Waste Management 27 (2007) 367374

www.elsevier.com/locate/wasman

Pilot installation for the thermo-chemical characterisation

of solid wastes
a,*
C. Marculescu , G. Antonini b, A. Badea a, T. Apostol a

a
U. POLITEHNICA Bucuresti, Spl. Independentei, 313, Bucuresti, Romania
b
U. de Technologie de Compiegne, BP 20529 60205 Compiegne cedex, France

Accepted 28 February 2006

Available online 17 April 2006

Abstract

The increasing production and the large variety of wastes require operators of thermal treatment units to continuously adapt the
installations or the functioning parameters to the dierent physical and chemical properties of the wastes. Usually, the treated waste
is encountered in the form of heterogeneous mixtures. The classical tests such as thermogravimetry and calorimetric bomb operate com-
ponent by component, separately. In addition to this, they can analyse only small quantities of waste at a time (a few grams). These
common tests are necessary but insucient in the global waste analysis in the view further thermal treatment.
This paper presents an experimental installation, which was designed and built at the CNRS Science Division, Department of Indus-
trial Methods, Compiegne University of Technology, France. It allows the determination of waste thermal and chemical properties by
means of thermal treatment. Also, it is capable of continuously analysing signicant quantities of waste (up to 50 kg/h) as compared to
the classical tests and it can work under various conditions:

 oxidant or reductive atmosphere (on choice);

 variable temperature between 400 and 1000 C;
 independently set residence time of treated sample in the installation and ow conditions.

The installation reproduces the process conditions from incinerators or pyrolysis reactors. It also provides complete information on
the kinetics of the waste thermal degradation and on the pollutant emissions. Using dierent mixtures of components present in the
municipal solid waste and also in the reconstituted MSW samples, we dened a series of criteria for characterising waste behaviour dur-
ing the stages of the main treatment process such as: feeding, devolatilisation/oxidation, advancement, solid residue evacuation, and pol-
lutants emission.
2006 Elsevier Ltd. All rights reserved.

1. Introduction density, as well as in physical and chemical characteristics.

The concept of using waste as a renewable source of
energy has become a priority in the eld of wastes treat-
ment. The goal of this research consists not only in waste
destruction but also in increased recovery of thermal
energy from the processes.
There are numerous and various sources of waste; there-
fore the resulting waste is heterogeneous in form and in
*
Corresponding author. Tel.: +40 745 133 713; fax: +40 21 411 31 61.
E-mail address: cosmin@cce.ab.pub.ro (C. Marculescu).
That is why there is a high number of treatment methods
for that waste, each one with its own particularities. In
order to maximize treatment eciency, it is necessary to
choose the appropriate method for the given waste. Waste
thermal treatment units encounter operational problems
due to waste variety. Selecting the optimal treatment
method requires the deep knowledge of many physical
chemical characteristics of the waste to be treated (Tillman,
1991). The only existing tests for characteristic diagnosis
are designated for solid wastes, especially coal. These vali-
dated tests, and we refer mainly to thermogravimetry,

0956-053X/$ - see front matter 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.wasman.2006.02.011
368 C. Marculescu et al. / Waste Management 27 (2007) 367374
require a high degree of sample preparation, because the
quantity of analysed substance must not exceed a few
grams and a micronic granularity. It is dicult to obtain
a representative sample for the waste to be analysed, due
to the low quantity required by the test and to the large
number of components in the sample. Furthermore, con-
cerning the texture of the waste, it is almost impossible to
achieve micron granularity. Applying these tests on wastes,
even reconstituted ones, generates more or less accurate
results, even less clear in the case of waste with a poorly
dened composition.
These are some of the reasons leading to the necessity of
developing of other tests that will provide the same accu-
rate results on the physicalchemical data, analysing
increased waste samples (measured in kilograms) with a
granularity in centimetres that do not require laborious
preliminary preparation.
At this moment there are no tests that provide global
accurate information on the thermal degradation charac-
teristics of a waste. The most common tests used for the
waste characterisation are the thermogravimetric analysis
(TGA) and the calorimetric bomb experiment. The rst
one provides data on the mass reduction rate of the sample
under oxidant or pyrolytic condition. The mass variation
speed can be converted in degradation kinetics as a func-
tion of temperature and liberated volatiles content (Coz-
zani et al., 1995b). The second test provides data only on
the low heating value (LHV) of the sample. The content
of volatile substances is an essential piece of information
both for oxidant and non-oxidant waste thermal treatment,
because the end of the liberation phase for these substances
marks either the beginning of xed carbon oxidation phase
in incineration, or the end of the pyrolysis process (Cozzani
et al., 1995a; Fontana et al., 1997). Several studies having
as subject the TGA applied to dierent products (Wu
et al., 1993; Bockhorn et al., 1999) indicated that it is di-
cult to compare experimental data from dierent studies
because no standardized way to perform TGA experiments
exists and because the obtained weight loss characteristics
strongly depend on the measurement conditions (Raveen-
dran et al., 1996; Varhegyi et al., 1997; Orfao et al., 1999;
Stengsen et al., 1999). In order to determine the shares of
single components in an unknown waste mixture, the
markers of individual components are correlated with the
one of the mixture. This is why the weighed sum method
(WSM) presented by Cozzani et al. (1995b) is applied.
According to the WSM, the devolatilisation behaviour of
a mixture can be expressed as the sum of its corresponding
single constituents. The work of Heikkinen et al. (2004) on
the sensitivity of the calculated mixture composition using
WSM when changing the position and height of the TGA
peak of a single component revealed that modifying the
peak height has a strong impact on the calculated compo-
sition. On the other hand, the LHV (low heat value) data of
a waste to be treated gives information only on the waste
combustion capacity, and nothing more. Nevertheless,
these tests are applicable to clearly dened and sampled
waste components. Not all the components are suitable
for sampling at the dimensions required by these tests,
and, furthermore, the components are not always known.
TGA and the LHV data are necessary but sucient to
qualify a product for further thermal treatment (Tillman,
1991). Therefore, thermal treatment units have strict oper-
ational specications and limitations related to: product
feeding and advancing systems, ignition delay and minimal
residence time, solid residue mechanical behaviour at the
unit evacuation section, and combustible reduction rate
eciency. These characteristics were not taken into
account by previous studies probably because they are spe-
cic to industrial installations, but as the aim of each
research in this eld is the industrial scale application
improvement, they should not be neglected.
For a proper analysis of the nature of the solid waste
that ts the requirements of thermal treatment installations
for their elimination, we dened a certain number of fea-
tures to be quantied by an oxidant or non-oxidant (pyro-
lytic) test. From now on, in this paper they will be called
criteria.
2. The experimental device
2.1. Installation description
The pilot installation used throughout this study was
developed in the laboratory of CNRS Science Division,
Department of Industrial Methods, Compiegne University
of Technology, France. It was designed to quantify hydro-
dynamics and heat exchanges in a vibro-uidized bed ther-
mal unit. Fig. 1 presents the simplied scheme of the
device. It consists of a tubular rectilinear reactor heated
by circulation of hot ue-gases (natural gas burner) within
an external double envelope. The solid continuously
advances by vibro-uidized transport. This type of trans-
portation corresponds to a piston-like ow. For instance,
if we consider a transversal section in the solid advancing
bed, the particle distribution and number in the bed will
be same all along the transport tube, with no sliding layers.
The speed of advancing solids is set by the installation
vibratory parameters (Socha and Kolze, 1992; Caudron,
1999).
Each solid particle within the tubular reactor undergoes
thermal treatment for the same duration. The tube is
160 mm in inner diameter and 4000 mm long. Heat
required by the process is supplied by the circulation of
hot gases counter-owing the circulation of solids in the
3300 mm active zone.
The ow rate (up to 50 kg/h) is determined by the feed-
ing ratio and the advancement speed of the product in the
installation, and it imposes dierent lling levels of the
tube. The feeding system consists of an Archimedes screw,
whose rotation is controlled by a frequency variator.
From the feeding system, the ground waste is pushed
forward by screw rotation and dropped into the vibrating
transport tube. There it advances in a piston-like move-
 C. Marculescu et al. / Waste Management 27 (2007) 367374
Natural gas
Alimentation with towards burner
divided solid
Heated case
Transport tube
Treatment gas inlet
(comburant or inert)
Fig. 1. Functional scheme of the pilot for solid divided waste characterisation.
ment, which is generated by the vertical vibration of the
tube; it goes through the heated zone, towards the tube
outlet, into a cyclone separator. The resulting sub-prod-
ucts, gases and solids, are collected using a cyclone that
performs the separation of solid residue from gases, so that
they can be analysed separately. The residue is collected at
the base of the separator in an ash/coke collector. After l-
tration and fast cooling in a quench (to avoid dioxin for-
mation), both process and heating gases are evacuated by
an exhauster on separate ways. A pressure-control device
inside the cyclone separator maintains a constant pressure
level inside the tubular reactor. Keeping a constant gas
layer over the solid bed in the tubular reactor allows not
only the control of the thermal treatment atmosphere (oxi-
dant or pyrolytic conditions) but also the evacuation of the
vapour and gases resulting from the solid treatment pro-
cess. The device operates at temperatures in the range of
4001100 C.
The adjustable installation parameters for transport
control of the divided solids inside the tube are the follow-
ing: tube vibration frequency, impulse, tube inclination and
feed ow rate (Akiyama and Yamaboshi, 1992).
In stationary regime, frequency increase involves ampli-
tude decrease (hence solid particle agitation diminishes)
and a shorter residence time for the treated waste. The
installation operates above resonance frequency. Tests
proved that an operation frequency below the resonance
point is insucient for initiating the movement of the
product.
The impulse is imposed by the load eccentricity of the
two electrical engines that power the installation. It is com-
mensurate with the eccentricity degree of the loads
attached to the engine shafts, which generate vibration of
the engines. As the engines are xed on each side at the
middle length of the tube, their vibration is transmitted
to the tubular reactor. The more the impulse increases,
the shorter the residence time.
The tube inclination refers to the angle between the tubu-
lar reactor and the horizontal plan. The residence time of a
particle in the reactor increases with the tube inclination
369
Process gases analysis
and treatment
Solid residue
Burned
gases analysis
because the particles have to advance on a higher slope
and the phenomenon of rear sliding in the transport tube
increases (Neddreman and Harding, 1990).
The feed ow inuences only the bed height and the par-
ticle agitation degree. The residence time variations due to
the feed ow are negligible. On the other hand, a higher
feed ow rate will diminish particle agitation due to the
increase of the bed height (Akiyama and Yamaboshi,
1992).
The hydrodynamic characteristics of the pilot installa-
tion had previously been studied (Porquet, 1994). The main
features of the divided hydrodynamic behaviour of the
divided solid are:
 For a xed frequency and eccentricity (impulse) imposed
by the engines, the residence time becomes longer if the
tube inclination angle increases.
 For a xed inclination angle and a constant frequency
the residence time becomes shorter if the load eccentric-
ity degree increases.
 For a xed inclination angle and a certain impulse, the
residence time becomes shorter if the frequency
increases.
2.2. The samples
Our study is based on thermal treatment by vibro-uid-
ized transport applied to divided solids. The concept of the
divided solid is, therefore, very important. No mechanical
device forces the solid waste to advance. Consequently,
only the acceleration forces generate particle movement,
causing the advancement of the product in the tubular
reactor. If the solid is not entirely divided (mud-like paste
consistency, cotton-like brous consistency), the movement
forces will be more or less compensated by the particle
cohesion forces in an arbitrary and non-quantiable way
(Jaraiz et al., 1992). Therefore, the piston-like ow condi-
tions will no longer be achieved. The installation is suitable
for solids that are easily broken down to granules, with no
370 C. Marculescu et al. / Waste Management 27 (2007) 367374
sticky or textile bre aspect. Tests showed that a fraction
up to 20% of textiles is acceptable in a mixture with other
solid products. Considering the transporting tube diameter
(U = 160 mm), the tolerated solid granularity may reach up
to 40 mm. Furthermore, the treated sample must have a
certain mass balance of granularity: at least 90% of its
quantity must have a granularity superior to 1 mm. If this
condition is not accomplished, the product will start to
compact at the feeding inlet, as a result of the vibrating
forces, like powder (Jaraiz et al., 1992).
2.3. Limitations
The technical limitations of this installation are related
to product characteristics. Some of them require high air
ows (enough oxygen for a complete oxidation reaction)
that can generate ing particles, as a consequence of the
increased air speed through the constant tube section
(ash, solid residue or even particles that did not undergo
the combustion process). Others have an important fusible
fraction in the mixture that can melt, generating a soldering
process. For the solids with a high percentage of combusti-
ble compounds, the required air for complete combustion
is very important, and, therefore, the gas ow speed above
the solid waste bed in the tube can sensibly increase. The
mixture components with a low specic weight start to
move with the gas ow, and, consequently, the homoge-
nous aspect of the treated sample begins to alter. There will
also be particles that will not undergo the same treatment
period. The only solution is to reduce the air ow rate in
order to decrease gas speed. The simplest method is to
decrease the feed ow rate at the installation inlet or to
mix the waste with a certain inert solid fraction in order
to decrease the LHV of the sample, so that the required
air quantity for complete oxidation will diminish. The inert
solid fraction solution also ts the second main limitation
introduced by the products containing a high fusible frac-
tion. In this case the melted components will stick to the
inert ones and will continue to move.
3. Analyses
We established a certain number of criteria for a min-
utely detailed characterization of the incineration capacity
of divided solid waste. They allow the identication of solid
waste behaviour between feeding and evacuation. That is
very important from the standpoint of choosing the most
appropriate type of incineration installation. The inuence
of global functioning parameters (humidity, temperature,
residence time, etc.) has also been studied. The tests have
been carried out both under pyrolytic and oxidant
conditions.
3.1. Reductive atmosphere
The SHM criterion (sample heat maneuverability) refers
to the sample maneuverability at high temperatures. The
purpose of this test is to highlight the problems encoun-
tered during the incinerator alimentation with solid waste.
Some types of waste to be treated (plastics) change the
aggregation phase or structure during the feeding proce-
dure when the temperature increases from the one in the
environment to the one inside the incinerator. Therefore
its mechanical characteristics change completely, impeding
the product advancement. If the product covers the critical
distance inside the heated tube (about 350 mm) until it
reaches the treatment temperature without agglomerations,
the SHM criterion is considered positive. Otherwise the
product composition is considered inappropriate for the
treatment parameters and has to be modied or the treat-
ment type has to be changed. These data can be grouped
in a ternary diagram as an equilateral triangle. The main
classes of components found in a heterogeneous mixture
of solid waste are placed in the corners (inert, melting
and non-melting combustible fractions). Each mixture is
represented by tokens placed on this triangle. The dierent
zones that result separate the compositions into favourably
or inappropriate maneuverable at elevated temperatures.
The tests were performed on a mixture whose composition
is given in mass percentage fractions: PET (granules of
2.5 mm medium diameter), pine bark (dried and grinded
to medium size less than 10 mm) and clay balls (5 mm).
Fig. 2 presents the SHM diagram for dierent mixtures
of the above components. It highlights the transition zones
corresponding to mixtures that meet the requirements of
the previously dened criterion.
The LVF criterion (liberated volatiles fraction) refers to
the volatile components in the gaseous substances that
are generated along the process. It is an intrinsic character-
istic of the solid, or the mixture, regarding its combustibil-
ity because under oxidant conditions the rst substances to
ignite are the volatiles. We conducted devolatilisation tests
on PET, wood, paper, forge coal and mixtures at tempera-
tures within the range of 400900 C in an inert atmosphere
created by constant nitrogen ow inside the transport tube.
PET
75% PET
25% pine bark
25% argyle
40% PET
20% argyle
50% argyle 40% pine bark
25% PET 50% pine bark
Transition zone 75% pine bark
75% argyle
ARGYLE Negative MAC criterion BARK
Positive MAC criterion
Fig. 2. MAC criterion ternary diagram for mixtures of wood/plastic/inert.
 C. Marculescu et al. / Waste Management 27 (2007) 367374
The measured volatile fraction of the mixtures was com-
pared with the calculated one using the weighed sum
method (WSM) presented by Cozzani et al. (1995b) and
volatile fraction of each component from the scientic
the literature. The errors were about 12% caused by inher-
ent experimental limitations.
The AFI criterion (agglutination and friability index)
describes the friable and agglutinant features of the solid
residue at the end of the treatment process. The product
resulting from the pyrolysis stage of the treatment may
be compact or powder-like, friable or not. For instance,
from a mechanical point of view, a waste whose residues
tend to agglutinate into a very resistant solid will not be
incinerated in a grate type installation (Martin, 2001)
because it is at risk to stick to the mobile parts. The test
is also applicable to the carbon residue from the pyrolysis
reactions. This residue is submitted to a compaction and
friability test, derived from the method in the Roga
(1974), using repeated 1 mm siftings and refusal centrifu-
gation (for about 5 min). The ratio between the sift refu-
sal weight of the residue and its initial weight gives the
AFI. For example it varies between 0 void compacting
capacity (anthracite coal) and 100 highly agglutinant
characteristics (forge coal with 95% of its mass retained
after siftings). The results of the test performed on paper,
PET, wood, and mixtures, indicated AFI values of: 11.5
for wood, 55 for a mixture of 50% plastic50% wood,
75 for a mixture of 50% plastic50% paper, 65 for a mix-
ture reconstituting a dried MSW sample containing plas-
tic 25%wood 20%paper 45%inert 10%. The AFI
criterion has to be correlated with the mechanical charac-
teristics of the target treatment unit. Generally speaking,
an AFI coecient below 70 is suitable for common indus-
trial installations.
3.2. Oxidant atmosphere
The GID (global inammation delay) quanties the time
interval passing from incinerator feeding to product igni-
tion. This criterion is essential in characterizing the com-
bustibility of a certain waste. It shows the capacity of
analysed product to ignite, rapidly or slowly, after being
introduced into the incinerator. In the case of our installa-
tion, taking into account the transport type (piston-like
ow), this delay can be estimated considering the position
of the ame front from the device inlet. This position can
be determined by measuring the temperature inside the
tube with a thermocouple. The tube reactor lengthwise
temperature prole has a peak, revealing the ame front
existence. The global inammation delay is the ratio
between the ame front distance to the transport tube inlet
and the advancing speed of the particle. It strongly depends
on the content of volatiles (LVF criterion) and the humid-
ity of the product. The test showed that LVF variation
from 15% to 90% involves GID decrease from 50 to 10 s.
The humidity variation within the 235% range increased
the GID by about 50%.
371
The CFR criterion (combustible fraction reduction) repre-
sents the percentage of the combustible fraction of the
product that is consumed by incineration. For a xed res-
idence time and operation temperature, only a part of the
combustible fraction of the waste is oxidized. This share
depends on waste composition. Therefore, this parameter
is important for the characterization of waste combustibil-
ity intrinsic properties and is presented in the second part
of this paper.
The LPC criterion (liberated pollutants content). This cri-
terion concerns the amount of pollutant gases produced by
the thermal treatment of one mass unit of waste and the
concentration of each toxic component in the gases. It
enables the researcher to choose the appropriate residual
gas treatment systems for the analyzed waste sample.
These tests had been validated on standard combustibles
(wood and coal) under reductive and oxidant atmosphere
before being applied to reconstituted waste. We present
the results of the main criteria applied on dierent products
treated in the installation.
3.3. Mechanical aspects
When generally speaking about waste, we assume it is
heterogeneous in size, form and specic weight. Therefore,
it is important to know if these characteristics have an
inuence on the residence time in the installation for the
same vibrating conditions. We conducted a series of tests
to establish the inuence of granularity and heterogeneity
(specic weight) on residence time.
Fig. 3 presents the variation of the residence time with
vibration frequency and the granularity of aluminium balls
for a xed impulse (the engines loads eccentricity degree of
35%). The average diameters of the aluminium balls ranged
between 1; 2 and 5 mm. The four curves, obtained for dif-
ferent vibrating frequencies, are practically horizontal and
prove that residence time does not depend on the treated
product granularity, but only on the installation operating
frequency. We also noticed that the residence time of the
sample in the tubular reactor decreases when the vibrating
frequency increases. This variation is asymptotic to a
certain value, and we assume it is the inferior time limit.
110
100
90 21 Hz 17 Hz
Residence time [sec]
80 25 Hz 19 Hz
70
60
50
40
30
20
10
1 1.5 2 2.5 3 3.5 4 4.5 5
Granularity [mm]
Fig. 3. Residence time variation with the vibration frequency for dierent
aluminium balls diameters.
372 C. Marculescu et al. / Waste Management 27 (2007) 367374

Considering the transport tube length (meaning the heated
zone of the tubular reactor the active zone, about
3300 mm), the residence time can be converted in transport
(advancement) speed.
A study on the particle segregation depending on their
specic weight showed the independence of the residence
time from the product heterogeneity. The tested product
was a mixture of paper pieces and aluminium balls with
an average diameter of about 2 mm. The measured resi-
dence time was the same using dierent samples of argyle
balls, pine bark, PET pieces, and mixtures of them.
3.4. Thermal aspects
A series of tests allowed us to characterise the combus-
tibility of reconstituted urban waste with the following
composition given by Martin (2001): vegetal compounds
35%; paper, cardboard 29%; plastics 14%; inert (glass)
22%.
The operating conditions that we imposed during the
tests mirror actual industrial scale units, with a tempera-
ture of 1050 C and air excess of 1.2 (20% over the stoichi-
ometric value). These tests were made in order to establish
waste incineration eciency. Therefore, CFR was mea-
sured for four dierent residence times
Xm
CFR S
% 1
zCOMB
m_ m  m_ 0m
X m T ; ts 100  % 2
m_ m
where zSCOMB is the combustible mass fraction of the waste,
Xm is the mass conversion rate (reduction) for xed temper-
ature and residence time, T is the xed temperature of the
ln1  CFR=100 B  ts 5
That leads to the conclusion that the logarithmic curve
function of the residence time (ts) is a line whose slope is
B.
The experimental curves presented in Fig. 4 allow estab-
lishing CFR as a function depending on waste advance-
ment in the treatment device, for example, along the
incineration grate of a Martin installation.
Marking the experimentally established points, we
notice two distinctive lines with dierent slopes. The rst
one corresponds to the volatile emission zone (components
that volatilise at temperatures below 450 C) and its slope
BMV represents the variation speed itself. Heikkinen et al.
(2004) showed that MSW components such as plastics
(excepting PVC) and cellulose (paper, cardboard, packag-
ing, wood and organic waste) liberate volatile substances
within the range of 400440 C. This phase duration is
about 280 s depending on the composition of the reconsti-
tuted MSW sample. The other line (slope BCF) corresponds
to the waste xed carbon oxidation. Knowing the residence
time on an incinerator grate, for instance, it is possible to
establish CFR evolution law under industrial conditions
relying only on the waste advancement status.
The installation also enables the researcher to quantify
CFR variations with the treatment temperature under oxi-
dant conditions for a certain residence time. Fig. 5 shows
this variation for a 600 s residence time for the same sample
of reconstituted municipal solid waste with the above-
0
BMV = 0,012 s-1
-2
ln(1-CFR/100)

process, ts is the xed residence time, m_ m is the waste mass BCF = 0,003 s-1
-4
feed ow and m_ 0m is the solid residue mass ow at the instal-
lation outlet. -6
As the waste has inert components as well as a combus-
-8
tible fraction, in order to establish the product reaction volatiles emission zone fixed carbon decomposing zone

kinetics, it is important to know the mass reduction rate -10

0 100 200 300 400 500 600
of its combustible fraction but not the global mass reduc-
Residence time [s]
tion rate. For that we have to take into account the
uncombusted (unoxidized) fraction in the slag. Knowing Fig. 4. CFR variation with the residence time for reconstituted urban
the initial inert fraction of the waste, we can calculate the waste.
combustible fraction reduction rate: Xc(T, ts). The two
parameters Xm and Xc are related by the equation:
95
Xm 93
Xc % with K the waste inert fraction. 3 91
1K 89
CFR [%]

Eq. (4) connects the CFR of collected solid residues to the 87

85
unoxidized fraction XIMB from the slag:
83
 
100 S zSIN 81
CFR S z  X IMB % 4 79
zCOMB COMB 100  X IMB 77
75
where zSIN
is the inert mass participation in the waste and 450 500 550 600 650 700 750 800 850 900 950

XIMB is the unoxidized fraction in the slag. Temperature [grd C]

The CFR denition can be transposed for a xed tem- Fig. 5. CFR variation with the reactor temperature (residence
perature in: time = 10 min).
 C. Marculescu et al. / Waste Management 27 (2007) 367374
100
95
CFR [%]
90
85
80
75
70
0 20 40 60
Inert fraction in the mixture [%]
Fig. 6. Inuence of the inert fraction in the sample on the CFR variation
(residence time = 10 min, temperature = 900 C).
mentioned composition. This kind of test can be accom-
plished for dierent residence periods within the opera-
tional limits of the installation. The residence period can
be adjusted by means of vibration parameters settings. If
the treatment period is not long enough compared to the
research conditions, the product can be repeatedly passed
through the installation. That will generate an error due
to the inevitable solid cooling in the re-feeding procedure.
The tests were made at temperatures between 500 and
900 C in stationary conditions. The temperature range
was above 500 C to ensure that no volatile components
are still present in the solid particles, even if their emission
temperature was lower.
We assumed that this variation is independent from the
inert fraction of the sample. We initiated a series of tests to
determinate the inuence of the inert on the CFR criteria.
The results presented in Fig. 6 prove that CFR variation is
independent from the percentage of inert materials in the
urban waste sample. The small variations are caused by
the inherent experimental conditions. Therefore the rela-
tion for the CFR calculation does not take into consider-
ation, as previously explained, the quantity of inert
component in the treated mixture (in our situation the
argyle balls, as a substitute for the glass and metallic com-
ponents in the urban waste).
In order to validate the results on the reconstituted
urban waste, similar tests were carried out for various com-
binations of products (wood, pine bark and PET) with
inert solids (argyle or glass balls).
Therefore, the inert can be used to surpass the installa-
tion exploitation limits described at 2.3 Limitations
without a notable inuence on the xed carbon
decomposition.
4. Conclusions
The vibro-uidised transport reactor enables the estab-
lishment of the thermo-chemical properties of waste in mix-
tures from the combustibility or pyrolytic standpoints. The
device also enables the analysis of the solid residue for each
homogeneous thermal treatment process that the solids
continuously undergo. Test accuracy is guaranteed by the
isothermal installation prole and the piston-like ow res-
373
idence time distribution, independent from the solid bed
properties. Tests showed the independence of the treated
product residence time (from 4 up to 40 min) from its gran-
ularity and specic weight if the product components with
less than 1 mm granularity do not exceed a maximum of
10% in the mixture. The mass ow ranges from about
1 kg/h up to 50 kg/h, depending on the product specic
weight. The treatment atmosphere is adjustable to research
80 conditions: from pyrolysis with less than 1% oxygen to
combustion with air excess up to 1.51.8. The dened cri-
teria for waste mechanical and thermo-chemical character-
isation are fully quantiable and appropriate for the
simulation of industrial scale thermal degradation of solid
heterogeneous mixtures.
References
Akiyama, T., Yamaboshi, H., 1992. Behaviour of vibrating beds of
irregular particles. Powder Technology 69, 163169.
Bockhorn, H., Hornung, A., Hornung, U., Schawaller, D., 1999. Kinetic
study on the thermal degradation of polypropylene and polyethylene.
Journal of Analytical and Applied Pyrolysis 48, 93109.
Caudron, J.C., 1999. Mises au point dune procedure de caracterisation de
lincinerabilite des dechets solides (Procedure for the characterization
of the solid waste incineration capacity), These UTC, Compiegne, pp.
5365.
Cozzani, V., Nicolella, C., Petarca, L., Rovatti, M., Tognotti, L.,
1995a. A fundamental study on conventional pyrolysis of a refuse-
derived fuel. Industrial and Engineering Chemistry Research 34,
20062020.
Cozzani, V., Petarca, L., Tognotti, L., 1995b. Devolatilization and
pyrolysis of refuse derived fuels: characterization and kinetic
modelling by a thermogravimetric and calorimetric approach. Fuel
74, 903912.
Fontana, A., Burzynska-Weis, B., Jung, G., Braekman-Danheux, C.,
Laurent, P., 1997. Pyrolysis of municipal solid waste: an additional
technique for waste management. IChemE Environment Protection
Bulletin 048, 1520.
Heikkinen, J.M., Hordijk, J.C., De Jong, W., Splietho, H., 2004.
Thermogravimetry as a tool to classify waste components to be used
for energy generation. Journal of Analytic Applied Pyrolysis 71, 883
900.
Jaraiz, E., Kimura, S., Levenspiel, O., 1992. Vibrating beds of ne
particles: estimation of interparticle forces from expansion and
pressure drop experiments. Powder Technology 72, 2330.
Martin, C., 2001. Modelisation et criteres de combustibilite en incineration
combinee de dechets menagers et de dechets industriels banals
(Simulation model and combustion criterions for the incineration of
the urban and industrial wastes mixtures), These UTC, Compiegne,
pp. 125127.
Neddreman, R.M., Harding, G.H.L., 1990. The ight-free conveyor: basic
theory and performance analysis. Transactions of the Institution of
Chemical Engineers 68 (Part A), 131138.
Orfao, J.J.M., Antunes, F.J.A., Figueiredo, J.L., 1999. Pyrolysis kinetics
of lignocellulosic materials three independent reactions model. Fuel
78, 349358.
Porquet, C., 1994. Rapport de DEA Master Degree in Pyrolysis of
Agricultural Wastes. Universite de Technologie de Compiegne.
Raveendran, K., Ganesh, A., Khilar, K.C., 1996. Pyrolysis characteristics
of biomass and biomass components. Fuel 75 (8), 987998.
Roga, C., 1974. International Norm ISO 335, Pitcoal: Agglutinant
Capacity Quantication Roga Test.
Socha, H., Kolze, K., 1992. Procedes de calcul pour lanalyse du
movement des convoyeurs vibrants (Computation model for the
vibrant conveyors movement). Maschinenbautechnic 39, 403407.
374 C. Marculescu et al. / Waste Management 27 (2007) 367374

Stengsen, M., Jensen, A., Dam-Johansen, K., Grnli, M., 1999. Exper-
imental investigation and kinetic modeling of biomass pyrolysis. In:
Proceedings of the Second Olle Lindstrom Symposium, Stockholm, 9
11 June.
Tillman, D.A., 1991. The combustion of solid fuels and wastes. Academic
Press, Inc., Boston, pp. 3645.
Varhegyi, G., Antal, M.J., Jakab, E., Szabo, P., 1997. Kinetic
modeling of biomass pyrolysis. Journal of Analytic Applied
Pyrolysis 42, 73.
Wu, C.-H., Chang, C.-Y., Hor, J.-L., Shih, S.-M., Chen, L.-W., Chang,
F.-W., 1993. On the thermal treatment of plastic mixtures of MSW:
pyrolysis kinetics. Waste Management 13, 221.