An Advancement in Steam Turbine Chemistry

Monitoring
03/30/2018
By Brad Buecker

Even minor traces of chloride, sulfate, and sodium hydroxide can

cause severe problems in steam systems, and especially turbines.
In the case of sodium hydroxide, stress corrosion cracking of
turbine blades and rotors may occur very rapidly. Chloride and
sulfate salts deposit in the last rows of the LP turbine (the phase
transition zone, PTZ) and can induce pitting that in turn leads to
stress corrosion cracking and corrosion fatigue.

The Electric Power Research Institute’s (EPRI) recommended steam

sodium limit is 2 parts-per-billion (ppb), with a concentration of 1
ppb “normally achievable in both drum and once-through units with
good control of mechanical carryover or operating on AVT or OT.”
The 2 ppb limit also applies to chloride and sulfate, again with the
understanding that concentrations should normally be much lower.

Reliable on-line sodium analyzers have been available for years,

but trace chloride and sulfate monitoring has been more difficult.
Ion chromatography is a valid technique, but the instrumentation is
expensive and requires much operator attention. The surrogate
measurement for these two impurities has been cation
conductivity, (or more properly termed, conductivity after cation
exchange [CACE] to emphasize the fact that the sample is routed
through a cation resin column to exchange all cations, e.g.
ammonium, sodium, calcium, etc., for hydrogen ions).

1/9
Chloride/sulfate analyzer. Photo courtesy: METTLER TOLEDO Thornton.
This ion exchange process creates a sample of a dilute acid
solution of chloride and sulfate, whose conductivity is then
measured. An advancement that has become popular is degasified
CACE, which utilizes a sample reboiler or nitrogen sparging
compartment to scrub CO2 from the sample and minimize
interference from this compound.

The common normal CACE or degassed CACE limit for steam is 0.2
µS/cm, but laboratory data has shown that the concentration of
chloride and/or sulfate can be considerably greater than 2 ppb with
CACE at the 0.2 µS/cm threshold. Thus, direct measurement of
these troublesome impurities offers significant advantages, and a
new instrument is now available to monitor trace concentrations
(down to 0.5 ppb). The analyzer combines two well-established
technologies to provide on-line measurements of these ions –
capillary electrophoresis to separate the ions and conductivity to
measure and calculate their concentration.

A sample and a known reagent are collected in a cartridge with a

capillary, and voltage is applied across the cartridge. The ions in
the mixture start separating and moving through the capillary
towards respective electrodes at the ends of the capillary, thus
causing separation of ions. Based on the size-to-charge ratio,
different ions move at different speeds, forming clusters of each ion
as they flow towards the electrode. Each cluster of ions passes a
conductivity sensor on the cartridge just before they reach the
electrode, and the sensor records a measurement based on the
concentration of the ion. By comparing the measurement for the
ions with the measurement for the known reagent, the analyzer
calculates the concentration of the ions in ppb. This measurement

2/9
can be repeated as frequently as every 13 minutes, drawing a fresh
sample each time.

This ability to accurately monitor trace concentrations of chloride

and sulfate offers an excellent enhancement to steam monitoring
capabilities. Looking towards the future, company personnel have
been asked if the instrument capabilities could be expanded to
analyze for other trace compounds, most notably two of the
primary decomposition products, formate and acetate, that are
generated in steam boilers if any organic compounds (neutralizing
or filming amines) are employed for pH control and corrosion
protection. Research is underway, but a definitive answer is not yet
available.

WHAT ABOUT THE REST OF THE STEAM GENERATING

NETWORK?
While protection of the steam system and turbine is extremely
important, maintaining proper chemistry in the other sections of
the steam generator is also vital to minimize corrosion and fouling.
Impurity ingress from condenser tube leaks or malfunctioning
makeup water systems can introduce impurities, including our old
friends, chloride and sulfate, that can cause rapid damage in
conventional boilers and HRSG evaporators. Poor
condensate/feedwater chemistry control will induce corrosion that
not only can cause catastrophic failures (flow-accelerated
corrosion, FAC) within these systems, but introduce corrosion
products to the boilers that then precipitate on waterwall tubes and
influence corrosion chemistry. So, with those themes in mind, the
following sections outline guidelines for continuous sampling of the
other systems within the steam generator.

MAKEUP TREATMENT SYSTEM

The core process of most power plant makeup water systems is
reverse osmosis (RO) followed by either mixed-bed ion exchange
(MBIX) or electrodeionization (EDI) to “polish” the RO effluent to
meet utility steam generator requirements. RO units typically are
equipped with a number of instruments to monitor system
performance, including pressure, temperature, flow, and specific
conductivity. We will focus upon the recommended analyses of the
final effluent from either a MBIX or EDI polisher.

Note: In this and several of the following sections, the normal limit
for each parameter is included.

 Specific conductivity: ≤0.1 µS/cm

3/9
  Silica: ≤10 parts per billion (ppb)

 Sodium: ≤2 ppb

Control within these guidelines ensures that high-purity water is

being distributed to the steam generator. A rise in any of the values
indicates that either the MBIX resin has reached exhaustion or that
a failure has occurred in the EDI unit. Prompt corrective action is
necessary.

Particle monitor. Photo courtesy: CHEMTRAC

Often the design specifications for new plants call for continuous
pH monitoring of makeup system effluent, but pH measurement of
high purity water is very difficult. The analyses listed above are
sufficient for evaluating process conditions.

CONDENSATE PUMP DISCHARGE

Condensate pump discharge (CPD) is an absolutely critical
monitoring point, particularly for systems with water-cooled
condensers, as this is the most likely source for major condensate
contamination. A condensate polisher will provide a buffer against
contaminant ingress, but unfortunately polishers are often not
considered necessary for drum units, when in fact they can be of
great benefit.

RECOMMENDED CPD ANALYSES ARE:

 CACE or degassed CACE: ≤0.2 µS/cm

4/9
  Specific Conductivity: Consistent with pH

 Sodium: ≤2 ppb

 Dissolved Oxygen: ≤20 ppb

 pH: 9.6 to 10.0 (This is the pH range for triple-pressure feed-

forward low-pressure [FFLP] HRSGs, where the LP circuit
basically serves as a feedwater heater for the intermediate-
pressure (IP) and high-pressure (HP) evaporators. The range
may be a bit lower for other HRSG designs.)

Sodium monitoring is very effective for detecting condenser tube

leaks. With a tight condenser, sodium levels in the condensate
should be below 2 ppb, and in many cases less than 1 ppb.
Excursions of course suggest a leaking tube(s).

As with sodium, a rise in CACE indicates impurity in-leakage,

although this measurement is also influenced by carbon dioxide
ingression, typically via air in-leakage at the condenser. The CACE
limit of 0.2 µS/cm is a standard requirement for implementation of
all-volatile treatment oxidizing [AVT(O)] chemistry, which is the
best choice for condensate/feedwater systems that do not have
copper alloys (virtually all HRSG systems have no copper alloy
components). No longer is oxygen scavenger feed recommended in
all-ferrous systems, as this chemistry can induce flow-accelerated
corrosion in HRSG low-pressure and intermediate-pressure
economizers and evaporators, and attemperator lines.

Dissolved oxygen (D.O.) monitoring is important for evaluating

condenser air in-leakage. A sudden increase in D.O. may indicate a
structural or equipment failure in the condenser shell, at
penetrations to the condenser, or even at remote locations such as
the gland steam condenser. However, some air in-leakage is
desired, as it provides the oxygen necessary for AVT(O) chemistry.
In this regard, the D.O. normal limit was increased from 10 ppb to
20 ppb several years ago.

5/9
Previously we noted that pH measurement of high-purity water is
difficult, and is not practical for demineralizer effluent. While direct
pH monitoring is recommended for condensate and feedwater, the
measurement is still difficult. However, in the absence of significant
condenser air in-leakage, ammonia concentration, pH, and specific
conductivity (S.C.) are directly related. S.C. is a very reliable
measurement, and thus is normally utilized to control ammonia
feed and pH in the condensate and feedwater.

LP ECONOMIZER INLET/BOILER FEED PUMP

DISCHARGE
These samples are necessary to monitor the feedwater before it
enters the evaporators to ensure that the chemistry is being
properly controlled to minimize FAC. Additionally, the
measurements provide backup to those of the condensate pump
discharge.

Recommended feedwater/economizer analyses are:

 CACE: ≤0.2 µS/cm

 S.C: Consistent with pH

 Sodium: ≤2 ppb

 Dissolved Oxygen (range): 5 to 10 ppb

 pH: 9.6 to 10.0 (This is the pH range for FFLP HRSGs. The
range may be a bit lower for other HRSG designs.)

 Iron: ≤2 ppb

The discussion for CACE, S.C., pH, and sodium mirrors that for the
condensate pump discharge, and also the first three, along with
dissolved oxygen, are critical measurements for AVT(O) chemistry.

6/9
 Integrated corrosion product sampler. Photo courtesy: Sentry

The reader will note the inclusion of iron in this set of parameters.
Iron monitoring provides a direct measure of flow-accelerated
corrosion and the effectiveness of the feedwater chemistry
program. Typically, 90 percent or greater of iron corrosion products
generated by FAC are particulate in nature. Several methods exist
to monitor steel corrosion, and include:

 Continuous particulate monitoring

 Corrosion product sampling

 Grab sample analysis

A particulate monitor simply measures particle count in real time,

and does not chemically differentiate what those particles might
be. But, in high-purity feedwater systems, and in the absence of
any copper alloy components, virtually all of the particles will be
iron oxide.

A corrosion product sampler utilizes both a filter and ion exchange

resin to capture suspended and dissolved iron. (It can also capture

7/9
other metals including copper, if necessary.) After a designated run
time, where the instrument also has a flow totalizer, analyses of
the filter and ion exchange resin reveal the amount of metal
captured. Straightforward calculations determine the corrosion rate
over the period of time the sample was collected.

Finally, improved grab sampling techniques are available, in which,

with proper sample treatment, iron measurements down to 1 ppb
are possible. This method can provide near real-time data of
corrosion rates, although on a snapshot basis.
Iron monitoring is often also recommended for CPD and boiler
water samples to determine corrosion rates in other locations. With
the on-line instruments, some form of sample sequencing can be
arranged. Grab samples for each location can obviously be
collected at any time.

EVAPORATOR (BOILER) WATER

Evaporator water sampling is critical for two primary reasons. First,
whether the steam generator is an HRSG or conventional unit, the
highest heat fluxes occur within the boiler. Thus, the effects of
impurity ingress or poor chemistry are magnified by the high
temperatures and pressures in these circuits. This issue is
magnified by the fact that corrosion products from the feedwater
system, most notably iron oxides, tend to precipitate on boiler
internals. Iron oxide deposits are typically porous in nature, which
allows impurities to concentrate underneath the deposits and cause
corrosion, sometimes very severe, that would not occur otherwise.

Spectrophotometer. Not shown is the sample digestion equipment to

convert particulate iron to soluble iron for total iron analysis. Photo
courtesy: Hach

Secondly, boiler water chemistry must be established and

monitored to ensure that steam purity matches the guidelines

8/9
previously shown. Excess impurities in the boiler water can lead to
problematic carryover to the steam circuits and turbine.

Recommended boiler water analyses include:

 pH (<8.0, immediate unit shutdown; typical range 9.0 to 9.8

with the precise range subject to steam generator design,
pressure, and chemical treatment program)

 CACE

 Specific Conductivity

 Chloride

 Silica

 Phosphate (for those units on phosphate treatment)

The reader will notice no absolute limits listed for many of these
parameters. This is due to the fact that allowable impurity
concentrations vary as a function of boiler pressure. As pressure
increases, and the density of water and steam converge, drum
moisture separators become less effective, which in turn will allow
greater carryover of water droplets into the saturated steam.
Within the steam generator(s) proper, temperature and pressure
influence corrosion, and in particular affect reactions underneath
iron oxide deposits on waterwall tubes and other boiler internals.
This influence is magnified at higher temperatures.

Also, the recommended chemistry control ranges will vary

depending upon the selected boiler water treatment program, i.e.,
tri-sodium phosphate, caustic, or AVT-only chemistry. (Note: A
concept known as sodium balancing is important for boiler water
chemistry control, particularly for units with AVT-only programs. A
discussion of sodium balancing is beyond the scope of this article.)
Charts and graphs for acceptable control ranges may be found in

9/9
EPRI guidelines or those from the International Association of the
Properties of Water and Steam (IAPWS). In addition, software
programs (including an excellent program developed by Mr. Randy
Turner of Swan Analytical) are now available to precisely calculate
boiler water conditions, and alert chemists and technical personnel
to chemistry upsets.

Acknowledgement
The author wishes to thank Mr. Akash Trivedi of Metter Toledo
Thornton for providing the detailed information regarding the
chloride/sulfate analyzer.

Brad Buecker is senior technical publicist at ChemTreat.

10/9