3

Selected Topics on Combustion

Processes

Combustion processes involve mass and energy transport and chemical reaction. Depending on the relative pace of these necessary steps, the process characteristics change. Such
changes can affect the degree of success of the combustion system in performing
according to expectations.
It is the goal of design engineering to apply an understanding of the physical and
chemical processes that govern systems of interest in order to quantitatively predict
performance. Alternatively, we may wish to translate performance goals into system
hardware designs and operating set points. A rst-principles approach to these tasks is
difcult and time-consuming. Yet exploration of the fundamental processes involved gives
valuable insight into the appropriateness of designs and into the interpretation and solution
of operating problems.
To this end, complete combustion processes for gaseous, liquid, and solid fuels are
described. Also, intentionally incomplete combustion (or pyrolysis) is described to aid in
understanding waste processing concepts that are based on this interesting but less
common concept in waste pyroprocessing.
I.

GASEOUS COMBUSTION

Although most wastes requiring incineration are not gaseous, most combustion is
completed through gas phase reactions. Many liquids vaporize in the hot furnace
environment, and it is the resulting gas molecules that actually engage in combustion.
Heavier liquids and solids are pyrolyzed by intense heating. This generates lowermolecular-weight volatile fragments that burn in the gas phase. Even carbon char
(appearing as the waste of interest or as a product of the pyrolysis process) is usually
gasied (by oxygen or water vapor) to carbon monoxide that burns in the gas phase.
In reviewing gas phase combustion, systems may be divided into those where fuel
and stoichiometric oxidant enter the combustion environment separately and where the
combustion rate is almost always mixing limited, and those where combustion is initiated
in homogeneous fueloxidant mixtures, with the ame reactions propagating through the

system in a substantially continuous (although, perhaps, smudged by turbulence) ame

front. The latter case is found primarily only in premixed gas burners. The former is by far
the most common.
The mechanism by which oxidant and fuel are brought together provides a second
classication of ame type. When molecular diffusion predominates, the ame is known
as a laminar diffusion ame. When eddy diffusion predominates, the ame is known as a
turbulent diffusion ame. The Reynolds number of the streama measure of the ratio of
momentum forces (scaling eddy mixing) to viscous forces (scaling molecular mixing)
can be a useful indicator of the regimes where one or the other ow conditions exist.
A.

The Premixed (Bunsen) Laminar Flame

Visual examination of simple, premixed conical hydrocarbonair ames indicates three

regions: preluminous, luminous, and postluminous. In the preluminous zone, the cold feed
gas undergoes preheating and some seeding with reactive species due to thermal and
mass diffusion against the direction of convective ow. At some point the combination of
temperature and mixture reactivity reaches a level where rapid heat release begins.
Activated radicals (e.g., OH, CC, and others) appear in high concentration and account
for much of the visible light released in the blue luminous zone. Gouy suggested in 1881
(14) that the conical shape of the luminous zone could be explained by assuming that the
ame propagates normal to itself at a constant rate, known as the burning velocity or ame
speed. This simple concept has withstood the testing of many investigators and appears to
be able to quantitatively explain ame front geometry. For most hydrocarbon fuels, the
ame speed falls in the range from 35 to 60 cm=sec at normal atmospheric temperatures
and pressures.
The combustion reaction is seldom complete within the luminous zone. Particularly
for the combustion of complex fuels, heat release from CO and H2 oxidation extends into
the postluminous zone (an afterburning region). In the postluminous zone, the temperature
falls due to radiation heat loss and to the assimilation of cool ambient uid into the ame
ow, and combustion is completed.
Figure 1 illustrates the heat release for a stoichiometric propaneair ame at 0.25 atm
(15). The low pressure increases the thickness of the luminous zone. A similar thickening
is observed as the degree of premix (air-to-fuel ratio) is decreased.
B.

The Diffusion Flame

In a diffusion ame, fuel combustion reactions must await the arrival of oxidant.
Preparatory steps such as pyrolysis, evaporation, or thermal cracking, however, may
proceed under the inuence of radiative heat transfer from the (typically) hot surroundings,
the heating due to turbulent mixing of hot burned gases into the fuel stream, and
conduction from the ame regions.
For most burner systems, the fuel is introduced into the combustion chamber in a jet,
with or without swirl. Since combustion is mixing limited, the ow characteristics
(especially entrainment) of the jet are prime determinants of combustion rate, ame
length, and ame shape. Since jet behavior is very important to diffusion ame and
combustor uid mechanics, it is discussed in detail below. One important ame quantity,
the ame length, will be mentioned here for the more common case, the turbulent diffusion
ame.

Figure 1 Heat release rate in stoichiometric propane-air ame at 0.25 atm. Curve calculated from
the temperature prole using estimated thermal conductivities.

Based on a simplied model for jet mixing, Hottel (16) proposed and tested the
following relationship dening the ame length L, corrected for the distance X from the
nozzle face to the break point where turbulence is initiated:



1=2

 Ma
L  X 5:3
TF

fs 1  fs
1
fs
d
M 0 y m T0
where
L ame length (meters)
X distance from nozzle face to initiation of turbulence (meters)
d nozzle diameter (meters)
fs mole fraction of nozzle gas in a stoichiometric gasair mixture
Ma ; M0 molecular weight of air and nozzle gas, respectively
TF adiabatic ame temperature (Kelvin) of a stoichiometric mixture
T0 nozzle gas temperature (Kelvin)
ym moles of reactants per mole of product in a stoichiometric mixture
Estimation of X is necessary if total ame length is to be determined. Generally,
X is small compared to L  X except for CO ames for which X =L  X is as
high as 0.4. On the average, X is about 0.05L, is less than 0.09L in the turbulent range,
and decreases with increasing nozzle Reynolds number.
This correlation is useful in that it allows anticipation of ame jet impingement
problems and allows estimates of the effect of fuel changes (changes in fs , M0 , TF , and ym )

and of preheat (changes in T0 ) on ame length. Its accuracy for a variety of gaseous fuels
ranged within 10% to 20% of the experimental values. It should be recognized that devices
that increase the rate of entrainment of oxidant will act to shorten the ame. Thus, swirl in
particular will signicantly shorten ame length.
EXAMPLE 1. Natural gas at 15
C (methane, with a ame temperature of 2170 K) is
to be burned in air from a port 4 cm in diameter at a velocity of 3 m=sec. The furnace is 8 m
wide. Flame impingement has been a problem at other plants, and so the plant engineer
suggests cutting back on the fuel ow by up to 20% to reduce ame length. Will this work?
What will?
First, the jet should be evaluated to determine if it is, indeed, turbulent.
du r
2
Reynolds no: NRe 0 0 0
m0

where
d0 nozzle diameter 4 cm
u0 discharge velocity 3 m=sec 300 cm=sec
r0 density of methane at 15
C 6:77  104 g=cm3
m0 viscosity of methane at 15
C 2:08  104 g=cm sec
NRe

43006:77  104
2:08  104

3900 dimensionless
Since the Reynolds number exceeds that required for turbulent ow (2100), the ame
length relationship should hold.
Since 1 mol of methane burns stoichiometrically with 2 mol of oxygen, and
2(79=21) 7.52 mol of nitrogen, fs 1=1 2 7:52 0:095.
Methane burns according to the reaction
CH4 2O2 ! CO2 2H2 O
3 mols

3 mols

so
ym 3=3 1:
Assuming that X is much less than L,



1=2
L
5:3
29
2170

0:095 1  0:095
d 0:095
16 1288
202
and
L 8:07 m
Will dropping the ring rate help? The lower inlet velocity will lower the air entrainment
rate in proportion, and the ame length will remain essentially constant.
What could be done?
1.

Use a 3-cm nozzle diameter. This will cut the ame length to 3=4 of the base
case with an increase in gas compression costs. Alternatively, use two or more
nozzles.

2.
3.
4.
5.

Preheat the fuel. If the methane is introduced at 250
C (523 K), the ame length
will decrease to 6 m.
Partially premix air with the gas (changing fs and My ), being careful to avoid
ashback problems.
Explore the possibility of tilting the burner.
Explore the use of swirl-type burners.

Use of the equation presented should cause no problems in many situations. If,
however, the burner discharges into hot furnace gases where the oxygen concentration is
depleted, or operates as part of a bank of fuel nozzles where the zone of inuence of fuel
jets overlaps, a lengthening of the ame can be anticipated.
II.

LIQUID COMBUSTION

The combustion of liquids begins with the accumulation of heat in the liquid mass from
radiation from the ame or from hot refractory walls, from recirculating burned gases, or
from a pilot burner. Eventually, the fuel approaches its boiling point and it begins to
vaporize. The fuel vapors then diffuse from the liquid surface, mixing with oxygen in the
air and increasing in temperature. At some point, the rate of oxidation reactions and
associated heat release is high enough that full ignition and combustion ensues.
In most situations it is desirable to subdivide the oil into small droplets in order to
enhance vaporization rate (increasing the combustion rate per unit volume), to minimize
smoking, and to ensure complete combustion. In some simple pot-type heaters, in pit
incinerators, and occasionally in highly unusual waste disposal situations, sludge or liquids
are burned in a pool.
A.

Pool Burning

Fires above horizontal liquid pools have been studied both experimentally and analytically
to clarify the relationship between burning rate and the properties and characteristics of the
system.
In small pools (up to 3 cm in diameter), burning is laminar and, unexplainedly, the
_ 00 decreases in approximate proportion as diameter increases. In
burning rate per unit area m
a 3- to 5-cm-diameter pool, the ame is still laminar but regular oscillations occur.
For diameters greater than 7 cm and less than 20 to 30 cm, a transition to turbulent
_ 00
burning takes place. Convection is the primary source of feedback energy inuencing m
00
_ is essentially independent of pool diameter. For
in this pool size range, and the overall m
_ 00 range extends to
ames that yield little radiation (e.g., methanol), the constant m
diameters over 100 cm.
_ 00 (average) for a range of pool diameters does not imply a constant
A constancy of m
_ with radius. Akita and Yumoto (17) report a 44% maximum radial mass ux variation
m
for a 14.4-cm diameter methanol re, and Blinov and Khudakov (18) report an 18%, 18%,
and 27% maximum radial variation for benzene, gasoline, and tractor kerosene, respectively, in a 30-cm tank. These variations, however, may be considered small.
Studies by DeRis and Orloff on pool burning rates in the convectively controlled
region (19) identied a relationship between burning rate and characteristics of the fuel
which appears to satisfactorily correlate several sets of data for values of B (a dimension_ 00 is
less mass transfer driving force) greater than unity. Below B 1, the burning rate m
lower than predicted, apparently due to the onset of unexplained extinction processes.

DeRis and Orloffs correlation is

!

2=3
l
gra  r3 1=3 ln1 B
1
00
_ 0:15
m
B
kg=sec-m2
B
Cp1
m 2 a2

3a

and
B

DHc
H
 s
yDHv DHv

an enthalpy driving force

3b

where
_ 00 burning rate (kg=sec m2 )
m
g acceleration of gravity (m=sec2 )
DHc heat of combustion (HHV) of fuel (kcal=kg)
y kilograms oxidant (air) per kilogram fuel (kcal=kg)
DHv heat of vaporization of fuel (kcal=kg)
Hs heat content (sensible) of 1 kg of oxidant (air) at the liquid temperature at the
surface of the burning pool (kcal=kg)
ra density of air at ambient temperature (kg=m3 )
and where the following are the properties of the gas mixture formed upon stoichiometric
combustion of the fuel and subscripted 1, 2, or 3 to correspond to evaluation at 25%, 50%,
and 100%, respectively, of the total enthalpy of a stoichiometric mixture:
l thermal conductivity (kcal m1 sec1
C1
Cp specic heat (kcal kg1
C1
r density kg=m3
m viscosity kg=m sec1
a thermal diffusivity m2 =sec
To use this correlation, the temperatures corresponding to various enthalpies are
easily determined, using a plot such as Fig. 4 in Chapter 2. Mixture properties at a
temperature T are calculated for mole fractions fi of components with molecular weights
MWi :
P
r
f i ri
4
i

fi MW i1=2 mi
i
m P
fi MW 1=2
i

fi MW i1=3 li
l P
fi MW 1=3
i
i

fi Cpi MW i
Cp P
fi MW i
i

l
a
rCp

As the pool diameter increases beyond the 20- to 30-cm range, radiation from the
ame plume plays an increasingly important role in setting the evaporation (burning) rate.
For these larger pools where the ame plume is optically thick, the burning rate of a liquid
with a density r g=cm3 is given by
_ 00
m

0:076r DHc
kg m2 sec1
DHv

where terms are dened as for Eq. (3). An analysis by Steward (20) of the data of several
investigators developed a relationship correlating the height of the visible ame plume
with the pool size and burning characteristics. Stewards relationship showed that the tip of
the visible ame plume corresponded to a condition of 400% excess air.
B.

Droplet Burning

For most liquid waste incineration and liquid fuel burning equipment, atomization is used
to subdivide the feed for more efcient and complete burning (see Fig. 2). Atomization is
effected by a number of means, including
Low-pressure (0.03 to 0.35 atm) air atomization
High-pressure (2 atm or more) air or steam atomization
Mechanical atomization of uid through special orices at pressures from 5 to 20
atmospheres
High-speed rotating conical metal cups
The power requirements and characteristics of each atomizing method are described
in the chapter relating to liquid waste incinerators.
Early analysis of droplet burning was based on analogies with coal combustion.
These studies viewed droplet combustion as involving the diffusion of oxidant to the
droplet surface followed by heterogeneous (two-phase) reaction. Later work led to the

Figure 2 Typical model of droplet burning.

presently accepted concept that combustion occurs in a homogeneous diffusion ame

surrounding an evaporating droplet (Fig. 2).
A review by Williams (20a) suggested the following as the burning time tb (sec) of a
droplet of initial diameter dr (cm) of a hydrocarbon oil of molecular weight MW at a
temperature T (K) in an atmosphere with an oxygen partial pressure pO2 (atm) as
tb

29;800Mw do2
pO2 T 1:75

10

A more detailed theoretical analysis of the evaporation and burning of individual droplets
was presented by Kanury (435) who considered the two-step process: gasication or
vaporization of a liquid waste or fuel followed by gas phase combustion. For lightweight
liquids, gasication is a purely physical (evaporation) process. For heavier fuels or wastes,
pyrolysis or destructive distillation may be involved.
One of the key steps in the analysis is the establishment of the conditions at the
surface of the droplet. For lighter fuels, it is useful and usually accurate to assume that the
surface temperature of the evaporating droplet approaches (but is always slightly lower
than) the boiling temperature of the liquid at the extant total pressure. If pyrolysis is
involved, there is often no discrete boiling temperature and the surface temperature is
usually much higher.
For the lighter liquid fuels or wastes, Kanurys analysis considered the heat transfer
and mass transfer processes involved in droplet evaporation. In most incineration
scenarios, the temperature of the gaseous environment into which the droplet is sprayed
(T1 ) is signicantly higher than the liquids boiling point TB . Under these conditions, the
mass transfer number (B) is given by
B

cg T1  TB
L c1 TB  TR

dimensionless

11

The evaporation constant lv is given by

lv

8rg ag
lnB 1 cm2 =sec
r1

12

And the droplet evaporation time tv is equal to

tv

d02
r1 d02
sec

lv 8 rg ag lnB 1

where
cg specic heat of the liquids vapor (cal=gm
C)
T1 temperature of the surrounding air (
C)
TB boiling point of the liquid (
C)
L heat of vaporization of the liquid (cal=g)
cl specic heat of the liquid (cal=gm
C)
TR bulk temperature of the liquid
rg density of the liquids vapor (gm=cm3 )
ag thermal diffusivity of the liquids vapor cm2 =sec kg =rg cg
kg thermal conductivity of the vapor at TB cal cm=sec cm2
C
rl density of the liquid (g=cm3 )
d0 initial diameter of the droplet (cm)

13

Kanurys extension of the analysis to include the effects of combustion leads to a

modication of the mass transfer number to
B

DHf Y01 cg T1  TB
L c1 TB  TR

dimensionless

14

where the denitions are as above and

DH heat of combustion of the liquid (cal=g)
f mass stoichiometric ratio (grams of liquid=g of O2 )
Y01 mass concentration of oxygen in the environmental gas (grams=gram) and
equal to 0.232 (23.2%) for air
The mass transfer driving force for several liquid fuels burning in air are illustrated in
Table 1.
For the burning constant lb dened by
lb

8rg ag
lnB 1 cm2 =sec
r1

15

Table 2 indicates the accuracy of prediction of lb in contrast to experimental determinations.

The burning time tb of an individual droplet is given by
tb

d02
sec
lb

16

Table 1 Mass Transfer Driving Force for Combustion of Liquids in Air

Liquid
n-Pentane
n-Hexane
n-Heptane
n-Decane
Benzene
Methanol
Ethanol
Kerosene
Light diesel
Medium diesel
Heavy diesel
Acetone
Toluene
Xylene
Assumed values:
cg 0:31 cal=gram
C
T1 20
C
YO1 assumed 0.232
TR assumed 20
C
Source: From (435).

cg T1  TB

DHf =Y01

cl TB  TR

TB

4:95
14:80
24:30
47:70
18:60
13:80
18:10
71:20
71:20
74:30
77:50
11:40
28:00
34:00

785
770
770
770
790
792
776
750
735
725
722
770
749
758

8.91
25.80
41.20
80.50
24.70
25.20
32.60
106.00
103.50
107.50
109.00
23.60
35.00
45.30

87.1
87.1
87.1
86.0
103.2
263.0
200.0
69.5
63.9
58.4
55.5
125.0
84.0
80.0

36.0
68.0
98.5
174.0
80.0
64.5
78.5
250.0
250.0
260.0
270.0
56.7
110.6
130.0

8.15
6.70
5.24
4.34
6.05
2.70
3.25
3.86
3.96
3.94
3.91
5.10
6.06
5.76

Table 2 Burning Constant for Various Hydrocarbons Oxidant

(Air at 20
C and 1 atm)

lcalculated
Fuel=Liquid
Benzene
Toluene
o-Xylene
Ethyl alcohol
n-Heptane
n-Heptane
iso-Octane
Petroleum ether
Kerosene
Diesel oil r 0:850

lmeasured

103 cm2 =sec

11.2
11.1
10.4
9.3
14.2
14.2
14.4

9.7
8.5

9.9
6.6
7.9
8.6
8.4
9.7
9.5
9.9
9.6
7.9

Reference
443
444
444
443
443
444
444
444
444
444

The application of burning time predictions based on Kanurys analysis must be

cautious and conservative. Extrapolating from the single droplet model to the analysis
of real, large-scale combustors requires addition of four rather complex renements to the
theory:
1.
2.
3.
4.

In dense sprays, evaporation and diffusion burning of the vapor become the
dominant mechanism.
Relative motion between droplet and the gaseous surrounding medium removes
the simplifying assumptions of radial symmetry.
Initial droplet particle sizes are not uniform, and the evaporation characteristics
are poorly described using models based on mean droplet sizes.
Mixing processes of fuel and oxidant take place in the aerodynamically complex ow in the burner and furnace (not in a simplied, one-dimensional
reactor).

Unfortunately, such studies have not yet produced quantitatively useful correlations for use
in combustion chamber design. However, the parameters in the equations give considerable insight as to the effect of user-controllable variables in enhancing the combustion rate.
For example, droplet combustion rate is increased by reducing drop size, increasing the
temperature of the ambient atmosphere, operating at high oxygen concentration, and
preheating the fuel.
As a starting point in burner analysis, the ame length relationship in Eq. (1) could
be used for jets of liquid fuel droplets. Application of Eq. (1) is subject to the qualication
that the ame will be longer than predicted with the difference in predicted and actual
decreasing as
The number of coarse droplets decreases.
The latent heat of evaporation (less any preheat) decreases. And=or
The mean droplet diameter decreases.
Also, the ame will be considerably shortened by swirl effects.

III.

SOLID COMBUSTION

The combustion of solid phase wastes, as with liquids, is largely a process of gasication
followed by combustion of the gaseous products (see Fig. 3). Even in the cases where
heterogeneous attack of the solid by a gaseous oxidant such as O2 , CO2 , or H2 O occurs
(e.g., in combustion of coke), the initial by-product is often carbon monoxide rather than
fully oxidized CO2 .
The development of furnaces to burn solids has let to three characteristic ring
modes:
1. Suspension burning. Solids burn while entrained in a fast-moving air stream. In
some embodiments, the particles of solids are nely divided and dry. This is the case for
the burning of pulverized coal, rice hulls, etc. In other instances, the feed is introduced as a
semiliquid sludge that dries, gasies, and burns in suspension. This is the case for
wastewater treatment plant sludges in a uid bed combustor. The gaseous environment is
hot and often oxygen rich. Combustion is completed while the particles are suspended in
the gas ow. In some congurations, the solids leaving the combustion zone are captured
and reinjected to improve burnout. Much of the residue leaves the system still entrained in
the ue gases.
2. Semisuspension burning. Coarsely subdivided solids are injected mechanically
or pneumatically into the combustion space. Drying and combustion occur partially while
the solids are airborne and are completed after the particles fall to a grate. This technology
is used for spreader stoker-red coal and prepared refuse derived fuel (RDF). Residue is
dumped periodically or is continuously withdrawn.
3. Mass burning. Unsubdivided or coarsely subdivided solids are moved into the
combustion space by mechanical means (e.g., by hydraulic ram, manual charging, or
drawn in on a grate) and are pushed, are dragged, or tumble through the furnace until a
satisfactory degree of burnout is obtained. In some small, low-capacity units, the feed
remains substantially undisturbed in the combustion space while it burns. Residue is
discharged periodically or continuously.
In waste disposal practice, each of these techniques has been used. Quantitative
descriptions of the processes involved are still under development and are costly to apply
in most conceptual and nal design efforts. However, several of the important steps in
combustion are subject to analysis, particularly if graphical tools and modern microcomputers are used. A summary of the results are given below to give insight into the
processes and their control.
In suspension burning, entering particles of fuel are almost instantly exposed to
intense radiant, convective, and conduction heat transfer as hot combustion products are
entrained into the jet of suspending uid. In mass burning, much of the mass is shielded
from radiant heating. Thus, fuel particles deep within the bed must await the arrival of the
ignition front before rapid temperature increases are experienced.
Semisuspension burning processes fall between the extremes. As the temperature of
a fuel particle increases, free moisture is lost, and then decomposition of organic matter
begins. The volatile matter includes a wide variety of combustible hydrocarbons, carbon
monoxide, carbon dioxide, water vapor, and hydrogen. The weight loss and composition of
the volatiles developed by gasication of the solid depend on the time and temperature
history. The heating rates in pulverized coal furnaces are very high (104
C=sec or more).
In such cases, decomposition is complete in less than 1 sec for particles smaller than about
100 mm.

Figure 3 Phase changes in the combustion of organic solids.

At the other extreme, a slower rate of heating (less than 10
C=min) characterizes
carbonization process in charcoal manufacture. Intermediate heating rates would be found
in waste incineration by mass burning (say, 1000
C=min) or the proximate analysis test
(about 300
C=min).
The signicance of these processes to the incinerator designer arises from their
impact on the spatial distribution of volatile release, and thus combustion air requirement
within the furnace. If, for example, volatile release is extremely rapid, the region near the
charging point of a mass burning system would behave as an intense area source of fuel
vapors, driven out of the fuel bed by undergrate air ow. Combustion air requirements,
therefore, would be largely in the overbed volume in this region of the furnace and the
high, localized heat release rate must be taken into account in developing the design of the
enclosure. If, on the other hand, volatilization is slow or limited, heat release will be
conned largely to the bed and the majority of the combustion air should be underre.
With a perversity not uncommon in the analysis of waste disposal problems, the
intermediate situation appears to be the case.
The introduction of oxidant only in localized areas of the incinerator environment
introduces a second complication. In the situation where a mass of ignited solid waste sits
upon a grate through which air is passing, the gaseous environment varies with distance
above the grate line. In the idealized case, the following processes occur:
At and near the grate, air supply is far above stoichiometric, and released volatiles
burn rapidly and completely to CO2 and water vapor. Heat release rates per unit
volume are high. One could call this the burning zone.
Just ahead of the fully involved burning zone and closer to the feed point, the waste
mass includes both wet and dry material. The ignited, burning dry material
generates heat and hot gases that supply the heat needed to dry and preheat the
wet material. A portion of the drying energy may be supplied by preheated
undergrate combustion air. This is the drying zone.
At some point away from the grate, the free oxygen is exhausted, but both CO2 and
H2 O act as oxidizing agents to produce CO and H2 in both heterogeneous and
homogeneous reactions. This is the gasication zone.
Further up in the bed, no oxidizing potential remains in the gases, and the effect is
one of immersion of solid waste in a hot gas ow with pyrolysis reactions
predominating. This is the pyrolysis zone.
In real situations, the bed is often split open in places due to the heterogeneous form
of solid wastes. This leads to channeling or bypassing effects. Also, some air may be
entrained into the bed from the overbed space. Further, differences in the point-to-point
underre air rate (m3 =sec m2 ) due to grate blinding and spatial variation in the density of
combustible throughout any vertical plane in the bed lead to irregularity in volatilization
patterns. The net result of these factors is to smudge the boundary between zones and to
produce, at the top of the bed, a gas ow that varies moment to moment from fuel-rich to
air-rich.
A.

Thermal Decomposition

The thermal decomposition or pyrolysis of waste solids in the absence of air or under
limited air supply occurs in most burning systems. Pyrolysis is a destructive distillation
process effected by the application of heat in an insufciency of air to yield gaseous, liquid

(after cooling), and solid products. In comparing suspension burning and mass burning
processes, both involve pyrolysis of incoming solids, but for suspension burning the
physical scale of the fuel-rich zone is smaller, and the pyrolysis products will differ due to
the differences in heating rate. Figure 4 illustrates the sequence of steps in pyrolyzing an
organic char-forming material.
Both physical and chemical changes occur in solids undergoing pyrolysis. The most
important physical change in materials such as bituminous coal and some plastics is a
softening effect resulting in a plastic mass, followed by resolidication. Cellulosic
materials increase in porosity and swell as volatiles are evolved.

Figure 4 Process sequence in pyrolysis.

As cellulose pyrolysis begins (at about 200
C, complex, partially oxidized tars are
evolved. As the temperature increases, these products further decompose or crack, forming
simpler, more hydrogen-rich gaseous compounds and solid carbon. The solid residue
approaches graphitic carbon in chemical composition and physical structure.
Whether the overall pyrolysis process of a given solid is endothermic or exothermic
depends on the ultimate temperature attained. For most materials, the process is
endothermic at lower temperatures and exothermic at higher temperatures.
The rate-controlling step in pyrolysis (the rate of heat transfer into the waste or the
pyrolysis reaction rate) is dependent upon the temperature and the physical dimensions of
the waste. Below 500
C, the pyrolysis reactions appear rate-controlling for waste pieces
smaller than 1 cm in size. Above 500
C, pyrolysis reactions proceed rapidly, and heat
transfer and product diffusion are rate-limiting. For large pieces (greater than 5 cm), heat
transfer probably dominates for all temperatures of practical interest.
The upper temperature limit for pyrolysis weight loss is a function of the material
being heated. For bituminous coal the weight loss achieved at 950
C is about as high as
will be obtained at any practical temperature. Cellulose pyrolysis is essentially completed
between 575
and 700
C.
1.

Pyrolysis Time

The time required for pyrolysis of solid wastes can be estimated by assuming that the rate
is controlled by the rate of heating. Neglecting energy absorption or generation by the
reactions, one may estimate pyrolysis time by assuming the refuse piece (a plate or a
sphere) is suddenly plunged into a hot reactor maintained at a high temperature. The
pyrolysis (heating) time is dened here as the time required for the center temperature to
rise by 95% of the initial temperature difference between the specimen and the surroundings. A thermal diffusivity of 3:6  104 m2 =hr was used; it is roughly equal to that of
paper, wood, coal, and many other carbonaceous solids. Radiation and contact conduction
are neglected. The results for thin plates or spheres (21) are shown in Figs. 5 and 6.
The effect of specimen size is evident, thus showing differences expected with size
reduction. The effect of gas velocity is also signicant and illustrates that in mass burning,
where relative gassolid velocities may range from 0.1 to 0.4 m=sec, slow pyrolysis times
would be expected. The relatively long pyrolysis time calculated for small particles at very
low (00.1 m=sec) relative velocity (as would be found in suspension burning) suggests
strongly that other heat transfer mechanisms (especially radiation) can become of primary
importance. The results for entirely radiant heating to the plate or sphere starting at an
initial temperature of 20
C is shown in Figs. 5 and 6 as the heating time at innite crossow velocity (V1 ). The indicated heating time requirement is seen to be large compared to
the residence time in the radiant section of a boiler (approximately 12 sec) and thus,
without considerable reduction in particle size of the feed, unburned material can enter the
convection tube banks.
2.

Pyrolysis Products

a. Liquid Products. The liquid product of pyrolysis is known to contain a complex

mixture of alcohols, oils, and tars known as pyroligneous acids. Pyrolysate liquids from
municipal refuse contain from 70% to 80% water with the remainder being a wide range of
(often oxygenated) organic compounds. Specic compounds found in the pyrolysate
include acetic acid, methanol, 2-methyl-1-propanol, 1-pentanol, 3-pentanol, 1,3-propane-

Figure 5 Radiative and convective heating time for a thin plate. (From Ref. 21.)
diol, and 1-hexanol. The composition is highly dependent upon refuse composition,
heating rate, and ultimate temperature (see Tables 3, 4, and 5). The yield of liquid is
approximately 50% to 60% of the air-dried, ash-free refuse, decreasing with increasing
ultimate temperature. The heat content of the liquid per pound of refuse decreases as the
pyrolysis temperature increases.

Figure 6 Radiative and convective heating time for a sphere. (From Ref. 21.)

A set of pyrolysis data were collected by Folks et al. (345) for thermophilic and
mesophilic biological wastewater treatment sludge from the Hyperion treatment plant in
Los Angeles. The pyrolysis experiments were run from 765
to 1004
C and produced char,
pyrolysis oil, and gaseous products in the proportions 48.0%, 4.73%, 24.44% and 41.35%,

Table 3 Yields of Pyrolysis Products from Different Refuse Components (Weights Percent of
Refuse)a
Component
Cord hardwood
Rubber
White pine sawdust
Balsam spruce
Hardwood leaf mixture
Newspaper I
Newspaper II
Corrugated box paper
Brown paper
Magazine paper I
Magazine paper II
Lawn grass
Citrus fruit waste
Vegetable food waste
Mean values
a

Gas

Water

Other
liquid

Char
(Ash-free)

Ash

17.30
17.29
20.41
29.98
22.29
25.82
29.30
26.32
20.89
19.53
21.96
26.15
31.21
27.55
24.25

30.93
3.91
32.78
21.03
31.87
33.92
31.36
35.93
43.10
25.94
25.91
24.73
29.99
27.15
23.50

20.80
42.45
24.50
28.61
12.27
10.15
10.80
5.79
2.88
10.84
10.17
11.46
17.50
20.24
22.67

29.54
27.50
22.17
17.31
29.75
28.68
27.11
26.90
32.12
21.22
19.49
31.47
18.12
20.17
24.72

0.43
8.85
0.14
3.07
3.82
1.43
1.43
5.06
1.01
22.47
22.47
6.19
3.18
4.89
11.30

Refuse was shredded, air-dried, and pyrolyzed in a retort at 815
C. From Ref. 22.

Table 4 Yields of Pyrolysis Products from Refuse at

Different Temperatures (Percent by Weight of Refuse
Combustibles)a
Temperature

C

Gases

Liquid
(including water)

Char

480
650
815
925

12.33
18.64
23.69
24.36

61.08
59.18
59.67
58.70

24.71
21.80
17.24
17.67

From Ref. 23.

3.45%, 20.81% for the thermophilic and mesophilic sludges respectively. Table 6 shows a
comparison between the characteristics of the pyrolysis oil in comparison to No. 6 fuel oil.
b. Gaseous Products. The yield, composition, and caloric value of gaseous pyrolysates
depend upon the type of material, the heating rate, and the ultimate temperature. Table 3
shows that for typical refuse components, gas yield ranges from 17% to 31% of the airdried feed. From Table 4 it can be seen that a doubling in gas yield occurs as the ultimate
temperature is raised from 480
to 925
C. Table 5 indicates that the yield decreases, then
increases with heating rate.
As the ultimate temperature increases, more of the combustion energy in the feed
material appears in the gas product, although the data indicate that the gas caloric value
does not show a regular variation (Table 7). Table 8 shows an increase in hydrogen and a

Table 5 Effect of Heating Rate on Pyrolysis Products and Heating Value of the Pyrolysis Gas from
Newspapera
Time taken
to heat
to 815
C
(min)

Gas

Water

Other
liquid

Char
(Ash-free)

Heating value
of gas,
kcal=kg of
newspaper

1
6
10
21
30
40
50
60
71

36.35
27.11
24.80
23.48
24.30
24.15
25.26
29.85
31.10

24.08
27.35
27.41
28.23
27.93
27.13
33.23
30.73
28.28

19.14
25.55
25.70
26.23
24.48
24.75
12.00
9.93
10.67

19.10
18.56
20.66
20.63
21.86
22.54
28.08
28.06
28.52

1136
792
671
607
662
627
739
961
871

Yield, weight percent of air-dried refuse

From Ref. 22.

Table 6 Sewage Sludge Pyrolysis Oil Propertiesa

Property

No. 6
fuel oil

Sludge
pyrolysis
oil

Sulfur content (%)

kcal=kg
Specic gravity
kg=liter
kcal=liter

0.73.5
10,100
0.98
0.985
9,939

0.721.35
1,0209,280
0.790.84
0.7960.846
5,2207,830

From Ref. 345.

Table 7 Caloric Value of Pyrolysis Gases Obtained by Pyrolyzing Refuse

at Different Temperaturesa
Caloric value
Temperature

C
480
650
815
925
a

From Ref. 23.

Gas yield per kg of

refuse combustibles,
m3

kcal=m3 gas

kcal=kg refuse
combustibles

0.118
0.173
0.226
0.211

2670
3346
3061
3124

316
581
692
661

Table 8 Composition of Pyrolysis Gases Obtained by Pyrolyzing Refuse at Different Temperaturesa

Gas composition, % by volume
Temperature

C

H2

CH4

CO

CO2

C2 H4

C 2 H6

480
650
815
925

5.56
16.58
28.55
32.48

12.43
15.91
13.73
10.45

33.50
30.49
34.12
35.25

44.77
31.78
20.59
18.31

0.45
2.18
2.24
2.43

3.03
3.06
0.77
1.07

From Ref. 23.

drop in carbon dioxide as the ultimate temperature increases. Table 5 suggests that the
heating value of cellulosic materials (represented by newsprint) rst decreases, then
increases with increasing heating rate. In general, pyrolysis of 1 kg of typical refuse
combustibles yields from 0.125 to 0:185 m3 of gas having a caloric value of about 3000
kcal=m3 ; about one-third the caloric value of natural gas. Table 9 shows the composition
of the gaseous products from pyrolysis of sewage sludge (345).
c. Solid Products. The solid product or char resulting from pyrolysis is an impure carbon.
The proximate analysis of the char is similar to coal, with the rank of the corresponding
coal increasing as the ultimate pyrolysis temperature increases. Char formed at 480
C is
comparable to certain bituminous coals, whereas pyrolysis at 925
C produces an
anthracite-like product. The yield of char ranges from 17% to 32% of the air-dried, ashfree feed, decreasing with increasing heating rate and ultimate temperature (Table 10). The
caloric value of refuse char is around 6600 kcal=kg of air-dried char and decreases slowly
as the ultimate temperature increases (Table 10). Data for the carbon and sulfur content and
the heating value of the char formed by pyrolysis of sewage sludge (345) are shown in
Table 11.
3.

Thermal Decomposition Kinetics

Although the details of the complex chemistry and heat transfer processes controlling
pyrolysis reactions in wastes (thick sections, often compounded of several materials,
Table 9 Composition of Pyrolytic Gases (Average Percent by Volume)
Temp.
C

Min.

H2

O2

N2

CH4

CO

C2 H6

CO2

C 2 H4

12.7
19.6
16.88
16.55
20.45

3.6
1.27
1.32
0.7
0.25

26.5
15.25
10.56
23.9
12.45

1.9
0.67
1.88
1.4
0.15

7.4
8.4

0.2
0.25

13.3
14.4

0
0.15

Thermophilic sludge
795
849
854
927
1004

105
75165
15165
1545
1545

25.0
31.5
30.6
26.6
30.0

2.2
7.4
6.82
0.9
0.85

18.9
14.3
13.9
15.25
14.8

18.8
12.55
14.1
13.3
12.85

Mesophilic sludge
867
860

3060
3060

13.4
17.95

1.75
1.8

61.2
53.2

3.25
4.65

Table 10 Comparison of Char Produced by Refuse Pyrolysis with Certain Coalsa

Proximate analysis of
air-dried material
Volatile matter %
Fixed carbon %
Ash %
kcal=kg of air-dried material
a

Char resulting from refuse

pyrolysis at different
temperatures

Bituminous coal
from lower
Freeport Seam

480
C

650
C

815
C

925
C

Pennsylvania
anthracite

23.84
65.36
8.61
7880

21.81
70.48
7.71
6730

15.05
70.67
14.28
6820

8.13
79.05
12.82
6410

8.30
77.23
14.47
6330

8.47
76.65
11.50
7565

From Ref. 21.

Table 11 Characteristics of Char from Pyrolysis of Sewage Sludgea

Pyrolysis
temperature
C

Percent
carbon

Percent
sulfur

kcal=kg
dry sludge

kcal=kg
char

626
576
628
547
541
494

1332
1193
1278
1152
1130
1018

253
261

613
633

Thermophilic sludge
593
765
849
854
927
1004

17.0
15.0
16.3
14.7
13.3
13.0

0.82
0.81
0.76
0.70
0.57
0.40

Mesophilic sludge
867
860

7.8
8.1

0.58
0.54

From Ref. 345.

anisotropic and heterogeneous in thermal and chemical properties) are not known,
pyrolysis has been extensively studied for wood and synthetic polymers. Much of the
general body of literature concerned with this topic is found under the heading Fire
Research and derives its support from an interest in the basic processes that start and feed
(or stop) conagrations in manmade and natural combustible materials.
Studies by Kanury (24) using an x-ray technique to monitor density changes during
the pyrolysis of wooden cylinders subjected to convective and radiative heating provide
useful insight into pyrolysis kinetics. As a consequence of the x-ray technique, Kanurys
analysis, summarized below, is based on density change rather than mass.
Based on convention (25, 27), surface pyrolysis of wood may be assumed to follow a
rst-order Arrhenius-type rate law



dr
DE
k1 r  rc exp
17
dt
RT

where
r instantaneous density (g=cm) and subscripts c char, v virgin solid
t time (min)
k1 pre-exponential factor (min1 )
DE activation energy (kcal=kg mol)
R universal gas constant (1:987 cal mol1 K1 )
t Absolute temperature (K) and subscript 0 indicates initial conditions
Kanury plotted r  rc 1 dr=dt versus 1=T on semi-log paper to test the validity
of Eq. (17) and found good agreement with an activation energy (DE) of about
19,000 kcal=kg mol and a pre-exponential frequency factor of about 106 min1 .
Observation of the surface temperature with a microthermocouple indicated an
approximately linear change with time:
T T0 a1 t
Combining Eq. (17) and (18) and integrating yield


z
r  rc
k1 DE
ln
g0 z  g0 z0 

g0 zdz
a1 R
r v  rc
z0

18

19

where
DE
RT
1
g0 z  expz
z
z

Integration of gy z yields

1 1n zn
P
g0 zdz ln z
n1 nn!

20

The right side of Eq. (20) converges for positive z, and the exponential integral of the rstorder g1 z is dened as


 1

1 1n zn
P
g0 zdz
g1 z 0:5772  ln z

n1 nn!
z
Therefore
1
g0 zdz g1 z0  g1 z

21

Substituting Eq. (21) into Eq. (19)


r  rc
k DE
1
g0 z  g0 z0 g1 z0  g1 z
ln
a1 R
rv  rc
Since z0 is usually 30, g0 z0 and g1 z0 are of the order of 1010 and may be neglected,
yielding



r  rc
k DE
 1
g0 z  g1 z
ln
22
a1 R
rv  rc
g0 z and g1 z are tabulated in standard texts.

Kanurys experimental data on the surface densitytime function was in excellent

agreement with a prediction based on the experimentally determined value for a1 of
50 K=min.
Kanurys data also showed two plateaus in the pyrolysis rate: at 100
and 350
C. The
plateau at 100
C is, most likely, associated with supply of the latent heat of evaporation of
moisture in the wood. The plateau at 350
C is suspected to be due to the resistance to mass
ow and diffusion of the pyrolysis gases offered by the char layer. It was noted, in support
of this hypothesis, that the 350
C plateau was absent in pyrolysis of the surface layers or
when major cracks and ssures appeared as the solid near the axis of the cylinder
pyrolyzed.
Kanury also observed that if the pyrolysis rate becomes constant (in the plateau
region), then from Eq. (18):
dr
constant a2 rv  rc
dt
A plot of the data from DE=R 9500 K gave satisfactory agreement for







r  rc
a
DE
DE
2 exp
1:436  107 exp
RT
RT
rv  rc
k1

23

The slope of the plot of the log of the l.h.s. of Eq. (23) versus 1=T K had a slope of unity,
thus providing additional conrmation of the rst-order reaction rate expression.
Another observation was that the pyrolysis rate, after attaining its maximum value,
fell in a manner somewhat like an exponential, namely,



r  rc dr
a3 expa4 t
24
rv  rc dt
Integration yields



r  rc
a
3 expa4 t a5
rv  rc
a4

25

Since, as t becomes large, the exponential tends to zero and r approaches rc , a5 is 0.

Plotting the data for the pyrolysis rate in the inner layers a3 1:404 min1 and
a4 0:216 min1 or



r  rc
6:5 exp0:216t
26
rv  rc
Eliminating the exponential between Eqs. (24) and (26), and substituting into Eq. (17),
yield



DE
 0:216 a constant
27
k1 exp
RT
For k1 106 min1 and DE=R 9500 K, T 619 K or 346
C, the observed plateau
temperature. In summary, then, it would appear that the wood cellulose pyrolysis process
consists rst of an essentially nonreacting thermal heating stage with moisture loss to
about 350
C. Then, pyrolysis occurs at a rapid rate, followed by further heating of the char
residue through conduction and radiation from the exposed surface.
Data by Shivadev and Emmons on thermal degradation of paper (28) also showed a
plateau in a temperature range about 350
C. Their experiments involved radiant heating of

lter paper samples. They deduced reaction rate constants and energy release terms by
radiating preweighed specimens for progressively longer time periods and then determining the total weight loss as a function of time. Temperatures in the specimen were
evaluated by considering the specimens to be thermally thin and thus rapidly attaining an
equilibrium temperature related to the radiative ux level, the absorptivity of the surface,
and material properties, all of which were either known or measured.
Shivardev and Emmons tted their mass-loss (surface density m) and temperature
data with two expressions. The rst, for temperatures below 655 K (382
C) is



dm
26;000
5:9  106 exp
g cm2 sec1
28a
dt
RT
For temperatures above 655 K they found



dm
54;000
1:9  1016 exp
g cm2 sec1
dt
RT

28b

They speculated that the expression for the high-temperature reaction may incorporate
combustion effects of the pyrolysis gases. It is noteworthy here that in associated studies of
ignition, the inection temperature associated with ignition was 680  15 K. The
numerical values obtained are in rough agreement with the Kanury data.
Chiang et al. (445) studied the kinetics of petrochemical sludge pyrolysis. In an
isothermal reactor, sludge was heated with nitrogen to temperatures from 400
to 800
C
for times from 1 to 60 minutes. Pyrolysis products (both solid residues and volatilized
liquids) were analyzed.
The pyrolysis process was characterized using the equation


dx
E
k1  xn A exp  a  1  xn
r
29
dt
RT
where
r reaction rate of pyrolysis
x transformation rate given as
x

W  Wf
W0  Wf

t pyrolysis time (sec)

k Arrhenius constant
W0 sludge weight at t 0
Wf sludge weight at t 1
a frequency factor sec1
Ea activation energy (kcal=kg mol)
R universal gas constant
T absolute temperature (K)
n apparent order of reaction
The composition of sludge and the solid residues are shown in Table 12. The
reaction order and activation energy were 2.5 and 2,640 kcal=kg mol, respectively. The
average Arrhenius constant was 0:0616 sec1 .

Table 12 Elemental Composition of Sludge at Differing Pyrolysis Temperaturesa

Temperature
C

Residue

%C

%O

%H

%N

%S

%Ash

105 (raw sludge)

400
500
600
700
800

18.9
13.8
10.7
9.7
8.2
8.6

39.0
42.4
41.5
40.6
39.3
38.5

24.4
24.4
18.1
11.6
9.17
8.54

5.7
2.79
2.36
1.92
1.62
1.50

4.75
6.38
3.42
2.59
2.00
1.92

1.18
1.87
2.41
1.96
2.45
2.50

24.97
22.16
32.21
41.33
45.46
47.04

Sludge dried at 105
C for 24 hours. Sludge pyrolysis time: 30 minutes.

Source: From (445).
a

Table 13 Kinematic Parameters of Scrap Tire Materials

Material
Processing oils in natural and polybutadiene rubber
Processing oils in styrenebutadiene rubber
Natural rubber
Butadiene rubber
Styrenebutadiene rubber

ln A min1

DE kJ=g mol

7.84
7.56
38.20
34.08
24.02

48.0
43.3
207.0
215.0
148.0

Source: From (448).

Scrap automobile tires were also subjected to pyrolytic conditions using derivative
thermogravimetric (DTG) and thermogravimetric analysis. In their investigations, Kim et
al. (446) studied the pyrolysis kinetics and mechanisms for the compositional components
of two different sections of scrap tire rubbers: sidewall and tread. They found that the
breakdown of each of the compounds comprising the tires followed a irreversible, one-step
decomposition mechanism.
The tires used in their experiments involved tread (comprised of two types of
styrene-butadiene rubber) and sidewalls (a mixture of natural rubber and polybutadiene
rubber). The resulting kinetic constants are shown in Table 13. The decomposition left a
residue approximating 34% of the initial weight: approximately 28% of the carbon black
originally in the tire compounds and not lost in the pyrolysis event, plus about 6% ascribed
to char residues of the thermal decomposition.

B.

Particle Burning Processes

The combustion of solids as nely divided solid particles is most commonly seen in
pulverized coal boilers. Also, however, many agricultural residues (e.g., sawdust) and other
wastes (e.g., dried sewage sludge) are pneumatically conveyed and blown into combustors.
In the combustion of most solid particles, the rate is most strongly inuenced by the
rate of heat transfer to the surface. This leads to surface heating, pyrolytic breakdown and
gasication, and combustion in a diffusion ame surrounding the particle. Surface heating
by radiation from the ame accelerates pyrolysis, nally yielding a residue devoid of

Table 14 Combustion Properties of Carbon and Metals

Fuel

Oxide

Fuel
density
g=cm3

Al
Be
B
Ca
C
C
Hf
La
Li
Mg
Pu
K
K
Si
Na
Th
Ti
U
U
Zr

Al2 O3
BeO
B2 O3
CaO
CO
CO2
HfO2
La2 O
Li2 O3
MgO
PuO2
K2 O
K2 O2
SiO2
Na2 O2
ThO2
TiO2
UO2
U3 O8
ZrO2

2.700
1.840
2.340
1.550
1.50a
1.50a
13.300
6.160
0.530
1.740
19.840
0.860
0.860
2.350
0.970
11.300
4.500
19.050
19.000
6.440

Mfuel
27.000
9.000
10.800
40.100
12.000
12.000
178.500
139.000
6.900
24.300
242.000
39.000
39.000
28.100
23.000
232.000
48.000
238.000
238.500
91.200

Moxide

B:P:fuel


C

B:P:oxide


C

Boxygen

101.900
25.000
69.600
56.100
28.000
44.000
210.600
325.800
29.900
40.300
274.000
94.200
110.200
60.100
78.000
264.100
79.900
270.000
842.000
123.200

2,467
2,970
2,550
1,487
4,827
4,827
5,400
3,469
1,317
1,107
3,235
774
774
2,355
892
4,000
3,260
3,818
3,818
3,578

2,980
3,900
1,860
2,850
(191)
(79)

4,200
1,200
3,600

2,230
657
4,400
2,750
2,500

5,000

1.120
0.564
0.451
2.500
0.750
0.375
5.590
5.790
0.865
1.520
7.560
4.890
2.440
0.879
1.437
7.260
1.497
7.430
5.580
2.850

1.120
0.564
0.451
2.500
0.750
0.375
5.590
5.790
0.865
1.520
7.560
4.890
2.440
0.879
1.437
7.260
1.497
7.430
5.580
2.850

*For 1 micron 104 cm) particles, rg D is assumed as 5  104 g=cm=sec.

a
1.301.70
b
0.580.76
c
1.972.58
d
1.011.32
e
3.784.95
Source: From (435).

Bair

tb  105 sec*
(in O2

tb  105 sec*
(in air)

0.260
0.131
0.105
0.580
0.174
0.087
1.300
1.341
0.200
0.353
1.755
1.134
0.567
0.204
0.334
1.685
0.347
1.726
1.294
0.662

0.900
1.040
1.570
31.000
0.670b
1.160d
1.760
0.810
0.210
0.470
2.310
0.120
0.170
0.930
0.270
1.340
1.230
2.230
2.520
1.200

2.930
3.790
6.150
0.850
2.280c
4.160e
3.980
1.820
0.730
1.450
4.900
0.280
0.480
3.230
0.850
2.870
3.750
4.740
5.730
3.190

volatiles. Oxygen diffuses to the solid surface, releasing heat and further raising the solid
temperature.
Kanury (435) analyzed the process of solid particle combustion to yield an estimate
of the time required to completely burn a particle of initial diameter d0 as
tb

rF d02
8rg D0 lnB 1

29

where
rF density of the particle (g=cm3 )
d0 initial particle diameter (cm)
rg D0 product of density and diffusivity of the surrounding gas (g=cm=sec)
B mass transfer number (dimensionless) f YO1
f normalized oxygen requirement of the waste (g O2 =g waste)
YO1 mass concentration of oxygen in the surrounding gas (0.232 for air)
Values of the key variables are shown in Table 14.
C.

Mass Burning Processes

In the idealized mass burning model described above it was postulated that in thick beds
the upper regions could behave as a true pyrolyzer. Evidence (29, 30) for coal and refuse
beds would tend either to discount the existence of the pyrolysis zone or, more probably, to
suggest that this zone does not appear in beds of practical thickness. Thus, the off-gas from
a bed would be a mixture of gases from both burning and gasication zones.
Studies (30, 377) of the off-gas from beds of burning refuse in a municipal
incinerator have conrmed the earlier data of Kaiser (31) and the hypothesis (21) that
the off-gas composition is controlled by the watergas shift equilibrium. This equilibrium
describes the relative concentration of reactants in the following:
H2 O CO ! CO2 H2
pH O pCO
Kp 2
pCO2 pH2

30

The importance of this equilibrium in mass burning is the incremental gasication

potential given to the underre air. In tests in Newton, Massachusetts (30), refuse and
off-gas stoichiometry were studied. Average refuse was given the mole ratio formula
CH1:585 O0:625 H2 O0:655 . For this formula and assuming that the watergas shift equilibrium holds, over 1.5 times as much refuse can be gasied by a given quantity of air as
would be predicted for stoichiometric combustion to CO2 and H2 O. Burning=gasication
rate data showed rates 1.7 to 2.1 times those corresponding to stoichiometric combustion.
A second result coming from the watergas shift reaction is that a denite and
relatively large combustion air requirement will necessarily be placed on the overre
volume. The air requirement for the CO, H2 , distilled tars, and light hydrocarbons can,
indeed, be as much as 30% to 40% of that expended in gasication, thus creating a need
for effective overre air injection and mixing.