3.1 Different Types of Controllers

1) On-Off Controller
The controller output can only have two positions;: fully open and fully closed. The
controlled variable is controlled within a band. When the variable falls below the lower
limit of the band, the controller output changes to open. The variable starts to rise
again. When the upper limit of the band is reached, the controller output changes back to
closed. The variable starts to fall again.
Example: Temperature control of a refrigerator.
2) Continuous controller (PID-controller)
The output can have any position between open and closed.
3) Adaptive controller
The controller adapts automatically its parameters to changing process characteristics. It
has to be differentiated between controllers which adapt themselves continuously and
controllers that adapt only when this is desired.
4) Predictive Controller
The predictive controller uses a mathematical model to predict the behavior of the
controlled process. With this information the controller can optimize the control actions.
3.2 PID - Controller

The most frequently used controller is the PID-Controller. The total control action is a
combination of three basic control actions:
1) P- Controller (proportional controller)
The output of the proportional controller is the error (e) multiplied by a constant factor.
This factor is called the proportional constant or the proportional gain.

Figure 8 Proportional Controller
The control algorithm:

y = Kp e
Kp: Proportional constant
Characteristic Features:
Increasing Kp causes stronger control actions and thus a faster return to the setpoint.
But with increasing Kp the tendency of the system to cycle increases (see figure 9).
Off-set: the P- controller is adjusted to work for one specific load of the controlled system
(in the cooling tower example: one specific flowrate of the hot inlet gas). If the load
changes, a permanent error.

2) I-Controller (integral or reset Controller)
As long as a deviation between setpoint and actual value exists, the I-controller sums the
deviation up with time and produces an output proportional to this sum.
Figure 10 Integral controller

y = KI edt
KI: integral constant or integral gain
KI can be expressed with the proportional constant
KI = Kp/TI
Kp: proportional constant
TI : integral action time - Time after which the output of the I-controller
is equal the output of a P-controller (for a step).
As a permanent error leads to a continuously increasing control signal, an I-controller is
very suitable to eliminate an off-set.
With increasing KI increasing tendency to oscillate

3) D-Controller (differential, derivative or rate controller)
The D-controller produces an output which is proportional to the rate of change of the
deviation between setpoint and actual value. The D-controller weights the speed of a
disturbance.
Figure 11: D-Controller

Y = KD de/dt
KD: derivative constant or derivative gain.
KD can be expressed with the proportional constant KD = Kp TD
Kp: proportional constant
TD: derivative action time. Time after which the output of the D-controller
is equal to the output of a P-controller (for a ramp).
D-controller reacts earlier because it detects slope errors which occur before real errors.
A D-controller cannot be used alone. It is used in combination with P or PI controller
actions.
As the D-controller reacts to slopes, it can be unstable for highly oscillatory systems.

4) PID - Controller
The three basic control actions can be used alone (whit the exception of the D-controller)
or as combinations like PI-Controller, PD-controller, PID-controller. In this case the total
control action is the sum of the different basic control actions which means the controllers
are arranged in parallel.
Example: PID - Controller (three mode controller)
Figure 12 PID Controller
The control algorithm when using the parameters Kp, KI, KD:
y = Kp e + KI edt + KD de/dt
The control algorithm when using the parameters Kp, TI, TD:
y = Kp e + ( Kp / TI ) edt + Kp TD de/dt

Figure 13 Comparison of the different controllers

Which controller to choose depends on the plant behavior. Below you will find some
guidelines on how to choose a controller for a given plant.
In a cement plant a lot of controllers are just PI - controllers because many of the processes
which have to be controlled are relatively slow and thus a derivative control action (D -
controller) is not necessary
Table 1 Guidelines how to choose a controller

3.3 Analog and Digital Controller
A PID - Controller can be realized as an analog or as a digital controller.
a) Analog Controller
The control loop is closed all the time. The control operations are executed at every point
of the time scale.
The control algorithms are realized on the hardware side with appropriate circuits. This
means: the more sophisticated the control algorithms the bigger the hardware.
Figure 14 Analog Controller
b) Digital Controller
The control loop is only closed at certain sampling points. The different arithmetical
operations of the control algorithm are not executed simultaneously but one after the
other.
The control algorithms are realized on the software side.
Figure 15 Digital Controller

3.4 Terminology
Sometimes it can be very confusing that in the control technique a lot of different
terminologies are used. Some of the terms used are listed below. We recommend strongly to
stick to the terms used in the other chapters of this paper.
Kp KI KD
Controller Kp *e + Kp/TI *edt + KpTD * de/dt
output =
100/PB (100/PB)*(R/M) (100/PB)*R
T
Kp = 100 / PB proportional gain or

proportional constant
PB = Xp proportional band
KI = Kp / TI integral gain or
integral constant
1 / TI = R/M reset gain or
reset rate
R/M reset per minute
TI integral action time
KD = KpTD = (100/PB) derivative gain or
*RT derivative constant
TD = RT derivative action time or rate time or
rate gain

4. HOW TO TURN THE KNOBS - OPTIMAL SETTINGS OF THE CONTROLLER
4.1 Stabilization Point

Two possibilities to interrupt the process:
changing the setpoint
disturbing the process
According to settings of the PID parameters Kp, TI, TD the reaction of the control system to
such an interaction varies:
Figure 16 Changing the setpoint
Figure 17 Disturbing

4.2 Adjustments of the PID Settings
the list and the figure below shall give you a better feeling about what happens if you change
one of the PID parameters Kp, TI, TD.
Adjustment Reaction
Increasing Increasing Kp Speeding up control action:
P action - smaller amplitude,
- smaller period of oscillation
Decreasing offset
Increasing tendency to oscillate
Decreasing Decreasing Kp Slowing down control action:
P action - bigger amplitude
- bigger period of oscillation
Increasing offset
Decreasing tendency to oscillate
Increasing Decreasing TI Bigger amplitude
I action Smaller period of oscillation
Faster elimination of the offset
Decreasing Increasing TI Smaller amplitude
I action Bigger period of oscillation
Slower elimination of the off-set
Increasing Increasing TD Smaller amplitude
D action Bigger Period
First degreasing but beyond a certain
point increasing tendency to oscillate
Decreasing Decreasing TD Bigger amplitude
D action Smaller period
First increasing, beyond a certain point
decreasing tendency to oscillate
Adjustment of the PID Settings Kp

Adjustment of the PID Settings Ti
Adjustment of the PID Settings Td

4.3 Optimal Settings of the Controller
Chien-Hrones Method
The optimal controller settings are determined from the reaction curve of the opened control
loop to a step disturbance. This method is also called
process reaction curve method
lag reaction rate method
1) open the loop (switch controller to manual)
2) create a step on the controller output signal
Figure 18 Step in control element position

3) record the answer of the process to this step
Figure 19 Process reaction curve

4) determine the following expressions:
Td: Dead time [min]
Ts: Time constant [min]
N: Slope = reaction rate in % of feedback signal range per minute [%min]
C D*100
N=
TS
P: % change in control element position [%]
A
p= *100
B
5) Optimal settings (rough estimation)
Kp de
Output = Kpe+
TI edt +Kp * TD *
dt
P-Controller Kp = p
N*Td
PI-Controller Kp = p
0.92*
N*Td
TI = 3.3 * Td
PID-Controller Kp = p
1.2*
N*Td
TI = 2 * Td
TD = 0.5 * Td

Example: Cooling tower control loop
A step is made in the position of the cooling water control value from 60 % to 50 %. The
measured outlet gas temperature signal shows the following picture (the range of the
temperature signal is from 100 C to 300 C):
Figure 20
Figure 21

Td = 5 min.
Ts = 4 min.
180150
* 100
300100
N= =3.75% / min .
4
5060
p= *100 =10%
100
Optimal Settings for PID-Controller
Kp = 1.2*p = 1.2*10 = 0.64%/%
N*Td 3.75*5
TI = 2 * Td = 2*5 = 10 min.
TD = 0.5 * Td = 0.5 * 5 = 2.5 min.

PRACTICAL EXERCISE - Fuzzy-Control
F. Fehr
98/6346/E
1. Introduction...................................................................................................................416
2. Exercise.........................................................................................................................417
2.1 Start / Set-up ....................................................................................................417
2.2 Input/output Fuzzyfication ................................................................................417
2.3 Implementing rules...........................................................................................418
2.4 Testing .............................................................................................................418
2.5 Save .................................................................................................................418
3. Miscellaneous ...............................................................................................................418

1. INTRODUCTION
This is an approach to teach the Fuzzy theories in a practical way, without a lot of paper and
formulas. On behalf of a Siemens Demo Software, a Fuzzy controller will be shown and
worked out empirically.
The goal of this exercise is to get a basic idea about
What is Fuzzy-Control
How does it work
There is a diskette included for every participant. Please take this diskette home as an
example that you can recall anytime, whenever you are in a position to work with fuzzy-logic
or explain to someone else what fuzzy is really about.
The task of this exercise is:
To set-up a temperature controller which uses a room and an outside temperature signal
thus actuating a heater and an air-conditioner.
Room Temperature
Controller Heater
Outside Temperature Air Conditioner

2. EXERCISE
2.1 Start / Set-up

Start Fuzzy.exe
Create new Project with the name "TEMPCON" ("File" "New...") with 2 Inputs and 2
Outputs.
Assign names to inputs and outputs (see pictures in 1. Introduction)
Define 3 temperature ranges (membership functions) for room temp (e.g. chilli,
comfortable, warm).
Define 3 temperature ranges for outside temp (e.g. cold, warm, hot).
Define 4 ranges for the heater (e.g. off, low, high, on) and 3 ranges for the air conditioner
(e.g. off, med, on).
Adjust the temperature scale to reasonable values (e.g. room temp. from 1030C).
2.2 Input/output Fuzzyfication

Perform the Fuzzyfication of the 2 inputs and 2 outputs by asking the following questions
and set the curves accordingly:
where is 0% cold
from where to where is 100% warm
Out of 100 people, how many people would say that is chilli (first curve) at what temperature.
Do the same for the other 2 curves. The number of people (in %) which is actually the
degree of fulfilment (or the percentage of truth) is on the vertical axis, the temperature on the
horizontal axis.
On the example in the next picture below, it is 50% chilli (50 out of 100 people would call it
chilli) at 20.5C and after 22C it is not chilly anymore.

2.3 Implementing rules
Now fill in the rules in "Ifthen".
2.4 Testing
Use the menu "View" to simulate input values as constants or curves (sub menu "Curve
Parameters") and test the controller (sub menu "Curve Recorder").
2.5 Save
Save your project (File Save or File Save As).
For those who could not work through the exercise, there is an example of this exercise on
the diskette, named exercise.txt. Rename it to *.fpl to be able to call it from the demo
software.
3. MISCELLANEOUS
On the Cement Course CD-Rom, the directory \MISC\FUZZY contains all files and
information being used for the practical work of fuzzy control logic described in the chapters
above. Below a list of all files of the directory \MISC\FUZZY
Files Description
FUZZY.EXE program file for the Fuzzy demo

FUZZY.ICO icon for the above file
README.TXT supplementary supplier information about the fuzzy package
SHUTTERS.FPL example "shutters controller"
PENDULUM.FPL example "inverse pendulum"
EXERCISE.TXT example "temperature control"
S7FUZ1EF.PPT Power Point slides about the fuzzy package

Gas Flow Measurement - Practical Work
H. Nyfenegger
PT 97/14231/E
1. Theoretical Basis..........................................................................................................420
1.1 The Law of Bernoulli ........................................................................................420
2. Local Velocity And Pressure Measurements .............................................................422
2.1 The Measurement of the Pressures.................................................................422
3. Gas Flow Measurement By The Prandtl Tube............................................................424
4. Integral Gas Flow Measurement With Nozzles And Orifices ....................................427
4.1 Standard Nozzles.............................................................................................429
4.2 Standard Orifices and Venturi-Nozzles ............................................................431
5. Calculation Of Density .................................................................................................433
5.1 Requests, Calculation of Real Conditions........................................................433
5.2 Gas Compositions............................................................................................434
5.2.1 Calculation of a Heterogeneous Gas............................................................434
6. Practical Work In Gas Flow Measurement .................................................................435
6.1 Test Equipment ................................................................................................435
6.2 The Prandtl Tube .............................................................................................436
6.2.1 Calculation of the Gas Density .....................................................................438
6.2.2 Numerical Calculation of Density..................................................................438
6.3 Calculation of the Gas Velocity ........................................................................440
6.4 Calculation of the Gas Flow Rate.....................................................................442
6.5 Anemometers...................................................................................................443
6.5.1 Type A (Vane Wheel Anemometers)............................................................443
6.5.2 Type B Thermal Anemometers.....................................................................443
6.5.3 Vortex Anemometers (not used in practical work)........................................444
6.6 Characteristic Curves of the Fan......................................................................444
6.6.1 Measurements..............................................................................................445
6.7 INLET NOZZLE................................................................................................446
6.8 Manometers .....................................................................................................448
6.8.1 Inclined Stationary Manometer.....................................................................448
7. Symbol List Of Gas Flow Measurement .....................................................................449

SUMMARY
This paper shows the basic equations for gas flow measurement. Furthermore, the principles
of gas flow measurement by the Prandtl tube as well as the Orifices and Nozzles are
explained.
Some of the presented principles and instruments will be used for the practical work carried
out on a test equipment in the laboratory hall.
1. THEORETICAL BASIS
1.1 The Law of Bernoulli

(Daniel Bernoulli, 1700..1782 Swiss Scientist)
In aeronautics, a law or theorem state that in a flow of incompressible fluid the sum of the
static pressure and the dynamic pressure along a streamline is constant if gravity and
frictional effects are disregarded.
Resultant from this law is that if there is a velocity increase in a fluid flow, there must be a
corresponding pressure decrease. Thus an airfoil, by increasing the velocity of the flow over
its upper surface, drives lift from the decreased pressure.
As originally formulated, a statement of the energy conservation (per unit mass) for a non-
viscous fluid in steady motion. The specific energy is composed of the kinetic energy v2/2,
where v is the speed of the fluid; the potential energy gz, where g is the acceleration of
gravity and z is the height above an arbitrary reference level; and the work done by the
pressure forces of a compressible fluid v dp, where p is the pressure, v is the specific
volume, and the integration is always with respect to value of p and v on the same parcel.
Thus, the relationship
v2/2 + gz + v dp = constant along a streamline
is valid for a compressible fluid in steady motion, since the streamline is also in path. If the
motion is irrational, the same constant holds for the entire fluid.
If the fluid is compressible If the fluid is incompressible:

& 1=m
m &2 1= 2
1 V&1= 2 V&2 V&1=V&2
1 A1 w 1= 2 A2 w 2 A1 w 1=A2 w 2

If the fluid is incompressible, that means is constant:
1= 2 =
2 2
P1 w 1 P2 w 2
+ = +
2 2
Symbols and units:
A Cross section area [m2]

d Diameter of the smallest aperture [m]
D Diameter of the tube [m]
m Cross section area ratio m = d2 / D2 [-]
m Mass flow rate [kg/s]
p Pressure N / m2]
Volume flow rate [m3 / s]
w Velocity [m / s]
Density [kg / m3]

2. LOCAL VELOCITY AND PRESSURE MEASUREMENTS
The above figure shows a fluid flowing around an obstacle. One of the streamlines reaches
the body at the point 2, called stagnation point. At this point the velocity of the fluid is zero.
The Bernoulli equation, for the undisturbed flow at point 1 to point 2,
2 2
P1 w 1 P2 w 2 2
+ = + or (w2 = 0) P2 =P1+ w1
2 2 2
where P2 is the total pressure

P1 is the static pressure
2
w1 is the dynamic pressure
2

p2, p1 and 2 w 2 are pressures called total, static and dynamic pressure.
2.1 The Measurement of the Pressures

The total pressure of a fluid can be measured at a stagnation point. For this purpose the so-
called Pitot tube is used.

At the point 2 a stagnation point is created where the total pressure can be measured.
A boring placed vertically to the flow (1) measures the pressure independent from the
direction. This pressure is called static pressure.

3. GAS FLOW MEASUREMENT BY THE PRANDTL TUBE
The dynamic pressure, defined as

w 12
2
can be expressed as the pressure difference between the total and the static pressure. The
relation is

p 2 p1 = w 12
2
This difference is measured with the Prandtl Tube.
The inner part of this instrument corresponds exactly to a Pitot Tube and the total pressure is
measured there. The outer part has the function of the static pressure measurement.
The difference between the two connections is the dynamic pressure.
The direction of the flow should correspond to the axis of the Prandtl Tube. Deviations until
10 however, do not essentially influence the measuring accuracy.
The error in the measurement of the pressure is below 1%.

With the dynamic pressure the fluid velocity can be calculated as follows:
2 (P2 P1 )
w 1=

The Prandtl Tube measures a local velocity only. In order to get an integral velocity (e.g. the
average value) the integration of several local velocities over the cross sectional area must
be made.
1
A
w= w dA
A
The gas flow rate is finally calculated by multiplying the average gas velocity with the cross
sectional area.

Specific information on the gas flow measurement by the Prandtl Tube is given in chapter 6
"Practical work in gas flow measurement.

4. INTEGRAL GAS FLOW MEASUREMENT WITH NOZZLES AND ORIFICES
The idea of this method is to create a change of cross section of the whole fluid stream. By
means of this change also the pressure and the velocity change and from the pressure
difference the velocity can be calculated.
A precondition for this measurement method is a more or less equalized velocity profile.
Disturbances like
Flow separation behind bends, valves, etc.
Pulsation behind a piston type compressor
High dust loading
etc.
falsify the measurement and make it useless in a lot of cases. As a rule of thumb it can be
said that prior to such a flow measurement a straight tube length of more than 10 tube
diameters is to be provided.
Between point 1 and 2 the continuity condition can be formulated as

A1 w1 = A2 w2
Between the same two points the Bernoulli equation (see 1.1) gives:
2 2
p1 w 1 p2 w 2
+ = +
2 2

From the two relations the following equation is derived:
2 (P1 P2 )
w 1=
A 2
1 1
A2
The velocity w1 can be calculated from the pressure difference between the undisturbed flow
(point 1) and the accelerated flow in the smallest cross-sectional area.
The pressure p2 is lower than pressure p1 and compared to the pressure p1, is p2 a negative
pressure.
The relation is
A 2
(P1 P2 )= W 1 1
1
2
2 A2
The pressure difference is a function of the ratio square of the cross sectional area. Great
ratios produce therefore extreme underpressures! An example of such a great ratio is the
reduction of the cross-sectional area during coating formation in preheaters.
The change of the cross-section causes, especially when it is done suddenly, a disturbance
of the flow. The fluid cannot follow the wall and a contraction of the stream is produced.
The effective cross section Ae is now smaller than the real one and the calculated velocity
must be corrected. The correction cannot be calculated, it must be measured. For this
reason several types of integral measuring instruments are standardized by DIN 1952.

In the following the 3 basic standard instruments are sketched:
4.1 Standard Nozzles

Standard nozzles are used for tubes with diameters from 50 to 500 mm and a cross-
sectional ratio (smallest area against tube area) of 0.1 to 0.64.
The velocity in the tube and the volume flow, respectively, are calculated by the following
relation:
2 p
w = m
1
2 p
V& = m D 2
4 1
The factor considers the change in the density of a compressible fluid. For incompressible
fluids (e.g. water or air until a maximum velocity of 100 [m/s]) this factor has the value 1,
otherwise it is tabulated in the DIN - Standards 1952.

The value is the correction factor for the contraction of the fluid. It is a function of the
Reynolds number and the ratio m of the cross-sectional areas. This factor is also tabulated
in the DIN - Standards 1952. The table below shows an output of these standards.
Correction factors for the contraction = f (m2, Re) for standard nozzles in tubes with
smooth surfaces, valid for diameters D between 50 and 500 mm. Between the indicated
values of m2 (not m) can be interpolated linearly.

4.2 Standard Orifices and Venturi-Nozzles
Orifices and Venturi-nozzles are standardized in the same way as above-mentioned
standard nozzles.
The following figure shows the two standard orifices distinguished only in the pressure taps.

The following figure shows the two used Venturi nozzles distinguished only in the length of
the diffuser part.

5. CALCULATION OF DENSITY
5.1 Requests, Calculation of Real Conditions

The methods based on the law of Bernoulli needs the knowledge of the density of gas. For
simple gases like dry air this request makes normally no problem. In case of kiln gases e.g.
the density must be evaluated or approximated to the real condition.
N
p0 =760[Torr ]=760[mmHg ]=1.01325*10 5 2 or [Pa]
m
T0 =273.16[K ]=0[C ]
N
1[mmHg ]133.322 2 or [Pa]
m
In case of air and in function of the relative humidity is almost invariant:
Table 1
[%] o [kg/m3]
100 1.290
80 1.290
60 1.291
Thus, the actual gas density G can be found with the following correlation:
T0 pG 273.16G
G = 0 =1.29
TG p0 TG 1.01325105
where pG = actual static gas pressure [N/m2]
TG = actual gas temperature [K]
PG can be measured by adding the differential pressure between tube and the ambient to the
barometric pressure.

5.2 Gas Compositions
Mainly in the field of kiln, the gas compositions can be very different so that their density
must be determining. The greatest influences to the density come from the components
carbondioxide (CO2) and water content (H2O, humidity). Often both components are
unknown and cant be measured easily, so that the user depend from approximations.
5.2.1 Calculation of a Heterogeneous Gas

Density of dry gas composition: ndr
Component Content [%] Factor Part of Unit
CO2 1.977 [kg/m3,N, dr.]
O2 1.429 [kg/m3,N, dr.]
CO 1.25 [kg/m3,N, dr.]
N2 1.257 [kg/m3,N, dr.]
1.000 ndr = [kg/m3,N, dr.]
Density of wet gas composition:
ndr + W
nw =
0
1+fdr
Where W0 = Water content [kg/m3 n,dry]
W0
fdr = [m3/m3 n,dry]
0.804

6. PRACTICAL WORK IN GAS FLOW MEASUREMENT
6.1 Test Equipment

The measurement of the velocity profile and the volume flow is shown on a special test
equipment which is sketched in the following figure.
The following measuring methods are demonstrated and applied:

1) Prandtl Tube: The velocity will be measured on several points of the cross sectional
area by means of which the average velocity and the volume rate will be calculated.
2) Anemometers: At the outlet of the test equipment the mean velocity will be measured
with two different anemometers.
3) Fan Characteristics: With the knowledge of the characteristic curves of the fan given by
the supplier, the volume rate is determined by the pressure difference of the fan and the
revolutions of the fan.
4) Inlet Nozzle (Piezometer): At the suction side of the fan an inlet nozzle is attached. This
nozzle is calibrated and the volume rate can be calculated by means of the pressure
difference between the nozzle inlet and ambient.

6.2 The Prandtl Tube
Prandtl tubes are mainly used for speed measurements of gaseous mediums. The
application of a Prandtl tube is suitable for following measurements:
Higher Temperatures (Until 500C, depending of construction materials)
Few dust loaded gas flows (approx. 50 g/m3 max.)
Only for higher gas velocities (approx. 10 m/s min.)
The density of the medium must be known
The Prandtl tube itself is a very simple and cheap measuring instrument. Because the
measuring is selectively for an integral gas flow measurement several measuring points
must be measured. Therefore, the time consumption for the measurement itself and the
evaluation is considerable.
Most application in cement industry for Prandtl tubes are:
Air output of grate coolers
Riser ducts
Tertiary air ducts
etc.
Generally, the application of Prandtl tubes are indicated when other instruments e.g.
anemometers as a result of high temperatures or other reasons cannot be used.
If a Prandtl tube is put into a flowing medium, the following pressures are produced:

The total pressure Ptot acts on the ball-shaped measuring head, composed of the static
pressure Pstat and the dynamic pressure. The Bernoulli-equation formulated from the point 0
(ambient point) to the point 1 (pressure point) states:
G
pstat + w 2 =ptot total pressure
2
dynamic pressure (1)
static pressure
The flow speed is only influenced by the dynamic pressure:
2
w= (p tot p stat ) (2)
G
Its value is obtained from the difference of the total pressure Ptot and the static pressure Pstat
read on a differential pressure gauge.
(p tot pstat )= W g h (3)
From equation (2) and (3)
w = 2g W
h (4)
G
PS: Prandtl tubes are insensitive to deviations from the flow direction up to an angle of about
10 only (See chapter 3, Fig 2)

6.2.1 Calculation of the Gas Density
Theoretically, the real gas density G, considering the humidity, should be calculated by
determination of the water content or the dew point. The influence of the humidity is only
relevant in case of high temperature and high water contents, e.g. exhaust gas of the kiln.
In the case at hand the density of moist air on standard conditions
N
p0 =760[Torr ]=760[mmHg]=1.0132510 5 2 or[Pa]
m
T0 =273.16[K ]=0[C]
N
1[mmHg]133.322 2 or[Pa]
m
N
1[mbar] 10 2 [ 2 ]or[Pa]
m
and in function of the relative humidity is almost invariant:
Table 1
[%] o [kg/m3]
100 1.290
80 1.290
60 1.291
Thus, the actual gas density G can be found with the following correlation:
T0 pG 273.16pG
G = 0 =1.29 (5)
T G p0 T G 1.0132510 5
where pG = actual static gas pressure [N/m2]
TG = actual gas temperature [K]
PG can be measured by adding the differential pressure Ps between the tube and the
ambient to the barometric pressure.
6.2.2 Numerical Calculation of Density
Barometric pressure Pa = ................ [mbar]

= ................ 105 [N/m2]
Static pressure Ps = ................ [mmW] )
= ................ 105 [N/m2]
Gas pressure PG = pa + Ps =................ 105 [N/m2]
Gas temperature tG = ................[C]
TG = ...............+ 273 [K]

273.16pG
G =1.29
TG 1.0132510 5
273.16...........10 5 kg
G =1.29 =.................... .. 3
..........1.0132510 5
m

6.3 Calculation of the Gas Velocity
With the knowledge of the gas density G the gas velocity can be calculated by the following
equation (6)
1000
......
The velocity in the given tube is not constant over the cross section and a certain velocity
profile appears:
For this reason the average velocity must be found. This is made by an integration of the
local velocities over the cross sectional area.
In order to split the integration in discrete values, the cross sectional area was shared in
equal-sized concentric sections, each including four measuring points (two on each
measuring axis). The locations where the pressure measurement are to be carried out
depend on the desired total number of measuring points to be made on each measuring axis
(see Table 2 and Figure 4).
Table 2
12 measuring 10 measuring 8 measuring 6 measuring

points points points points
1 0.022 x D 0.025 x D 0.032 x D 0.043 x D
2 0.067 x D 0.080 x D 0.105 x D 0.146 x D
3 0.118 x D 0.145 x D 0.194 x D 0.296 x D
4 0.177 x D 0.225 x D 0.323 x D 0.704 x D
5 0.250 x D 0.340 x D 0.677 x D 0.853 x D
6 0.356 x D 0.660 x D 0.806 x D 0.957 x D
7 0.644 x D 0.775 x D 0.895 x D
8 0.750 x D 0.855 x D 0.968 x D
9 0.823 x D 0.920 x D
10 0.882 x D 0.975 x D
11 0.933 x D
12 0.979 x D
(D = Diameter)

For this measurement six points are taken
Distance from inside wall

s1 = 9.043 D = 11 mm
s2 = 0.146 D = 37 mm
Distance from a reference mark
x1 = S1 + a = 235 mm
x2 = S2 + a = 261 mm

Table 3
Axis Point Distance from Distance from h* h Velocity (see 6)

inside wall reference [mm W] [m W] w = .... h
[cm] mark [m/s]
[cm]
A 1 1.1 23.5
A 2 3.7 26.1
A 3 7.4 29.8
A 4 17.6 40.0
A 5 21.3 43.7
A 6 23.9 46.3
B 1 1.1 23.5
B 2 3.7 26.1
B 3 7.4 29.8
B 4 17.6 40.0
B 5 21.3 43.7
B 6 23.9 46.3
w=
w=
12
6.4 Calculation of the Gas Flow Rate

a) Operating conditions
m3
V=A T w=0.0491..............=.............
s
b) Standard conditions
............. Nm 3
V 0 =VG G =.............. =.............
0 1.29 s

6.5 Anemometers
6.5.1 Type A (Vane Wheel Anemometers)
The revolutions of the vane, being proportional to the gas velocity, have directly been
measured by an electronic sensor. The instrument integrates the number of revolutions by a
certain time and the indication is in meter per second:
m
w A =.............
s
m3
V=A T w A =0.0491.......... ...=.............
s
Vane wheel anemometers are independent of gas condition.
6.5.2 Type B Thermal Anemometers

Thermal gas flow meters were based on the principle of heat transfer from an electrical
heated body to the ambient medium (gas) depending from the relative velocity between body
and medium. The probe consists of a heated temperature sensor and two additional
temperature sensors. The first sensor measures the temperature of the heated sensor and
the second the temperature of the medium. The temperature difference between heated
sensor and the medium contains the information for calculating the gas velocity with the aid
of an integrated micro processor.
The indication of thermal gasflow anemometers is depending from gas conditions:
specific heat
conductivity
density

The calibration is normally based on the medium air. For other gases (kiln gases) correction
factors should be used. Tables with such factors are available from the suppliers.
6.5.3 Vortex Anemometers (not used in practical work)

When a object is inserted in a flowing fluid then a vortex tail, known as Karman vortex street,
arises. Karman found (about 1912) out, that depending on the velocity a constant relation
exists between vortex interval and vortex sequence behind a cylindrical body in a current. He
also observed that at constant flow velocity the vortex frequency downstream remains
practically constant.
Vortex anemometers are independent of gas conditions like density, pressure and
temperature until 180C. Vortex measuring probes have no moving parts. They are robust,
reliable and corrosion-resistant. Even in rough conditions as dust loaded fluids a high fatigue
strength is ensured.
6.6 Characteristic Curves of the Fan

The supplier establishes the fan characteristics by measuring the difference between the
static pressure at the fan outlet and the total pressure at the fan inlet in function of the
revolutions of the fan and the volume rate. This measurement is given by the following
diagram.
PS: For most fans in cement plants, it is satisfactory to assume that the fans static pressure
is the difference between the static pressure at the inlet (here atmospheric pressure) and
outlet of the fans.

6.6.1 Measurements
n = ............. [rpm]
p = ............. [mbar]
From the given diagram by the supplier: Vf = ........... [m3/s]

6.7 INLET NOZZLE
Equation of Bernoulli:
p0 w 02 p1 w 12
w0= + = +
G 2 G 2
(7) (8)
2(p0 p1 )
w 1=
G
Continuity condition:
2(p0 p1 )
Vf = A Nw 1= A N (9)
G

U - Tube manometer:
(p0 -p1) = W . g . h (10)
From equations (9) and (10) :
W
Vf =AN 2g h (11)
G
= 0.93 [-] (measured by the

supplier)
AN = 0.1244 [m2]
W = 1000 [kg/m3)
G = ........... [kg/m3) (see 2.2)
h = ........... [mm W] . 0.001 [m WC]

Vf = 1000
0.930.1244 29.81 ..........
.......
Vf = ........................................... [m3/s]

6.8 Manometers
6.8.1 Inclined Stationary Manometer
PS: On the instrument itself mostly not the angle is indicated but directly the value of sin ,
e.g.
1 : 5 = 0.2 = 11.5
1 : 2 = 0.5 = 30.0
The angle of the used manometer is fixed and the scale is calibrated in mm W

7. SYMBOL LIST OF GAS FLOW MEASUREMENT
A [m2] Cross section area

d [m] Diameter of the smallest aperture
D [m] Diameter of the tube
h [m] Height of Liquidcolumn
m [-] Cross section area ratio m = d2 / D2
&
m [kg / s] Mass flow rate
2
p [N / m ] Pressure
v& [m3 / s] Volume flow rate
w [m / s] Velocity
[-] Correction factor for the contraction
[%] Humidity
[-] Correction factor if fluid is compressible
[kg / m3] Density
fdr [m3/m3 dry] Water content (Volume)
3
W0 [kg/m n,dry] Water content (Mass)

V5 Process Technology 3

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

V5 Process Technology 3

Hochgeladen von

Copyright:

Verfügbare Formate

Volume 5

Process Technology III

Cement Manufacturing Course Version 2005 Volume 5 - Page 1

Cement Manufacturing Course Version 2005 Volume 5 - Page 2

Cement Manufacturing Course Version 2005 Volume 5 - Page 3

Cement Manufacturing Course Version 2005 Volume 5 - Page 4

Cement Manufacturing Course Version 2005 Volume 5 - Page 5

2. CHEMICAL AND PHYSICAL PROPERTIES AND ENVIRONMENTAL ASPECTS

2.1 Nitrogen (N2)

2.1.1 Physical Properties [1]

Cement Manufacturing Course Version 2005 Volume 5 - Page 6

2.2 Nitrogen Oxides (NO, NO2 N2O)

2.2.1 Physical Properties [1]

Cement Manufacturing Course Version 2005 Volume 5 - Page 7

Compound N2O NO NO2 / N2O4 N2O3 N2O5

Cement Manufacturing Course Version 2005 Volume 5 - Page 8

Cement Manufacturing Course Version 2005 Volume 5 - Page 9

Cement Manufacturing Course Version 2005 Volume 5 - Page 10

2.3.1 Physical Properties

2.3.2 Chemical Properties

Cement Manufacturing Course Version 2005 Volume 5 - Page 11

Cement Manufacturing Course Version 2005 Volume 5 - Page 12

Heat capacity, gas

Cement Manufacturing Course Version 2005 Volume 5 - Page 13

2.4.3 Formation of Trophospheric Ozone

Cement Manufacturing Course Version 2005 Volume 5 - Page 14

3.1 Emission Limit [16]

Emission Limits: Nm3dry = m3 at 273 K, 101300 Pa and 0% water

Cement Manufacturing Course Version 2005 Volume 5 - Page 15

[ppm NO] to [mg NO2/Nm3]

[mg NO2/Nm3] = 2.05 * [ppm NO]

221 [ppm NO] * 2.05 = 453 [mg NO2/Nm3 at 7% O2]

0.88 [g NO2/kg cli] * 2.204 = 1.94 [lbs NO2/t cli]

Cement Manufacturing Course Version 2005 Volume 5 - Page 16

5. BEHAVIOR OF NITROGEN IN THE PROCESS

5.1.1 Formation of "Thermal NO"

Cement Manufacturing Course Version 2005 Volume 5 - Page 17

The pronounced dependence of K1 on temperature can be seen in Fig. 3. Thus, at high

Figure 3: Dependence of K1 on temperature

Cement Manufacturing Course Version 2005 Volume 5 - Page 18

Hence, the formation of "thermal NO" is governed by the following parameters:

Cement Manufacturing Course Version 2005 Volume 5 - Page 19

Cement Manufacturing Course Version 2005 Volume 5 - Page 20

For a significant NO-decomposition according to (10) high hydrocarbon concentration and

5.2 Formation of Nitrogen Monoxide in a Cement Kiln

Cement Manufacturing Course Version 2005 Volume 5 - Page 21

Kilns without secondary firing

Cement Manufacturing Course Version 2005 Volume 5 - Page 22

Cement Manufacturing Course Version 2005 Volume 5 - Page 23

Figure 8: Influence of the sintering zone temperature on the NO concentration in

5.3.2 Temperature Peaks [6]

Cement Manufacturing Course Version 2005 Volume 5 - Page 24

5.3.3 Excess Air

Cement Manufacturing Course Version 2005 Volume 5 - Page 25

5.3.4 Residence Time

5.3.5 Burner Operating Parameters

5.3.5.1 In-flame Air Level at Ignition

5.3.5.2 Total Axial Momentum

5.3.5.3 Tangential Momentum and Swirl Level on the Swirling Channel

Cement Manufacturing Course Version 2005 Volume 5 - Page 26

5.3.5.5 Influence of Cooler System on NOx Emissions

Figure 19: NOx emission for different cooler types

Cement Manufacturing Course Version 2005 Volume 5 - Page 27

6.1 Reduction of Nitrogen Input