-------
hg =\ heat transfer coefficient, assumed equal to both
gas and solid /
The rate of heat generated by reaction, assuming chemical kinet-
ics control, is,
Qgen £ ndAyD e-E/RTAHn (92)
U2 °2
where AHQ2 is the heat of reaction of oxygen with carbon mon-
oxide per unit weight of 02. For sustained combustion
Qgen > Qloss (93)
or
AyS2AH02e"E/RT - hs(Tf - V + hs(Tf - V (94)
Silver showed that the term on §.he right-hand side of equation
(94) was in fact equal to $YQ N1- and concluded that for a fire
to exist the temperatures 2 had to be great enough for
the chemical kinetics not ~ ITd to be a controlling factor.
This appears to have been thu first quantitative statement
that has been published concerning the phenomenon of combustion
stability in a fuel bed. Silver's development will not be
covered in detail here as a discussion (which is somewhat more
general than this approach) on fuel bed combustion stability
will follow on pages 113-118. Briefly, the development required
that hs be related to 3 using the Reynolds analogy for'the re-
lationship between hs and the friction factor, and that T
could be approximated as a linear function of Ts.
Experimental and Theoretical Work on Refuse Combustion
The Experimental Studies at the Bureau of Mines (26, 27, 28).
The Bureau of Mines pioneered some of the first small-scale
experimental work in incineration about fifteen years ago.
Their work has been concerned primarily with the testing and
development of the vortex incinerator (27, 28), so called be-
cause the combustion air (almost all supplied* above the bed)
is fed tangentially into the overfire region. The main con-
cern of the Bureau of Mines was the complete combustion of
the volatiles distilling off the bed, and thus the major ex-
perimental effort was concentrated in the gas phase overfire
region. From their most recent data (28), it appears that
burning rates of 16 - 30 Ib hr=^"ft~2 of bed area are
achievable with this system when using a standard synthetic
refuse containing 29% moisture,, 56% volatile matter and 12%
fixed carbon. Burning rates at the lower end of this range
-------
I
were those most commonly reported. In these tests a very small
fraction of the total air requirements for combustion (approx-
imately 5%) was supplied as underfire air, but from the com-
position and temperature profiles above the bed it was likely
that a portion of the cold overfire air, still with some oxy-
gen content, spiralled down the side of the unit and then
flowed back up through the center of the bed. The extent to
which this will occur can be determined from a momentum and
buoyancy force balance and is roughly proportional to L1/* and
to Tc 1/2 where T is the cold gas temperature at the
1 ~ xT walls, Th iS the hot gas temperature at the center
of the unit, and L is a characteristic height.
The only work that the Bureau of Mines has done on fuel bed
combustion has been with the equipment shown in Figure 26. The
unit was divided into two sections which were joined together
with a water seal. The bottom section, which contained the
fuel bed, could be weighed throughout the course of combustion
with the scale,, The temperature profiles throughout the bed
were measured with Chromel-AIumel thermocouples inserted through
the insulating walls ct regular intervals above the grate. The
underfire air (sometimes not used) was supplied by a blower
through the grate supporting the fuel bed. The overfire air
was supplied tawgentially front ports at a fixed height above
the bed.^ To perform a test, the two sections were separated;
the bottom section was charged with fuel and the thermo-
couples inserted into the bed. During this time, the top sec-
tion was preheated with the gas burners. When the temperatures
in the top section refractories had stabilized, the two sec-
tions were joiaed together arid the progress of combustion was
monitored by noting the change in weight of the bottom section
and taking gas samples from the stack. No auxiliary fuel was
used in any of the tests.
The fuel used in these experiments was composed of wood chips
(3/4 in. sg-uaife) , newspaper, corrugated carboard (1/4 - 2 in.
strips), and leafy vegetables such as lettuce and cabbage (3/4
in. square), The moisture content of the fuel was adjusted to
either 25 or 50 percent moisture by varying the proportion of
leafy vegetables.
Typical teaperature-tirae profiles from one of their tests are
illustrated In Figure 27, which shows distinct plateaus when
the temperature &t eacl'i thermocouple reached 212°F, correspond-
ing to the v&porisatiori of theroisture. When vaporization was
nearly completed, the temperature rose quickly and the fuel
ignited rapidly. From this type of data the rate of travel of
the evaporation front and the ignition could be calculated; a
typical result is shown in Figure 28, along with the corre-
sponding weight loss. For these calculations, the rate of
travel of the drying front was determined from the position of
the leading edge of the drying front and the rate of travel of
- 90 -
-------
Tangential air -
Gas
Gas
Gas sample
Thermocouples
Gas ports
Tangential airs lot
Hood in heating position
Hood in test position
Fig. 26. Apparatus Used by the U.S. Bureau of Mines
(26) for Studying Combustion Characteristics,*
of Refuse.
- 91 -
-------
1,800
1,6001-
1,400 h~
1,200 I—
j 1,000 \-
OE
§
cc
LJ
Q.
400
200
Thermocouple Distance from
number top, inches
800h-
600 h-
5 8 10
TIME, minutes
Fig. 27.
Bed Thermocouple History, Incineration of Syn-
thetic Feed with 50 percent moisture, total
air rate 61 cfm, roof temperature 1800°Ff (27),
- 92 -
-------
201 1 1 1 r
Travel of evaporation front
10
TEST TIME, min
15
Fig. 28. Travel of Evaporation and Burning Fronts and
Total Weight Loss as Functions of Test Time
in a Typical Bed of Synthetic Refuse Having a
Moisture Content of 50 Percent and a Bed Depth
of 18 Inches (27).
- 93 -
-------
the ignition front from the propagation rate of the 500°F plane
through the bed. From Figure 28, it appears that the burning
rate was at a maximum at the beginning of the experiment and
continuously decreased throughout the duration of the
run. However, this high initial burning rate is not possible
on a physical basis, for the thermal wave could not have pene-
trated very deeply into the bed initially and thus pyrolysis
and "burning could not have occurred to any appreciable extent.
The data show that the distance between the drying and ignition
fronts increased initially and then remained almost constant.
This is roughly consistent with a drying limited combustion
process.
This study provided useful data on a specialized vortex unit
and some valuable qualitative insights into the combustion of a
high moisture content fuel, but the experiments do not lend
themselves to 'quantitative analysis. The major deficiency of
the study was tha inability to measure the quantity of air that
spiralled down the cold walls and up through the bed. The
quantity of air that is drawn through the bed in this manner
will depend in a complex way on the refractory temperatures,
the firing rate of the tangential air, the geometry of the
system, and on the burning rate itself.
Research of £gsejrifoigh et a 1. Essenhigh and his co-workers at
Pennsylvania Sti/Ee University have presented a number of papers
(30-47) in the general incineration area and have attempted in
some of them to develop theoretical equations for design pur-
poses (32, 36, 37., 3J5_, 41 „ 43, 47) . Only those papers and sec-
tions of~~papers which deal with 'the solid bed combustion pro-
cesses will be discussed here? many of the papers are also
concerned with modelling conditions in the overfire region.
Some of the p&p-ars (37_,, 3_3) which dealt with the combustion of
the solid fuel developed a simplified Mayers-type model for the
rate of propagation of the Ignition wave through a fuel bed
simulated by a eat of dowel rods placed vertically in a com-
bustion pot. The analysis encountered the same problems as
Mayers' did in estimating the physical properties of the fuel
bed and was essentially a semi-empirical approach. The authors
concluded, as did Mayers, that radiation played an important
part in the ignrtion process, since the observed ignition
rates were much higher than those predicted from their theory
on the assumption that all of the heat fed forward for ignition
was by conduction alone0 When a contribution for radiation was
added into the tliamal conductivity term using a mean radiating
length of a reasonable magnitude, the theory could be made to
fit the experimental results adequately. The results from such
experiments must be taken as being only qualitative as this
analysis ha® even less merit than Mayers', since the experi-
mental set-up could not be considered a model of any real sys-
tem.
- 94 -
-------
The other papers by Essenhigh et al. that deal with solid bed
combustion attempt to develop theory directed towards predict-
ing burning rates and maximum temperature levels in beds as
well as the effect of inert content on the burning rates.
Shieh and Essenhigh (4_7J tried to unify much of the previous
theoretical work of the group and tested the theory against
experimental data taken on their test incinerator. As this
paper covered nearly all the group's work connected with the
fuel bed combustion processes, it will be the only one dis-
cussed here in detail <,
Shieh and Essenhigh (47) proposed that the burning rate of a
fuel bed per unit area~(F^J could be given by an expression of
the form
FA = (0.75FRCS) rn (1_v) (g.A.M) (95)
where F^g is the relative carbon saturation factor (see pages
75-82), mQ2 the mass fraction of air in the underfire air, 6
the air supply rate per unit, area, V the volatile fraction of
the fuel on a dry-ash-free basis, and A and M the ash and
moisture fractions of the fuel, respectively. The derivation
of this equation was based on the simplified physical model of
an incinerator shown in Figure 29. It was assumed that the
fuel was fed in an over-feed mode and that the material was
fully pyrolyzed (i.,e.- it consisted only of carbon) before it
entered the combustion and cj&sification zone. The rate of sup-
ply of oxygen tp the combustion zone by the underfire air was
mQ-G Ib hr~lft~^0 Assuming all the carbon dioxide formed in
the combustion zone is reduced to carbon monoxide by the time
the gases reach the top of sone I(A) of Figure 29, the carbon
burning rate will be (2M 5ra G where M and M are the
molecular weights of 2 carbon and 2 oxygen, re-
spectively. If the MQ bed is not deep enough to per-
mit complete reduction 2 of carbon dioxideeto carbon
monoxide, the carbon burning rate becomes 0.75m0?GF c Ib hr~^
ft" . TO calculate the total fuel burning rate from rhis car-
bon burning rate, the volatile, moisture and ash content of
the fuel must be considered. If the volatile fraction of the
fuel on a dry-ash-free basis is V, each pound of carbon is
produced from 1 pounds of dry and inert-free fuel; and one
pound of char (1-V} is produced from 1 pounds of as-
fired fuel. (1-V)(1-A-M)
Essenhigh and Shieh related the RCS factor to the combustion
bed depth Lr by the equation
"I
J
T - (96)
Lc -
- 95 -
-------
INCINERATION PROCESS
INCINERATOR
BURN-OUT
SECTION
.OVERFED
IGSJIYIOM
SECTION
-4-
SOLID PYROLYSIS
SECTION
OVERFIRE
SOLID BED
(COMBUSTION &ND GASIFICATION)
SECTION
UNDERFIRE
AIR
n IB)
OVERFIRE
AIR
Fig. 29. Schematic of Test Incinerator Used by
Essenhicjh et al. (47)
- 96 -
-------
where Ln was defined by the authors as being a characteristic
reaction depth and K a semi-empirical constant of approximately
1.10. Equation (96) was derived in an earlier paper by Kuwata,
Kuo and Essenhigh (31) f where the RCS factor was related to
depth assuming that the rates of the two reactions C + 02 ->• 2CO
and C + CC>2 -*• 2CO were controlled by the diffusion of reactants
to the solid char surface and that the diffusivities of the
reactants and the temperature dependence of the diffusivities
were the same. In the case of the carbon-oxygen reaction, there
is a good body of evidence to suggest that it is diffusion con-
trolled under the conditions in a fuel bed; but there is strong
evidence to suggest that the Boudouard reaction is kinetically
controlled under fuel bed conditions unless ash build-up hinders
the reaction. These points have been fully discussed earlier
on pages 75-82. The form of equation (96) is very similar to
the empirical expression suggested by Thring (110), namely
FRCS ' *«-al/D> (">
where A and a are empirical constants, L the fuel bed height at
which the RCS factor is required and D the mean diameter of the
particles in the fuel bed* The RCS factor calculated from
equation (97) with suitable values of A and a increases far
more rapidly with depth than that experimentally observed? as
discussed on pages . Equation (96) would be expected to
have the same deficiency as it is not materially different from
equation (97). The simple exponential dependence of the RCS
factor can only ba obtained using the above assumptions of
Essenhigh and Shieh and, as it is established that the Boudouard
reaction is much slower (about five orders of magnitude) than
the carbon-oxygen reaction at the same temperature and reactant
concentration,, it is not surprising to find that the calculated
RCS factor increases more rapidly with depth than that experi-
mentally observed.
The Kuwata, Kuo and Essenhigh development of the variation of
the RCS factor with fuel bed depth does suggest bounds that can
be placed on the constant K but it provides very little insight
into the selection of LQ, which, although it has some theo-
retical basis (given the assumptions involved), cannot be con-
fidently calculated, from first principles. Although LQ can be
estimated from experimental data, the underlying assumptions of
the theory are unsound and the use of a value of LQ determined
from one set of experimental results in another system would be
unwise. In addition to the fuel bed material balance, the
authors wrote a simple energy balance for a differential element
of the fuel bed and used the equation so developed to predict
the temperature profile within the fuel bed. This energy bal-
ance was decoupled from the mass balance equation because the
RCS factor could not be used to give an indication of the heat
release rate in the fuel bed as a function of height. The
- 97 -
-------
equation used to describe the temperature variation within the
fuel bed was stated as
K
.£ d2T
az
where for zone I
QR .
G - F
-------
major assumptions. First, the authors assumed that all reactions
in this zone were exothermic; this is not strictly correct, as
near the top of this zone carbon monoxide would be produced via
the endothermic Boudouard reaction. Second, no account was
taken of changes in gas composition (given by r) with depth as
r was assumed to be a constant. Obviously the heat release at
different points within the bed was fixed by the local gas com-
position and temperature. The authors have completely neglected
this and have essentially assumed a fixed heat release rate
scaled with depth by the factor .1 exp(-z/LQ) . There is no
physical significance to this L approach. Equation (99) can
be broken down into three parts: (a) 3/4GmQ , which, as shown
before, gives the pounds of carbon carried by the combustion
gases assuming carbon monoxide is the only product of combus-
tion; (b) (l-6.72r)AHc represents the proportion of char which
reacted to give carbon monoxide. The net heat release is then
given by
c
AH = - + (1 - r)AHC (104)
n
Q
where AH is the heat of combustion of carbon to carbon dioxide,
and n is the ratio of the heats of combustion of carbon to car-
bon dioxide and to carbon monoxide (approximately 94/26.4 =
3.55). Part (c) , <5exp(-z/L0) f is ^te factor that scales the
heat release rate L_ with depth; it is highest when
z=0 and decays exponentially with height up to z=Lc. This term
gives only a crude approximation to the heat release as it is
based neither on the appropriate kinetics of the reactions in-
volved, nor on any experimental gas composition profiles within
the bed.
Equation (100) may be considered reasonably valid for the
simplified sequential pyrolysis-combustion system of Figure 29.
However, in the experimental set-up used in this research, con-
siderable quantities of oxygen were known to have spiralled
down into the fuel bed from the overfire region (46) and it
must be concluded that some exothermic reactions dTd take place
in zone I (B) .
The three boundary conditions used are open to serious critic-
ism. The first boundary condition states that the temperature
of the fuel bed at z=0 is the same as the ambient, yet the
heat release as given from equation (99) is at a maximum; it
was physically impossible for the temperature at z=0 to be
held at To in the experimental apparatus employed. Using the
boundary condition that d_T = 0 at z=L forces the temperature
profile to peak at z=L dz and in essence also forces the
profile to fit the experimentally observed form, once a suit-
able temperature at the top of the bed is specified. The tem-
perature at the top of the bed was calculated from the remaining
- 99 -
-------
boundary condition given in equation (103). This equation is
a heat balance taken on a differential section at the top of
the fuel bed; the effect of convective heating of the incom-
ing fuel by the combustion gases has been neglected and no-
where in the paper do the authors say how Qr, the net radiant
heat flux to the top of the bed, was obtained.
The experimental results, against which the above theory was
tested, were obtained on a batch incinerator fully discussed
in another pape^ by Shieh and Essenhigh (46) and Figure 29 will
suffice as a rough schematic of the apparatus. Briefly, the
unit consists of a refractory box of 2-ft-square cross-section
standing about 10 ft high with 9-in. heavy-duty refractory
walls. Computer cards were used as the fuel and were fed from
an aperture in one wall near the top of the unit so that they
fell onto the grate, which was made of a perforated steel
plate. The bed was normally maintained about 10 in. deep with
a fuel feed rate of 120 Ib hr"1 (30 Ib hr-1ft~2). Air was
divided between overfire air and underfire air in varying pro-
portions, with the total air running from deficient to excess.
When overfire air was used, it was injected through two sets
of four nozzles located so that two vortices rotating in
opposite senses were produced above the bed. This produced a
stirred section — zone I (A) — roughly 10 in. to 20 in. above
the top of the fuel bed. Above that, another 50 in. was avail-
able for the (plug flow) burn-out region — zone II(B) . Ex-
perimental data included temperature and oxygen measurements
in the fuel bed; wall temperatures; gas temperatures, and
measurements of CO, CO2, 02 and H20 in the overbed region. The
procedure adopted for comparing theory and experiment was by
way of the temperature profile through the bed. As indicated
earlier, the bed temperatures were found to increase to a peak,
Tmax» at a height above the grate designated by the authors as
Lc. Normalisation of the temperature profile was based on
these two parameters which were obtained both experimentally
and through calculations based on the theory previously dis-
cussed. Table 7 shows the comparison of predicted and experi-
mental values of Tmaxand Lc and Figure 30 shows the theoretical
and experimental normalized temperature through the fuel bed.
The calculations were performed for a value of LQ = 0.5 ft.
No values of the other model parameters Cp, r, BQE, Qp and Q
were given. Tha theoretical value of Lc Was calculated usinf
equations (955 and (96) as FA was known from the rate of fuel
addition to the test apparatus that was required to keep the
bed depth constant during an experiment.
The variation of Lp with underfire air rate, as calculated by
Shieh and Essenhigh and reproduced in Table 7, is unusal. As
the underfire air rate increased (runs 1 to 4) the calculated
value of Lc decreased, as expected from the equations. How-
ever, with the underfire air flow rate still increasing, but
- 100 -
-------
Comparison of Predicted and Experimental Values
of iMmximiurri C~->d Temperature
and Combt'stiO'- Zone 'C-i^n !i-c) in the Solid B«d (47)
. i r ._ . ^ n
Run
No.
1
2
3
4
. 5
6
Air ii'p|.i|y '/jts
Clill 1
Gver-IUriUCT-
tire I fire
Lx.ess
I'UX
''re
Air ',!•:< ;d
fits Air 1 r/'
80
80
80
80
40
40
7 I 40
8 I 40
10 29 1230
55
20
25
30
40
50
60
36 MO!
43 i<>:0
50 1&07
0 ',035
14 998
29
973
43 953
{•"•, geri-
mcntal
900
950
980
iOOO
940
950
960
980
i L ,.!i)
1
H ,
1 Pro- I 1'xpen-
! dieted mentaJ
i j
'
0.67 i) 58
O.t>3 ' 058
(159 ' 058
! 0 56 : 0.58
; 0.90 i 0.60
[076 ' 0 60
j 0.67 j (173
| 060 1 073
-------
1.2
1.0
Predicted Totol
Bed Depth
-------
this time coupled with a step decrease in the overfire air rate,
the calculated value of LC was reported to rise suddenly (the
overfire air rate was not considered in developing either equa-
tion (95) or (96). This effect can be seen by comparing the
values of Lc calculated for runs 4 and 5. With the overfire
air maintained at the higher value, the value of L^ again de-
creased with an increasing underfire air rate. This jump in
the calculated value .of Lc is anomalous, because equations (95)
and (96) do not take into account the effect of the overfire
air flow rate since they constitute only a simple material bal-
ance on the fuel bed; the authors offer no explanation for this
behavior. The experimental values of Lc were found to be
approximately constant at 0.58 ft for the high overfire air
rate and did not depend on the underfire air rate. For the low
overfire air rate the value of Lc increased with underfire air
rate, contrary to the theory presented by the authors.
Another astonishing result emerges from the air supply data of
Table 7. The burning rate of the cards was not increased by
augmenting the underfire air rate. On the assumption that the
composition of the computer cards used was approximately the
same as cellulose, the burning rate for runs 1 through 8 can be
calculated; the calculations show that the burning rate for all
the runs was the same, about 16.5 Ib hr ft"2. This value is
mucn lower than the value of 30 Ib hr~^ft~2, which the authors
quoted as being typical. Previous experience with coal beds
burning in the overfeed mode has shown (23) that the burning
rate should be roughly proportional to the underfire air rate.
Departures from strict proportionality are caused by changes of
the RCS factor; out over a wide range of conditions it was
found that the higher fuel bed temperatures, and therefore the
greater reaction rates associated with increased underfire air
flow rates, compensated for the shorter residence time of the
gases in the fuel bed.
The major reason the theory predicts the opposite effect of
underfire air rate on the value of Lc from that experimentally
observed stems from the fact that the area firing rate, FA,
appears to have remained constant with increasing underfire
air; this forced the value of Lc, as predicted from equations
(95) and (96), to decrease. Physically, this means that a
given value of the RCS factor was reached in a progressively
shorter distance as the underfire air was increased. This is
completely contrary to all other experiences with overfeed fuel
beds, as well as to the authors' own theory. Shieh and
Essenhigh state that the characteristic depth L can be given
as
LQ = v6d/4D0_N (105)
- 103 -
-------
where v is the average gas velocity through the bed, d is the
average pore or channel diameter in the bed, D is the coef-
efficient of oxygen diffusing through nitrogen, and 6 is an
effective diffusion path length or diffusion-layer thickness.
The analysis leading to equation (105) is discussed by Kuwata,
Kuo and Essenhigh (37j and contains, like the development of
equation (96), a number of approximations. However, given
that equation (105) is valid, it would follow that LQ would
increase as the underfire air flow rate increased for the en-
suing reasons. The film thickness 6 can be expressed as (113)
0.5
L
where Dp is the diameter of the fuel particle and M and p the
viscosity and density of the combustion gases. Assuming that
d and Dp do not vary with underfire air rate
L a .JL- v- --£ <107>
0 LP*DJ
Equation (107) indicates that LQ should increase slowly with
the underfire air rate. An offsetting factor is the rise in
temperature associated with the increase of underfire air.
The Schmidt number! _u__ 1 is a weak function of temperature
since y a /T,p a I p D , and D a T1'75; hence
frL 9 j
N = V aT'0-25 (108)
be p D
so that LO a /G . Using the data from runs 1 and 4 on the
maximum bed V T temperature and the amount of underfire air,
LO would be expected, on the basis of equations (107) and
(108), to increase about 70% from run 1 to run 4; thus the
depth required to reach a given value of the RCS factor would
increase, a result compatible with experiment.
In summary, the two problems with the Essenhigh and Shieh
theory rest with the unvarying firing rate FA with underfire
air and the assumed constancy of LQ. It is likely that chan-
neling may have caused the burning rate not to increase with
underfire air. The method used in this study, of feeding the
computer cards onto the fuel bed, and the low pressure drop
grate, almost definitely would lead to bad channeling, espe-
cially near the walls. The problem was probably aggravated by
the use as a fuel of computer cards, which would tend to pack
together in layers inaccessible to the underfire air.
- 104 -
-------
this time coupled with a step decrease in the overfire air rate,
the calculated value of LC was reported to rise suddenly (the
overfire air rate was not considered in developing either equa-
tion (95) or (96). This effect can be seen by comparing the
values of Lc calculated for runs 4 and 5. With the overfire
air maintained at the higher value, the value of L,, again de-
creased with an increasing underfire air rate. This jump in
the calculated value of Lc is anomalous, because equations (95)
and (96) do not take into account the effect of the overfire
air flow rate since they constitute only a simple material bal-
ance on the fuel bed; the authors offer no explanation for this
behavior. The experimental values of Lc were found to be
approximately constant at 0.58 ft for the high overfire air
rate and did riot depend on the underfire air rate. For the low
overfire air rate the value of Lc increased with underfire air
rate, contrary to the theory presented by the authors.
Another astonishing result emerges from the air supply data of
Table 7. The burning rate of the cards was not increased by
augmenting the underfire air rate. On the assumption that the
composition of the computer cards used was approximately the
same as cellulose, the burning rate for runs 1 through 8 can be
calculated; the calculations show that the burning rate for all
the runs was the same, about 16.5 Ib hr ft~2. This value is
mucn lower than the value of 30 Ib hr~^ft~2, which the authors
quoted as being typical. Previous experience with coal beds
burning in the overfeed mode has shown (23) that the burning
rate should be roughly proportional to the underfire air rate.
Departures from strict proportionality are caused by changes of
the RCS factor; out over a wide range of conditions it was
found that the higher fuel bed temperatures, and therefore the
greater reaction rates associated with increased underfire air
flow rates, compensated for the shorter residence time of the
gases in the fuel bed.
The major reason the theory predicts the opposite effect of
underfire air rate on the value of Lc from that experimentally
observed stems from the fact that the area firing rate, FA,
appears to have remained constant with increasing underfire
air; this forced the value of Lc, as predicted from equations
(95) and (96), to decrease. Physically, this means that a
given value of the RCS factor was reached in a progressively
shorter distance as the underfire air was increased. This is
completely contrary to ail other experiences with overfeed fuel
beds, as well as to the authors' own theory- Shieh and
Essenhigh state that the characteristic depth L can be given
as
LQ = v6d/4D0_N (105)
- 103 -
-------
where v is the average gas velocity through the bed, d is the
average pore or channel diameter in the bed, D is the coef-
efficient of oxygen diffusing through nitrogen, and 6 is an
effective diffusion path length or diffusion-layer thickness.
The analysis leading" to equation (105) is discussed by Kuwata,
Kuo and Essenhigh {_3?) and contains, like the development of
equation (96}, a number of approximations. However, given
that equation (105) is valid, it would follow that LQ would
increase as the underfire air flow rate increased for the en-
suing reasons. The film tliickness 6 can be expressed as (113)
-,,0.5
(10€)
where Dp is the diameter of the fuel particle and n and p the
viscosity and density of the combustion gases. Assuming that
d and Dp do not vary with imderfire air rate ;
(107)
Equation (107) indicates that LQ should increase slowly with
the underfice sir rate. Ar, offsetting factor is the rise in
temperature associated with the increase of underfire air.
The Schmidt m'aruberf _j£ 1 is a weak function of temperature
since p a /T~,p a Is P D ? and D a T1*75; hence
i L 9 J
so that Lo a /G . Using the data from runs 1 and 4 on the
maximum bed \? f temperature and the amount of underfire air,
LQ would be expected, on the basis of equations (107) and
(108), to increase about 70% from run 1 to run 4; thus the
depth required to reach a given value of the RCS factor would
increase, a result compatible with experiment.
In summary, the. two problems with the Essenhigh and Shieh
theory rest with the unvarying firing rate F» with underfire
air and the assumed constancy of LQ. It is likely that chan-
neling may have caused the burning rate not to increase with
underfire air. The method used in this study, of feeding the
computer cards onto the fuel bed, and the low pressure drop
grate, almost definitely would lead to bad channeling, espe-
cially near the walls . The problem was probably aggravated by
the use as a fuel of computer cards, which would tend to pack
together in layers inaccessible to the underfire air.
- 104 -
-------
Another point which can be raised in discussing the material
balance equations is the validity of comparing the calculated
value of Lc with the value found from experiment. First, it
will be recalled that the experimental LC was taken as the
depth where the fuel bed temperature peaked; this was also the
point where the oxygen concentration reached its minimum. This
point does not compare physically with the value of LC as cal-
culated in the theory* The experimental value of LC reflects
the degree of back-mixing of the overfire air into the fuel bed
and Che rate of reaction of the oxygen in this air with the
pyrolysis products in zone 1(3) of Figure 29. The amount of
back-mixing will decrease as the underfire air rate is in-
creased and therefore Lc will increase, a conclusion in accord
with the experimental observations. Lc as used in equation
(96) has a different physical significance, namely the depth
of fuel bed required to reach a given RCS factor. As shown
earlier, Lc will be expected to increase as the underfire air
increases, but not for the same reasons that the experimental
values increased. Second, it was assumed that there was sub-
stantially no net hydrogen in the char throughout the entire
depth of the so-called combustion zone. This assumes that char
combustion is the controlling step in the burning process; this
may not be correct as the pyrolysis and drying stages may be
much slower than the combustion stage, being heat transfer
limited in many cases. It is possible that the computer cards
are thin enough for this assumption to hold, but this assump-
tion would need to be abrogated for a fuel with a larger char-
acteristic dimension.
The heat balance results, for the value of Tmax, exhibit the
same anomalous results as those for Lc, as they also show a
dependence on the overfire air rate (compare the calculated
values of Tmax for runs 4 and 5 in Table 7). It is possible
that the overfire air rate has some effect on Qr, the net
radiant heat flux to the top of the bed, but as the authors do
not state how this value was calculated it is impossible to
comment on it. However, if one assumes that Qr is dependent
on the amount of excess air, runs I and 7 and 3 and 8 would
have equivalent heat fluxes to the top of the fuel-bed and
hence equal temperatures TL, since FA is a constant. The tem-
perature at the top of the bed will directly influence Tmax
but as can be seen from Table 7 the values of Tmax for runs 1
and 7 and 3 and 8 are very different. In addition to this
anomalous behavior, the calculated value of Tmax progressively
decreased with increasing underfire air rate (at constant over-
fire air rate). This is in direct contrast to the experimental
values, which increased, in agreement with other experimental
evidence (23).
Table 7 also shows that the experimental values of Tj.- were
affected by the overfire air rate and decreased by 60*C between
- 105 -
-------
runs 4 and 5. It is felt that this reflects the quantity of
air that back-mixed into the fuel bed, which would diminish as
the overfire air was decreased and with increasing underfire
air. The reason that the predicted values of Tmax exhibited
the opposite behavior to the experimental values probably lay
with the unchanging burning rate with underfire air rate. In-
creases in G would on the basis of equation (99) increase the
heat release within the bed (Qr) , but the major effect will be
felt near the grate at z=0 where the boundary condition of
equation (101| forced the temperature to T0. At the same time,
increases in G will increase convective heat losses without
compensation from the convective heating of the fuel^flow, and
thus the predicted Tmax will drop with increases in G. In
short, despits the superficially good agreement between theory
and experiment, as shown in Figure 30, the heat balance has
many shortcomings and it can only be assumed that the theory
was forced to fit experiment.
The authors conclude in their paper that their theory has been
adequately substantiated by experiment and needs minor refine-
ments for it to have a broader range of applicability. The
above discussion has highlighted the shortcomings of the theory
and suggests that the authors" conclusion is perhaps premature.
In addition, it; may be questioned if this theory will be valid
for fuel bed conditions where large diffusional resistances to
heat transfer are encountered within the fuel elements.
The Theory of Jf 0 R. Jliessen, A. F. Sarofim et al. (19) . The
authors have p'resenticT~ir~s imp;li fled global picture of re fuse
pyrolysis and gasification on the assumption that the water
gas shift reaction (CO2 * ^2 + co + H2°) is equilibrated at
the top of the fuel bed. This assumption was based on Kaiser's
data (114) from the Oceanside Incinerator on Long Island, New
York. The assumption of the equilibration of the water gas
shift reaction is often made in calculations on coal gasifica-
tion (115) when temperatures above 2000°F are encountered. The
water gas shift reaction is primarily considered as a hetero-
geneous phenomenon, although it can also take place in the gas
spaces in the fuel bed via a free radical mechanism,
,H0O + H * H_ + OH (109)
iff £
CO + OH t CO- + H (110)
There is no direct experimental evidence concerning the rate of
the water gas shift reaction, but it will obviously depend on
temperature and may well be strongly affected by fuel reactiv-
ity and catalytic effects of the fuel ash. There is, however,
a good body of experimental evidence (115) which suggests that
the reaction is rapidly equilibrated above 2000°F. The authors
- 106 -
-------
assume that the gasification reactions occurring within the
bed produce mainly CO, CC»2, ^2 • an<* H2° witn only small quan-
tities of CH4, tars and soot. Nicholls' (24) results show
that this is a reasonable assumption (see Figure 13) and
Kaiser's data (114) also substantiate this assumption.
The refuse was assumed, on a dry/inert and ash free basis, to
have the composition of cellulose. The composition could then
be represented by C(H2O)n where n had the value 5/6 for dry
cellulose. The following overall material balance, which
assumes no ignition restrictions, was written for the gasifica-
tion of the process
00 + 3.76N0 + xC(H00) ->• aCO + bC00 + cH0 + dH00 + eC + 3.76N0
2 2 2 n 222 2
where x is the number of moles of C(H20)n gasified per mole of
oxygen in the underfire air. Equation (111) contains six un-
knowns. Five relationships between these unknowns can be
found from consideration of the three element balances, an
overall energy balance, and by assuming that the water gas
shift reaction is equilibrated. The energy balance, of course,
introduces one more unknown T, the temperature of the gases
leaving the bed. These relationships are:
Element Balances
Carbon: a •*- b + e = x (112)
Hydrogen: c + d = nx (113)
Oxygen: b + (a+d)/2 = 1 + nx/2 (114)
Water Gas Shift Equilibrium
[C02] [H23 bc
[H20] [COT = da" ~ 'StfGS (115)
The amount of energy released in the bed per Ib mole of oxygen
(QR) is given by the difference between the heats of formation
of the products arid react an ts
products reactant moisture^vaporization
QR = '47,560a + 169,290b + 140,2406* - 320,940(|) - 104,240 (^^Ox
(CO) (C02) (H20) (C6(H20)5) (H20)
(116)
- 107 -
-------
where the last term represents the endothermic vaporization of
the moisture in the fuel exclusive of the 5/6 mole per C atom
that is chemically bound. The energy lost from the bed in the
sensible heat of the off gases (Qp) assuming that the temper-
ature of these gases is 2000°F is given by
(117)
Qn - (aC +bC +cC +dC +3.76C ) (1940) + (18,000) (n - |)
P PCO PC02 PH2 PH20 ?N2 6
where the average heat capacities (Btu mole"1 "F"1) between
60°F and 20QO°F are C = 13.8, C - 8.3, C^ Q - 11.0,
C =7.6 and C =8.2.
P"2
With the rate of heat loss from the bed through side-wall and
radiant-heat losses being QL/ the following balance is obtained
QR = Qp + QL (1U)
which provides the fifth independent equation relating the six
unknowns. The temperature of 2000°F was chosen by Niessen et
al. (19) as it was representative of the data of Nicholls1
(24) and Kaiser's (114) results on the Oceanside Incinerator.
Although a somewhat arbitrary temperature, the 2000°F figure
may not be unreasonable, as the endothermic reactions which
occur after all the oxygen has been consumed may well be ef-
fectively quenched at temperatures below this value. The only
other way the temperatures in the bed could drop below this
level would be through radiative and side-wall losses.
Another assumption had to be made in order for the set of
equations (112) and (118) to be solved. This assumption con-
cerned the amount of char produced per mole of oxygen (e),
which will be a function of the rates of ignition and burning
and as such will depend on the under fire air rate. From the
previous,discussion on pages 46-59, it can be concluded that
at high air rates, e will be zero and will increase as the
air rate is decreased.
Results were calculated for different moisture contents of the
fuel, fraction carbon gasified and heat lost or gained by the
bed. With all other parameters held constant, increases in
moisture content were shown to decrease the amount of CO and
H2 formed; this resulted in a decrease in the percentage of
the total energy release that was given off above the bed (for
- 108 -
-------
0% moisture, 745 of total heat release occurred above the bed,
while for 33% moisture this value dropped to 44%). The same
trend was observed when the amount of heat lost by the bed
increased (all other parameters being held constant). The
explanation of these results is evident when gasification is
considered to proceed via a net exothermic step (reaction of
the cellulose to give char, OC>2 and H2O and the vaporization
of any free moisture) followed by an endothermic step con-
sisting of the reactions C -5- H20 -»• CO + H2 and C + CO2 •*• 2CO.
With the assumption of a fixed final gas temperature the
smaller the quantity of heat that is available for the endo-
thermic reactions, the lower the concentrations of CO and H2
above the bed. The amount of heat available for these re-
actions depends on the moisture content of the fuel and the
heat lost from the bed. Heat losses from the bed will depend
in part on the furnace configuration and would be greater
with a water-walled unit than with a conventional refractory-
lined furnace.
This simple model showed the necessity of overfire air for
the burn-out of the combustibles issuing from the bed and how
the air requirements would vary with changes in the feed mois-
ture content and heat loss froia the bed. Much of the combust-
ible would be expected to be mainly CO and H2, with tar and
soot and 014 only present in any quantity near the inlet of
the furnace where ignition was achieved. The model further
showed that for conditions where the fuel moisture content
increased the burning rate could be kept up if the underfire
air rate was increased; the air helps provide more energy for
water vaporization., This function would obviously be helped
if the air was preheatedo This result concurs with what is
often done in practice and gives a quantitative basis for
a guideline which has been disputed by some incinerator oper-
ators.
The major shortcoming of the model is the simple one-step char
production assumption of equation (111) and the somewhat
arbitrary assumption about the magnitude of e. Obviously in
a fuel bed there will be ss&jor diffusional resistances to
heat transfer in many particles and therefore pyrolysis will
continue long after the ignition front has passed over the
particle. The modification of th,is model, to take into
account events in a real bed, wilfL have to await a better
understanding of the bed burning processes if the model is to
be truly!predictive. The development of a more realistic
model, followed by a set of arbitrary assumptions that would
have to be made, at this stage of our understanding of bed-
burning processes, concerning the varying C/H/0 ratios in the
fuel as a function of combustion extent would negate the attractive
- 109 -
-------
features of the model as it now stands.
It is possible, on an order-of -magnitude basis, to check the
validity of -tfce 2000°F quench temperature by considering tne
rates of the two reactions C + H2O * CO + H? and C + C02 -*
CO? which are the predominant reactions taking place in tne
endothewic *eg&» of the fuel bed. A simple heat balance
on a differential section of the bed, assuming equality ot
the gas and solid temperatures and negligible heat capacity
of the solid phase, gives,
-
where c* is the molar concentration of A(lb mole ft ) and
A0A the*heat of reaction (Btu Ib mole"-1) . Equation (119) can
be conveniently changed to a time base by substituting
t = zpg/G , giving,
St
C p
Pg
(120)
For the reaction C + CO2 -»• 2CO, the data of Wu (111) and
Lewis, Gillllsnd and McBride (116) for electrode carbon of
30 - 40 raesH sis®, as indicated by von Predersdorff and
Elliott (115) i suggest a reaction rate of the form
co
For the reaction C -t- H20 * C02 + H2, the rate and mechanism
of which i& understood less well than the Boudouard reaction,
the rate is given by von Fredersdorff and Elliott (115) for
coconut sheii charcoal as,
-62,300'1
exp - — P _
r = - ,- - — • — j - ^Ai' , A^«; - 9 mole g min atm
" ^ y
6.9 x IOSexp
1 + 0'. 014exp
-4?,60'0
RT j
^5,000 ^
RT
Pco2
PCO + 0-21exp -^r-
g mole
"I (g) (miri) (atm)
i co
1.58 x
"H20
-16 lOexp
(122)
This rate is exclusive of the water gas shift reaction, CO +
H2O * CO2
- 110 -
-------
The rates given in equations (121) and (122) are on the basis
of unit weight of carbon and it is necessary to convert these
rates into rates based on a unit volume of bed for use in
equation (120). There is no exact way of doing this as there
is no experimental evidence given by von Fredersdorff and
Elliott as to the effective surface areas of the different
carbon types used. For simplicity the weight to unit volume
of bed ratio will be taken as (1 - <5)ps although the treat-
ment of the data in this manner may grossly misrepresent the
kinetics under fuel bed conditions. The rate can then be
given by
x 60 x (1-6)
lb mole ft~3hr~1atm~1
(123)
= °-1 atm* PHoO =0.15 atm, 6 = 0.4
ps = 0 lb ft ~ for temperatures of 2500°F and 2000°
F the rates of the two reactions are found to be
Assuming pco ~ Pc02 = §H
and ps = 30 lb ft ~3
dc
co
T=2500°F = 39 lb mole
dc
CO,
dt
dcH
T=2000UF
2 2 lb mole hr ft
(124)
dt
T=2500VF =
1 13
lb mole hr xft
o
dt
T=2000 F
-
^ 42 lb mole hr xft
Equation (124) indicates that the rates of both reactions
drop by more than an order of magnitude from the maximum ex-
pected rate at the end of the exothermic zone to the 2000°F
temperature level. At temperatures a little below 2000°F
the reactions will be slow enough that they can be considered
quenched. Moreover, the above calculations show that at tem-
peratures above 2000°F the steam-carbon reaction may well be
mass transfer controlled.
The rate at which the temperature of the gases falls can be
calculated using equations (120) and (124) . As the steam-
carbon reaction is very fast at 2500.°F, it will be assumed
that it is mass transfer controlled and that the rate of
steam conversion is 40 lb mole hr^ft"3. On this basis, a
rough estimate of the possible rate of change of temperature
is
- Ill -
-------
AT -v, JlJLj[1^02_L-i° x ^6:5°° £ 4500°F sec"1 (125)
It max = ~ ~3bl5D x CT2 x 0.4
This calculation has neglected the heat effects associated
with the solid? but since generally UCpg < GC , the order of
magnitude of (AT/At)i:flax will not be g greatly
affected. This result suggests that the temperature drop in
the endothermic £1300G - 350°F/sec) , confirming that -the rates
of the eadothsrmie reactions at this point in the bed are
indeed quite high.
Ignition anJ Combustion Stability of a Fuel Bed
Background. The problem of ignition and combustion stability
of a fuel "bed can be sonvsniently divided into three parts:
the ignition of the top layers of fuel in the virgin bed by
radiation received from the hot combustion gases &ad the re-
fractory lining of the furnace; the propagation of the igni-
tion wave down into the fuel? and the stability of the fully
ignited bed*. The i'irst of these problems has been extensively
studied in t!ie literature on fire research and flame spread.
Under most operating conditions incipient ignition will cause
little difficulty is an incinerator as long as the underfire
air rate at the front of the grate, where the fuel is ignited,
is kept low enough. The reason for this is covered below.
Simras and Lav i±±^'i have skoun that radiation densities of
the order of 1. - 1.7 cal &.a~* sec"1 (13 - 22 x 103 Btu hr"1
ft~^) are sufficient -'cor the spontaneous ignition of typical
woods (= I 1/2 OM in thickness) containing 20 to 60 percent
weight of moist-ore, ita ignition criterion based solely on a
flux density saust be used with discretion, as other factors
(thermal properties of tho material and size, for example)
can markedly affect the amount of heat that must be trans-
ferred to the ;.;peciriisii to achieve ignition. However, the
ignition of the types of material used by Simms and Law would
probably be acre difficult than that of most objects in refuse
because of their high moisture contents and thickness. Radi-
ant flux densities under siost conditions within an incinerator
would b£ in excess of 50,000 Btu hr"1 ft"2, showing"that in-
cipient ignition Is not likely to be a problem unless the con-
vective cooling of the fuel by the underfire air is excessive.
- 112 -
-------
The remainder of this section will be devoted to studying the
process of ignition through the bed and the stability of the
fully ignited bed. The latter problem will be discussed first,
as it is the leas complicated of the two and more studies have
been addressed to it. These studies have been concerned with
the stability of catalytic packed-bed reactors but the basic
principles developed can be expected to apply to a fuel bed; a
catalytic packed-bed reactor can be considered to be a highly
idealized refuse bed. The literature in this field (e.g.,
143, 144, 145, 146, 147 and references contained therein) has
shown that a multiplicity of steady states can be achieved in
a reactor. Two different physical phenomena are apparently
responsible for this multiplicity of steady states. First,
the fact that the heat must be conducted in the opposite di-
rection to the fluid flow gives rise to multiple steady states
for the reactor system, and second, because reaction takes
place on the surface of or inside a catalyst pellet (fuel
element) each pellet may be in one of two steady states. For
a fuel bed, at least two steady states are immediately recog-
nizable: (a) the virgin fuel bed at ambient temperature, (b)
active combustion?
-------
where it has been assumed that all the heat is generated
within the Solifi end that the rate of heat generation may be
controlled by both mass transfer and kinetic factors. Equa-
tion (126) m&y £»e made dimeasionless by dividing through by
the maximum possible heat generation rate giving
T " (127)
k m_ AHC
m °2
A simple heat balance ? on an element of the fuel bed con-
sidered as a tube with carbon walls for adiabatic steady-
state operation, gives
T - T = M- —-— I IT , - T^l (128)
Xg A0
L '"°2j
where To is the entering air temperature, mQ2 the mass frac-
tion of oxygen in the entering air and Ta
-------
2/3
Tad
Fig. 31. Sketch of Dimensionless Heat Generation and Heat
Lost Curves Q represents heat lost curve for
no radiative losses. Q,.... represents heat lost
curve with radiative losses. Points (•) are
stable points; points ( 0); are unstable).
- 115 -
-------
(101) . Inspection of Figure 31 shows that once a coal par-
ticle has started to bum it is inherently very stable as a
consequence of the Lewis number being so close to unity (N^g
for an 03-N 2 mixture is about 1.15). The stability of the
system increases with the gas temperature (Tg); T~ will, of
course, increase as the gases pass through the bed. As Tg
increases, point A moves downward and the temperature of the
upper stable intersection points moves towards Ta(j. The re-
action can 'be quenched by increasing the gas-flow, which
shifts the heat,generation curves to the right. The critical
gas flow rate (Gc) and temperature (Tc) for this condition to
occur can be found by using equation (130) and the fact that
the slopes of the two carves must be equal at this point/ the
latter condition is given by
kmEexp(E/RTc)
K
k Eexp(E/RT
r\j in v
2 = 2
c
N
2/3
Le
(131)
ad
_ T
where the approximation can be made since Figure 31 indicates
that at Tc the chemical kinetics will be faster than the mass
transfer. Squations (130) and (131) can be used to eliminate
km, giving the following approximate equation for Tc
fP _ fp *T1 ^.
c g c
E(T
,
a
- T )
RN
Le
Tc -
ETad
-------
rp — £
c 2R
4R i 12AHC
E ) I6C
pg
U3I)
in which 12AHC/16C rti + T was quoted to bti of the Order
o ^ 2 9-
of 1720 K. Spalding*9 analysis, although based on a
approach, uses much the same reasoning given above. The author
feels that although Spalding reaches the same general conclu-
sion as above, he overstated the importance of posaibi*
quenching of the reaction by high gas flow rate* as » bMifie-
quence of a poor resumption used in deriving equation (133) ;
the value of T calculated using this equation is too lov.
Employing a value of T calculated from equation (133) Spal-
ding suggested that the minimum diameter particle whioh could
burn in air was 0.256 cm; and as this phenomenon is ftever ob-
served experimentally he suggested that some of the cttettlcal
and physical constants used in deriving this figure iNM?4 pro-
bably wrong. The result derived above, that TC i« of tft» order
of T ,, indicates that the decrease in size is not nearly so
important as Spaldiog'a analysis claims.
The treatment given here covers a very restricted CaSft where
no radiant heat los-ees are taken into account. The affect of
considering radiant heat losses is shown in Figure 31 with
the heat loss curve- Qj°;8s. For this case, PQVJt A 4* the same
as before but point B has moved up to B1 = NEe j 1 * **Md/
h (Tad- Tg) } . The slope of the line, instead Of bein^f
linear, is now proportional to T|. Without resorting to working
out the mathematics, it can be readily deduced that beat losses
by radiation tend to make the fuel bed less stable art* th4 bed
could be quenched with a somewhat lower G than that required
in the absence of radiation heat losses. Radiant h*4fc *nd side-
wall losses account for the quenching of fuel beds at
values of air rate; this behavior cannot be explail»*tf-
on the basis of cOnvective h£at losses.
In concluding this section, it will be shown that the Solid
temperature, in the case where convection heat losses dominate,
is handly affected by the degree of conversion, and i* ftOt
dependent on the gge flow rate G. Since the combustion ttte
is entirely controlled- by ma.ss transfer and assumiftg
- *g '
S
- 117 -
-------
(134)
Le
Eliminating Tg from equation (134) , using equation (128) gives
Ts - T0 = (Tad - V
which shows that (T§-TO) is hardly affected by the degree of
conversion and is of the order (Ta2. However, the heat of re-
action AHCmay £>@ adjusted to compensate for endothermic
effects. An assumption has to be made concerning the func-
tional dependence of the reaction term r§|f on distance and
of the oxygen consumption distance ZQ on the im-
posed gas flow rate (G).
Vortmeyer assutaed that ZQ could be given by an equation de-
veloped by Wicke (122) for mass transfer controlled condi-
tions,
- 118 -
-------
=Q.2 3
Tad
Fig. 32. Variation of Gas Composition, Temperature and
Reaction Rate in the Vicinity of Ignition Front.
After Vortmeyer (119).
- 119 -
-------
0
* A - 0.185
(137)
where A is the distance that the oxygen concentration has
dropped by e factor of e (called "Abklinglange" by Wicke).
The deveiopiJisat by Wicke for A follows Mayer's two-zone
theory discussed on pages 61-75 and results in an infinite
depth being required for complete oxygen consumption. As
shown on pages 75-82, the three-zone theory is required to
predict the correct order of zo, but Vortmeyer (119) implic-
itly corrected for this by assuming ZQ = A.
To obtain a reaction term insensitive to changes in gas flow
rate and oxygen concentration .Vortmeyf retook an average heat
generation rate defined by (r^*)
function
ave
and defined the
f(z) =
eff
°2
ave
ef f
Where
dz
ave
= z in which z
is an
average
value of the oxygen consumption length and is assumed to be
independent of G and mg Introducing the dimensionless
variables *
T - T
9 =
O
Tign T0
(T
ign
P + % >
*g *s
(138)
(T.
v ign
into equation
•<22e
. %
(C) = 0
(139)
The appropriate boundary conditions are,
- 120 -
-------
9 = 0 at ? = -«»
(140)
+ 0 as C - C
where Co is defined as
J
03 eff
(rFV
v O0 'ave
(141)
This is an eigenvalue problem, but,. with the transformation
given above, C is now dependent on G and Dp as well as the
initial mass fraction of oxygen in the air (m§2) . Using the
relationships for K| proposed on pages 25-31, the dependence
of t on G and Dp is found to be very weak [Ca(DpG) ]•
Assuming that m does not change, and that only saall vari-
ations in G and Dp are considered, for a given f(C)
s 00
• • II 4 a 2 clVG / T A n \
GC + UC * X(mJ: ) / r^ _ T t (142)
Pg ps 2 y v ign O'
E
Using the relationships for Ks, equation (142) may be written
approximately as
c ~~
pg
-1.55 (Pa6h
^ A^R?
where A is a constant, dependent upon mn
U2
as a function of
A plot of G +
- 121 ,-
-------
is shown in Figure 33 for the data of P. Nicholls (24) on the
underfeed combustion and ignition of different sizes of high
temperature coke. The expected linear behavior is quite
closely followed, but the effect of particle size is not fully
taken into account by the theory- It should be noticed that
this method of plotting the data is not very, sensitive to
values of U but the deviations from linearity shewn for any
particle sise in Figure 33 can probably be attributed more to
experimental arror than to insensitivity of the plot.
Given the many assumptions that went into the development of
equation (143), the general agreement between theory and ex-
periment shown by Figure 33 is quite good; but a major short-
coming is the inability of the theory to predict the upper
and lower ignition extinction limits. The theory suggests
that there is no upper limit to extinction, for it is possi-
£le to have the ignition aon^ travel in the same direction as
G (i.e., negative values of U) . The only limit would be
reached when the ignition zone was blown out of the reactor
or fuel bedo This type of behavior was noticed by Vortmeyer
(119) in his sxperiraents. This behavior has never been ob-
served in a fuel bed and points to an essential difference
between the two systems.
Vortmeyer's experiments were conducted with 0.4 cm diameter
activated charcoal particles burning in O2-N2 mixtures; the
oxygen concentration of the mixing was kept in the range 5% -
9% to minimize the amount of combustion that occurred. The
particles were contained in a quartz tube which was surrounded
by a heating mantle to keep heat losses from the burning zone
small. The ...particles ^are ignited in an underfeed mode. The
low oxygen concentrations used in this study explain why it
was possible to observe the ignition front traveling with the
gas flow; at nigh gas velocities, there was enough fuel re-
maining in the "burnt™ bed to sustain ignition as the ignition
front reversed direction and traveled back up the bed. This
behavior would not be expected to be observed in a fuel bed
because there would be very little fuel left behind the igni-
tion plane.
An attempt was ma
-------
U
U
60O
550r-
500K
450K
400r-
350r-
300J—
250 r-
200h-
150
100
50
iots are for ' /
^gs*»r -1 -2 /
I/'T^ \ . * 4.4 ib hr ft v'
4O 80
Data for High Temperature
Coke ; P Nicholls (24J
• Dp = 2"-2i"
£i Dp = 17" - 2 "
I ^ I
t60 200 240 280 32O
.1.55
0.45 ( 0.8
Fig. 33. Comparison of Vortmeyer's Q.19) Ignition
Theory with the Experimentar~Data of P.
Nicholls (24).
-------
cannot be described, as it can in laminar flame propagation
theory, by a simple function of temperature; thus the varia-
tion of the rate with distance through the bed must also be
considered. This makes the problem rather intractable, but
an approximate solution can be found; the development follows.
dT
Assuming that -T-* can be represented as f
dz
-dm.
D
i? *-»i ^ 0
« ™_ , and that the only exothermic reaction
is C + 02 •* CO2, then for steady-state conditions,
d1
H = 0 (144)
'g ~" "2
and
dm
* -£ - -*S" (i«>
where
m c*
eff As °2 , Ps
r0o = I—• 1 r and a = ^-2. (146)
~ + pKexp(-E/RT) Pg
T 0
Putting Q => T , m = mh (1-C) gives from equations (144)
max U2 2
and (145)
AHCm° (I-,) ,
°2
- G ^ > = 0 (147)
max I
and
(148)
- 124 -
-------
where
pKexp(-E/RTmaxe)
(149)
Using the boundary condition that
gives
de
-»• 0 as 6 -»• 1 and C -*- 1
(Ua + G)C (6 - 1) + GAH m~ (l-£)/T
t> U«-* IllcLA
12 ?_
K!
(150)
(151)
Equation (151) cannot be integrated directly as
But
d; _ dc; dz
d9 ~ d^ ' ~
(1 - O
(e .
(152)
max
•p
which can be integrated if a mean value of the product KgE is
taXen. Putting,
1 - C
(153)
and
C T
(6 - 1)
(154)
"O.
and taking Kgk at,its maximum value, when 6=1 gives
dw
a?
max
w
(155)
- 125 -
-------
The solution to equation (155) , which passes through w
is given by
w =
+ 1
0,
(156)
Using equation {156) in equation (152) gives, upon integrating
the resulting expression for dC/dO.from O to 1, the following
result for the ignition velocity (U)
u =
G j (Ksk)max
a \ 2
i 6 c
? p
( 9
1 1
/ f R
-------
LJ
a:
O
0
70
60
50
40
3O
20
10
*
Particle Size = l"- 1 y"
Fuel Type : High Temperature Coke
E = 54,000 Btu Ib mole'1 (93)
K . 4.13 x10ft ft hr"1 (93)
km=Mayers Empirical Eq. (90)
A Hc= 520O Btu Ib ' Oxygen
"\ .^Mayers' Theoretical Relationship (90)
\ ^--Present Theory Equation ( HI -156)
Experimental Observation
of P Nicholls ( 24 )
100 200 300 400 500 600
UNDERFIRE AIR RATE , G ( Ib hr ft )
Fig. 34. Comparison of Ignition Theories with Exper-
imental Data of P. Nicholls (24).
- 127 -
-------
implicitly corrected by the use of Mayers' empirical relation-
ships; this forced the reaction zone length to be of the right
order of magnitude. Using a standard mass transfer correla-
tion would, with this two-sone theory, broaden the reaction
zone to such ait extent that only very small ignition veloci-
ties (= 5 Ib hr"1£t~2} would be obtained.
Figure 34 shows that the present theory is not much of an
improvement ovar Mayers' results, but the trend in the correct
direction is encouraging. In addition, all physical and
chemical constant a, except for km, were chosen from what are
known to be representative values; this is unlike Mayers'
theory which used, for exaaaple, a heat transfer coefficient
many times greater than physically reasonable. The curve in
Figure 34 is therefore in no sense a best fit. The present
theory predicts shifts in the ignition rates with changes in
particle diair.etez' iTi the correct direction but of a magnitude
more of the ordar showa by Mayers' results than those ex-
perimentally oosarved. The reason for this is simply that
Mayers' empirical correlations had to be used to calculate,
km. The theory also predicts a much sharper variation of U
with G than experimentally observed, and there is nothing in
the mathematical development which prevents U from becoming
negative. This is, of course, physically impossible in a
fuel bed.
Both Vortmeye.;:8 s (119} theory and the present theory fail to
predict for the fueTTbed (although, in fairness, Vortmeyer's
theory was not developed for this application) the upper and
lower limit to ignition. The lower limit probably depends
on the magnitude of the heat losses from the bed and as such
will depend on the type of apparatus used in the experimental
study. Such h'-itit losses are not taken into account by either
of the theories presented here. The failure to predict the
upper extinction point is, this author believes, a result of
the internal nature of the problem. This has been hinted at
in the discussion or4 pages 46-59 on Nicholls' (24) work. As
the cooling effect of the underfire air increases, it becomes
more difficult to obtain the mean temperature level within
the particle ineeesjsry for ignition to occur. Presumably,
for coke or coal particles,, ignition is quantized to each
particle b^cau.-ja of tha large thermal sink of the particle
and"its good theraal conductivity. Thus it is postulated
that the localized ignition of one spot on the particle is
an inherently unstable condition. For ignition to proceed,
the whole iayar of virgin fuel particles must ignite and this
condition will only be reached when the whole layer reaches a
mean temperature level high enough for this to occur. Once
the cooling effect of the underfire air reaches a level where
- 128 -
-------
this mean temperature cannot be reached, ignition stops. The
ignition plane cannot now be driven backwards and the fuel
that has been ignited simply burns away and eventually dies
out.
For the case of refuse, similar trends would be expected even
though the characteristic dimension of the fuel elements will
vary greatly and thus the amount of heat that is required to
be transferred to each element for it to reach an ignition
temperature will also vary. Those elements (i.e., paper,
cardboard, etc.) which ignite easily will tend to propagate
flame ahead of the main ignition plane. It is likely that the
effect of the thermal sinks for both conductive and radiative
heat transfer surrounding these tongues of flame will cause
them to be damped out and the ignition front will in all like-
lihood propagate relatively evenly. The effect of tumbling
the refuse (which has not been considered in this discussion
up to now, although it commonly occurs in practice) will help
the ignition processes if the tumbling is mild enough to pre-
vent ignited material being plunged into an environment where
it can lose enough heat to colder materials to quench active
burning. The tumbling action can be considered an effective
way of increasing the heat transfer in the bed to the cold
unignited parts from the active burning part.
If, as suggested above, the ignition problem is an in-
ternal one, the simple theories presented here will not be
able to adequately predict ignition behavior. A far more
complex model would have to be developed, taking into account
the thermal history of each layer of particles. In a simplis-
tic way, some progress may be made by considering the unsteady-
state heat transfer into semi-infinite slabs of material with
suitable boundary conditions.
- 129 -
-------
PROCESSES OCCURRING WITHIN A FUEL BED
Introduction
The conditions within the bed of a traveling grate have been
qualitatively determined for coal-firing furnaces (123, 124,
94) . The processes within a refuse bed are markedly more
complex than those in a coal or coke bed because of the higher
moisture and volatile content, as well as the greater hetero-
geneity, of the fuel? the various processes occurring at dif-
ferent positions on the traveling grate of an incinerator have
not been adequately described. The purpose of this section is
to provide a description of these processes as a reference
frame for the discussion of the experimental results which will
be presented on pages 184-261. The discussion given here will
draw heavily on the results of the survey on pages 21-130.
The one area of refuse combustion not covered in the survey on
pages 21-130 is that concerning the role of drying and pyrolysis
of the fueli to date there has been no work in this field of
direct application to incinerator conditions. This section
will therefore begin with a description of the effect of mois-
ture content on the tisKe required for ignition and on the rate
of combustion of the fuel particles.
Effect of Moisture on Ignition and Combustion Times
Outline of Problem Considered. The heat effects associated
with drying are greater than those associated with pyrolysis
and, therefore, for a high moisture content fuel such as ref-
use, the drying heat load will be assumed to be the only one
of importance.. The moisture content of a fuel particle can
obviously affect the time required for the particle to ignite,
and the drying of the particle may provide the limiting step
in the overall combustion process. The drying behavior of a
particular particle will depend on many factors (amount of
moisture, both free and bound? micro- and macroscopic struc-
ture of the particle; its diathermicity; and the heating rate
itself). None of these factors has been studied in enough
detail to permit anything but a crude picture to be established
for a "typical" refuse particle. There will, naturally, be a
wide range of behavior in the drying characteristics of refuse
components; only the two extremes of this range will be dis-
cussed here. These limits correspond to the cases where (1)
the rate of heat transfer to the particle is much higher than
the rate of internal diffusion of water and (2) the rate of
internal diffusion of water is high enough for the surface
temperature to remain in the vicinity of the vaporization
- 130 -
-------
temperature of water (212°F).
For the first case it is assumed that, at the high heating
rates associated with a fuel bed, the solid exhibits drying
characteristics that resemble the "falling rate period" in dry-
ing operations, and that a vaporization plane will retreat into
the material as it is heated. This model, at least for wet
wood irradiated with intensities from 2.0 to 3.1 cal cm"2 sec"1
(2.7 to 4.2 x 104 Btu hr~1ft~2), has been observed to fit ex-
perimental results reasonably well (125, 126). The progress
of the drying front is controlled by the conduction of heat to
the vaporization plane, and other mechanisms of moisture trans-
port (diffusion of bound water through cell walls and the
migration of free water under capillary effects) are slow in
comparison to the rate of travel of the vaporization plane.
The effect of the intertracheid motion of the steam has been
shown by both Williams (125) and Garden (126) to have little
effect on the drying process. This model of the drying be-
havior of the fuel particle reduces mathematically to the so-
called Stefan problem, which despite its conceptual simplicity
is mathematically complex because of the nonlinearity intro-
duced by the boundary condition at the vaporization plane.
For the second extreme of behavior, it is assumed that the in-
ternal diffusion of water is very rapid. In this case, the
drying time can be calculated from the amount of heat that
must be transferred to the particle to supply all the heat for
drying under the assumption that the surface remains at 212°F.
This type of behavior might be exhibited by, for example,
oranges and watermelons.
Development of Drying Models. For the first type of limiting
behavior, the ignition time may be calculated using the fol-
lowing mathematical description of the drying process. The
particle is considered semi-infinite, and initially at the
vaporization temperature of the liquid (T^); all thermal prop-
erties of the material are uniform and constant. The equations
describing the process of vaporization, assuming that the con-
vective flow of the steam has a negligible effect on the tem-
perature distribution within the solid, are then as follows.
Within the dry phase, the temperature is determined by
2 1 9Ts
v TS - -a- air (160)
as
where a^ is the thermal diffusivity of the dry solid. The
required boundary conditions are
- 131 -
-------
(i) T = Tv at x = X, the dry-wet interface
S £>
(ii) -ks
* e
d s
x , x - ps AHv ar
(161)
f"^ rn 1
t • • • » i ^"* *«5 i — V» / rn ^ rn \
(111) 'ks "dx" I x = 0 ~ hs(Ts T»J
d dTs
or (iv) "ks -35F I x = 0 = Q
where Tg is the temperature at the drying front; kg the thermal
conductivity of the dry solid; hs the gas-solid heat transfer
coefficient per unit area of solid surface; T^ the temperature
of the gases flowing past the solid (assumed constant); Q the
net heat flux per unit area of solid surface; and
/ pw \
AH = AH 1 -4 - 1 I , where AH is the latent heat of vaporiza-
v v \ d / v
Us '„
d w
tion of water and ps and ps the dry and wet density of the
solid. The boundary condition at the dry-wet interface is
obtained by a heat balance which equates the rate at which
heat is absorbed at the interphase by vaporization of the water
to the rate at which heat is transferred to the interface by
conduction. The, boundary condition at the surface can be ex-
pressed in two forms, by postulating either radiant heat flux
or convective heating of the solid. Kreith and Romie (127)
and Goodman (128) have presented graphical results of the
solution to the above set of equations. These are shown in
Figures 35 and 36.
A rough check on this model may be made by calculating ignition
times for different moisture contents and comparing them to
the data of Sirams and Law (117) . Using Q = 0.5 cal cm"2 sec"1
(6^6 x 103 Btu hr'^-ft"2) , ~^ pg = 50 Ib ft"3, Cp =0.5 Btu
Ib"1 °F~1, ks = 0,2 Btu hr""1ft~2 <>F-1^ -^3 g^ ^s ignition
temperature suitable for piloted ignition, (T°). = 644°F
gives, from Figure 36, the time required for igASSion of wood
with a moisture content of 20% as 2.5 min. The data of Sirams
and Law show that, for the above conditions, times of between
4-5 min were needed for the piloted ignition of European
- 132 -
-------
CO
UJ
OOQC! 000)
OO
dh2 +
<*s M
Jpa
(kds>2
ilk
T.O-V
o5
01
Fig. 35. Surface Temperature (T °) and Depth of Penetration (x) of the Vapor-
ization Plane in a Semi-infinite Slab for Convective Heating of
Surface by Gas at
Results of Kreith and Romie (127) ,
AH
C (T^ + Too)
ps
-------
I
u>
10.0 CT
Ix*
U
O
L.
O
(SI
if
U
0.1 -
0.01
O.01
0.1
1.0
1O.O
1OO.O
or
Q(t)
1000.0
Fig. 36
Surface Temperature (T ) and Depth of Penetration (X) of the
Vaporization Plane in Semi-infinite Slab for Constant Heat Flux
(Q) Surface Condition. Results of Kreith and Romie (127) and
Goodman (128).
-------
whitewood. For Q = 0.7 cal cm~2 sec"1 (9.4 x 103 Btu hr~1ft~ )
and a moisture content of 60%, the data of Simras and Law indi-
cate piloted ignition times of about 4.5-5 min. Using Figure
36, an ignition time of about 3 min is obtained. The agreement
is quite reasonable considering the approximate nature of the
thermal properties used. The data of Simras and Law show that
the ignition time increased with decreasing heat flux at a rate
greater than that predicted by the theory presented here, which
indicates that the ignition time is proportional to Q"2. The
explanation for this discrepancy probably lies in the increas-
ing importance of convective cooling of the surface of the
material at the lower radiant heat flux densities; this cooling
effect was not considered in the above theory.
The flux level in a fuel bed in the vicinity of the ignition
plane can be readily evaluated using
^ ~ ~ s dx i x = ignition front
E
where Ks can be calculated using equation (13) . With the fol-
lowing approximate values s DD = 0.125 ft, ks * 0.2 Btu hr"1
ft'1 °F~1, 6 = 0.4, e = 1.0, T = 1800°F (2260°R) , K| is found
to be approximately =4.2 Btu hr ft"1 °F~1. The temperature
gradients in the vicinity of the ignition front were found in
this study (see pages 203-208} to be in the range 6-20 x 103 °F
ft"1, while in Nicholis" experiments temperature gradients in
the range 10-30 x 103 °F ft"1 were observed. The flux level
in the vicinity of the ignition zone is therefore expected to
be of the order of 25-100 x 103 Btu hr^ft"2. On this basis,
ignition times from a few seconds (for the high fluxes) to
about half a minute (for the low fluxes) can be expected within
a fuel bed even with particle moisture contents as high as 60%-
70% by weight.
The flux level within the bed will depend on the size of the
particles, the bed voidage and the temperature level of the
bed. Following Rosseland8s (129) treatment of radiation as a
diffusion process
160T3 dT
q = -- jj- g-
— 8
where a is the Stefan Boltzraann constant (0.1713 x 10 Btu
°R~4) and A the absorption coefficient (I/A is equiv-
alent to a mean free path or mean beam length) . Comparing
equations (162) and (163) and using Schotte's relationship,
equation (13) , for K|J gives A = 3 . Thus as the particle
- 135 -
-------
size (D ) or the voidage (6) increases, the mean beam length
(1/X) will increase and the unignited fuel will be exposed to
a greater heat flux as it will "see" more of the hot combus-
tion zone.
r>
The most important dependence of Ks will be on the temperature,
but the results of this study (see pages 204-206)and Nicholls1
(24) data show that the peak temperature at the ignition front
is not strongly affected either by changes in underfire air
rate or by fairly wide variations of bed composition. For fuels
which approximate refuse^ temperatures of the order of 1800°F
are typically observed near the ignition front (26, 27, 47) and
this finding has been substantiated by this study. Evaluating
the thermal conductivity on the basis of a temperature of 1800°F
therefore appears reasonable for incinerator fuel bed conditions.
The next consideration is how the moisture content of the fuel
hinders the combustion rate. This may be estimated as follows.
Assuming a fixed surface temperature (T|) and a linear tempera-
ture distribution in the particle (which will be valid for
high moisture contents), the progress of the drying front can
be readily shown to be given by
Equation (164) shows that the rate of propagation of the vapor-
ization plane decreases as t""^* , as expected, since the poten-
tial for heat transfer CdT/dx) decreases as t"*->. Further as-
suming that the pyrolysis reactions occur instantaneously above
a critical temperature T§, the rate of generation of pyrolysis
products is
(165)
where Wp is the mass of pyrolysis products per unit volume of
fuel. Bamford, Crank and Malan (130) indicated that a pyroly-
sis product generation rate greater than 1.84 Ib hr'^-ft"2 is
necessary for spontaneous combustion to occur. For the con-
ditions in a fuel bed, where some flame is present, this
pyrolysis product generation rate should provide an upper
bound for the rate required to sustain combustion. Inserting
- 136 -
-------
typical values into equation (174) — p^= 50 Ib ft"3, k^~ 0.2
Btu hr^ft"1 °F~1, T°=< 1800°F, TV* 212°F, T°= 600°F, w = 30
-3 — s _i s s P
Ib ft , AHv= 600 Btu Ib (for 60% moisture content) —
gives R = — - - . Using the Bamford, Crank and Malan criter-
P /t
ia for sustaining combustion, the rate of pyrolysis product
generation would fall below the threshold level only after
a period of 0.2 hr. It can be hypothesized, therefore, that
drying may limit the pyrolysis rate to the extent that it
will hinder active flaming. Some further elementary calcul-
ations show that the heat load associated with the movement
of the drying wave will not act as a severe local heat sink,
as the enthalpy required can be readily supplied by convective
transfer from the combustion gas and radiative heat transfer
from surrounding elements.
For elements whose drying characteristics are of the first
limiting type, the time required to dry a particle of known
moisture content and size can be estimated following Tao
(131) . The equations solved in this case are the same as
equations (160) and (161) , except they are written for spher-
ical geometry and for a constant surface temperature, so that
s
becomes
2 3T 3 T 1 3T
and
T° = a at r = r (168)
s o
This model assumes that there are no reactions taking place
within the solid. Wear the ignition zone the reaction of
oxygen on the surface of the particle will have a strong exo-
thermic effect while, when all the oxygen has been depleted,
the reactions C + C02 •*• 2CO and C + H2
-------
U,'
CO
or
U
,">
7
6
5
1
< 3
I I I
I I I
I I I
Extrapolation Based on Hnal
Slope Being That Predicted
/ ~
From Unsteady - State Heat /
//
Conduction in a Sphere
Numerical Solution
of Too (13D
Straight Line Extrapolation"
Based on Last Two Points
Calculated byTao(131) _j
ll
I I
0.4
0.5
0.6
*w
-1
Fig. 37. Drying Times for Spherical Particles of Different Moisture
Content. Numerical results of Tao (131)
-------
where TF is the final dyring time (dimensionless) and y the
dimensionless moisture loading. Tao's results for the rate
of propagation of the drying wave appear to agree (within 5%)
with the analytical solution to equations (160) and (161) for
short times, but there is no known analytical solution that
can be used to check his numerical values of Tp\
As attempt was made in this study to develop an approximate
analytical method for calculating drying times for spherical
particles. The first method follow that of Shih and Chou (132),
who indicated that their iterative integral technique for
spherical geometry produced results very similar to those re-
ported by Tao (131). From these authors' study of the Shih and
Chou method, which was originally proposed by Siegel and Sa-
vino (133), it is concluded that, although the method gave
results which agreed closely (within 5% - 10%) with Tao's cal-
culation for nearly all times in the range 0 < TD < T$, the
method failed at times close to T^. It was found that, in
the limit as TD approached T?, the method gave times, when the
drying front reached a certain radius„ which started to de-
crease as the drying front moved towards the center of th~e~ par-
ticle. This is obviously impossible on a physical basis. Shih
and Chou made no comment about this behavior in their paper.
Furthermore, this author also found that even after four it-
erations had been used (about the maximum that could be ob-
tained without an enormous amount of algebraic manipulation),
the solution gave a linear dependence of the drying time (tp)
on the initial moisture content (y). A linear dependence is
not shown by Tao's calculations and would not be expected on
physical grounds. These authors suggest on the basis of this
study, that the usefulness of this iterative integral technique,
at least in spherical geometry, is limited.
A second approximate integral technique suggested by Goodman
(128) was tried. This technique has been shown to give good
results with the semi-infinite slab with various boundary con-
ditions, but it has not, to these authors' knowledge, been used
before in spherical geometry. The spherical heat conduction
equation (167? was transformed into the linear form using 6 =
Tr, on the suggestion of Goodman (134) that the method worked
best for the. Laplace equation. The results showed that at
zero moisture content,, the method gave the correct dimension-
less drying time of 1/6 (see Figure 37) but that, as the mois-
ture content increased, the drying time increased exponentially
and approached infinite at values of the dimensionless heat
load greater than 5» The only explanation for the failure of
this method that can be offered lies in the different form
of the boundary conditions at the interface, which results
when the transformation 9 = Tr is used, to that form for the
semi-infinite slab given in equation (161). On the basis of
•- 139 -
-------
this study and a brief literature review it appears that at
present there are no approximate analytical techniques suitable
for the Stefan problem in spherical geometry.
Tao's calculations therefore represent the only generalized
method of caluclating drying times for spherical particles.
Using his results, the drying time for a 3-in.-diameter
sphere of wood containing 60% moisture is found to be about
25-30 min. The drying time scales as the radius squared
(r2) so that for a particle of 6 in. diameter containing
60$ moisture the drying time would be greater than a typical
residence time for traveling-grate incinerators, which is
about one hour.
On the basis of the foregoing discussion the following con-
clusions can be drawn for combustible elements which exhibit
the first type of drying behavior (i.e., where the rate of
heat transfer to the particle is much higher than the rate of
internal diffusion of water). The presence of moisture in-
creases the time required for ignition over that required
for a dry element; however, at the flux levels typically
obtained within a fuel bed, the ignition delay is not expec-
ted to be more than about half a minute even for particles
containing 60-70% by weight of water. From Tao's generalized
calculation of drying times, active drying and pyrolysis will,
for large particles, continue for extended periods after the
particles have been ignited. Using equation (165) and typical
values of the physical properties of wood, it may be hypo-
thesized that the drying rate may, with high moisture content
fuel elements, provide the rate-limiting step for combustion.
Finally, the heat load associated with the propagation of the
drying wave is not expected to act as a severe local heat
sink, and therefore the drying process will not affect the
combustion behavior of the drying particles' neighbors.
For the other case of limiting behavior, where the internal
diffusion of moisture is high, the calculation of an ignition
time is somewhat more complex because of the effect of the
local heat sink on the fuel bed surrounding the particle.
A simple calculation shows that, at the heat transfer rates
expected in a fuel bed, particles of up to a few inches (say,
the size of an orange) would be expected to dry in times well
under one minute. However, it is common experience that ma-
terials of this type can remain uncharred in a fuel bed for
much longer periods. The explanation lies in the
strong local heat sink created by this type of particle because
of its low surface temperature. It can easily be shown that
the strength of this heat sink can be expected to be far great-
er than any possible local heat generation (except very close
to the ignition front), and it can therefore be concluded
- 140 -
-------
that these particles essentially quench all the surrounding
materials (see pages 113-118 and the discussion of radiant
heat loss on ignition stability). Fuel elements of this
type are dried mainly by the convective heating of the hot
gases flowing past them. The drying time, based on the con-
vective heating of the particle with gases at 2000°F,' is about
10-15 min. for a particle the size of an orange.
The conclusion that, for refuse materials which exhibit this
type of drying behavior, the fuel bed is locally quenched is
a strong argument for effecting some tumbling action on the
grate so that these particles will be exposed to the flame and
refractory above the bed. For this type of refuse particle,
the drying time plays the dominant role in the overall com-
bustion time, as the combustion of the fully dried material
(which will only be a small fraction of the original weight
of the particle) will be a very small fraction of the total
"combustion" time.
Description of Fuel Bed Conditions on a Traveling Grate
Ignition of the Fuel Bed. Traveling grates are used in many
incinerators, and therefore the ensuing discussion will focus
mainly on conditions in the fuel bed of such a system. Des-
pite the heterogeneity of the bed and the different drying
and combustion characteristics of the various fuel elements,
it is convenient forpurposes of discussion to subdivide the
burning fuel bed on the traveling grate into well-defined
zones as shown in Figure 38.
A combustible element in a fuel bed may receive the thermal
energy required for drying, pyrolysis and ignition via a
number of different mechanisms. In the overfeed bed the ther-
mal energy is supplied by radiation from the overfire region
(from both the hot combustion gases and the refractory walls)
by the convective heating of the combustion gases flowing up
through the fuel bed and by radiative heating from the com-
bustion zone of the fuel bed. In the underfeed be<3 the major
portion of the thermal energy is supplied via radiation from
the combustion zone directly behind the ignition front, while
the remainder is supplied by conduction through the fuel (97).
In the case of the traveling-grate stoker the fuel at the top
of the bed near the feed end of the grate is heated solely by
radiation from the overfire region. Once the ignition plane
progresses down into the fuel bed, as shown in Figure l(c),
the thermal energy
- 141 -
-------
RAW
REFUSE-t—
UNDERFEED
— BURNING -
I i
CHANGE
'OVER '
OVERFEED
I
I
•EVOLUTION OF COMBUSTIBLE—H
Drying
Pyrolysis
Ignition
Zero Free O2
ttt
Underfire Air
I
Discharge End
Fig. 38. Simplified Schematic of Processes Occurring in the Fuel
Bed on a Traveling Grate
-------
required is transferred to the virgin fuel in a manner anal-
ogous to the underfeed case.
As discussed on pages 112-113 and pages 118-129 , it is im-
portant to keep the underfire air rate low at the entrance
to the traveling-grate stoker in order to insure that igni-
tion is achieved for the larger particles and for those with
high moisture contents. Subsequently, as the ignition plane
propagates through the bed, it is again necessary to supply
the right amount of air through the grate so as not to hin-
der the progress of the ignition wave. Ideally the under-
fire air should be set at the value that gives the maximum
possible ignition rate. Practically this is impossible, be-
cause not enough is known about .the processes of ignition
propagation for this value to be determined on an a priori
basis. Field tests would be the only method of selecting
the correct underfire air rate. Qualitatively it can be ex-
pected that the ignition rate will be a function of under-
fire air preheat (24) and supply rate (£4_) , particle size
(24_) , fuel type (23T and moisture content (26,27). For re-
fuse, Kaiser's (iXT) tests on the Oceans ideTncTnerator in-
dicated that the ignition rate varied from 0.3 ft/jnin for wet
material to 0.5 ft/min for average material. The Ignition
rate would be expected to decrease as the moisture content of
the fuel increased since there will be a decrease in the to-
tal heat liberated within the bed that is available for heat-
ing up fresh fuel (see pages 118-129 ) . Present theories -
do not give an indication of how the underfire air rate should
be varied to take into account the different fuel moisture
contents; it would seem reasonable to expect that lower air
rates might be necessary with increasing fuel moisture con-
tents. Finally, the hindrance of ignition by having too little
underfire air may be of little practical importance; since
in reality the underfire air is likely to be augmented by
air induced through the bed by temperature gradients between
the center of the furnace and the "cold" walls.
Combustion of the Fuel Bed. After the ignition plane has
passed over a fuel particle, the course of combustion will be
dictated by the nature of the particle and its moisture con-
tent. Two types of limiting behavior, controlled by the dry-
ing characteristics of the fuel, were discussed on pages 131-
141. When the internal rate of diffusion of moisture is low,
the surface temperature of the particle will increase slowly
until ignition is achieved. Depending on the particle's size
and moisutre content, this may take place over a period of
a few minutes after the main ignition front has passed by.
After the surface reaches an ignition temperature, the par-
ticle will continue to pyrolyze and burn until either all the
- 143 -
-------
oxygen surrounding the element is consumed or the rate of
generation of pyrolysis products falls below the level re-
quired to sustain ignition; the latter condition would be
achieved if the heat supplied to the fuel element was not
great enough to drive the vaporization plane at a rate suf-
ficient to supply the required amount of dry pyrolyzable
material. This point was discussed on pages 131-141 . The
surface of a fuel element, shortly after the ignition plane
has passed it, will begin to char; it will then be oxidized
by any available 0- or will react with C02 and H20, accord-
ing to the reactions C + CO- -»• 2CO and C + H2O •> H- + CO.
Only the very small particles will burn out completely at this
stage, and the larger ones will still be drying and pyroly-
zing when all oxygen in the underfire air is consumed.
For fuel elements in which the rate of diffusion of moisture
is rapid, the heat sink will cause, as discussed on pages
131-141 , local quenching of ignition and combustion; these
particles and those surrounding them will then remain in this
quenched state until all the moisture from the fuel element
has been removed and there is enough oxygen and heat supplied
to the element for it to ignite and burn. If the residence
time in the furnace is not long enough, these elements will
be discharged incompletely combusted and may cause some of
the surrounding particles to be discharged in a similar state.
The important reactions occurring within the fuel bed, ex-
cluding pyrolysis reactions and the combustion of tarry py-
rolysis products, are given in Table 8 along with their heats
of reaction. The equilibrium constants of these reactions
are plotted as a function of temperature in Figure 39 and their
relative reaction rates are compared in Table 9. Table 8
shows that, except for the almost neutral water gas shift
reaction (vi) and the slow and thermodynamically unfavora-
ble reaction (vii)„ the only exothermic reactions taking place
within the bed will consist of the reaction of oxygen with
char and pyrolysis products. The very high rate of reaction
of oxygen with both char and pyrolysis products, as indicated
by Table 9 suggests that all the oxygen in the underfire air
will be quickly consumed within a small zone just behind
the ignition front; this behavior was observed in the coal
bed studies of Nicholls (24) and Kreisinger et al. (23) —
see pages 9 - 14 and pages 18-21 . The heat released with-
in this zone provides the only source of energy within the
fuel bed to dry and pyrolyze the fuel and to sustain ignition.
Active burning will take place in a zone de-
termined by the rate of consumption of underfire oxygen in
both volatile and char combustion. Burning rates in coal
beds were found by Nicholls (24) to be approximately propor-
tional to underfire air supply rate, unless the burning rate
- 144 -
-------
TABLE 8
Common Fuel Bed Reactions and Their Heats of Reaction (135)
Reaction
(i) CffU + °? •*• C05
(*' (g) (g)
(ii) C + ^0 + CO
(6) 2 2(g) (g)
l
(iii) CO , N + -p. •* CO
(g) 2 2(g) 2(g)
(iv) C, . + CO., •* 2 CO, v
(6) 2 (g)
(g)
(v) C, . + HO. . -* CO, . + II
(6) 2 (g) (g) 2
(vi) CO, . + HO. . ->• CO^ + H_
(g) 2 (g) 2(g) 2(g)
(vii) c/0. +2 H -*" CH
(6) 2(g) 4(g)
(viii) H +|o - no
(g) (g)
(ix) en +20,, •*• con + 2 ii o
42 22
(x) C II 0 + 2 On -»• 2 H 0_ N + 2 CO,
Standard Heat of
Reaction* (Btu Ib
mole"1 x 10~3)
-169.29
-47.55
-121.74
74.19
56.48
-17-71
-32.20
-104.04
-345.17
-358.16
(g) (g)
(acetic acid)
2(g) 2
(g)
* Negative sign indicates exothermicity. 3-qraph.itG assumed to have
zero heat of formation; various types of amorphous carbon arc ronorted
to have positive heats of formation ranging from 3.06 to 4.68 Rtu Ib
mole"1 (136)
- 145 -
-------
TEMPERATURE (° F)
2600 2000 1600
1200 100O
3.0
Fin. 39.
70
4.0 5.0 6.0
RECIPROCAL TEMPERATURE (
Equilibrium Constants of Common Fuel Bed Reactions
(92) . (All compounds as gases; carbon as 8-
graphite.)
- 146 -
-------
TABLE 9
Relative Rates, of Common Fuel Bed Reactions
Reaction
Approximate Relative Rate at 1800°F
and 0.1 Atm. Reactant Pressure
C + O •* CO
10"
C + HO -+ CO + H
3-10
C + CO •* 2 CO
C + 2 H •* CH
10
-3
- 147 -
-------
was restricted by the ignition rate (see pages 46 - 59 )•
Roughly the same behavior can be expected in a refuse bed
if the model of Niessen et a_l. (1!9) is correct (see pages
106-112 )• Beyond the point of oxygen depletion, additional
release of volatiles will occur from the larger particles
and the char will be gasified by the C02 and H20 rising from
the burning zone. Up to the point where the ignition wave
reaches the grate, the burning action will be of the unres-
tricted underfeed type. After the ignition plane has reached
the grate, combustion of the residual char will be limited
by the supply of oxygen; the CO- released on combustion in
this region will partially react with the remaining char to
yield CO.
Burning rates may be estimated from the results of Niessen
et al. (19) or by considering that the reactions C + C02 •*
30, C + H70 -*• H_ + CO and C + 2H_ -»• CH. are all fast enough
for them to equilibrate by the time the top of the fuel bed
is reached. As discussed below, this concept of total equi-
librium is a poor one for a refuse bed, as the energy require-
ments are too great^ and fuel bed depths too shallow, for these
reactions to be equilibrated.
Figure 40 shows equilibrium gas concentrations, for differ-
ent fuel compositions, calculated under the assumption that
the above three reactions are equilibrated. These gasifica-
tion calculations were performed using values of the equil-
ibrium constants at chosen temperatures; hence, no energy
constraints were imposed upon the system. For typical fuel
bed temperatures (1600 F-2000 F), it is common to observe
the following ranges of gas compositions at the top of the
bed (assuming no bypass oxygen): CO, 10%-20%; CO-, 10%-20%;
H2, 5%-10%; H20, 10%-15%; and traces of methane and pyroly-
sis products. These values are quite different from those
expected on the basis of total equilibration, where the equi-
librium constants for the reactions indicate that mainly CO
and H2 should be observed at temperatures above 1600 F. This
discrepancy can readily be explained since, for the complete
conversion of C02 and H20 to CO and H- more energy would be
required than coGld be Supplied by the exothermic reactions
taking place within the bed, and therefore in an incinerator
the degree of approach to equilibrium will be constrained
by energy considerations (see pages 106-112).
Figure 41 shows the amount of fuel gasified per Ib mole of
underfire air as a function of fuel bed temperature and fuel
composition on the assumption of total equilibration and no
energy constraints. The caluclated gasification rates (at
typical refuse bed temperatures) are very much higher than
those observed in practice (about 60 Ib hr ft ), reflecting
- 148 -
-------
I 1 1
CH4«0.012 at 5OO*F
8OO 1OOO 12OO MOO 16OO 18OO 2OOO 220O
eoo 1000 1200 1400 ieoo woo 2000 2200
80O XXX) 12OO 14OO 16OO 18OO 2OOO 22OO
TEMPERATURE CF )
Fig. 40. Equilibrium Gas Compositions at Various Assumed
Fuel Bed Temperatures. Different fuels gasified
with air at 1 atm pressure: (a) pure cellulose,
(b) Kaiser's (49) approximate composition (dry),
(c) Kaiser's (4"!T) approximate composition
(15% moisture).
- 149 -
-------
12O -
ct
0:
U-
CE
u
o
z
z>
u
-J
O
CD
a
u
a.
a
LJ
1OO -
ui
;D
u
to
CO
20
O
8OO
1200 16OO 2CXX)
TEMPERATURE (°F)
2400
Fig, 41. Variation of Burning Rates with Temperature for
Different Fuels. 1 - dry cellulose; 2 - Kaiser's
composite, dry ash sulphur and nitrogen free;
3 Kaiser's composite, 15% moisture; 4 - Kai-
ser's composite. 30% moisture.
- 150 -
-------
the greater "saturation" of the underfire air with respect
to carbon and hydrogen when enough energy can be supplied
to the bed so that H_ and CO are the only products of com-
bustion. Figure 41 shows an interesting characteristic of a
fuel such as refuse, which contains relatively high amounts
of oxygen and hydrogen. As the moisture content of the fuel
is increased, more of the oxygen required to convert the fuel
carbon to CO can be supplied by the fuel and less is needed
from the underfire air. Figure 41 shows that for Kaiser's
composite refuse (see pages 9-14 ) with 30% moisture,
the gasification rate becomes infinite as the tem-
perature increases, so favoring only H~ and CO as products;
obviously, under these conditions, all the energy require-
ments would have to be supplied from an external source.
This qualitative description of the four major processes —
heating, drying, pyrolysis, and the gas reactions of carbon
and carbonaceous materials —- occuring within a fuel bed will
provide a framework for the discussion in subsequent sections
of this author's experimental work.
- 151 -
-------
EXPERIMENTAL EQUIPMENT AND PROCEDURE
Choice of Experimental Equipment
Brief Outline of Equipment Selected. The experimental equip-
ment chosen for this study was an extension of the underfeed
combustion pot equipment used by the early workers [e.g.,
Kreisinger et eO_. (23) and Nicholls (24) ] in the field of
solid fuel Bed combustion. The simple combustion pot was mod-
ified so that the radiant heat flux condition which would be
experienced by the top of the fuel bed in a traveling grate
unit could be simulated under these batch conditions and so
that burning rates could be determined by measuring the
weight loss of the fuel bed as a function of time. In prin-
ciple an instantaneous fuel bed burning rate can be calcul-
ated if it is possible to close a complete material balance
around the unit at any particular time. However, the cum-
ulative errors associated with this approach are greater
than those connected with a weighing technique and the latter
method is therefore preferable.
The modification adoped here follows the approach used by
Weintraub et al. (26), in which the apparatus is constructed
in two sections, tHe overfire combustion region, or top sec-
tion, and the fuel bed section, or bottom section. Overfire
air was supplied at a number of different positions in the
top section, while the underfire air was fed through a grate
which supported the fuel bed in the bottom section. The de-
sign allowed the top section to be isolated from the bottom
section and then preheated with gas burners. When the tem-
perature gradients in the top section refractories had stab-
ilized, the two sections were joined together and the fuel bed
was ignited by exposure to the radiant heat flux from the hot
brickwork of the top section. The initial radiant heat flux
density used to ignite the fuel was controlled by the final
temperature level of the refractories in the top section at
the end of the preheat stage. During a run the radiant heat
tiux to the top of the fuel bed could be roughly controlled
by adjusting the overfire air flow rate, which in turn changed
the temperature level of the refractories. The course of
combustion was followed as a function of time by measuring
the change in weight of the bottom section.
This type of apparatus provided a satisfactory model of the
conditions in the fuel bed on a traveling grate, in that time
- 152 -
-------
in this batch-type experiment was equivalent to position on
a traveling grate (94, 137, 138). The major criterion that
needed to be satisfied for adequancy of this model was the
similarity of the fuels in the two systems. This analogy has
had certain limitations which are discussed below.
Discussion on Limitations of Batch Systems. In the moving-
grate system a continuous belt carries the burning bed in one
direction. The bed is ignited at the feed end of the grate
by the absorption of radiant heat and the ignition zone grad-
ually penetrates the bed until it reaches the grate further
towards the discharge end of the grate. Referring back to
Figure l(c), it will be recalled that H is the length qf the
fuel bed burning dn what Nicholls (24) called "the unrestric-
ted ignition underfeed principle," and 0 is the lengthjof the
fuel bed burning in a manner similar to the overfeed principle,
It should again be stressed that length 0 of the fuel bed is
burning only in a way similar to the overfeed principle. The
burning rate in this section will increase roughly in propor-
tion to the quantity of underfire air, but there will be no
pyrolysis products issuing from the top of the bed, unlike
a continuous overfeed operation. The actual process of com-
bustion can be further complicated since the bed is in many
cases agitated by a reciprocating grate. This has the advan-
tageous effect of mixing and breaking up agglomerates in the
bed, but the back-mixing can also lead to increased losses
in unburnt material to the ash pit and in increased fly-ash
carry-over.
The development of methematical and experimental models of the
combustion processes within the fuel bed can proceed from the
stand-point of an external or internal observer. From any
external position, the conditions on a grate would appear
to be steady-state, resulting in a steady-state, two-dim-
ensional (length and depth) problem. The justification for
only considering two of the spatial directions in any math-
ematical representation lies in the fact that many traveling
grates are wider than 6 feet and therefore property variations
across the width of the grate can be ignored. From any fixed
position on the grate the conditions in the fuel bed would
be changing with time, and, if one were to model a small
section of this fuel bed as it traveled across the grate at
the prevailing grate speed, an unsteady-state, one-dimensional
(depth) problem would result. The property variations across
the width of the grate can also be ignored in this formulation
for the same reason as stated above. In this model, the time
coordinate is proportional to distance traveled across the
grate, the proportionality constant being the grate speed.
These considerations indicate that a small batch incinerator
- 153 -
-------
with no property variations in either the length or breadth
(or radial or angular) directions (i.e., one-dimensional in
depth) could be used to simulate a small section of the bed
as it travels across the length of the incinerator. A cri-
ticism of this formulation is that the inclination of the
ignition zone in the traveling grate creates both horizontal
and vertical components for the ignition wave propagation
velocity as indicated in Figure 42. The angle of inclination
of the ignition zone to the grate will be a function of both
the grate velocity, V , the underfeed ignition propagation
velocity, V , and the crossfeed ignition propagation velocity,
V . The relationship being
tan 3 = VN/(VG - V ) (168)
or
tan 6 = V../V- for V_ » V_ (169)
N la G P
where 3 is the angle of inclination of the ignition front to
the grate. Typical values of VN and Vp (4 in/min) and V
(1 ft/rain) indicate that 3 will in general be a small angle
(< 20 ) and suggest that under normal operating conditions
a small segment of the bed, taken perpendicular to the grate,
could be adequately treated as being one-dimensional in depth.
Further complications with this model formulation ensue, since,
during the course of conbustion, the resistance of the bed
varies and as a result, on a traveling grate, there will be
a tendency for greater underfire air flow through the portions
of the bed with low resistance (i.e., near the discharge end).
This problem is partly circumvented in practice by dividing
the wind box into compartments and controlling the flow of
air to each compartment. In the batch system this effect
can be simulated by varying the underfire air as a function
of time. A drawback with the batch mode of operation is that,
as the fuel bed burns down, cold walls are constantly being
uncovered in the bottom section and thus the isotherms in this
type of bed cannot be expected to follow those in a traveling
grate. This cooling effect can be minimized in the experimen-
tal model by carefully insulating the bottom section, or by
providing a movable grate which can be used to hold the ig-
nition front at the same relative position in the apparatus.
The latter expedient is difficult to accomplish experimentally
and all investigators using the combustion pot type of appa-
ratus have opted for the former approach. A somewhat simpli-
fied analysis of the cooling effect expected when using care-
fully insulated side walls is given in Appendix B and shows
- 154 -
-------
IGNITION
FRONT
UNBURNT FUEL
BURNING FUEL
en
en
UNDER FIRE AIR
Fig. 42. Diagrammatic Representation of Direction of Ignition Wave
Propagation Relative to Directions of Grate Travel and
Underfire Air Flow. V = ignition velocity parallel to
grate; VN = ignition velocity normal to grate; V = result-
ing ignition velocity; V = grate velocity; = angle of
ignition front to grate
-------
that the heat losses are small compared with the heat gener-
ated during combustion.
Up to this point the discussion has focused on the modeling
of the fuel bad itself, but it is also .necessary to consider
the similarities and dissimilarities in the processes occur-
ring in the overfire region of the continuous and batch systems
and any advantages or disadvantages that the batch system
might have. Generally, the flue gas compositions in batch
experiments would not be expected to compare with those found
in continuous units, unless a period of "pseudo steady-state"
operation were achieved. In fact, the gas composition from a
continuous unit would more closely agree with the gas compos-
ition found from a batch experiment if all the flue gases
from such an experiment were collected and mixed together.
Further difficulties are also apparent, since it would be
very unlikely that the degree of mixing of the overfire air and
combustion products would be alike in the two types of ex-
periments, as the geometry of the system plays an important
role in this process. The gas temperatures in the batch sys-
tem would tend to be lower than those in a continuous system,
since the latter would be physically larger and heat losses
less signifleant. Consequently, it is impossible to match
weight-averaged temperature histories of the gas in the dif-
ferent apparatus, and thus difficult to develop quantitative
theories which would be equally applicable to both. However,
as serious as these shortcomings appear, the batch system pro-
vides a very convenient and compact way of studying the qual-
itative effects of overfire mixing and its effect on the burn-
out of soot particles, smoke and carbon monoxide. Further,
in a batch system,, the changes in composition of the flue
gases may be studied during the different burning regimes of
the fuel bed, and effects detected which may well be dis-
guised by gross mixing in a continuous unit.
The above discussion has shown that many of the shortcomings
of this approach at simulating a traveling grate can be over-
come by careful design of the experiment. The advantages of
using a batch system over a scaled-down traveling grate far
override the disadvantages for the following reasons:
(a) The distribution of air, both above and below the
bed, can be closely controlled in a small experi-
ment by the use of a high-pressure drop grid below
the bed and by careful selection of jet size and
air momentum above the bed.
(b) A unidimensional system is approached so that a
fairly complete map of temperature and composition
distribution may be obtained both within and above
- 156 -
-------
the fuel bed. The problem of obtaining gas and tem-
perature profiles is greatly reduced.
(c) Instantaneous values of the rate of incineration may
be obtained. The measurement of a burning rate at a
particular time corresponds to that at a given posi-
tion on a gratei the start of a run corresponds to
the discharge from the feed-hopper, and the end of a
run to the point of discharge into the ash-quench
tank.
Description of Experimental Apparatus and
Peripheral Support Equipment
Incinerator Hardware. A brief description of the apparatus
is given in this sectione a fuller description with complete
engineering drawings being left to Appendix A. The descrip-
tion given here covers all the modifications that have been
made to the equipment since it was first completed in May
1971. Many of the experimental runs were not performed with
all the system modifications described below; the evolutionary
development will be fully covered in the section on Results
and Discussion (see pages 184-261).
A sketch of the apparatus indicating all the salient features
is shown in Figure 43. Figures 44 and 45 are photographs of,
respectively, the apparatus with some of its peripheral equip-
ment and the interior of the fuel bed section. The principle
of design for the apparatus closely followed that used by
Weintraub et al. (26), the top section (overfire region)
was joined to~~the bottom section (fuel bed region) so
that the bottom section could be weighted independently. A
flexible steel diaphragm, mounted in the "no flex" position
(which exerted the minimum drag on the bottom section), pro-
vided a seal between the two sections. The top section was
suspended from the ceiling of the test cell, while the fuel
bed section was supported on a load cell which measured this
section's weight using the strain gauge principle. As shown
in Figure 43, the fuel bed section was mounted on a base
plate which had a central shaft protruding from the under-
surface. This shaft, made from special case hardened steel,
passed through two linear ball bushings and rested on the
load cell.
There were a number of possible arrangements for weighing the
fuel bed section using load cell techniques, but the above
method was considered the most satisfactory as it eliminated
the effect of eccentric loads and kept inherent load cell
errors to a minimum. A more detailed discussion of this
- 157 -
-------
1 Levelmg Jocks (or Grate Support
2 Movable Base
3 Hydraulic Jack
4 Support Jack
5 Load Cell
G Load Cell Housing
7 Support Shaft
8 Linear Ball Bearing Bushing
9 Spirit Level
1O Track
11 Support Brace s
12 Support Jack Locating Rings
13 Probe Locating Hole
14 Protec live Fin
15 Thermocouple DAS Junction Bo*
16 Gra te
17 Thermocouple Probe
18 Flexible Diaphragm Seal
19 Heat Shield
20 Movable Refractory Shield
21 Movable Shield Support Structure
22 Roller Bearing Housing
23 Gas Burner
24 Gos Burner Pilot
25 Eclipse Zero Governor
26 Main Gas Supply Line
27 Pilot Gas Supply Line
28 Air Ports ( Capped )
29 Air Ports with Nozzle
3O Cooling Jacket
31 Cooling Coils
32 Stack
33 Gos , Smoke and Temperature Probes
Jsffiffl Castable Ceramic
[JS53 Lightweight Agglomerate
L^HJ Insulating fir* Bnc*
FigQ 43. Sketch of Experimental Incinerator.
- 158 -
-------
cn
vo
Fig. 44. Experimental Incinerator, Gas Analysis
Train and Peripheral Equipment.
-------
Fig. 45. Interior of Fuel Bed Section with Thermocouple
Probes Installed.
This page is reproduced at the
back of the report by a different
reproduction method to provide
— 160 - better detail.
-------
point and details of the load cell are given in Appendix C.
When the apparatus was not in use, the dead load of the fuel
bed section was lifted off the load cell using the support
jacks. Before using the weighing system the bottom section
was accurately leveled with the leveling jacks and the fuel
bed section was then lowered onto the load cell.
The two sections of the incinerator were separated by a slid-
ing refractory shield which was mounted just above the dia-
phragm seal. The shield moved in a plane perpendicular to
the rest of the incinerator and was constructed in two
refractory-lined sections,, One half had a cooling coil set
into its base while the other half had an 18-inch-diameter hole
cut through the middle of the refractory lining. When the
shield was in the "off" or closed position, the section con-
taining the cooling coils sealed the top section from the bot-
tom section and thus permitted the top section to be preheated
without any excess heat lealiage to the bottom section. In the
"on" or open position, the hole in the shield aligned with
the top and bottom section combustion areas, thus connecting
the two sections and permitting the radiant heat flux from the
top section to ignite the fuel bed.
Overfire air was supplied to the top section via any combina-
tion of twelve nozzles at various vertical and radial posi-
tions. The nozzle ports permitted the installation of a wide
range of nozzle configurations. The fuel bed was built on a
grate supported by a hydraulic jack which was used to set the
grate at different vertical positions so that the burning
characteristics of beds of different depths could be studied.
The underfire air was supplied through the grate, the design of
which insured an even air distribution. The pressure drop
across the grata was kept large in comparison to the pressure
drop across the fuel bed, thus minimizing the probability of
having the underfire air channel through the fuel bed.
The bottom section of the apparatus had three sets of nine
sample ports spaced at 120-degree intervals around its cir-
cumference. The vertical arrangement of the twenty-seven
sample locations allowed measurements to be made at one-inch
intervals throughout the bed. The maximum bed depth that could
be studied was thirty inches, equivalent to a fuel bed volume
of 4.2 ft^. The thermocouples (Chromel-Alumel) and the gas
probes were designed so that they were mutually compatible with
the sample port fittings. Three "heat loss" probes, each con-
sisting of three thermocouples at different radial positions,
were located at different vertical and angular positions with-
in the insulating walls of the fuel bed section. The tempera-
ture histories of the three thermocouples in each probe gave
- 161 -
-------
an indication of the rate of heat leakage through the insul-
ating walls of the fule bed. The top section had eight
thermocouples imbedded in the refractory lining near the sur-
face and three thermocouples on the outside steel shell.
These were used to monitor the temperature and temperature
gradients of the rafractory, both during the preheat stage
and during an experimental test. Gas and smoke samples were
taken from the flue gases in the stack, and the temperature
of the flue gas was recorded using a shielded Chromel-Alumel
thermocouple.
For the purposes of calculating material balances, it was
essential to know all air flow rates into the equipment.
Since it was very difficult to seal the whole unit against
air leaks, particularly around the sliding refractory shield,
the leakage rate was determined by feeding a known flow rate
of helium into the overfire air. Batch samples were taken
during a run from the stack gas probe and the helium concen-
tration in the sample subsequently measured using a Beckman
gas chromatograph.
The measurements taken during a typical experiment are dis-
cussed below. Twelve thermocouples were used within the fuel
bed, ten located at the center, the other two near the edges
of the fuel bed. The thermocouples were roughly evenly
distributed throughout the depth of the fuel bed. Two gas
probes were used in the bed, one 20 inches and the other 12
inches above the grate. These probes were used to analyze
for oxygen, carbon monoxide, carbon dioxide, nitrogen, meth-
ane and hydrogen. All three fuel bed heat loss probes were
employed, as were all the thermocouples in the top section.
The flue gas temperature and its water, carbon dioxide, car-
bon monoxide and oxygen content were recorded, and the con-
centration of the helium leak tracer was measured. The fuel
bed weight loss was followed by recording the load cell output,
and the overfire and underfire air flow rates were measured
at frequent intervals. The ambient conditions of temperature,
pressure and relative humidity were noted both prior to and
immediately following the experiment. Fuller details of the
measurements and the measurement techniques are given in sub-
sequent sections and appendices. All of the measurements
made with equipment and instruments that gave response sig-
nals in a voltage form were recorded on punched paper tape
using a Hewlett-Packard 2014B Data Acquisition System. The
data collection and analysis procedures are covered fully in
a later section.
Air Supply. The overfire and underfire air, as well as the
air necessary for running the gas burners and pilots, was
- 169 -
-------
supplied by a Hoffman blower capable of delivering 470 cfm
at 2 psig. A back-up air supply was available from an Inger-
soll-Rand 125 psig oil-free compressor capable of delivering
up to 450 cfm. The air supply system is shown schematically
in Figure 46. The air from the Hoffman blower and the 125
psig compressor was fed to a distribution plenum 18 inches
in diameter and 8 feet high, the air pressure from the com-
pressor having been previously reduced to 2 psig by a large
capacity pressure regulator. The air from the above blower
was introduced into the plenum,, a filter screen which helped
eliminate extraneous matter. The air flow rate to the plenum
from the compressor was controlled by globe valve V]_, while
that from the flower was controlled pneumatically by valve
V2« From the plenum, the air was distributed to the gas bur-
ners and the lines supplying the overfire and underfire air.
For general operation, the air for the gas burners and their
pilots was taken from the plenum below the filter screen; it
was then fed to the two burners via two 2-inch flexible hoses
and to the pilots via one 1-inch flexible hose. The two
globe valves, V3 and V4, provided an air shut-off for the gas
burners in addition to the butterfly valves on the gas burners.
An air shut-off for the pilots was supplied at the pilots
with a gas cock. In special situations, when it was impera-
tive to have detiled knowledge of the air flow rates to the
two burners, the air was fed through one of the lines avail-
able for supplying the overfire and underfire air, and the flow
rate was measured using an orifice plate.
Six lines were available for supplying the overfire and under-
fire air,the flow rates being measured with orifice plates.
The orifice plates were mounted between 300-lb flanges at the
ends of two lengths of nominal 2-inch-diameter shcedule-40
mild steel pipe that were mounted horizontally along one wall
of the test cell. The pipes extended six feet upstream of the
orifices, with four feet of pipe downstream of the orifice
plates before the flow control valves, vs to v10, as shown
in Figure 46. Downstream of each flow control valve there
was a short length of pipe followed by an elbow and nipple.
Flexible hose was used as a connector between these nipples
and the overfire air nozzles and the grate. At the far up-
stream end of each 6-foot length of pipe, a 6-inch section of
honeycomb was placed inside the mild steel pipe to help
straighten and calm the air flow. The pipes were connected
to suitable 2-inch mild steel nipples at different vertical
and radial positions in the distribution plenum with flexible
hoses.
Flange taps, set in the 300-lb flanges, were employed for
measuring the pressure drop across the orifice plates. The
pressure drops (inches of ^0) were measured on a bank of man-
ometers while the upstream static pressures were recorded in
mm of Hg on another manometer bank. The temperature of the
- 163 -
-------
6 Stove
I
H
Line 6 -C^b
Line 5-&h;
Line 4-C&J
Line
Line 2-tJj
•2" Globe Valves
r -
-------
air inside the plenum was measured with a Chromel-Alumel ther-
mocouple.
Two orifice sizes were employed. The smaller size (0.8535-
inch diameter orifice) was used for measuring the overfire
air rates_^nd_^he underfire air rate when it was set below
100 Ib hr ft . The larger size (1.312-inch diameter ori-
fice) was used when the underfire air rate was set above this
value and for measuring the air supply to the two burners,
when this was necessary. In the latter case, the one orifice
sufficed for both burners, the air supply to the burners be-
ing split downstream of the orifice plate. The smaller ori-
fice gave a pressure differential of 24 inches of H-O for a
flow of 50 cfm of air measured at 14.70 psia and 60 °F, while
the larger orifice gave the same pressure differential for a
flow of 120 cfm of air measured under the same conditions.
The orifice plates were used as a primary measuring device and
therefore great care was taken in installing them in hydro-
dynamic conditions which met American Gas Association Standards
(139). A fuller description of the orifice plate installation
and of the procedures and computations undertaken to calcul-
ate the air flow rates is given in Appendix D.
Gas Burners and Gas Supply. The overfire combustion section
was preheated with the two Eclipse 838-36 PMP sealed tunnel
burners shown in Figure 43j they were located 90 degrees a-
part and just above the sliding refractory shield. The bur-
ners consisted of a combustion block cemented to a flanged
cast-iron block holder, and were mounted to the outer shell of
the overfire section using the four corner mounting holes in
the block holder. Each burner had its own gas-air mixing system
and was supplied with low pressure air from the distribution
plenum and city gas at atmospheric pressure from an Eclipse
Zero Governor. The low pressure air was directed into a ven-
turi-shaped burner throat producing a pressure low enough to
draw in the gas from the Zero Governor. The gas-air ratio could
be varied, using the adjustable gas orifice located in the mix-
ing tee. Each burner was provided with a peep sight and a high-
pressure, manually ignited blast pilot. The air for the pilots
was taken from the air distribution plenum. The gas line to
the Zero Governor had an emergency "quick shut-off" gate valve
installed in it so that both burners could be rapidly shut down
in case of a flame-out. A pressure tap was located at the air
entrance side of each burner to measure the air delivery pres-
sure. The pressure was read on the manometer bank in inches
of H-O. This allowed both burners to be run under balanced
conditions (i.e., the heat output from both were equal). The
burners had a dual purpose in that they could also have been
used as after-burners when it was not possible to burn out the
smoke using overfire air alone.
Cooling Water Supply. Cooling water was used for the top
- 165 -
-------
section cooling coils shown in Figure 43, the cooling coils
in the sliding refractory shield, the stack gas sample probe,
and the fuel bed gas sampling probes. Each cooling circuit,
except for the fuel line probes, had a mercury pressure switch
installed in it to indicate when water was flowing through the
circuit and as a warning system to indicate a loss of water
main pressure. Visual checks were used to verify water flow
through the fuel bed probes. Provision was made for flow-
ing the water out of the top section cooling coils and the
stack gas probe after an experiment. This prevented the lines
from rupturing during winter months.
Fuel Bed Grate and Support System. The grate measured approx-
imately 3 Tnches~"deep~ and 17.5 inches on its outside diameter
at the top. The side walls were conical in shape and narrowed
down the cross section of the grate to roughly 15 inches OD
at its base. This method of construction minimized the chance
of binding between the grate and the fuel bed walls.
The inside of the grate was divided into two layers, each 4
inches deep. The first layer acted as a distribution plenum
for the air that yras supplied through the base of the grate.
A baffle was placed over the entry port to the plenum to help
distribution. The upper layer consisted of a 1/2 inch-thick
piece of high-temperature felt placed across the entire dia-
meter of the grate and carefully sealed at the edges, with the
remaining depth of the layer filled with small granite stones.
The stones prevented the felt, which supplied the major pres-
sure dropf from being blown out of the grate and were an ef-
fective packing for good air distribution. Under typical
operating conditions., the pressure drop across the grate was
approximately 5 inches H^O, while that through the bed was
of the order of 0.1 inch H20. The high pressure drop across
the grate relative to that across the fuel bed gave (at lease
in the lower portions of the fuel bed) good air distribution
across the entire fuel bed. The grate was mounted on a hy-
draulic jack- An automobile power-steering pump
utilizing a 3/4-horsepower, 1750 rpm electric motor was the
driving source. A four-way valve controlled the direction
of the flow of the hydraulic fluid to the jack, allowing the
grate to be ei-cher raised or lowered. A pump bypass regulated
the flow of fluid to the jack and was used to control the
speed at which the jack was moved.
Miscellaneous Equipment. Two audio warning systems helped to
provide ease and safety in operating the equipment. One in-
dicated when tha Data Acquisition System was about to run out
of paper tape while the other registered a low air pressure
in the large air distribution plenum. All the thermocouples
located in the top section of the apparatus were wired in
parallel with a Honeywell Electronik 16 recorder and the Data
- 166 -
-------
Acquisition System. The visual display of the recorder was
useful in controlling the refractory temperatures.
One of the channels through to the Data Acquisition System
was reserved as an "event indication" channel. By connecting
this channel to different known voltage sources, the time and
duration of a particular event (such as the start of a run
or when the underfire air was set at its final value) could
be accurately recorded.
Gas Analysis System
Previous Experimental Techniques. Apart from the notable
early work of Kreisinger et al7~(23) in 1916, Nicholls (24)
in 1934, and the later work of Kolodtsev (91) in 1945, there
have been relatively few studies reported HT the literature
of experimental measurements of gas concentrations within a
solid fuel bed. The three studies cited dealt with fairly
homogeneous fuel beds of coke, coal or carbon particles of
well-defined sizes and attempted to measure the oxygen, car-
bon dioxide, carbon monoxide and, in the case of Nicholls
and Kreisinger et al., hydrogen, methane, tar and soot com-
position at different depths throughout the bed. The samples
were collected on a batch basis during an experimental run
by drawing samples under vacuum with water-cooled probes
from different positions in the fuel bed into collection ves-
sels. The samples were analyzed using wet chemistry tech-
niques. The probes used were of large (approximately 0.3 inch)
internal bore, which helped prevent problems of clogging with
tarry pyrolysis products.
Analysis of the data obtained from these fuel beds indicates
that there was a fair degree of error associated with them,
a shortcoming which is understandable, considering the exper-
imental difficulties inherent in this type of study and the
pioneering aspects of these investigations. For example, the
oxygen concentration fell off very rapidly through the fuel
bed and once the ignition plane was reached it was difficult
to locate accurately the point of oxygen extinction; dif-
ficulties were also encountered in freezing the CO + 1/2 02 -*•
C02 reaction in the probes. Even in these relatively well
defined systems, the experimental difficulties associated
with obtaining good data were great. The experimental chal-
lenge offered in the study of a heterogeneous refuse bed
where the fuel is undergoing drying, pyrolysis and combus-
tion, and where relatively large concentrations of tarry
- 167 -
-------
pyrolysis products may be expected, is enormous. Only one
previous investigation (46) has been reported on gas compos-
itions with a simulated refuse bed, all other investigations
on refuse systems having been concerned exclusively with the
gas-phase combustion regime, where experimental techniques
have been developed.
Gas Sample Traip. Figures 47 and 48 show, schematically, the
arrangement used" for gas sampling and analysis. Gas samples
were taken from three positions in the apparatus; two water-
cooled probes were employed in the fuel bed section and one
water-cooled probe in the stack. Figure 48(b) shows the me-
thod of construction of the water-cooled probes. The gas sample
was drawn through the center tube while water passed down the
inside annulus formed by the other two tubes and out through
the outer annulus. The plug at the top of the probe between
the center and outside tubes was made with silver solder. The
fittings at the other end were made from 1/4-inch Sawgelok el-
bows .
The samples from the fuel bed were analyzed for oxygen, carbon
monoxide and carbon dioxide at frequent intervals throughout
the run with on-line instruments. Batch samples were also
taken from the fuel bed frequently during the run and analyzed
afterwards using a gas chroinatograph for hydrogen, oxygen, ni-
trogen, carbon dioxide, carbon monoxide and methane. The
samples from the water-cooled probe in the stack were also
analyzed with the on-line instruments at frequent intervals
for oxygen, carbon dioxide and carbon monoxide; in addition,
batch samples were taken regularly for the helium tracer
analysis (see Figure 47). The sample from this probe, which
was made from a 1/4-inch stainless-steel tube, was split
into two portions. One portion was cooled, the condensate re-
moved, and the dry gas tested with a Von Brank filtering re-
corder to obtain a rough estimate of the amount of particu-
late and smoke emissions. At regular intervals throughout
a run, a known volume of the second portion was drawn through
a steam-traced, line to a U-tube trap placed in an acetone-Dry
Ice mixture where the water content was removed by freezing.
The U-tubes used for these measurements were made from 6 mm
Pyrex tubing and measured about six inches from the top of the
legs to the bottom of the U. The volume of gas drawn through
the U-tube was measured with a wet test meter, and the amount
of condensate was found from the difference between the initial
and final weights of the U-tube. The weighings were performed
only after the U-tiabes had equilibrated with the ambient con-
ditions in the balance room. The details of the computational
procedure used to calculate the water content of the stack
gases from the above measurements are given in Appendix F.
The on-line instrument tain was set up to measure carbon
- 168 -
-------
H
cn
Cooling Water
CftttQLttt
Solenoid Valves for Sock Flushing
.Sotenoid Voives tor Controlling Ftaw to
Anolysrs Trotn
Both
Bleed
r-
Chamber
_j
To Atmosphere
Vacuum Pump ( P-2)
Oxygen
Analyzer
( Paramagnetic)
\
Carbon
Monoxide
Analyzer
NDIR)
Carbon
Dioxide
Analyier
(NDIR)
CaSO4
Drying
Tube
Diaphragm
Pump (P-1)
•*• To Atmosphere
Gas Sample Bottle
On-Line Instrument Tram
Thermostatically Controlled
at 100 *F
vent
Valve( V-5 )
Fig. 47 Gas Sampling and Analysis Train
-------
To Smoke Meter
i
H
O
I
1/2 P'pe
1/4 Copper Tuning
Steam Inlet
Rubber Tubing Wound
With Electrical Tape
Rubber Tubing
Pyrex Tube Packed
With steel Wool
Acetone and
Dry Ice
/ V-• To Atmosphere
'X
Wet Test
Water
VI: Flow Control Valve
V? -- Pressure Beliel v
-------
monoxide (0-50%) , carbon dioxide (0-30%) using nondispersive
infrared analyzers, and oxygen (0-21%), using a paramagnetic
analyzer. The analyzer sections of the oxygen, carbon mono-
xide andQcarbon dioxide units were situated in a closed box
in a 100 F thermostated environment. The temperature was •
held constant in order to minimize errors in the temperature-
compensating systems of the analytical instruments.
The batch samples were collected by attaching sample bottles
to the downstream end of the instrument train on the oxygen
bypass line. The samples taken from the fuel bed were passed
through a glass wool filter to remove the major portion of the
high boiling pyrolysis products, and then passed on to two banks
of three-way solenoid valves. The first bank was set up so
that the probes could be cleaned at regular intervals with
pulses of high pressure nitrogen. When a button was depressed
on the operator's control box, one solenoid valve was activated;
the valve would shut off the line leading to the instru-
ment train and open up the probe to the high pressure nitrogen
line. Under most operating situations, this technique of back-
flushing the probe was adequate in keeping the probes clean
if performed at regular intervals. On release of the button,
the solenoid valve would return to its normal open position.
After the first bank of solenoid valves, the fuel bed probe
sample lines joined the stack probe sample line and the three
lines entered the second bank of solenoid valves. These valves
were situated as close as possible to the instruments and were
activated from the operator's box. In the closed position,
the gas sample flow was directed to a "dump chamber" connected
to a vacuum pump. The dump chamber pressure was controlled by
adjusting a bleed valve leading into it. In the open position,
the valve directed the flow through to the instrument train and
under operation conditions, therefore, only one valve was open
at a time. As a safeguard against opening two valves simultan-
eously, a light next to the control switch for each valve was
activated whenever the valve was opened. From the opened sole-
noid valve, the gas stream passed first through a water trap
placed in a freezing ice-water mixture and then through a small
diaphragm pump. The pump pushed the sample through a drying
tube filled with calcium sulphate and then through the instru-
ment train. Two calcium sulphate drying tubes were mounted in
parallel in the instrument train, permitting a tube with expen-
ded calcium sulphate to be removed from the circuit and refilled
without disrupting the flow to the instruments. The bypass
to the oxygen meter was manually controlled to give the re-
guired flow through the sensor.
- 171 -
-------
The pressure inside the instrument train (generally around 5
to 6 inches of H?O) was recorded on a manometer. The mano-
meter had a dual purpose: to indicate blockage of a probe and
to act as a warning device when the batch samples were being
taken. The latter purpose needs some explanation. When the
sample bottles were attached to the end of the instrument train,
sloppy operation could result in a pressure surge through the
train. All the instruments measured the absolute number of
molecules in a sample cell and pressure fluctuations could cause
spurious results. The most critical instrument in this respect
is the oxygen meter, where a surge in pressure in the bypass
line could send too much of the sample gas through the sensor,
causing erratic swinging of the delicately suspended magnetic
dumbbells.
The length of line between the solenoid valve and the outlet
of the last instrument was kept as short as possible to mini-
mize the "dead, volume" that had to be swept out every time a
different probe was switched through to the instrument train.
Operating experience showed that it usually took about 45 se-
conds to flush out this "dead volume" adequately.
Data Collection
Almost all of the data taken during a run was collected on
punched paper tape with a Data Acquisition System, the only
data not being recorded this way being the stack gas water con-
centrations and, of course, the gas compositions determined by
gas chromatography. The Data Acquisition System (Hewlett-
Packard System model 2014B) had lyV resolution and an input
impedance of 10 megohms. Eight readings could be recorded
per second and the system provided five voltage ranges from
100 yV through 1000V with five significant figures recorded
on each range. The paper tape was punched in standard IBM
8 level code. Some data, for which a visual display was
required for control purposes, were collected on a Honeywell
Electronik 16 multichannel recorder. These data consisted
of the top section temperatures, stack gas temperatures, in-
let air temperature in the distribution plenum, and inlet and
outlet cooling water temperatures.
Figure 49 shows, schematically, the connections from the various
recording devices to the Data Acquisition System ;
Appendix E contains a table that shows the order in which the
instruments were connected to the Data Acquisition System
and the multichannel recorder. The switchboard shown in the
- 172 -
-------
u>
i
DATA
ACQUISITION
SYSTEM
COOLING WATER,
REFRACTORY,
AND STACK
THERMOCOUPLES
HONEYWELL
RECORDER
CONNECTION
BOX
HONEYWELL
RECORDER
FUEL BED
THERMOCOUPLES [
FUEL BED
THERMOCOUPLES
CONNECTION
PANEL
»A a «^T -£-j «£a =&> «A •>?«> «t j ^> •£* "Aii =.?=
4- 4=V 4-4-S-4-444-44-
•f I- 4 4- 4= 4- -i
4» 4- 4J- 4- 4" 4- 4> -{5 -!-' 4- 4- 4>
O 0/0 00000/000
O6OOOO Oy6 O O O O
O O O O O O €T O O O O O
oooooooooooo
7
INSTRUMENTS
IN THE GAS
ANALYSIS
TRAIN
GAS ANALYSIS
CONTROL
BOX
f ACQUISITION
MASTER
SWITCH
BOARD
INDICATION
SIGNAL
SYSTEM
Fir. 49. Instrument to Data Acouisition System. Connections
-------
center of Figure 49 provided a useful device for grouping sets
of data together. The top section, or upper switchboard,
contained sixty permanent connections to the Data Acquisition
System storage registers 140 through 199- The bottom section,
or lower switchboard, held lengths of cable with appropriate
electrical connections which were wired to the various re-
cording devices. This arrangement allowed the output from
any instrument or thermocouple to be connected to any of the
sixty storage locations available on the Data Acquisition System.
About forty separate thermocouple and instrument outputs were
monitored during any one run and measurements of these, outputs
were taken at ten-second intervals during the course of the
experiment.
The primary reason for taking readings so frequently was to
permit the recording of an adequate number of readings of
the gas compositions at the different probe locations. As
indicated in the previous section, it was only possible to
analyze the oxygen, carbon monoxide and carbon dioxide composi-
tions at one probe at a time with the on-line instrument train,
and the operator, therefore, had to switch each probe through
to the instrument train in sequence. After switching the sam-
ple flow from a new probe to the train there was a period of
approximately 45 seconds before a "pseudo steadystate" sample
concentration was achieved. The pseudo steady-state condition
applied after the instrument train had been adequately flushed
off the gas from the previous probe; true steady-state condi-
tions were never achieved in this type of batch operation.
To check the pseudo steady-state condition it was necessary to
obtain at least three readings of the sample oxygen, carbon
monoxide and carbon dioxide concentrations before switching
onto the next probe. With the Data Acquisition System taking
readings every ten seconds, the complete oepration of record-
ing the sample composition from one probe required seventy to
ninety seconds. This was about the maximum time permitted
if one were to map out a semicontineous plot of gas composi-
tions as a function of time throughout a run.
It was desirable to store the three sets of data (oxygen, car-
bon dioxide and carbon monoxide compositions) obtained from
each gas probe in separate storage registers in the Data Ac-
quisition System. A control box performed the task of direc-
ting the three outputs of the on-line instruments to three
sets of separate storage locations allocated. Each probe was
assigned one set of three storage locations. For exampl^, the
lower switchboard addresses 50, 51 and 52 held the outputs
from the oxygen, carbon dioxide and carbon monoxide analyzers
for samples taken from the probe located in the stack; addresses
53, 54 and 55 held the oxygen, carbon dioxide and carbon monoxide
- 174 -
-------
outputs for samples taken from fuel bed probe number one, et
cetera. It was of course still possible to store the infor-
mation held in the lower switchboard addresses 50 through 55
on any of the Data Acquisition System channels, simply by
making the appropriate connections through the upper switch-
board.
The control box had one master switch for each probe. When
the switch was in the "on" position, the signal from the on-
line gas analyzers was switched to the Data Acquisition
System; in the "off" position the switch short-circuited the
signal to the Data Acquisition System. At any particular time
during a run only the switch that was associated with the probe
being sampled was in the "on" position. This arrangement
made it very easy to tell from the data which probe was being
used during any time period .
One channel through to the Data Acquisition System was kept
as an "indication channel" and was used to indicate the start
of various events. For example, when the sliding refractory
shield was opened at the start of an experimental run, a
switch was closed„ completing an electrical circuit which sent
a four-volt signal through the indication channel. Voltages
of different magnitudes were used to signify the beginning
of other operations.
Sequence of Operations During an Experimental Run
The sequence of operations undertaken in preparing for, and
during, an actual test is outlined below. Full details on
the operation of all the equipment are presented in Appendix
E.
Feedstock Preparation. The feedstock was made up to the spec-
ifications required on moisture and inert content by combin-
ing the necessary weights of wet and air-dried wood and tin
cans. The moisture content of the air-dried wood was typically
around 7 percent, while that of the wet wood was around 50 to
60 percent. The individual constituents of the feedstock
were weighed and stored in separate containers, care being
taken to cover the wet blocks so that they did not lose any
moisture content during the period prior to the fuel bed being
charged. A priori it was difficult to determine what the bulk
density of the feedstock would be. Experience showed that
roughly 60 pounds of material were required for filling the
- 175 -
-------
fuel bed section when the maximum fuel bed depth of 30 inches
was used. This3corresponded to a bulk density of approximately
14 to 15 Ib ft" .
Top Section Preheat. After the feedstock had been weighed, the
overfire section was prepared for preheating. The fuel bed
section was pulled out along its tracks from underneath the
top section; this facilitated working on both sections and
was essential for charging the fuel. The four overfire jets
were then installed in their required positions. For all the
runs, four overfire jets in the second level of ports were
used. This configuration proved very satisfactory and the
jets were therefore installed semipermanently in these pos-
itions. With the jets installed, the sliding refractory
shield was pushed closed and the counterweight placed on the
hook of the draw wire used to pull open the refractory shield
at the start of the run.
The cooling water for the refractory shield, the top section
and the stack gas probe was then turned on. The overfire and
underfire air line valves were shut off, as were the butterfly
valves to the burners and the pneumatic valve on the inlet
to the Hoffman blower; this procedure minimized the load on the
blower motor during start-up. The oil level in the two bear-
ings on the blower were checked before the blower was
started and, once the impeller had reached its operating rpmf
the pneumatic valve was opened slightly. The control valves
on the overfire air lines in use were opened and the air
flow rates through the jets were set at the approximate values
to be used during the run. The valves for one of the burners
and its pilot burner were opened slightly and the gas line
valves on both the main supply line and the pilot line were
opened. The pilot burner was then ignited. After the pilot
flame had stabilized, the main burner was ignited, the heat
output being kept low until the combustion block warmed up.
Once this stage was reached, the major work in the preheat
stage had been done and the top section was slowly heated
over a period-of four to five hours by increasing the heat
output of the one burner until the refractories had attained a
temperature of around 1000 F, and then by using the other burner
to help boost the temperature to 2000 -2200 F. The re-
fractories were maintained at this temperature for at least
half an hour before the run was started so that temperature
gradients within tile refractories would have stabilized.
Fuel Bed Section preparation. While the top section was being
preheated,the fuel bed section was prepared. The section
was first cleaned by removing any unburnt remains from the
previous run with a vacuum cleaner. Concurrently, the condi-
tion of the fuel bed thermocouples was checked, badly corrod-
ed ones removed, new ones installed, and bent ones straightened.
17*; -
-------
If possible, the thermocouples were pulled out so that just
the tips were protuding into the fuel bed; this facilitated
charging the feedstock. Sometimes extensive carburi-
zation of the sheath over the thermocouple did not permit the
thermocouple to be pulled out. In this case, the thermocouple
was left in place and care was taken in loading the fuel bed
not to bend it out of position.
After the thermocouples had been checked out, the fuel bed
gas probes were removed and disconnected from the water-cool-
ing lines and sample lines? they were then thoroughly cleaned,
both inside and out. The cleaning of the probe sample tubes
reduced the chance of blocking during a run and was per-
formed by drawing a thin wire back and forth through each
probe and flushing it out, at frequent intervals, with ace-
tone. The outside of the probe was easily cleaned with a
wire brush. After cleaning, the probes were reconnected to
their associated cooling-water and gas sample lines.
At this stage the grate was set at the required height using
the hydraulic pump, and a seal was made between the bottom of
the grate and the fuel bed section to prevent air leakage
between the grate and the containing walls of the fuel bed.
The seal was made from suitable pieces of heavy-duty alumunum
foil held in place with electrical tape.
Gas Analysis Train Preparation. Following the preparation of
the fuel bed section, the instruments in the gas analysis
train were checked out and calibrated using the manufacturer's
standard procedures. In addition to following the manufac-
turer's procedures, each on-line instrument was checked against
at least one other gas of known concentration; this provided
a useful double check on instrument operation. THe procedure
adopted for these checks was as follows. The tip of one of the
fuel bed probes was pushed through one hole of a two-hole
rubber stopper placed in the mouth of a side-arm flask. A
1/4-inch line from the requisite gas source was pushed through
the other hole in the stopper and a hose from the side arm
dipped into a few inches of water. The calibration gas supply
was opened; the solenoid valve which directed the flow from
the probe being used to the instrument train was opened; and
the gas sample pump turned on. The gas supply rate was ad-
justed and balanced against the sampling rate so that just
a small gas flow was seen to bubble through the water trap.
The procedure insured that only the desire gas was sampled,
that a minimum of the gas sample was used in the process, and
that thte pressure in the system was as close as possible to
that found under actual operating conditions.
Miscellaneous Equipment Preparation. Once the instrument train
- 177 -
-------
was checked, the cold junctions for the thermocouples—five
cold junctions were used for the fuel bed thermocouples and
one for all the other thermocouples—were placed in Dewar
flasks filled with ice-water mixtures. The final step before
the fuel bed section was charged with fuel was the preparation
of the stack gas water determination equipment and the connec-
tions to the Data Acquisition System. For the water deter-
mination the dry U-tube traps were weighed (they were dried in
a vacuum over night) with rubber stoppers covering each arm
of the U's. The weights were noted and the tubes numbered
and laid out in order next to the cold trap, which was a large
vacuum flask filled with a Dry Ice — acetone mixture. The
wet test meter was then leveled and zeroes and the steam to
the sample line was turned on.
The connections for the Data Acquisition System were made
after the number of fuel bed thermocouples to be used had been
decided. The connections were made using the switchboard
arrangement shewn in Figure 49 so that sets of data were con-
veniently grouped together on adjacent registers on the Data
Acquisition System. The only guideline for this process was
that all the thermocouple registers had to be grouped together,
as the computer programs developed for analyzing the data
were written on this basis. The switchboard set-up typically
used was as follows: event indication channel, fuel bed ther-
mocouples arranged so that the thermocouple that was closest
to the top of the bed was first and the lowest one last, heat
loss probes, top section thermocouples, other thermocouple
outputs, load cell output and input, and finally gas analysis
outputs. Once tha connections were made, the Data Acquisition
System clock was synchronized with the clock in the test cell
and the Data Acquisition System was set up to take readings
every ten seconds through all the registers in use. The data
Acquisition System was then left in a stand-by mode of opera-
tion.
Loading the Feedstock. The first step in loading the fuel was
to obtain a reacting of the weight of the fuel bed section
without the fuel and with all thermocouple and gas probe con-
nections made. To do this, the section was leveled with the
appropriate jacks and then the support jacks were slowly lo-
wered until the full weight of the fuel bed section was on the
load cell.
A two-volt signal was switched to the event indication channel
and nitrogen zero gas was passed through the oxygen meter.
When the zero of the oxygen meter had been adjusted and
stabilized, five to ten readings were taken with the Data
Acquisition System., The support jacks were raised until the
- 178 -
-------
the bottom section weight was completely off the load cell,
and the fuel was loaded into the fuel bed section. The fuel
bed was loaded in a pseudo-random fashion with care taken to
distribute the wet and dry blocks of wood and tin cans evenly
throughout the depth of the fuel bed. The various thermocouple
and gas probes were pulled into place as the fuel bed was built
up, care again taken to spread the fuel evenly across the fuel
bed, as otherwise large gaps tended to develop at the sides.
The fuel bed was charged so that the top of the bed was even
with the top fo the brick doughnut of the bottom section;
fuel was added or subtracted in order to do this and these
additions or subtractions were carefully weighted and taken
into account in determining the final weight and composition
of the fuel.
With all the fuel charged, the support jacks were again lowered
carefully and a five-volt signal was switched to the event
indication channel. Air was passed through the oxygen meter
and the up-scale reading adjusted; five to ten more readings
were taken with the Data Acquisition System. The weight dif-
ferential betwen the empty and full weights of the fuel bed
section was calculated from the load cell outputs recorded
by the Data Acquisition System and compared with the known
weight of fuel added. The largest error between the calculated
and actual weight of the fuel was always less than one pound.
The support jacks were then raised and the bottom section
moved into place underneath the top section. The bottom sec-
tion was carefully leveled with the leveling jacks and then
raised using these jacks until the flange on the fuel bed
section just mated with the flexible diaphragm; the mating
was indicated by locating probes positioned on the fuel bed
section flange (see Appendix A). The section was raised by
rotating all three leveling jacks simultaneously one half-turn
at a time. The close fit between the two mating doughnuts
of the top and bottom sections (see Appendix A) made it
imperative that the bottom section be raised in this manner,
as the doughnuts would have bound together if the bottom sec-
tion were not vertical. Once this operation was performed
all electrical connections from the fuel bed thermocouples
to the control switchboard were made. The support jacks were
lowered and the weight taken on the load cell. At this point
nothing was allowed to touch or bump the bottom section, as
spurious readings from the load cell would have resulted.
At this juncture, another set of readings of all instrument
outputs was taken with the Data Acquisition System. The weight
of the fuel bed was computed from the load cell output obtained
- 179 -
-------
from this set of readings and compared with the previous value.
This gave a good check that there was no binding between the
top and bottom sections. This set of readings was also used
to check the outputs of all the thermocouples to make sure
that they were all operating correctly. Once all systems
were operating properly, the run could be started.
The Start and_jgperation of a Run. A run was commenced by
switching the solenoid valves for the gas analysis train so
that a gas sample stream was taken from the stack gas probe.
The Data Acquisition System was then switched on to a mode
where it took readings through the total set of inputs every
10 seconds. The underfire air flow was turned on slightly
(about 1/10 its final value) and the gas and air supply to the
gas burners and pilots was shut off. Immediately after the
burners were shut off, the sliding refractory shield was
gently pulled open using the draw wire. This latter opera-
tion needed to be done carefully as the shield could cause
considerable vibrations (with subsequent load cell errors) if
it hit the restraining stops too hard. The shield was pulled
out far enough to activate a switch which sent a six-volt
signal into the event indication channel and activated a light
on the instrument panel. This signal indicated the start of
the run.
Shortly after the run had begun, stack gas water concentration
measurements were begun and the underfire air rate was slowly
raised to its final value. About three to four minutes were
taken to raise the underfire air to the desired level, at which
time a twelve-volt signal was sent to the event indication
channel. Readings were taken of the orifice-plate pressure
drops and upstream air pressure on all the overfire air lines
and on the underfire air line every time they were changed.
The times at which these changes occurred were noted on a
log sheet.
Samples were taken from the three gas probes in a cyclic
fashion. However, samples were not taken from the fuel bed
until the ignition front had reached the first gas probe. Up
to that time, samples were taken continuously from the stack
with only brief breaks to check conditions at the first fuel
bed probe. Once the ignition front reached the first fuel
bed probe, as indicated by a dramatic increase in carbon dio-
xide level, a sampling sequence of gas probe /fuel bed probe
was followed as rapidly as this operation permitted. Typi-
cally it took one and a half minutes to obtain a sample, as it
took 45 seconds to attain a :s^oady-state reading after switch-
ing over the probes and at j&tst^ three readings were taken
before making the next switch. Every so often, the second fuel
— i <
JL i
-------
bed probe was tested and once active burning was detected
at this level a sampling sequence of stack/fuel bed probe
one/stack/fuel bed probe two/stack was followed. Near the
end of the run, when the carbon monoxide concentration in the
fuel bed began to decay and the oxygen content started to
increase, samples were again taken solely from the stack.
During these sampling operations batch samples were taken
from all the probes at those intervals throughout the run
when the probes were switched to the gas analysis train. The
probe location and the time at which the sample was taken were
noted on the sample bottles which were analyzed after th run
(hydrogen, oxygen, carbon monoxide, carbon dioxide, methane and
nitrogen for the stack samples).
Stack gas water determinations were performed at regular in-
tervals throughout the run, about every three to four minutes.
The procedure for this was as follows. One of the U-tubes
was connected between the two lengths of rubber hose shown in
Figure 48(a) and then the vacuum flask filled with the Dry
Ice — acetone mixture was raised so that the top level of
the liquid was as close as possible to the rubber/glass joint.
After a period of a few seconds to let the glass and steel-
wool packing cool, the valve V]_ of Figure 48 (a) was opened
slightly (the pump had been switched on and left running be-
fore the run was started) and a low flow rate of air was
drawn from the probe. After 0.3 cubic feet of air had been
pulled through the U-tube (three revolutions in the dial of
the wet test meter) valve v^ was rapidly closed and valve V2
quickly opened to equalize the pressure on both sides of the
U-tube. This procedure prevented the partial vacuum created
on the pump side by the high pressure drop across the ice-
packed U-tube from allowing any unaccounted-for air to be
pulled through the U-tube. Once the pressure had equalized,
the vacuum flask was lowered, the U-tube rapidly removed and
both ends stoppered. The time of the sample, U-tube number
and exact volume of gas passed through the U-tube were noted,
as well as the time taken to collect the sample, generally
one to two minutes. A new tube was mounted in position as
soon as possible and the process repeated. Water determina-
tions were started as soon as the run had begun and were con-
tinued until the end of the run.
Other than working the gas analysis train, taking batch gas
samples, doing the water determination and changing the punch
paper tape on the Data Acquisition System, the only additional
operation necessary during a run was adjusting the overfire air
rates in an attempt to keep the top section refractory
- 181 -
-------
temperatures around 1400°-1500°F. This method of operation
was necessary in this batch system if the net heat flux to
the top of the fuel bed was to be kept approximately constant.
In practice, it had been found that the best way to achieve
this goal was to keep the oxygen content in the stack gases
around 10 to 15 percent. The large thermal sink effect of
the refractories and the slow recording speed of the multi-
point recorder meant that there was a large response time asso-
ciated with any control system based on the refractory tempera-
tures .
The runs was ended when the concentration of oxygen in the
stack gases reached around 20^5 percent. At this stage the
Data Acquisition System was switched off, the support jacks
were raised until all weight was off the load cell and the
load cell power supply was shut off. The Hoffman blower
was switched off, all city gas lines closed at the main valves,
the instrument train and control system switched .off and the
fuel bed probes removed„ All cooling water circuits were turned
off except those for the roof and for the stack gas probe.
The cooling water circuits to these sections were left on for
a period of at least four to six hours.
Data Processing
Three computer programs were used to process the raw data,
which had been stored on punched paper tape. The first step
in the data processing was to transfer the Data Acquisition
System output from paper tape to magnetic tape. The first
program then read the magnetic tape data and decoded the IBM
8 level code format to give an output in punched card form.
This output deck could be checked for any errors which might
have been caused by malfunctions of the high-speed punch on
the Data Acquisition System. The deck also provided a con-
venient form for storing the raw data. The card output from
the first progra, coupled with a few control cards, was used
as input for the second program, which performed the task of
converting the voltage readings into corresponding tempera-
tures, gas compositions, and weight loss using conversion
tables and calibration curves. The card output from the second
program, which contained arrays of temperature, gas comppsition,
probe number,
- 182 -
-------
and weight loss at the corresponding elapsed time from the start
of a run formed the input data to the third program. This
program plotted the data as a function of elapsed time using
a Stromberg Carlson 4020 plotter. Complete details of the
three programs are given in Appendix G and sample outputs from
the programs are included in Appendix H.
In addition to the programs for handling and processing the
raw data a program was developed for calculating instantan-
eous material balances around the unit. This program used the
stack gas measurements, the overfire and underfire air flow
rates and the helium tracer flow rate as input data. The
development of the equations used for this program and a list-
ing are given in Appendix F.
- 183 -
-------
RESULTS AND DISCUSSION
Introduction
Summary of Experiments Performed. The experiments that were
conducted can be conveniently divided into four groups of
runs: Runs 1-2, Runs 3-10, Runs 11-16, Runs 17-18. A break-
down of the fuel type and composition used in all the tests
is given in Table 10; the conditions used in the experiments
and the measurements taken are summarized in Table 11. Runs
1 and 2 were of a preliminary nature, designed to check the
overall operation of the test incinerator. The fuel bed grate
used in these tests consisted of an 18-in.-diameter piece of
sheet metal drilled with 1/8-in.-diameter holes on 3/16in.
centers. No effort was made to control the underfire air, the
air supply being set by the prevailing natural draft. The
results from these two experiments showed that the overall
operation of the equipment was satisfactory, but that in order
to obtain the detailed information desired, a number of changes
would have to be made. These improvements included
removing intermittent short circuits in the fuel bed thermo-
couple electrical system and improving the method of joining
the two sections of the apparatus in the interest of minimi-
zing drag on the weighing system.
For Runs 3 through 10, an improved grate was added to the sys-
tem, which permitted a controlled amount of underfire
air to be introduced through the fuel bed. In these experi-
ments, the fuel composition was maintained approximately con-
stant (17-15% inerts, 28-33% moisture, and the remainder com-
bustible) and the underfire air was varied from 127 to 275
Ib hr Ift~Z.
In most of this second group of experiments, difficulty was
encountered in igniting the fuel bed properly, although for
the majority of experiments the underfire air was slowly in-
creased to its final level over a period of a few minutes
after the start of the run. Under these circumstances, the
bed initially burnt vigorously but then slowly died down; this
resulted in a severe decrease in the temperature level of the
refractory lining of the overfire section and consequently a
decrease in the radiant heat flux to the top surface of the
bed. The fuel bed temperatures showed that in the experiments
where ignition was not properly achieved, the temperatures rose
above the ignition temperature in the first few inches (%5-7
in.) of the bed and then the ignition process stagnated. The
- 184 -
-------
TMU 10
Synthetic fuel Composition and Analysis
I
H
00
tn
I
Run
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Hood Analysis
ultiaate Analysis
(Percent by Height
-Bono Dry Hood)
tH
tc
to
NOT ANALYZED
6.08
5.96
52.24
51.30
41.68
42.74
As-Received
Moisture
Content
of Hood
"*8
7.96
6.90
Sites of Fuel Constituents
(Dimensions in Inches)
Hood
2>ixm
[] 3/8
2Hr.m
X3/4
2>sx.m
i! 3/8
1 3/4
ilx3/4
-}!„•> 1
*Vti%
xl",
ISxlS
xu
Psper1
_
-
-
5»iK2
6:;3
4!1H6
S
ul
z>
R
Metal
Cans
h-height;
-
-
-
h=4,
d»3 5/8
h»3.
d«2 3/8
n-5.
d-4
h«3 1/8,
d«2 3/4
Glass
Jars3
dKi2i6fteter
_
-
-
h=l 7/B,
d-2 1/8
h»4,
d-2(i
3
Ul
3
R
Conposition of Fuel
(Percent by Height)
Moisture
~-B
~17
-
27.0
33.0
29.9
27.8
31.2
30.7
31.8
24.0
25.9
23.9
14.7
26.4
24.9
25.4
25.1
Coffibuotible
Hood
(Bone Dry)
"92
•^83
-
48.4
41.9
46.5
47.9
44.9
46.0
47.6
60. 5
61.3
62.3
70.2
73.6
58.9
59.4
59.9
Paper
-
-
-
8.8
7.3
8.4
8.7
8.5
8.3
3.6
-
-
-
.
-
-
-
-
Inert
-
-
-
15.8
17.8
15.2
15.6
15.4
15.0
17.1
15.5
12.8
13.8
15.1
0.0
16.2
15.2
15.0
Ultimate Analysis2
of Puel
(Percent
bv Height)
tc
-
-
-
29.2
25.2
28.0
28.9
27.2
27.7
26.4
31.0
31.5
32.0
36.0
37.8
30.2
30.5
30.7
tH
-
-
-
6.5
6.6
6.7
6.5
6.7
6.7
6.6
6.3
6.5
6.4
S.8
7.3
6.6
«.3
6.4
to
-
-
-
4S.5
50.4
50.1
49.0
50.7
50.6
49.9
47.2
49.2
47.8
43.1
54.9
47.0
48.0
47.9
1 Random configurations -- cut up, crumpled and twisted
2 Assuming analysis for paper was the same as for cellulose, which from Kaiser's data ( 49)
was a good approximation of the average composition of the mixture of papers used
3 When glass was used the ine^c was made up of approximately 55% glass and 45% tin cans.
-------
TABLE 11
Summary of Experimental Conditions5
Run
No.
b,P
la, r
,b,p
^q-r
a^ff ,P
q.r
4b,f
r,o
5b,f
r,q
6b,f
r,q
?b,f,r
a,s
8b'£
r,o
Fuel Analysis
(% by weight)
Combus-
tible
-92
-83
-75
57.2
49.2
54.9
56.6
53.4
Inert
-
-
-
15.8
17.8
15.2
15.6
15.4
Mois-
ture
-8
-17
-25
27.0
33.0
29.9
27.8
31.2
Actual
Amount
of Fuel
Charged
(Ib)
60.0
57.0
68.3
59.4
57.6
58.6
Amount
of Fuel
Charged ,
from Load
Cell Read-
ino (Ib)
Determined
o
53.0
69.9
X
58.0
61.5
Dnderfire
Air Ratedb
hr'lff2)
at (elap-
sed time,
sec)
Natural
Draft
137C
146°
93(1600-
2200)
145(2200-
4000)
127(4000-)
X
146°
210(0-120)
130(3400-
5260)
150(5260-
6220)
210(6220-)
Burning
Rate(lb
hr-ift"2)
at (elap-
sed time,
sec)
68° (960)
57c,d
_JJL200J__
d
359(1000)
64 (1900)
62(3800)
X
289(2000)
67 (2750)
d
Ignition
Ratedb
hr-lff2)
at (elap-
sed time,
sec)
110C (960)
e
e
m
125h(3800)
X
79h(1200)
y
Max.
Temp.
Obser
ved
(°F)
e
e
e
m
19001
X
2300:
y
Comments
ixploratory experiments. Data Acquisi-
sition System malfunction in Run 2 pre-
i/ented detailed analysis of data.
Demonstration experiment.
Burning appeared normal.
Bed initially ignited, then smoldered
for first 2/3 of run. At 2600 sec burn-
ina began to improve. At 3600 sec act-
ive burning started and bed burnt rapid-
ly for 800 sec until all fuel was con-
sumed.
Bed ignited but could not sustain burn-
ing, despite varying underfire air and
reigniting gas burners. Bed smoldered
for 5 hr and almost complete combustion
achieved. No data taken after first
20-25 min of runT
Behavior of bed similar to that obser-
ved in Run 5. Data indicated probability
that bed had not been allowed to ignite.
properly before underfire air was
raised to final level.
Bed did not ignite properly using under-
fire air at 210 Ibhr-lft"2. Air rate lo-
wered and then increased slowly back to
210 Ibhr'^ft"2 over 5 min period. Fire
still did not improve and underfire air
rate again decreased. Bed eventually
•sfrarl-iart to hum.
-------
Table 11 (Cont'd)
I
H-
00
I
*.•
r,q
10b,f,r
q,s
HT7kTT
q,t,s
„, ^
i-p-.q
t, s
!—_____
^pf-
I4i,q
t,s
IS1'"
t,s
16l'Q-
t,s
1?l,u
v,w
IB1'"
54.3
____
51.2
60.5
61,3
62.3
70,2
73.«
58.9
59.4
59.9
1
15:.0
•^^a^**™.
17.1
15.5
12.8
13.8
15,1
0.0
16.2
15.2
15.0
30.7
-onuau-^aun ._;
31,8
24.0
25.9
23.9
14,7
26.4
24.9
25.4
25.1
60.0
t— •--••— wuau
70.1
69.3
68.0
64.9
59.7
82.3
61.6
62.6
66.5
-i
58.4
69.7
70.8
69.5
65.6
60.5
81.7
63.1
62.7
67.2
275(180-
900)
88(900-
1020)
175(2400-
4200)
175(5000-)
275(120-
600)
175(600-)
155C
155C(600-)
240(0-900)
varying
(900-2400)
150(2400-
4200)
175(4200-)
160C
156C
118C
140C
86C
5(1250)
27(3200)
27(5200)
22(2200)
35(3600)
40C
n
n
n
n
n
39C
30(0-
1700)
41(1700-
3200)
64^(4600)
34e|256b'j
50C
y
y
48C
54C
41C ,
38°
43C
1800
22001
24001
2000
22001
24001
21001
2400i
23001
23501
Underfire air raised over 180 sec to
275 lbhr~1ft~2. Fire died down, under-
fire air was decreased, then fire caught
and underfire air increased back to 175
lbhr~1ft-2. Again fire died down, air was
lowered, and, when fire caught again, rais-
ed to 175 Ibhr-iff2 for rest of run.
Data only taken during later stages of
run when active burning was visually
observed.
At high initial underfire air ra'te.fir'e
began to go out; under fire air rate was
decreased aRd burning improved.
Gas probas plugged during test. Poor
electrical connection between oxygen
raster apd D . A . S .
Bed ignited but fire died down quickly.
Under fire air decreased and then, after
bed caught. was increased to final value.
Underfire air slowly raised to 2401b hf-1
ft" 2 over 15 min period. Fire died down.
Underfire air rate lowered and then, when
fire had caught, raised to 2101bhr~1ft~2.
Fire again died down. Underfire air rate
decreased and, when fire caught, was set
at constant level. Top of gas sampling
probe blew off during run twice .
»
Burning appeared normal .
-------
FOOTNOTES:
(a) All bed depths studied w.re 30 in. deep (4.2 ft3).
(b) Operating difficulties with Nondispersive Infrared CO meter and Sectarian 715 amperornetric oxygen sensor.
(c) Rate constant over the major portion of test
(d) Considerable scatter in data because diaphragm tightened too much.
(e) Intermittent short circuits in the thermocouple electrical system gave poor fuel bed temperature data.
(f) Runs 3 through 10 inclusive were performed with the low pressure drop grate.
(g) Oxygen present at both fuel bed probe positions
(h) Ignition rate increased with time.
(i) Observed towards end of run
(j) Runs 11 through 18 inclusive were performed with the high pressure drop grate.
(k) From Run 11 through Run 18 two temperatures were recorded at the edge of the bed in addition to the ten
measurements taken in the center.
(1) Paramagnetic oxygen analyzer used instead of amperometric sensor
(m) Problem with fuel bed thermocouple electrical system gave spurious results.
(n) Building vibrations and other difficulties with the load cell gave very poor weight loss data.
(o) Binding between top and bottom sections caused spurious results from weighing system.
(p) No fuel bed gas samples taken
(q) No air leakage rate measurements taken
(r) No measurements of moisture content in stack gas
(s) Some batch samples analyzed for CO, C0_, 02/ CH., and N,
(t) Measurement of moisture content in stack gas but readings all low because of poor water trap design
(u) Complete analysis made for fuel bed batch gas sample - CO, C02/ H_, 02, CH4, N
(w) Air leakage rate measurement satisfactory
(x) Incomplete data precluded analysis.
(y) Difficulties in igniting bed negated value of these measurements.
-------
maximum bed temperature noted under these conditions (1000-
1400 F) was very much less than the maximum observed when the
bed was burning properly (1800 -2200°F). When this behavior
was obseryed^the underfire air was decreased to around 60-
70 Ib hr ft and held at this level until the fire started
to burn vigorously again. At this time the underfire air was
raised back up to the desired level. On some occasions the
underfire air was pulsed in an attempt to revive the fire,
and the refractory brickwork reheated with the gas burners
so that the heat flux to the top of the bed would be increased
Runs 6,9 and 10). None of these methods proved to be very
satisfactory, although they only failed once (Run 6) to re-
vive the combustion process.
In this second group of experiments (Runs 3-10) data collec-
tion was extended to include gas samples from the fuel bed,
as well as from the stack. These samples were taken from two
positions in the bed and were analyzed using the on-line
gas analysis train. Some batch samples were taken from the
feul bed, and a limited analysis was performed on a gas chroma-
tograph for O?, CO, C0?f N? and CH.. A considerable amount
of operational difficulty was encountered in this new proce-
dure; most of the difficulty concerned the poor operation and
response time of the Beckman 715 Amperometric oxygen Analy-
zer originally used for the oxygen measurements. Additional
equipment problems were related to the electrical circuitry
of the non-dispersive infra-red CO analyzer. Finally, the
distortion of the flexible diaphragm, where the top and bottom
portion of the equipment were connected together, continued
to cause weighing errors„ The data from this group of runs
are therefore rather sketchy, although they did provide some
insights into the ignition behavior of the bed. These runs
further showed (1) that the underfire air tended to channel
through the fuel bed along the walls, (2) that the concept
of drying and pyrolysis waves propagating through the fuel
bed was unreasonable for the size and type of fuel used in
the tests, and (3) that the closure of material balances
coupled with the measurement of the concentration of water
and hydrogen within tha fuel bed would provide a valuable
insight into the progress of drying and gasification.
For all experiments subsequent to Run 10, the glass content
of the fuel was omitted. It was found that, when used, the
glass readily melted at the high temperatures of the bed,
coating the grate and insulating walls of the fuel bed. Great
difficulty was encountered in cleaning out the fuel bed, be-
cause the solidified glass clogged the holes in the grate and
covered the thermocouples. It was felt that, although remov-
ing the glass meant eliminating an element characteristic
of refuse, the tin cans would suffice to simulate the inert
content.
- 189 -
-------
For the remainder of the experiments (Runs 11-18), a para-
magnetic oxygen analyzer replaced the amperometric sensor
and the gas analysis train was modified to decrease the time
required to take a sample and to make its operation easier.
The method previously used to join the two sections of the
apparatus together (i.e., clamping the two sections together)
was abandoned and the sections were joined by butting them
together and relying on the rubber gasket between them to ef-
fect a seal. The rubber gaskets on the flange of the bottom
section were built up so as to minimize gaps between the flange
and the flexible diaphragm when two sections were butted to-
gether. For calculation of material balances ,the stack-gas-
water content had to be measured. The equipment necessary for
this measurement was built and installed for Run 11. A des-
cription of this equipment is given (see pages 168-172 ).
Runs 11 through 16 attempted to fill the information gap sug-
gested by Runs 3 through 10. For these runs (Runs 11-16),
both the underfire air and fuel composition were varied around
a base case. For the base case (Runs 11-12), the fuel compos-
ition was - 25% moisture, 15% inert, and 60%_combustible, and
the underfire air was fixed at 155-160 Ib hr ft . In Runs
13_and_16 the underfire air was varied from 118 to 240 Ib
hr ft while the fuel composition was held constant; in
Runs 14 and 15 the fuel compositions were 0% inert, 25% mois-
ture, 75% combustible and 15% inert, 15% moisutre, 70% com-
bustible, respectively, and the underfire air was held to the
base case value of 155-160 Ib hr ft" .
Problems were encountered in measuring the moisture content
in the stack (from which the moisture content in the bed was
calculated) and with the weighing system. The first problem
was attributable to faulty design of the measuring system used
and was readily corrected. The latter difficulty was caused
by excessive vibrations set up by other operating equipment
elsewhere in the building and could not be cured; however,
by scheduling runs during relatively quiet times, the pro-
blem was largely, circumvented.
Starting with Run 11, the grate design was modified so that
the pressure drop across the grate was markedly higher than
the drop in Runs 3 through 10; the modification appeared to
decrease the amount of channeling that occurred through the
fuel bed. Attempts to close material balances for Runs 11-16
showed that there was considerable air leakage into the test
incinerator, probably through the spaces around the sliding
refractory shield. This air leakage did not affect the mea-
sured burning rates because this air did not pass through
the fuel bed. For material balance purposes, however, the air
-------
leak had to be measured and a helium tracer method, described
on pages 168-172, was selected for this purpose.
The improved operation of the gas analysis train made it pos-
sible to obtain more gas samples from the fuel bed; in addi-
tion a greater number of batch samples were taken and analyzed
more completely than previously (i.e., samples were analyzed
for H2 in addition to CO, C02, 02, N2, and CE^). Although
the additional measurement of hydrogen provided more detail
of fuel bed conditions, it was not possible to check whether
the water-gas-shift reaction was equilibrated at the top of
the bed because of the poor stack-gas-water measurement and
the unknown quantity of air leakage into the equipment. (The
moisture content of the gases at the top of the bed could have
been calculated from the known underfire and the overfire air
flow rates and the stack gas moisture measurement.)
Before progressing with the final group of experiments (Runs
17 and 18), the apparatus, the gas measuring equipment, and
the leak measuring technique were all thoroughly tested by
closing material balances around the unit while burning a known
amount of city gas. For these tests, the two sections of the
apparatus were joined together in the manner used for a regu-
lar run, and overfire and underfire air was supplied at about
the ratios and rates used in a typical test. The combustion
air supplied to the gas burners was metered with a suitable
orifice plate. The amount of city gas used was measured by
a dry test meter specially installed in the main supply line
to the Fuels Research Laboratory for this purpose.
Runs 17 and 18 provided the last set of results obtained
and were the most complete runs performed. For both of these
runs, all equipment worked well and good closure of material
balances was achieved. The material balance gave an inde-
pendent check on the load cell results and provided insight
into the C/H and O/H ratios of the fuel that was being con-
sumed. Using the material balances and the measurements of
the concentration of hydrogen within the fuel bed made it pos-
sible to check the validity of the equilibrium of the water-
gas-shift reaction at the top of the fuel bed. The base case
fuel composition was used in both of these runs and the under-
fire air was varied from 140 Ib hr~lft in Run 17 to 86
Ib hr~1ft~2 in Run 18.
Outline of Discussion of Results. The ensuing discussion
will be based primarily on the results obtained from Runs 5
through 18, as the results prior to Run 5 were of a prelimin-
ary nature and experimental conditions were ill-defined. The
- 191 -
-------
complete data set for one experiment (Run 17) is contained in
Appendix H. It was not considered feasible for easons of
space to include the complete data sets for all experiments;
thus, the following discussion will be based on selected data
from typical runs. Rather than dealing with the results
from each of the experiments in sequence, the following sec-
tions will each be devoted to considering a different aspect
of the results (for example, Bed Temperatures, Material Bal-
ances, Fuel Bed Gas Compositions, etc.).
Fuel Bed Temperatures
Fuel Bed Temperature Distributions for Active Burning Runs.
The temperature"at the center of the fuel bed was measured
at ten different vertical positions spaced, on the average,
three inches apart. For the later runs (Runs 11 through 18)
two thermocouples were used to measure the bed temperatures
close to the side wall. The measurements taken in Run 10 at
five positions in the bottom half of the bed are shown in
Figure 50. It should be emphasized that these measurements
were made with a thermocouple at the center of the bed, and
that they represent a weighted mean of the temperature of
the gas that flowed over the couple and the temperatures of
the surfaces with which the couple was in radiative exchange.
The recorded temperature therefore depended on the drying
characteristics of the fuel surrounding the termocouple to
the extent that these characteristics affected the surface
temperature of the drying fuel element.
Despite the above qualifications, the temperatures in most
runs showed, as in Run 10, a regular progression of the burn-
ing front from the top surface of the bed to the grate. In
a few experiments the temperatures were observed to plateau
at 212 F ang about 500 F. Peak temperatures were typically
around 2200 F,, These maximum temperatures give a measure of
the degree of melting, clinkering and ash fusion that could
be expected on a traveling grate. The temperatures observed
here, which appear typical for most refuse beds, are high enough
for all soft glass to melt (as was experimentally observed)
and were high enough to suggest that aluminum would also melt.
The temperatures were never high enough to melt the tin cans,
but the bright metal coating disappeared as a probable' conse-
quence of both vaporization and diffusion into the substrata;
the steel was badly carburized. Figure 50 shows, from the
temperature registered by thermocouple No. 5, that the ignition
192 -
-------
240O
20OO
16OO
IL
O
LJ
Q
U
cr
o:
LJ
Q.
2
LJ
12OO —
BOO
40O —
3OOO
40OO
5000
ELAPSED TIME, SECONDS
Fig. 50. Fuel Bed Temperatures (Run 10). Thermocouples
1, 2, 3,4 and 5 were located 11, 9, 7, 2, 0
in. from the grate. Initial fuel bed depth 30
in.; underfire air rate 175 Ib hr~l ft~2; bulk
density of fuel 16.7 Ib
-------
plane reached the grate 4400 sec after the start of the run.
The temperature recorded by this thermocouple can be considered
to be the grate temperature and as such gave a measure of the
degree to which the grate was heated.
Care has to be taken when interpreting the fuel bed thermo-
couple data because of the continuously changing nature of
the fuel bed,., In this regard, two problems were encountered.
First, at the ei\d of a few runs, some thermocouples were found
to be severely bent out of position. The following tentative
explanation for this observation can be offered as no direct
evidence was available from the data. The consumption of the
combustible resulted in a continuous change in the density
of fuel bedf and therefore the ash and inert content of the
fuel had to settle slowly towards the grate over the course
of an experiment. The mechanics of the settling of the bed
would have depended on the difference between the ignition
and burning rates,. When the ignition rate was not much grea-
ter than the burning rate of the combustible content of the
fuel, it can be hypothesized that the fuel was consumed in a
thin reaction zone which moved through the bed. Under these
circumstances, the inert and ash content of the fuel would
have moved with the burning zone, and within this burning
zone the inert and ash content would have increased with
time. The settling of the bed in this fashion would have
caused little drag on the fuel bed thermocouples. When the
ignition rate was much faster than the burning rate, it can
be conjectured that most of the fuel was consumed after the
ignition front reached the grate; after this time, the major
portion of the remaining fuel would have been consumed by
the oxygen in the underfire air in the space of a few in. a-
bove the grata, Under these conditions, it can be expected
that the whole of ;he remaining bed would have slowly set-
tled towards the grate as the combustible was consumed; the
resulting drag on the fuel bed thermocouples under these con-
ditions would have been considerable.
An attempt was made to minimize the effect of the settling
fuel bed on the thermocouples by supporting the thermocouples
inside quartz tubing held in place by metal sheaths cemented
into the side wall of the fuel bed. These efforts were largely
successful but there were occasions when the quartz tubing
broke and the thermocouple casing was rapidly carburized
and weakened,, allowing the thermocouple to be dragged down
by the settling bed. (The rapid carburization of steel within
a fuel bed is one of the primary difficulties associated with
obtaining good fuel bed temperature data with thermocouples.
It was found that 304 Stainless was rapidly carburized at the
temperature i-avels of the bed; and of the more common high
nickel content alloys only Inconcl 600 appeared to have a
- 194 -
-------
a reasonable lifetime.) A further difficulty raised by the
consumption of the fuel was the relative position of the ther-
mocouples within the bed as a function of time. As the bed
burned down, the thermocouples were, since their position
with respect to the apparatus was fixed, eventually exposed
above the bed; At this point, the thermocouples were no long-
er recording bed temperatures but instead recorded a weighted
mean of the temperatures of the refractory lining, the over-
fire combustion gases, and the gases issuing from the top of
the bed.
The second problem assumed to have been encountered, but never
proved explicitly, concerned the shielding of the thermocou-
ples by the inert material (tin cans). It was possible that
some of the thermocouples were covered by the inert material when
the bed was being built up prior to the start of the run; in
this case, it is conceivable that the affected thermocouples
did not respond as rapidly as the unscreened thermocouples
when the ignition front passed over them. In addition, it was
likely that some thermocouples were shielded during the run as
the inert material settled towards the grate. (This may, in part,
be the reason for the erratic behavior of thermocouple No. 2
in Figure 50).
The thermocouple history can, on the basis of the above dis-
cussion, be expected to give a reasonable representation of
the temperatures within the fuel bed up to the point when the
ignition plane reached the grate. After this time, the inter-
pretation of the temperatures becomes difficult because of the
uncertainty introduced by the bed movement.
From the temperature histories of the twelve fuel bed thermo-
couples, bed temperature profiles were constructed. Figure
51 shows the temperature profiles throughout the bed at dif-
ferent times during Run 18. The plot is typical of the re-
sults of those runs in which sustained active burning was
achieved. The heterogeneous nature of the fuel bed created
a rather poorly defined ignition front, but this restriction
aside, Figure 51 shows that an ignition front of approximately
constant shape propagated through the bed from top to bottom.
Near the end of the run, the bed temperatures increased (typ-
ically by 400-500 F)„ It is thought that this increase in
temperature occurred subsequent to the drying and pyrolysis
of the major portion of the bed (all but the last layer or so
of fuel particles), and did so because at this late stage of
the run more of the energy liberated by chemical reaction
was available to raise the sensible heat of the bed and less
was required to provide the energy requirements of drying and
pyrolysis.- This result indicates that the grate links of a
traveling-grate stoker could receive their major punishment
- 195 -
-------
towards the discharge end of the furnace; in practice the under-
fire air rate to the last grate is high and provides the neces-
sary convective cooling of the links. Some care needs to be
exercised in interpreting the profiles shown in Figure 51,
as the location of the top of the fuel bed at different times
was not determined experimentally- The profiles indicated
are strictly, therefore, composite profiles of the entire
fuel bed section measuring temperature both in and above the
bed. The initial and final depths of the fuel bed are indi-
cated in Figure 51; the final depth is based on measurement
of the volume occupied by the quantity of inert material pre-
sent at the end of the experiment. The rate of recession of
the top of the bed was not measured experimentally and any
theoretical estimate would have to depend on a somewhat crude
model of the mechanics of fuel burn-out. This question will
be reconsidered on pages 213-230. This drawback affects
neither the above discussion of the temperature profiles at
the ignition front nor that concerning the temperature increase
at the end of the run; however, the shortcoming does not permit
any reasonable estimate of the heat lost or gained by the bed
to be obtained from the profiles of Figure 51. From the earlier
discussion of the mechanics of the settling of the bed as it
burnt out, it may be hypothesized that under the conditions
of Run 18 (approximately equal burning and ignition rates) the
inert content probably accumulated above the burning zone as
it moved through the bed, insulating the bed from radiative
heating from the overfire section.
The slight temperature decline from the peak temperature at
the ignition front to that measured at the top of the bed
observed in this study over most of the duration of the runs
parallels the results of Nicholls (24). Nicholls1results also
indicated that the temperature decline in an underfeed bed
was rather less than observed in an overfeed bed. The de-
crease in temperature can be explained by the occurence of the
endothermic reactions C + C02 -> 2CO and C + H_O -> H2 + CO in
in this region of the bed. Figure 51 shows tnat the tempera-
ture drop through the bed increased near the end of the run,
probably due to the augmented rates of the endothermic reac-
tions as a consequence of the general temperature increase
through the bed at this time.
In the early experiments (Runs 1-10) it was noticed that the
temperatures at the side wall appeared to be higher than at
the center of the fuel bed. Although side wall temperatures
were not measured directly, this trend was deduced from the dis-
coloration of the thermocouple sheaths extending
for about two inches from the side wall towards the cen-
ter of the bed. The higher side wall temperatures were attri-
buted to channeling of the underfire air, resulting in a
localized heat release. For the later runs (Runs 11-18) ,
- 196 -
-------
L-
o
u
QL
15
DC
LJ
Q-
Z
Ld
1800 -
1600 -
1400
12OO -
Approximate Location
of Top of Bed at End
of Test
Top of Fuel
Bed at Start
of Test
16OO sees
V 1800 sees
® 2200 sees
o 2300 sees
© 2800 sees
A 3200 sees
1OOO —
200 -
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
DISTANCE ABOVE GRATE (IN)
Fig. 51. Fuel Bed Temperature Profiles at Different
Times throughout Run. Data from Run 18.
- 197 -
-------
where the improved high pressure drop grate was used, the occur-
rence of channeling as manifested by high side wall tempera-
tures was not noticed; this was confirmed by the two thermo-
couples used in these later experiments to measure the fuel
bed temperatures at two different vertical and angular loca-
tions about one inch from the side wall (see Appendix H, Fig-
ures H.I and H.2). These edge temperature measurements also
showed that the ignition front appeared to propagate evenly
across the width of the bed (see pages 210-211 , Figures 56
and 57).
Fuel Bed Temperature Distributions under Poor Ignition Con-
ditions!In those runs where difficulty was experienced in
igniting the bed at the start of the experiment, the bed tem-
perature profiles were markedly different from those shown in
Figure 50. Although this problem concerns mainly those runs
in the second group of experiments (namely, Runs 5-9), similar
experiences were found in Runs 12 and 13. The temperature pro-
files observed depended on the methods used in promoting ac-
tive burning. A discussion will follow on the bed tempera-
tures observed in Runs 5, 6, 7 and 9, as they represent the
spectrum of behavior.
The fuel bed temperature data for Run 5, shown in Figure 52,
indicated that the bed smoldered until enough of the mois-
ture content of the bed had been removed for satisfactory burn-
ing to proceed. This was shown by the initial rise in the first
6-7 in. of the bed to 900°-1400°F by 900 sec, followed by a
slow decrease in the temperature at this position in the bed
and a slow rise of the temperatures through most of the remain-
der of the bed. The temperatures in the bed about 6 in.
below the ignited portion of the bed plateaued around 200 F.
During this period (0-2800 sec) the underfire air had been
raised in an attempt to kindle the fire. After 2800 sec the
temperatures started to rise slowly throughout the bed with
the temperatures toward the top of the bed leading those
lower down. By 4000 sec the whole bed had ignited and the
temperatures throughout the bed steadily increased until they
reached a peak of 2000 F at 4600 sec. At this time combustion
was nearly completed and the temperatures started to fall. The
underfire air was decreased at 4000 sec because visually the
bed did not appear to be burning satisfactorily at this time.
The data indicated that burning would have proceeded satis-
factorily if the underfire air had been left at its previous
level of 145 Ib hr xft .
No satisfactory ignition of the bed could be obtained in Run
6 despite a number of underfire air changes and reignition of
the gas burners so as to increase the heat flux to the top
of the bed. Visual observation of the fire showed that it
- 198 -
-------
2000
I
H
VO
160O-
1200 -
cc
ID
o:
800 -
400 -
0
Fig.
5600
52.
Fuel Bed Temperatures (Run 5). Thermocouples 1, 2, 3, 4. 5, 6, 7, and 8 were located
25, 23, 19, 16, 11, 9, 7, and 2 in. from the grate. See Table VI.2 for experimental
conditions.
-------
1
w
o
I
Ld
CC
ID
t—
<
CK
U
CL
2
LJ
2400
2000(—
ieooh-
1200h
eool-
400h
Fig. 53,
400
800
2000
2400
2800
1200 1600
ELAPSED TIME (SECONDS)
Fuel Bed Temperatures (Run 7). Thermocouples 1, 2, 3, 4, 5, 6, 7, and 8 were located
23, 19, 18, 13, 11, 9, 7, and 0 in. from the grate. See Table VI.2 for experimental
conditions.
-------
I
o
I
1800 -
1600 -
14OO -
- 12OO-
LJ
CC
ID
I-
<
ce
LJ
Q.
5
LJ
100O -
800 -
600 -
400 -
' 2OO
400 8OO
12OO 14OO 2000 240O 280O 32OO 3600 4OOO 440O 48OO 5 2OO 56OO 6OOO
ELAPSED TIME (SECONDS)
Fig. 54. Fuel Bed Temperatures (Run 9)
Thermocouples 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 were located 25, 23,
19, 18, 13, 11, 9, 7, 2, and 0 inches from the grate. See Table 11,
for experimental conditions.
-------
appeared to be almost totally quenched about 25 min after
the run had started. Data taking was stopped at this time
and a complete record does not exist for this run. It was
observed that the bed continued to smolder for about 5 hr
after the underfire air was cut off, and almost complete burn-
out of the fuel was obtained. This result provided striking
evidence of the tendency for air to be drawn down through the
bed by natural draft. The bottom of the bed had been completely
sealed by the cfrate and the combustion air could only have
been supplied by air sinking down through the bed at the "cold"
wall. The burning rate under these conditions was very low.
The data for Run 7 depicted a different picture of the bed
temperatures,, which are shown in Figure 53. In the first
few inches of the bed, the temperatures rose quickly to well
above ignition temperature, but at depths below 10 in. the
temperatures rose slowly over a period of about 10 min to
1000 F. About halfway through the run (1500 sec) the whole bed
had reached a temperature of around 1000 F; from this time un-
til complete burn-out was achieved, the temperatures roseQ
slowly throughout the bed over a period of 10 min to 2200 F
(the maximum temperature observed, was achieved about
10 min before complete burn-out). The burning rate of the bed
increased raarkely shortly after the bed temperatures reached
their maxiim.ua and remained relatively constant until final
burn-out. The underfira air washeld-constant for the major
portion of this run at 146 Ib hr ft since, visually, the
bed appeared to be burning reasonably well.
For Run 9, yet another type of behavior was observed. The
bed temperature profiles for this experiment are shown in
Figure 54. The bed ignited rapidly down to 5 in. (about 400
sec,into the run) despite the underfire air rate of 275 Ib
hr ft , and then the temperatures in this ignited sec-
tion of the bed fell to around 600 F (at 1200 sec). The gas
burners were raignited at 1500 sec to keep the top section
refractories above 1400 F, and the underfire air was decreased
since it was visually observed at this time that there had
been poor ignition. From this time up to 2400 sec into the run,
the Jemperature in the top few inches of the bed rose slowly to
1800 F, after which the ignition plane started to move through
the bed until about two thirds of the bed was ignited. Vis-
ual observation showed that the bed appeared to be burning
properly at th|s time and the underfire air was raised to
175 Ib hr ft . The temperature data indicated that the
underfire air rate must have been increased too quickly, for
at 3000 sec the bed temperatures in the ignited portion of
the ^ed began to decrease, dropping to a minimum of 800°-
1400 F at 4400 sec. The temperature profile was still pro-
pagating into the virgin fuel at this time but temperatures
- 202 -
-------
did not reach the ignition level. The underfire air was again
decreased and at 4400 sec the temperatures in the top two
thirds of the bed began to increase. The remainder of the bed
began to ignite, slowly at first and then more rapidly. The
bed was fully ignited at 5000 sec, at wbich time the under-
fire air was again raised to 175 Ib hr^ff2 since the bed
appeared to be burning satisfactorily. The burning rate was
observed to increase in the later stages of the run but not
to the extent noticed in Run 4.
Drying Front. As discussed on pages 192-208 , the drying rate
of a fuel element is limited by the rate of transfer of
energy from the burning front and by the rate of conduction
of energy in the object being dried. When the latter resis-
tance is low, for example in thin or porous objects, the tem-
perature will plateau at 212°F and the drying wave will pro-
pagate ahead of the ignition wave. This was the case in the
Bureau of Mines study (2J>_,2_7_) , where the moisture was intro-
duced with leafy vegetable material. In the present study
the moisture was introduced in blocks of wood, in sizes such
that the heating of the center of a block lagged behind the
surface heating by times of the order of 1000 sec. For this
case only a slight plateau was observed when the temperature
measured within the bed reached 212 F and drying continued
within the wood blocks even after the surface had ignited.
The peak temperatures within the fuel bed observed toward
the end of a run probably occurred subsequent to the dry-
ing of the entire bed.
Additional evidence for the protracted drying of elements of
the fuel bed is provided by the data from runs in which dif-
ficulty was experienced igniting the bed (e.g., Run 5). In
this case the temperature wave from the ignition front dis-
sipated slowly enough that marked plateaus were observed when
the temperature ahead of the ignited zone reached the vicin-
ity of 212 F (see Figure 52). Under normal burning conditions
this 212 F plateau would not be expected because of the much
higher heating rates to which the fuel bed elements would be
subjected (on pages 192-208 ) . Under the normal burning con-
ditions encountered in this study, the quenching effect of the
wet wood blocks, where the internal diffusion of water was
low relative to the heating rate, would not have been a hin-
drance to ignition and combustion.
Pyrolysis Front. In a number of runs a pl?.teau at about 500 F
was observed, corresponding roughly to the temperature of py-
rolysis of cellulosic materials. It is difficult to deter-
mine if this was a real effect or was caused by temporary
shielding of the thermocouples by inert material. The heat
- 203 -
-------
effects of pyrolysis will generally be rather lower than
the drying heat effects and at the higher fuel bed heating
rates it is unlikely that the pyrolysis "wave" would have
been noticed. However, as in the case of drying, the pyroly-
sis of larger wood blocks was limited by the rate of pro-
pagation of energy into the center of the blocks. Pyrolysis
would therefore have been initiated as the surface tempera-
ture of a block approached 500 F and continued internally
within the block for extended periods.
Temperatures at the ignition Front. The maximum temperatures
observed at the ignition front for the different runs in
which it was possible to obtain reasonable values are shown
in Table 12,-there was considerable scatter in the data, and
selecting a range of values was often arbitrary. There was
surprisingly little variation in the ignition front tempera-
tures despite fairly considerable changes in the underfire
air flow rate (86-175 hr~lft~2) and the moisture content of
the fuel (15-32%). The lowest temperatures observed occurred
in Run 1, but these temperatures were quite possibly erroneous
as they seem very low for the type of active burning usually
observed. It appeared that the lower temperatures were asso-
ciated with the lower air flow rates. A simple explanation
for this is the greater effect of heat losses at these lower
heat generation rates. There were no distinct trends with
increasing moisture content although the theory developed ear-
lier (a simple overall heat balance) suggested that the maxi-
mum temperature should decrease with moisture content, accor-
ding to
net
T = s (54)
max GC + UC
P P
rg *s
This result may be attributed to the drying characteristics
of the material used. The wood was known to dry over a pro-
tracted period of time, and the amount of moisture released
up to the point when the temperature at the ignition front
reached a maximum may have remained roughly constant, regard-
less of the moisture content. For material where the inter-
nal diffusion of water is high the effect of moisture on the
maximum temperature will be much more pronounced. This will
have a strong adverse effect on the ignition rate.
On the basis of equation (54) , the maximum temperature for Run
15, would have been expected to be lower than in Runs 11 and
12 because of the higher ignition rate (based on the rate of
- 204 -
-------
TABLE 12
Maximum Temperature Observed at Ignition Front with
Varying. Underfire Air Rates and Fuel Compositions
Run
No.
1
10
11
12
14
15
16
17
18
Fuel Analysis
» Inert
-
17.10
15.51
12.79
15.08
-
16.23
15.18
15.04
% Moisture
8
31.76
23.99
25.89
14.73
26.40
24.85
25.38
25.06
% Com-
bustible
92
51.14
60.50
61.32
70.19
73.60
58.92
59.49
59.90
Underfire
Air Rate
(Ib hr~1ft~2)
Natural Draft
175
155
155
160
156
118
140
86
Maximum Temper-
ature Observed at
Ignition Front <°F)
1100 - 1400
1600 - 1800*
1800 - 2000
1800 - 2000
1800 - 2000*
1800 - 2200
1800 - 2100
1600 - 2000
1600 1800*
1600 - 2200
1700 1900*
1600 - 1900
1500 - 1700*
1600 - 1900
(a) Observed in top two thirds of bed
(b) Observed in bottom third of bed
- 205 -
-------
ignition of both fuel and combustible). The range of maxi-
mum temperatures was observed to be slightly lower than in
Runs 11 and 12 but the difference was hardly significant.
Temperature Profiles Below Ignition Front. From the temper-
ature history of the different thermocouples in the bed and
a knowledge of the ignition rate, the fuel bed temperature
could be plotted as a function of distance below the igni-
tion plane. The results for one test were not all the same
because of the heterogeneous nature of the bed surrounding
the thermocouples;however, it was possible to construct an
average curve for the temperature-distance profile. These
profiles are shown in Figure 55.
From equation 58 the temperature below the ignition plane is
given by
- (a + 3)z
ign
(58)
where
a =
6 =
Under most practical conditions h - 200 Btu hr ft F ,
~1~1°~1 ~lo~1
- 4 Btu hr
pg
= 0.5 Btu lb
50 Ib hr'1 ft~2 and G = 100 Ib hr~1ft~2
U *
giving
T ^ T. e
~ ign
K
v
E
(170)
Equation (170) predicts that as the volumetric heat transfer
coefficient, h , igcreases or the effective thermal conducti-
vity of the bed, K , decreases the profile below the ignition
front will become steeper; conversely, as the heat transfer
coefficient decreases or the effective thermal conductivity
increases, the profile below the ignition front will become
shallower. This prediction is confirmed by the results of
Runs 18, 16 and 11 shown in Figure 55, in which progressively
- 206 -
-------
M
O
•z.
O
CC
U.
Z
CD
LJ
CL
Ct
UJ
Q_
Z
LJ
600
500
= 40O
O
Q 300
Run 16
Run 18
Run
No
1
Underfire
Air Rate
Lb HrVt?
1
Ignition
Rate
Lb Hr Ft2
1
Fuel
Composition
200
100
0
I
0
Fig. 55.
0.1
DISTANCE
0.2 0.3
BELOW IGNITION FRONT (FT)
Fuel Bed Temperatures Below the Ignition Front for Different Fuel
Types and Underfire Air Rates
-------
greater underfire air flow rates were used (h scales approxi-
mately as G0-63). The temperature profiles of Run 15 also
follow the trends indicated by equation (170). In this run
the absence of the inert material (the components of which were
slightly larger than those of the combustible portion of the
fuel) very likely decreased the effective mean beam radiating
length through the bed (see pages 198-203 ) and thus decreased
the effective thermal conductivity.
Using the heat transfer correlation of Bradshaw fit &1. (64)
and parameter values as follows: Cp - 0.4 Btu Ib"10?"1 ,
N2/3 = 0.81, y = 0.1 Ib hr^ft"1, D - 0.125 ft, 6 = 0.4,
the heat transfer coefficient based on a unit volume of bed
is found to be, = 10<460.625 -1-30 Figure
55 shows theoretical curves, calculated fgom equation (170)
using the above relationship for h and K^ = 4 Btu hr-lft"1
F'1 (see pages 198-203 } , for two different values of G
corresponding to conditions for Runs 18 and 11. The theore-
tical profiles are of the correct order but the variation with
changing underfire air rate is not as large as that observed
experimentally. However, the theory is only approximate and
the assumptions of constancy of effective thermal conductivity
and particle isothrrmality, which are known not to be strictly
true for many fuel bed conditions, will mitigate against
precise prediction of the temperature profile below the igni-
tion front.
From the results of the theory presented above, the tempera-
ture profiles of Run 14 should have been close to those of
Run 11 as the only difference between the two runs lay in
the moisture content of the fuel. The fact that the data
fell between those obtained for Runs 16 and 18 is anomalous
and no satisfactory explanation can be offered.
Ignition Rates
The ignition of the bed can be conveniently divided into two
stages — the initial ignition of the bed at the start of the
run by the radiant heat from the hot refractory brickwork and
the propagation of the ignition front through the bed. For
all the runs performed, the fuel bed ignited at the start of
the run, even when the top surface contained more than a pro-
portionate amount of wet wood blocks, based on the total mois-
ture content of the fuel. The initial radiant heat fluxes
- 208 -
-------
to the top of the bed were of the order 10 Btu hr ft ,
very much greater than required for ignition, even for blocks
of wood containing 60% moisture (117) , which was the maximum
moisture content of any of the wet wood blocks used in this
study.
The major difficulty with the ignition process lay in adjust-
ing the underfire air rate so that the ignition front would
propagate through the bed; this has been described on pages
198-203 for the cases of Runs 5-7 and 9. At the time these
runs were conducted, it was felt that burning rates typical-
ly observed in a traveling-grate incinerator (60 Ib hr ft~2)
could be obtained with underfire air rates around 150 Ib hr~l
ft~2 and it was thought that at these underfire air rates
the burning rate would be close to its maximum. It was
for this reason that underfire air rates varying from 120 to
275 Ib hr ft~2 were employed. The range of underfire air
rates was weighted to rates higher than 150 Ib hr-1ft~2 be-
cause of the desire to map out the ignition-limited regime
of burning.
Runs 5-9 clearly showed that raising the underfire air too
quickly at the beginning of a run (on a traveling grate this
would be equivalent of having too high an underfire air rate
through the first few windboxes) seriously hampered the igni-
tion rate; this behavior was not observed by Nicholls (24)
in his studies on coke and coal as in his experiments tEe"
underfire air was raised to its final level very quickly
after the fuel bed had been ignited. In addition, once the
ignition front had propagated a few inches into the bed, the
effect of changing the radiant heat flux to the top of the bed
was found to be minimal because of the insulating effect of
the inert and charred fuel. All the energy required to drive
the ignition front forward had to come solely from the heat
liberated by the reaction of oxygen with the char and pyrolv-
sis products in the partially ignited bed. Once the tempera-
tures in the ignited zone dropped low enough to provide a
significant chemical resistance to combustion, it was very dif-
ficult to increase the heat generation rate to get the bed
back into an active burning state because of heat losses to
the cold fuel bed wall and the unignited fuel.
As discussed on pages 192-208, ignition temperatures differ
between materials and with experimental conditions. For cel-
luosic materials, piloted ignition occurs at temperatures of
550° to 600°F. In this study, a temperature of 600 F was se-
lected for the derivation of curves, such as that shown in
Figures 56 and 57 for the propagation of the ignition front.
In active burning runs the temperature gradient near the ig-
nition front was sufficiently steep that the values of the
- 209 -
-------
26
LJ
X
(J
(L
UJ
O
2O
16
12
8
4
2000 3OOO
ELAPSED TIME, SECONDS
4OOO
Fig. 56. Ignition Wave Propagation (Run 10). Ignition
temperature assumed equal to 600° F; underfire
air flow 175 Ib hr"1 ft"2; bulk density of fuel
16.7 Ib ft""3; initial bed depth 30 in. $ meas-
ured at center of bed; gj measured at edge
of bed.
-------
I
ro
D Measured at Edge
of Bed
O Measured at Center
of Bed
Measured at Edge
of Bed
Measured at Center
of Bed
u 20 —
en
(D
UJ
>
O
CD
UJ
U
(f)
Q
0
1000 20OO 3OOO
ELAPSED TIME (SECONDS)
Fig. 57. Ignition Wave Propagation (Runs 17 and 18)
4OOO
-------
ignition rate were found to be insensitive to the value of
ignition temperature selected. The ignition rates were com-
puted as the product of the rate of travel of the ignition front
and the initial bed density. For active burning runs the rate
of travel of the ignition front was found to remain approxi-
mately constant over the duration of the experiment; the varia-
tion of the data shown in Figures 56 and 57 was typical of the
deviation from linearity in the plot of depth of ignition
front penetration versus elapsed time. Ignition rates were
found to vary from 125 Ib hr'^-ft"2 for an underfire air rate
of 127 Ib hr~1ft~ (Run 5) to 34 Ib hr-1ft" for an underfire
air rate of 175 Ib hr~1ft~2 (Run 10). A summary of the igni-
tion rates found in the experiments, the underfire air rates,
and fuel compositions in given in Table 51.
The high ignition rates observed in the later stages of the
runs where difficulty had been experienced in igniting the
bed (e.g., Runs 5, 7 and 9) were undoubtedly much too high
for the initial fuel composition employed. This descrepancy
occurred because,, at the times when these ignition rates
were calculated, the bed had been partially dried (see page
203 and Figures 52, 53 and 54). The high ignition rate ob-
served in Run 1 (110 Ib hr~-*-ft~2) can be attributed to the
low moisture content of the fuel. This ignition rate was pro-
bably close to the maximum achievable for the inert-free
fuel used in this study- As clearly demonstrated in Nicholls1
(24) experiments arid by the embryonic ignition theory developed
on pages 118-129, the ignition rate would be expected to pass
through a maximum as the underfire air rate is increased.
It is hypothesized that under natural draft conditions the
underfire air rate increases to a level which gives an igni-
tion rate close to its maximum. If this hypothesis is cor-
rect it lends credence to the belief that the ignition rate
for Run 5 was measured for a fuel which had been almost com-
pletely dried. For Runs 11 through 18, where the calculated
ignition rates were more truly representative of the initial
fuel composition and of the conditions of the test, the ob-
served ignition rates lay in the range 34-50 Ib hr~1ft~2.
The ignition rates calculated for Runs 10, 11, 16, 17 and 18
which were all conducted with a fuel of approximately constant
composition, show (with the exception of Run 11) a monotonic
.increase with decreasing underfire air rate (see Table 11).
This result follows the general trend in ignition rates ob-
served by Nicholls (24) in the equilibrium burning regime (e-
qual ignition and burning rates). It may be postulated, there-
fore, that if underfire air rates below 86 Ib hr~1ft"2 had
been employed the ignition rate would have been observed to
reach a maximum and then to decrease as the underfire air
— • O 1 *>
-------
was further decreased. The ignition behavior observed in Run
11 is anomalous and no satisfactory explanation can be offered.
More experiments are required to extend the range of under-
fire air rates and to duplicate the results presented here
before a more quantitative picture of the ignition behavior
can be drawn.
Camparison of the ignition rates obtained in Runs 11, 14 and
15 indicates that fuel composition changes had very little
effect on the ignition rates. Within the limits of experi-
mental error in determining the ignition rates, the rates
were the same for the base case composition (15% inert, 25%
moisture and 60% combustible) and for the cases where the inert
and moisture content were both 15%, and where no inert was
present and the moisture content was 25%. This is a some-
what surprising result, as decreasing both the inert and
moisture fraction of the fuel would be expected to enhance
ignition propagation. Further experiments are needed cover-
ing a wider range of inert (0-30%) and moisture 0-40%) con-
tents before trends can be determined.
It is thought that the maximum underfire flow rate at which
ignition will still be sustained for the typeof fuel used
in this study is in the vicinity of 275 Ib hr ft". In Run
10 the underfire air rate was increased over a period of around
2 min to the 275 Ib hr-1ft~2 level only after the bed had ig-
nited and started to burn vigorously- About 10 min after
the air rate had reached 275 Ib hr~^-ft~2 the bed appeared to
have been almost totally quenched. The underfire air rate
was then reduced to 175 Ib hr-1ft~2 and visual observation
of the fuel bed indicated that, within the space of a few
minutes, the ignition and burning processes had started again.
This behavior was borne out by the fuel bed temperature pro-
files, the first thermocouple (5 in. from the top of the bed)
not reaching the ignition temperature until after 900 sec
into the run; typically the time for the ignition plane to
reach this level was around 300-400 sec. Similar experiences
with early quenching were obtained from Run 9, where the un-
derfire air rate was also increased to 275 Ib hr~1ft""1^ over
the first few minutes of the run.
Burning Rates
Plots of weight loss of the fuel as a function of elapsed
time for Runs 10, 17 and 18 are shown in Figures 58, 59 and
60. These plots were obtained using the load cell weighing
- 213 -
-------
n 4O
CO
O
I—
X
O
LU
2O
1O
1OOO 2OOO 3OOO 4OOO
ELAPSED TIME , SECONDS
5OOO 6OOO
Fig. 58. Weight Loss as a Function of time (Run 10)
Underfire air rate 175 Ib hr"1 ft"2 and
bulk density of the fuel 16.7 Ib ft"3;
initial fuel bed depth 30 in.
- 214 -
-------
tn
I
Burning Rate Over Constant
Burning Period = 39 Ibs./Hr"1 Ft
4OO 8OO 12OO
1600 2OOO 24OO 28OO
TIME IN SECONDS
3200 36OO 4OOO 44OO
Fig. 59. Weight Loss as a Function of Time (Run 17)
Underfire air rate 140 Ib hr"1 ft"1; bulk density of fuel
15 Ib ff3; initial fuel bed depth 30 in.
-------
CD
60
50
40
to
to
P 30
20
ui
10
Burning Rate
30 Lb HrVt"2
11 I
-1 -2
Burning Rate = 41 Lb Hr Ft
I I I I
I I I I I
1000
2OOO 3000
TIME (SEC)
40OO
Fig. 60. Weight Loss as a Function of Time (Run 18)
Underfire Air Rate 86 Ib hr"1 ft~2, bulk density of fuel 16 Ib ft" ,
initial bed depth 30 in.
-------
system and are representative of the burning rate curves
found for those runs where active burning was observed for
the major portion of the test. These weight loss curves
have been smoothed; a representative unsmoothed curve is shown
in Figure H.7 in Appendix H. For Runs 17 and 18 independent
checks on the burning rates were available from the material
balance calculations (see pages 230-237 ) . The slight offset
in the weight loss curve at zero time in Run 17 was caused
by slight differences in the recorded weight of the fuel bed
section before and after it was butted up to the top section
of the apparatus. This offset did not affect the weight loss
results and the "weight loss" of 1.5 Ib at time zero may
simply be taken as the origin for the ordinate. The burning
rates were obtained for the different runs by taking the slopes
of the weight loss curve during the time periods that the burn-
ing rates were constant.
For Run 14 no load cell data were available because of equip-
ment difficulties, but the burning rate was estimated from
closure of material balances at different times during the
run. Since no accurate measurements were available for the
stack gas water content and the air leakage rate was not mea-
sured, the material balance was forced to close by selec-
ting air leakage rates and stack gas moisture content correc-
tion factors that closed the carbon and hydrogen balances.
The burning rates for Runs 15 and 16 were estimated by assum-
ing that almost complete combustion had been achieved by the
time the ignition front had reached the grate and by further
assuming that the burning rate was constant over this period
of time. This estimate is expected to be a little low. For
comparison, the same technique was used to,calculate the burn-
ing rate for Run 14j the result, 43 Ib hr ft" , agreed close-
ly with that calculated from the material balances.
The burning rates for all the runs in which representative
values could be obtained are shown as a function of underfire
air rate in Figure 61,- for comparison, the ignition rates cal-
culated-for these runs (on the basis of Ib of combustible
hr ft ) are also indicated. For burning rates lying on the
solid line AB, the underfire air and fuel consumption rates
were in stolchiometric proportions for a fuel of a composition
approximating the average fuel composition used in Runs 11-
13 and 16-18. Although the rates were in stoichiometric pro-
portions the compositions of the gases leaving the top of the
bed were not restricted to being only CO- and H20; the off-
gases could have contained any mixture of 0~, CO, CO_, H_,
H_0, N« and trace pyrolysis products as long as the 6- and
combustibles were in stoichiometric proportions. For burning
rates to the left of this line, secondary air was required for
complete combustion; for burning rates to the right of
- 217 -
-------
AIR DEFICIENT (PERCENT)
'cc
i
ui
CD
2
O
u
CO
LU
(-
<
CC
z
O
I—
I/)
3
CO
z
O
CJ
ce
O
z
O
75
,50
.25
75
50
25
75
50
25
/
80
70
60
50
D Burning Rate
O'Q^ition Rate
^Burning Rate Estimated
From CO and H2 Cone. At
Top of Bed stoichiometric Line
Runs 11 -13 , 16-18
— Stoichiometric Line
Run 10
30 -
2O —
10 —
/50
/ 25
/°/
Air
Deficient
/25
/
25
/
25
/25
(2) (3)
m
x
O
m
x
o
m
z
Combustible
30
56
25
60
20
65
15
70
Stoichiometic Line Based on the
Following Analysis of Combustibles
( Wt "/„ )
Stoichiometric Line Run 14 C-042
H = 0.06
O= 0.52
-Stoichiometric Line Run 15
_J I I L_
O 100 2OO 3OO
-1 -2
UNDERFIRE AIR RATE ( LB HR FT )
Fig. 61. Ignition and Burning Rates as a Function
of Underfire Air Rate
- 218 -
-------
the line, excess oxygen passed through the bed and ftccoft4*ry
air was not strictly required if adequate mixing wa» Achieved
above the bed. The scales on the top and sides of Fi>
the burning bed depth was not sufficient for all >th
-------
This finding was confirmed by an estimation of the burning
rate from the air deficiency calculated from the measured H2
and CO concentrations within the fuel bed; this ppint is shown
as a triangle in Figure 61.
For Run 18, the carbon burning rates calculated from fuel bed
and stack gas measurements gave a different picture. From
the material balance and the stack gas C02 content, an aver-
age carbon burning rate (from 2400 - 3200 sec) of 1.4 Ib
moles hr-lft-2 was calculated,, while the carbon burning rate
estimated from the fuel bed gas composition was about 1.45
Ib moles hr-lff2. In this run, therefore, channeling did
not appear to be a problem. Further confirmation that no
channeling occurred came from an estimation of the burning
rate by calculating the air deficiency from the fuel bed gas
compositions. The point shown as the triangle in Figure 61
was calculated from the gas compositions obtained from the top-
most fuel bed probe at 2800 sec, at which time it was esti-
mated that the probe was close to the top of the bed. The re-
sult agreed closely with the burning rate obtained from the
load cell, which is an average rate for the whole bed. Nic-
holl's data (24) indicated that in his experiments channeling
became more serious as the underfire air rate was increased.
The results of this study indicated at similar trend.
For all runs the early burning rates increased as the run
progressed. The shape of weight loss/time plots took on two
basic forms as shown by Run 17, Figure 59 and Runs 10 and 18,
Figures 58 and 60. For Run 17, the burning rate slowly in-
creased with time until a constant rate was attained which
was then maintained up to about the time the ignition front
reached the grate. Similar behavior was seen in Runs 1 and
2. For Run 18 the burning rate started at a constant level
and about halfway through the run suddenly increased to ano-
ther constant level. This sudden break in the burning rate
was confirmed by the material balance calculations. Similar
trends were noticed in Run 10, but the burning rate appeared
not to make such an abrupt change.
For Runs 10, 14, and 15-18, the data showed that, during the
early stages of these runs, the burning rates were less than
the ignition rate of the combustible content of the fuel,1
while during the later stages the burning rates were greater
than the ignition rates. These findings were in marked con-
trast to Nicholls' (24) results, where equilibrium burning
(equal ignition and burning rates) was observed at high under-
fire rates over the major portion of an underfeed burning
test. Nicholls did not report any data on the developing thick-
ness of the burning zone as burning rates were only measured
(from stack gas compositions) after equilibrium burning was
established. The results presented here, by contrast, were taken
-220-
-------
during the period when the depth of the active fuel bed was
developing. It is obvious that a situation where the igni-
tion rate is lower than the combustion rate is unstable
since under these conditions, the active bed thickness will
decrease until the burning and ignition rates equilibrate.
Before discussing these trends it will be worthwhile to indi-
cate exactly what constituted the measured burning rate. The
burning rate was made up of contributions from the drying and
pyrolysis (or carbonization) of the combustible, char oxi-
dation (C + 1/2O2 + CO and C + 02 -»• CC>2) and char gasification
(C + H20 -» H2 + CO and C + CO^ -> 2 CO) throughout a depth
slightly greater than the ignited zone (because some drying
and to a lesser extent pyrolysis takes place ahead of the
ignition front). Vaporization of metal coating and oxidation
of steel are small enough to be neglected.
The observed increase in the burning rate as the runs pro-
gressed was in accord with the development of a burning depth.
As the distance between the ignition front and the top of the
bed increased, more time was available for the various gas
phase and gas-solid reactions to occur. In addition, more
depth was provided in which drying and pyrolysis could take
place. (As indicated above, the observed burning rate was
the sum of the weight losses associated with drying, pyrolysis
and solid char consumption.)
As the burning zone developed, an increasing amount of the oxy-
gen in the underfire air was consumed and in most experiments
the burning zone eventually increased to the extent that all
the oxygen was consumed. The depth required depended on the
rate of mass transfer of the oxygen to the char surface and
the rate of reaction of the oxygen with pyrolysis products.
The oxygen consumption depth may be theoretically calculated
from the correlation presented on pages 61 - 75, Figure 19,
although the calculation must only be considered approximate.
For this study the Reynolds numbers DpG/u used lay in the range
80-200 and from Figure 19 the "effective" jD factor was there-
fore about 0.15-0.20= On the assumption that the effective
mass transfer area was 30 hr~2ff3, the oxygen consumption dis-
tance (z0), where the oxygen concentration fell to 0.1%, was
of the order of 0.5 ft. This value of ZQ is of the same
order of magnitude as that observed experimentally - see pages
75 - 82 • As the depth of the burning zone increased, and
more oxygen was consumed, the burning rate would be expected
to increase partly as a result of increased char oxidation
and partly as a consequence of the greater heat release
within the bed. (From the analysis on pagesm increas-
ing the underfire air rate increases ZQ (zo G°-41) and
-221-
-------
therefore low underfire air rates are preferable while the
burning depth develops.)
Once enough depth had become available for all the oxygen to
be consumed, further increases in depth provided a region in
which the gasification reactions could proceed. The degree
to which the gasification reactions occurred depended on there
being adequate carbon left for reaction and high enough tem-
peratures in the fuel bed (ca. 2000 F) for the reactions to
have proceeded at reasonable rates. Burning rates in this
regime would, on the basis of the theory of Niessen et al.
(19) , and on the assumption of no ignition rate limitations,
be~~expected to increase proportionately with increases in the
underfire air rate and would increase with decreasing heat
losses from the bed (e.g., side wall losses or radiant heat
losses from the top of the bed). The burning rate will obvious-
ly be dependant on the amount of channeling that takes placef
but an effective burning rate could be calculated on the hy-
pothesis that there are regions of the bed that act as gasi-
fyers and regions where bypass of the underfire air occurs.
For the fuel particles used in this study, where there was
a major resistance to heat transfer within the elements, the
weight loss that was associated with the drying and pyroly-
sis of each of the combustible fuel elements would have de-
pended on the size and moisture content of the element and
the length of time it had been in the ignited zone.
Towards the end of a run, after the ignition front had reached
the grate the thickness of the burning zone started to de-
crease. Eventually the burning zone became thin enough for
oxygen to break through the bed; the' depth at which this occurred
can be estimated using the methods outlined above for calcul-
lating z , but the effective mass transfer coefficient would
be expected to be rather less in this case because of the lo-
wer probability of any reaction between oxygen and pyrolysis
products. The rate of oxygen breakthrough and the burning
rates at this stage in the combustion process may be esti-
mated as follows. Assume that the oxygen concentration at
the top fo the bed can be given by
-k
f " m T 1
= 0.21 exp -^51 z^t) J (171)
T
where PQ (t) is the mole fraction of oxygen at the top of the
bed; km the effective mass transfer coefficient, lb hr~1ft~3;
G the underfire air flow rate, lb hr-ift'2; and zT(t) the height
of the bed at any time t, ft. The burning rate, assuming a
fuel comprising only carbon and inerts and that the reaction
- 222 -
-------
C02 predominates, is then given by
Burning Rate = i| G 0.21 - P£ (t)
(172)
The rate of recession of the top of the bed can be approxi-
mately expressed as,
-Ap
12 -
2~9" G
6 To.
21 -
(t)
(173)
where Ap is the density difference between the carbon and
inert fuel and the inert alone. Substituting equation (172)
into equation (173) , T integrating and substituting the re-
sulting equation for z as a function of time back into equa-
tion (171) gives, with some approximation in the final an-
swer, which becomes more correct as complete burn-out is
achieved.
= o.21 exp
-k
m
2G
log!1-expj
T
-Vo)
G j
Yl-0.0869Gt/l
/J Ap
(174)
(J
where ZQ is the height of the bed when oxygen just starts to
break through; since the oxygen concentration never reaches
zero using the representation of equation (171), ZQ is taken as
when the oxygen concentration reaches an arbitrarily small
number. From equation (174) the time required for the oxygen
concentration to reach 21% after incipient breakthrough (t )
is given by
t «
T1 0.0869G
(175)
-3 T
For typical values of Ap - 3 Ib ft , z - 0.7 ft , G
- 100 Ib hr-lff2 and k - 560 Ib hr-lff3 , t is approxi-
mately of the order of the1 times experimentally observed for
the burning rates to decrease to zero from their "steady
state" values prior to oxygen breakthrough (Figures 58 - 60)
and for the times it took the fuel bed oxygen concentrations
to increase from 0% to 21% at the end of a test — see pages
237-253 ; from the assumptions involved in this simplistic
theory, the estimate of t would be expected to be too high.
- 223 -
-------
(The above analysis indicates that underfire air rates should
be decreased once the bed thickness has decreased to the point
tht oxygen breakthrough occurs, for the following reasons:
firstly decreasing the underfire air will decrease the excess
air above the bed and secondly will help keep bed temperatures
high and will, therefore, prevent premature quenching of the
bed.)
The rate at which the active fuel bed depth (defined as the
distance from the ignition front to the plane of zero combust-
ible) increased was not strictly proportional to the differ-
ence between the ignition and burning rates as the measured
weight loss (somewhat loosely called the burning rate here)
included, as mentioned above, weight loss that did not affect
the depth of the active bed (i.e., drying and pyrolysis).
A distinction must be drawn here between the active fuel bed
depth as defined above and the distance from the ignition
front to the top of the bed. This latter distance will,
under certain circumstances, be rather greater than the former
because of the accumulation of inert material as the bed burns.
Inert accumulation will depend on the relative rates of igni-
tion and burning and has already been qualitatively discussed
in regard to the work of Marskell and Miller (94) and on pages
192-198 . The inert content at the top of the~^ed insulated
the burning zone from the radiant heat flux from the overfire
flame and the hot refractory brickwork of the top section and
helped occlude the oxygen present in the overfire section
from the combustible in the bed. Presumably gas phase reac-
tions took place in the void spaces in this "dead zone" of
the bed; for instance, the oxygen in the underfire air could
have reacted with CO, H9, and pyrolysis products in this region
of the bed. ^
No experimental determinations were made of the depth of either
the active fuel bed or the accumulated inert. However, an
approximate calculation was used to provide an estimate of
the depth of the active burning zone and the rate of accumu-
lation of inert above this zone; these calculations will be
discussed below. The depth of the active fuel bed can only
be calculated reasonable accurately if it is assumed that no
volume change of the fuel particles takes place during the
time required to completely dry and pyrolyze them; and if
it is possible to estimate from the fuel bed gas composition
what fraction of their carbon content can be attributed to
char combustion. For a fuel, such as that used in this study
with a large volatile content (- 75%) in comparison to fixed
char content (= 25%); this latter estimation is particularly
difficult. Paced with these difficulties more approximate
methods were employed.
Figures 62 and 63 show the calculated depth of the active
224 -
-------
NJ
ro
U
\-
<
o:
(D
LJ
>
o
CO
LJ
U
Ul
Q
Depth of Accumulated
Inert
Location of
ition Front
Location of Top
of Bed
Active Burning
Depth
0.8 —
0.4 —
0
1000
2000 3000
TIME ( SEC.)
4000
Fig. 62. Estimate of Fuel Bed Thickness and Accumulated Inert Layer
(Run 17)
-------
to
K>
UJ
I-
<
cr
LU
O
CD
LJ
U
Z
<
—
Q
I I
Location of Top of Bed
Location of
— Ignition Front
th of Accumulated
n
-------
burning zone and the depth of accumulated inert over the course
of Runs 17 and 18, respectively. For these calculations it
was assumed that the fuel maintained its original density (po)
until it was completely consumed, at which time its density
dropped to that of the inert material (pj - 9 Ib ft"3) . For
the case where the density of the active fuel bed varies
linearly from pQ to pj over the depth of active burning zone
and where the inert fraction within this zone is constant,
active burning depths about 30% greater than those shown in
Figures 62 and 63 would have been calculated. For an alter-
native case, where the inert fraction (X^) changes as the
density changes, if it were assumed that the desnity of the non-
inerts (i.e., wood blocks) remains unchanged during the course
of combustion, the calculated active burning depths would
have been about twice those shown in Figures 62 and 63. For
this case the depth of the inert layer (where xj - 1) would
be much less than those indicated for early times in Figure
62 and 63. These models are expected to provide reasonable
bounds on the active burning depth.
Figures 62 and 63 indicate how the depth of the active burning
layer changed with variations in the burning and ignition
rates. Figure 62 indicates that because of the varying igni-
tion rate, the depth of the active fuel bed changed markedly
during the course of the experiment.
For Run 18, Figure 63 shows that the active fuel bed depth
changed gradually over the course of the test? the depth
reached a maximum at the time the burning rate increased.
In addition, a comparison between Figures 62 and 63 indicates
that the active fuel bed depth in Run 17 was greater than
that for Run 18. This result coupled with the fact that the
underfire air rate used in Run 17 was greater than that used
in Run 18 suggests that the burning rate for Run 17 should
have been greater than that for Run 18. The observed aver-
age burning rate for Run 17 (39 Ib combustible hr-1ft~2)
was actually lower than that observed in the later stages of
Run 18 (41 Ib combustible hr-lft~2). This apparent anomaly
was caused by channeling in Run 17 and the burning rates
based on the gas measurements at the centerof the bed for
this test were rather higher (53 Ib combustible hr~lft~2)
than the average burning rate.
Figures 64 and 65 present ignition rates and burning rates
obtained from both the load cell and material balances (see
pages 192-198) for Runs 17 and 18 respectively- The varying
depth of the burning zone can be deduced from the different
areas under the burning and ignition rate curves.
The increase in the burning rate in Runs 10 and 18 cannot be
- 227 -
-------
K)
K)
00
80
40
ICC
I
HI
_J
UJ
30
o:
O
z
z 2O
cr
ID
CD
LL
O
10 -
Burning
Rate
(Load Cell )
Burning Rate
( Material Balance)
I
10OO
20OO 30OO
TIME ( SEC)
1 ^1
40OO
Fig. 64. Burning and Ignition Rates (Run 17)
-------
KJ
NJ
VO
o
(D
Burning Rate
Load Cell)
Burning Rate
/(Material Balance)
1O 00 2OOO 3OOO
TIME ( SEC.)
Fig. 65. Burning and Ignition Rates (Run 18)
4OOO
5OOO
-------
explained strictly by the increase in fuel bed temperature
observed towards the end of each run, as originally proposed
by Rogers, Sarofim and Howard (140)> since the increase in
temperature always followed the increase in burning rate.
No satisfactory explanation can be offered at this time to
explain these results,, It is likely that the degree of chan-
neling changes as the bed burns and will give rise to
different burning rates. Local changes in the way the fuel
and inert were charged to the bed may also have affected the
degree of channeling. This effect was probably aggravated
by the larger size of the inert with respect to the fuel but
could not be readily altered as the tin cans used were the
closest to the fuel in size that were easily obtainable in
bulk quantities. However, these effects would need to be
very non-linear to explain the magnitude of the changes in
burning rate observed. Although they probably did affect
the burning rate to a certain extent it is very unlikely
that they were the main cause of the observed increase. It
is also possible that the burning rates changed because of
changes in the rate of heat loss from the bed (e.g., by accu-
mulation of inert on top of the bed) but this cannot be exper-
imentally proven.
Material Balances
Complete material balances for Runs 17 and 18 are given in
Tables 13 and 14. The symbols used in Tables 13 and 14 have
the^following meanings: F = total fuel burning rate, Ib
hr ; FC = carbon burning rate, Ib hr"1; FH = hydrogen burning
rate, Ib hr"1; FO = oxygen burning rate, Ib hr~l; ALPHA =
Ib carbon/lb oxygen burnt; BETA = Ib hydrogen/lb oxygen
burnt; GAMMA = Ib carbon/lb hydrogen burnt; G = stack gas
flow rate, Ib moles hr"1; LEAK = air leakage rate, Ib moles
hr"1; SUM = sum of CO., and H20 concentrations within the
stack gases, mole fraction; C = cumulative carbon loss, Ib;
0 = cumulative oxygen loss, Ib; H = cumulative hydrogen loss,
Ib; TLOSS = total cumulative loss, Ib.
Tables 13 and 14 show good agreement on the overall material
balance as indicated by the calculated carbon, hydrogen, and
oxygen loss compared to the know weights of these materials
in the fuel. The largest discrepancy is shown in the carbon
balance for Run 17, and can be explained by either loss of
carbon as particulates in the stack gases or by incomplete
combustion. The former is very unlikely considering previous
- 230 -
-------
Table 13. Material Balance Calculations (Run 17)
BURNING SATES AND INTEGRATED HEIGHT LOSt FOR RUN 17
TOTAL FUEL CHARGED • 62.J.30 WT. FRACTION MOISTURE . 0.2517 WT. FRACTION INERT • 0.1516
A^BUNT CONDITIONS. Y02A » 0.20T6 YN2A . 0.7008 YH20A • 0.0115
INTEGRATION STARTED 200.0 SECONDS INTO RU«4
TIHF
200.0
100.0
400.0
900.0
600.0
700.0
800.0
900.0
1000.0
1100.0
1200.0
1100.0
1400.0
110C.O
1150.0
17IC.O
1*00.0
1-3C.C
?w"»P.r
21T. )
22C PA'IO • 0.114
OLCIILATFT FllFL C/1 PATIO . 0.180
ACTUAL FUEL C/H RATIO • 4.793
ACTUAL FUEL H/0 RATIO > 0.112
ACTUAL FUEL C/0 RATIO • 0.635
BU»NING RATF OVER CONSTANT BURNING PERIOD IFROM 1300.0 TO 3700.0 SECONDS! • 62.6*7 LBS/HR
- 231 -
-------
Table 14. Material Balance Calculations (Run 18)
OURMIN6 RATES MO JMTE6RATEO XC16MT LOSS FO* RUN II
TOTAL FUEL CHARGED • 66.900 BT.FRACTION MOISTURE • 0.2306 WT.FRACTION INERT • O.UOJ
AMBIENT CONDITIONS Y02A ° 0.2099 YN2A » 0.77*7 YH20A • 0.0194
INTEGRATION STARTED 100.0 SECONDS INTO RUN
TIME
100.0
200.0
300.0
400.0
300.0
600.0
700.0
600.0
900.0
1000.0
1100.0
1200. C
1300.0
1400.0
! f 30 .0
iuoo.o
noo. )
1FOO.O
19PO.O
7000.0
2 100.0
27PO.P
2300.0
24HP.P
2100.0
?600.P
7700.0
2S1C.C
2900. C
icro.o
nor.c
32"0.0
'33P.;:
3400.?
? ^19.1
1n">P.O
1'OO.C
1H30O
?93C .'J
4010.0
4 i?p . p
42or.c
43CC.O
4400.0
43CO.C
4630.0
4700.0
T
49.441
30.923
34.721
96.90*
96.367
37.866
32.400
S3.B14
90.4R9
34.1)4
53.288
33.247
32.906
34.086
53.332
34.63B
6C.170
64.964
63.271
66.786
69.349
72.124
77.729
77.921
71.P75
68.432
70.326
66.364
66.872
63.842
93.904
48.02C
43.291
34.73B
23.392
12.597
9.216
7.269
3.84P
4.908
3.196
2.112
1.073
2.738
2.714
2.692
FC
11.370
14.273
16.695
16.939
16.612
17.707
17.369
17.606
17.330
19.99))
16. 4U,
16.346
15.890
15.776
15.988
16.87.!)
19.269
20.394
71.677
27.874
2S.770
24.904
24.S79
24.541
75.137
24.709
?5.40»
25.646
25.374
24.491
21.810
27.. 310
21.303
18.539
15.097
11.60?
".908
7.21S
.638
.629
.367
.736
.724
.412
.376
.398
FM FO
2.4Si« 25. 330
3.JM 32. OSS
S.»S2 30.975
3.487 S*oO'»S
4.007 34.234
4«:80 34.479
4.0'Ji 30.204
S.835 30.90B
1.849 10.614
3.717 30.779
S.46S 3'J. OSS
jcSOO 33.359
1.629 S3.7JS
3.7'>1 )3.389
3. 964 34.113
4.234 1J,S'J3
4.6S3 31.934
4.S16 3&.085
3.076 39.333
5»")63 38.230
3»98
0.3323
0.49VO
0.342B
0.31)3
0.4S92
0,5691,
1.1672
0.5194
0. 4619
0.4900
0.4711
0.4724
0.4636
0.5200
0.3776
0.5339
0.3223
0.5471
0.6026
Oo699*>
0.9963
0.3930
0.977)
0.6093
0.6463
0.6446
0.7263
0.6342
0.7075
0.7729
0.9S43
1.1140
1.3223
2.21&2
3D. 9133
-6.6722
•73.2(33
4.8793
3.1640
3.3037
5.4341
1.3117
1.0781
1.0997
1.0*42
BETA
0*0973
0.0963
0. 1 il*»
0.1093
0.1U6
0.1213
o.usp
0.1237
0.1263
0.1507
0.1076
C.1073
0.1079
0.1120
0.1166
0.1301
0.1393
0.1S34
0.1290
0.1403
0.1440
0.1327
0.1310
0.12Q6
0.1296
0.1321
0.1437
0.1396
0.1331
3.U07
0.1376
0.1391
0.1632
0.1476
C.1336
O.J779
3.3397
-0.4619
-1.1&66
0.2635
0.1236
0.1206
0.1377
0.0269
0.0266
0.02C.S
0.0339
GAMMA
4.6108
4.3237
4.7781
4.6075
4.6*42
4.2361
4.2922
4.3314
9.3610
4.30U6
4.&7R 1
4.5663
4.3793
4.2167
4.0123
4.0021
4.1463
4.0305
4.0491
4.0413
4.1834
4.S169
4.350V
4.6399
4.4403
4.6127
4.4973
4.6174
4.7417
4.9633
9.1417
3.9339
6.0293
7.5476
8.4986
9.7241
11.6316
13.8444
J9.9600
19.3119
2J. 7699
27.409)
50.6122
*§.7447
40.9027
40.S220
40.2269
G
29.64)3
30.4917
30.9269
31.30)2
31.64*3
12.16*)
32.3129
32.4762
31.6*09
32.7)6)
32.8233
32.6394
32.3U02
32.3370
32.182B
31.8692
11.7030
31.6493
31.6*92
31. 4674
31.3216
31.2200
31.3332
31.2698
31.3047
31.2073
11.1243
31.1636
31.0118
30.9VO)
30.6308
30.6*38
30.1676
29.7015
29.2234
28.3846
27.6922
27.3621
26.8902
26.4047
29.7669
23.4042
24.6877
34.4*67
24.0291
23.411)
23.0970
LEAK
10.3*1*
10.9619
11.9972
11.6169
12.072*
17.3631
li.6996
12.840)
13.0212
13.1*6*
13.1*64
1J.02S2
12.840)
12.6396
13.6120
14.49U1
14.2207
13.9389
13.772)
13.7932
14.6110
14.9974
14.9039
14.4307
14.4907
14.3976
14.3976
14.3976
14.3976
14.1976
14.4507
14.4907
14.3976
14.2)96
14.0634
13.8268
13.4793
13.0643
12.6110
12.1938
11.9396
11.201)
10.7104
10.2182
9.820*
9.201*
6.869*
SUM C
0.211 0.197
0.219 0.914
0.211 0.9*1
0.212 1.406
0.210 1.9UU
0.210 2.40!.
0.212 <6Vi
0.209 .38)
0.208 .669
0.210 .132
0.212 .782
0.212 .217
0.21) .689
0.212 .129
0.210 .366
0.206 .-122
U.206 .311
0.207 .03*
0.207 .307
0.20* .17)
0.20) .792
0.201 10.440
0.207 11.116
3.207 11.808
0.20B 12.494
0.207 13.U4
3.20) 13.876
0.204 14.57}
0.202 19.282
0.203 13.VU6
0.206 16.674
0.206 17.149
0.203 17.981
0.206 18.981
0.206 19.136
0.203 19.60)
0.201 19.97)
0.203 20.298
0.236 20.412
0.206 20.6*7
0.209 20.762
0.206 20.8*9
0.203 20. 9J)
0.210 20.931
0.210 20.999
0.210 21.0)1
0.210 21*071
0 H TLOSS
0.391 0.01* 0.3*)
1.1*6 0.112 1.77)
2.021 0.20* 3.167
2.92* 0.303 0.61*
3.871 0.41U 6.184
4.826 0.32) 7.796
9.810 0.6)8 V.)**
6.742 0.7*9 10.676
7.17* 0.82V 12.07}
8.00} U.9U6 11.2**
6.909 1.009 14.697
9.9*2
10.77)
11.706
12.6*)
11.566
14.436
19.400
16.447
17.323
16.S6)
19.610
20.691
21.894
2). «27
24.190
23.294
26.572
27.410
28.40V
29.399
JO. 307
31.0*3
31.622
32.002
32.371
)2.*6V
32.433
32.4)2
32.441
12.470
12.496
12.310
)2.3))
12.370
.109 16.169
.209 17.669
•112 19.1*3
.419 20.629
.9)) 22.121
.69) 23.621
.761 23.216
.919 26.V14
.06* 29.763
.21* 30.990
«)»6 32.418
•31V 34.327
•670 )6.))2
.621 )».)4)
.972 *0.)*7
.12* 42.295
.277 44.22)
.429 46.121
.376 *7.»72
.713 49.767
.8)9 31.492
.9*9 32.V77
.0)9 3*. 2*3
.109 35.326
.160 36.1)9
. 1V6 56.»)V
.218 »t.9)3
.232 97.1*6
.241 97.3)0
.246 97.479
.250 97.992
.131 37.663
.2*2 97.7)7
.21) 37.816
12.606 4.234 37.894
12.6*2 4.233 37.970
IHTfr.RATFD C«SBOM LOSS » 21.071 LBS
I'!TF.F«ATfP HYOROGFM LOSS • 4.253 LBS
INTEGRATFC OXYiE'i LOSS • 32.642 LSS
I'iTEGRATEP '.vEISHT LOSS
47.970 LBS
THEORETICAL CARBON CHARGED • 20.439 LBS
THEORETICAL HYDROGEN CHARGED • 4.226 LBS
THEORETICAL OXYGEN CHARGED • 3U6*o LBS
TOTAL COMHuSTlBLE CHARGED
• 36.499 LBS
CALCULATED FUFL C/H <5aT10 « 4.931
fALCULATFP FUEL H/0 RATIO • 0.130
CALCULATED FUFL C/0 RATIO • 0.643
PUNNING RATE OVER CON5TA1T BURNING PERIOD
1U9NINC, RATE OVER CONSTANT BURNING PERIOD
ACTUAL FUEL C/H RATIO • 4.8)9
ACTUAL FUEL H/O RATIO • 0.1)2
ACTUAL FUEL C/0 RATIO • 0.6*1
(FROM 100.0 TO 1700.0 SECONDS) • 51.4)9 LBS/Hd
(FROM 1700.0 TO 3200.0 SECONDS! • 66.749 LB4/HR
- 232 -
-------
experience with the smoke meter (not used in this test) which
showed only trace participates in the stack gases. The latter
explanation is more likely and is confirmed by the weight loss
plot of Figure 59, which shows a final weight loss of close
to 51 Ib. The incomplete combustion can be explained by the
fact that small pieces of wood tended to fall inside the tin
cans, where they burnt slowly and were incompletely burnt
by the end of the run when fuel bed temperatures fell low
enough for the carbon-oxygen reaction to be quenched.
The accumulative carbon, hydrogen and oxygen losses and their
sum, TLOSS, for Runs 17 and 18 are plotted in Figures 66
and 67. Comparing the values of TLOSS versus the load cell
data of Figures 59 and 60 gives an independent check on the
accuracy of the weight loss determination; the agreement be-
tween the material balance calculations and the load cell
data was very close for both these experiments. For Run 17
the average burning rate over the constant burning period
(1300 to 3700 sec) calculated from the material balance is
38 Ib hr'-'-ft" , which is within 3 percent of the load cell
result of 39 Ib hr-1ft~2» For Run 18 two constant burning
rate periods were observed, one lasting from about 100 sec
to 1700 sec into the run and the other from 1700 sec to 3200
sec. For the former constant burning,period, the load cell
data gave a burning rate of 30 Ib hr ft~2 and the material
balance 31 Ib hr~1ft~2 while for the latter the load cell
data gave a burning rate of 41 Ib hr~lft~2 and the material
balance 40 Ib hr ft
The material balance calculations were sensitive to the O
and CO2 stack concentrations because they involved the compu-
tation of differences between large numbers; this sensitivity
was most marked near the beginning and end of a test. The
anomalous behavior indicated in both Tables 13 and 14 near
the end of Runs 17 and 18 can be attributed to inaccurate
data on these concentrations. Small changes in the measur-
ing instrument calibration could have accounted for these
errors. A check on the O.., and CC>2 compositions measured in
the stack was given by the SUM column; for the complete com-
bustion of cellulose to CO, and H-0 the mole fraction of
these two components in the stack should have equaled the
mole fraction of O2 in the combustion air (0.21). For all
the points taken for the material balance this was found; the
worst point at 2200 sec in Run 18 being less than 5 percent
in error. (Actually, the fuel used contained net hydrogen so
that the value of SUM should have been slightly less than
0.21.)
Figure 68 shows the values of the ratio of hydrogen to oxy-
gen (BETA) and the ratio of carbon to hydrogen (GAMMA) in
- 233 -
-------
to
u>
CO
CO
(A
to
O
' i-
I
O
u
£
HYDROGEN LOSS
TOTAL WEIGHT LOSS
0
1OOO
4OOO
2000 3000
ELAPSED TIME (SEC)
Fig. 66. Accumulated Carbon, Hydrogen and Oxygen Weight Loss of Fuel (Run 17)
-------
CO
OJ
en
HYDROGEN
LOSS
CARBON LOSS
TOTAL
WEIGHT
LOSS
OXYGEN LOSS
LJ
4OOO
1000 2OOO 30OO
TIME (SECONDS)
Fig. 67. Accumulated Carbon, Hydrogen and Oxygen Weight Loss of Fuel (Run 18)
-------
to
u>
Q
LJ
0 ID
|8
_j
U LL.
^U
15
10
5
0
^ y Ignition Front |
— —18 at Grate Runs |
~ 17 and 18 — |
vC/H Ratio of Fuel (4.8)
X^ "'' ''
^/ >•'' '
s'
I I I I I I I I I
O
P
O
I b-
Q O.3
u
0.2
0.1
0
• 17
•18
,O/H Ratio of Fuel ( 0.13)
I
I
1
0
0.2 0.4 0.6 0.8
FRACTION OF COMBUSTIBLE CONSUMED
Fig. 68. Hydrogen/Oxygen and Carbon/Hydrogen Ratio of Fuel
Burnt as a Function of Fraction of Combustible
Consumed (Runs 17 and 18)
.•I
I
1.O
-------
the fuel burnt for Runs 17 and 18; in this figure the regions
where there was the most uncertainty in the calculated values
are shown as dashed lines. For Run 18 some anomalous values
of the H/O ratio were calculated for the times between 3700
and 3900 sec; these inconsistencies were probably caused by
incorrect stack gas composition measurements.
The constancy of the H/O ratio over the major portion of
these experiments confirms the view that drying and pyrolysis
occur over protracted time periods.
Gas Compositions
Fuel Bed Gas Compositions. Representative compositions of
the gases withdrawn from the center of the bed at two posi-
tions 20 in. and 12 in above the grate are shown, for Run 17,
in Figures 69 and 70, and for Run 18 in Figures 71 and 72.
Measurements of fuel bed gas compositions were taken, as des-
cribed on pages 152-184, on an intermittent basis. The smooth
curves joining the data points must therefore be considered
only to represent the average trend in gas compositions with
time.
The bias introduced by radial concentration gradients was
checked for a number of runs by comparing the carbon contents
of the gases at the sample points with those in the stack.
For the gas samples from Run 18 there was little bias, but
for other runs where larger quantities of underfire air were
used, the degree of channeling increased; this has been fully
discussed on pages 213-230.
Figures 73, 74 and 75 show the smooth gas composition curves
replotted versus distance above the ignition front. These
plots were obtained in a manner analogous to that used in cal-
culating the temperature profiles below the ignition front
and the reader is referred to on pages 206-208 for a more com-
plete description of the method used. The gas compositions
measured during Run 17 from Fuel Bed Probe 2 were not plot-
ted in the above manner because of the incompleteness of the
data on H2 and CH4 concentrations. (For Figures 70, 71 and
72, note the Fuel Bed Probe 1 and Fuel Bed Probe 2 were 20
in. (1.67 ft) and 12 in. (1 ft) above the grate; this was
the maximum that these probes could be above the ignition
front.)
- 237 -
-------
to
U)
CO
in
CD
cc
Q
o
o
o
CL
2
O
U
I 1
Time Ignition Front
Reached Probe
at Grate "*|
0
1400
2000
30OO
TIME (SEC)
400O
Fig. 69. Gas Compositions from Fuel Bed Probe 1 (Run 17)
-------
(0
u>
vo
(f)
CO
CC
Q
o
>
Z
O
cr
u
u
z
O
U
I I I I I
Time Ignition Front
Reached Probe
4
Ignition Front
f at Grate
8 —
O
2000
3000
TIME (SEC)
40OO
4500
Fig. 70. Gas Compositions from Fuel Bed Probe 2 (Run 17)
-------
CO
CO
<
CO
o
>
CO
O
CL
2
O
u
2O P"**^ Time Ignition Front
\xReachcd Probe
16 -
1000
Ignition Front S_
at Grate •
2000 3000
ELAPSED TIME (SEC)
4OOO
CO
2 2
1 3
U
Fig. 71. Gas Compositions from Fuel Bed Probe 1 (Run 18)
-------
C/)
CD
o:
Q
Z
O
to
O
CL
O
u
4 -
0
Ignition Front
at Grate f
Time Ignition Front
'"Reached Probe
2000
Fig. 72
or
O
O
>
30OO 4000
ELAPSED TIME(SEC)
Gas Compositions from Fuel Bed Probe 2 (Run 18)
0 O
u
-------
V
ft.
(ft
<
CD
o:
Q
z
Q
*
oc.
UJ
o
z
o
u
0
O
O.2
O.4 0.6 Q8 1;O 1.2 1.4
DISTANCE ABOVE IGNITION FRONT (FT)
Fig. 73. Gas Compositions as a Function of Distance Above Ignition
Front (Run 17)
-------
I
NJ
CD
Q
O
O
H
<
or
K
Z
u
u
z
O
u
O.2 0.4 0.6 0.8 1.0 1.2 1.4
DISTANCE ABOVE IGNITION FRONT (FT.)
1.6
Fig. 74. Gas Compositions as a Function of Distance Above Ignition
Front (Run 18)
-------
<
or
Q
o
z
O
o:
8-
Z
UJ
u
z
o
u
0.2
0.6 0.8 1.0
8GNITION FRONT (FT)
Fig. 75. Gas Compositions as a Function of
Distance Above Ignition Front (Run 18)
(Calculated using ignition velocity =
0.083 ft sec ); fuel bed probe located
12.n. above grate.
- 244 -
-------
For2these runs where low underfire air rates (150 Ib fer
ft ) were employed, the oxygen concentration fell rapidly to
zero as the ignition plane passed the probe and typically was
consumed in a distance of 0.25 - 0.8 ft. The point at 2300
sec in Figure 68 appears to be spurious as there was no de-
crease in combustibles (CO and H2) at this time. The decay
of the oxygen concentration with distance below the ignition
front for Runs 15, 17 and 18, where fairly complete records
were available, is shown in Figure 76. Some of the variation
of the shape of these profiles can be attributed to inaccur-
ate determination of the time the ignition front reached the
sampling probe. In addition, local variations of tfte igni-
tion velocity could have affected the shape of the profile
as the distance-time conversion was performed using a con-
stant average ignition velocity.
As the O, concentration decreased,the concentrations of CO2,
CO, H2 and CH increased. The increase in C02 concentration
always led the increases in concentration of the other gases;
the same behavior was observed by Nicholls (24). Peak concen-
trations in both C02 and CH. were observed near the point of
oxygen extinction, which aglin parallels Nicholls1 (24) ex-
perimental results. Peak concentration of CO- were E^pically
16-20%; while those for CH4 were 3-4%.
After all of the O« had been consumed, the conditions within
the fuel bed appeared to differ between experiments. From
the earlier discussions on the various processes occurring
within a fuel bed, it would be postulated that due to the Bou-
douard reaction (C + C0~ -*• 2CO) the C02 concentration would
decrease, while the CO concentration would increase with dis-
tance from the plane of zero oxygen concentration. Similarly
the H_ and CO concentrations would be expected to increase,
at the expense of the H2O because of the occurrence of the
reaction C + H,0 -»• H2 + CO. In addition, if the assumption of
the equilibration of the water-gas-shift reaction (except in
the vicinity of the ignition front) is correct, increases in
the CO concentration should be followed by increases in the H2
concentration and vice versa.
The above behavior was exhibited in Run 18 (Figures 71 and 72)
with the exception of the H? concentration measured with Fuel
Bed Probe 1, at times greater than 2800 sec. (Figure 71 shows
that, after this time, the H2 concentration decreased while
the CO concentration increased.) However, the exact location
of the H2 concentration profile was not known because of the
paucity of the data for this species and it is possible that
the H2 concentration profile could have followed a similar
shape to the CO profile. The gas compositions obtained from
Run 17 exhibited behavior different from that observed in Run
18, as the CO and H2 concentrations decreased with tine (or,
- 245 -
-------
I
to
c\
I
z
UJ
u
o:
UJ
O
O
Z
O
O
LU
O
>-
X
O
o
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
DISTANCE BELOW IGNITION FRONT (FT)
Fig. 76. Oxygen Concentration as a Function of Distance Below Ignition Front
-------
equivalently, distance above the ignition front). This var-
iation of gas composition cannot be explained by bed tempera-
ture changes as these temperatures increased from 2000 sec
to 2800 sec by about 200°F and would have favored the forma-
tion of higher concentrations of CO and H20. The decrease
in CO and H2 in this run could be explained by mixing of by-
pass oxygen back into the core of the bed; at the higher tem-
peratures which prevailed in this region, the oxygen would
have reacted rapidly with the combustible and no oxygen would
have been observed. The high amount of bypass oxygen that
apparently channeled up the side wall during this run (see
pages 213-230) makes this explanation plausible. Further
confirmation of this hypothesis is supplied by the lower
than normal CO and H2 concentrations measured at position
12 in. above the grate. The CH4 concentrations measured with
this probe were also very small (<0.1%).
For both Runs 17 and 18, as well as others, the CO2 concen-
tration within the bed fell off quite sharply near the end
of active combustion (3800 sec for Run 17 and 3600 sec for Run
18 — see Figures 59 and 60, which show that the major por-
tion of the total weight loss was achieved by these times).
Similar results were obtained from the stack gas measure-
ments for these two experiments; this is discussed on pages
248-253 • The rapid decrease in CO-> at the time when bed
burnout is nearly complete may provide a relatively simple
criterion for the burnout of the fuel bed on a travling grate
incinerator as long as there is no gross back-mixing of the
overfire combustion gas at the discharge end of the grate.
The gas compositions obtained from Fuel Bed Probe 2 in Run
18 (Figure 72) indicated that, near the end of the run, the
C02 concentration started to increase as the CO concentra-
tion began to decrease. The CO2 concentration continued to
increase up to the time the CO concentration fell to zero
at which point it started to fall.
From Figure 63 it is seen that even with a generous estimate
for the depth of the active bed, the gas sampling probe was
probably well above the top of the bed at the time the CO
concentration started to decrease. This observation may be
explained by the presence of a diffusion flame above the bed;
the fuel to the flame was supplied by the combustible rising
from the top of the bed and the oxidant was supplied by the
oxygen from the overfire region. As the top of the bed re-
ceded from the probe, samples were first taken from the fuel
regime of the diffusion flame, and then as the probe passed
-247-
-------
through the flame envelope from the oxidant side of the flame.
The turbulence above the bed would make the location of the
flame envelope ill-defined. Similar behavior would have been
expected from the other gas compositions measured, but the
intermittent sampling procedure may have obscured this trend.
Tables 15 and 16 (a) and (b) show the calculated ratio of
the product of the CO and H20 concentrations to the product
of the C02 and H2 concentrations within the fuel bed at var-
ious times during Runs 17 and 18. For both these runs, batch
samples were taken from the fuel bed probes and analyzed for
H2. The H20 content within the bed was back-calculated from
the measured t^O content in the stack gases. Table 15 shows
that the water gas shift equilibrium was closely followed
when oxygen began to penetrate through the bed (at 3760 sec).
The two points, at 2080 sec and 2240 sec, where the calculated
values are lower than would be predicted from equilibration
of the water gas shift reactions could well be caused by inter-
ference from pyrolysis products (see Figure 66 or 70). Similar
results were obtained from the two different sampling posi-
tions in Run 18, as shown in Table 16 (a) and (b) for two dif-
ferent locations within the fuel bed.
From the method of calculating the H20 content in the bed the
estimates of these concentrations were probably higher than
those representative of conditions at the top of the bed for
those cases where the sample was taken some distance from the
top of the bed? this the estimate of (CO)(H2O)/(CO2)(H2) at
these positions would have been higher than that observed at
the top of the bed. These results although they must be con-
sidered of a preliminary nature tend to support the theore-
tical model of Niessen et al. (19) (see pages 106-112), and
give a rough guideline for determining overfire air require-
ments.
Stack Gas Compositions. Figures 77 and 78 show the stack
gas compositions for Runs 17 and 18. Although the burning
rate was constant over the major portion of Run 17, there was
some variation in the stack gas concentrations. This was
caused by the varying air leakage into the top section. The
air leakage would be expected to be proportional to the
temperature differential between the ambient and the inside
of the test incinerator. Figure 77 shows that the air leak-
age reached a maximum at around 3000 sec; at which time the
inside refractory temperatures were at a maximum (see Appendix
H, Figure H.4). The line marked "Start of Integration" at
200 sec in Figure 77 indicates the time when the first mater-
ial balance was calculated — see pages 230-237^t
- 248 -
-------
TABLE 15
Clieck oa • Water -XJaa- . JEa.ulJ.JJuJ.um .wl.t±Lin
vo
I
(Fuel Bed Probe 1, 20 in. above grate)
(.Run -17)
Elapsed
Time
(sec)
1570
1930
2080
2240
2540
2930
3350
3760
Temperature
(°F)
just at ignition
1700
1700
1750
1800
1900
1750
1800
0=2}
PJ
0.2
0.4
1.0
1.0
0.75
0.52
0.59
0.09
[H20]
[H2]
77.0
6.6
0.3
0.66
2.9
4.3
1.7
OB
M C«2°]
Cco J f HJ
(Experimental )
15.0
2.6
0.3
0.66
2.2
2.2
1.0
00
[co]fi2o]
[C°21 EH21
(Calculated assuming
water gas shift
equilibrium)
> 0.10
1.42
1.42
1.54
1.65
* 1.86
1.65
1.65
[A] denotes concentration of A
-------
TABLK 16 (a)
Check on Water Gas Shift Equilibrium within Fuel Bed (Run 18)
(Fuel Bed Probe 1, 20 in. above grate)
Elapsed
Time
(sec)
1450
1600
1850
2830
3550
Temperature
(°F)
?
7
•v 1400
1500
1650
[col
0.1
0.4
0.7
1.0
0.57
[H20]
KT
90
59
4.7
1.2
2.1
[C0l fa O]
I J L 2 J
PJ [H2]
(Experimental )
9.0
24.0
3.3
1.2
1.2
[CO] [H20]
[C°J [»2]
(Calculated assuming
water gas shift
equilibrium)
—
—
0.8
1.0
1.3
TABLE 16 (b)
Check on Water Gas Shift Equilibrium within Fuel Bed (Run 18)
(Fuel Bed Probe 2, 12 in. above grate)
Elapsed
Time
(sec)
2550
3320
Temperature
1100-1500
1700
[CO]
Kl
0.74
2.0
[H20]
[H2]
4.1
1.8
[CO] [H20]
On 1 fu T
LC02] 1H2 J
(Experimental )
3.1
3.6
[CO] [H20]
[C02]H2]
(Calculated assuming
water gas shift
equilibrium)
~ 0.5
1.6
[A] denotes concentration of A
- 250 -
-------
NJ
cn
CO
CE
Q
Z
O
O
o_
2
O
u
Carbon Dioxide
4 —
0
400 800 12OO 1600 2000 2400 2800 320O 3600 4000
TIME IN SECONDS
Fig. 77. Stack Gas Compositions (Run 17)
4400
-------
K)
Ol
K)
to
CD
cr
Q
O
z
O
CO
O
Q.
2
O
u
16
12
hUO
I
400 80O 12OO 1600 2000 2400 2800 3200 36OO 40OO 44OO 46OO
TIME ( SEC)
Fig. 78. Stack Gas Compositions (Run 18)
-------
For Run 18 the temperatures inside the overfire section did
not change as markedly as those for Run 17; the leak rate was
23Qr?f?re relativelY constant during this test (see pages
~^J7 ) and the CO-, O and H.,0 concentrations did not
cnange appreciably during periods when the burning rate was
steady. The increase in the CO9 concentration that was ob-
served around 1600-1800 sec was consistent with the increase
in the burning rate observed at this time (see Figure 60).
Figures 77 and 78 show the same rapid decrease of
tne concentrations of CO- towards the end of tne run as in-
dicated by Figure 66 through 72. For Run 17 a comparison be-
tween Figures 77 and 59 shows that at the point where the con-
centrations of C02 and H0O began to fall (3600 sec) about 95
percent of the combustible fuel content had been consumed.
Similar results were obtained for Run 18.
For both Runs 17 and 18 the H70 concentration in the stack
decreased at the same point that the CO- concentration de-
creased. This result confirmed the hypothesis of protracted
drying and pyrolysis of the fuel.
Energy Balances
Differential energy balances taken at different times dur-
ing an experiment can provide information on the total heat
release rate (the sum of the heat release rates within and
above the bed) at those times, while an integral heat balance
taken over the duration of active burning can give an esti-
mate of the total heat released which can be checked against
that calculated from the known fuel composition and amount
consumed. Attempts were made to calculate both types of
balance but only integral heat balances could be calculated
with any certainty because of the unsteady-state nature of the
combustion processes and the "thermal flywheel" provided
by the top section refractory lining. The difficulties en-
countered with the energy balance calculations will be dis-
cussed below.
Prior to the start of an experiment a steady temperature pro-
file was established in the refractory brickwork of thg over-
fire section; typically, inside temperatures were 2100 F and
outside temperatures 275°F. At the start of an experiment,
a few seconds after the refractory shield was opened the tem-
perature dropped sharply (generally from 2100 F to 1700 F
in about 200 sec) as the energy stored in the refractory sec-
tion was transferred to heat the fuel bed surface and the
- 253 -
-------
under fire and overfire air. Stack gas temperatures typically
fell a few hundred degrees Fahrenheit during this "ignition"
period. Once active burning was established the temperatures
of the refractories changed slowly to an "equilibrium" level.
For those runs where adjustments were made in the amount of
overfire air employed or where the burning rate varied,
the temperatures changed continually during the course of the
experiment.
For the purposes of calculating differential energy balances
it was important to be able to estimate the effective heat
sources and sinks associated with the changes in temperature
level of the refractories. This was particularly important
at the beginning and end of a run if heat release rates were
to be determined for these times but no reasonable method for
calculating this effect was available. Assuming a linear
temperature profile within the top section refractories.
the change in total energy content of this section per F
change in the temperature of the inside surface AO/ATS was
calculated to be
- 90 Btu °F-1 (176)
This equation can be used to calculate the integral heat
source or sink effects but clearly would give an erroneous
result under conditions of rapid heat transfer. (When the
temperature of the top section dropped rapidly at the start
of a test, the temperature profile within the refractories
reached a maximum some distance from the inside surface.)
With the temperature measurements taken, there was no simple
method available which could be used to improve the calcula-
tion of the source sink effect of the refractories and there-
fore limited the use of differential energy balances.
The other difficulty encountered concerned the estimation of
the heat loss through the feul bed wall. Heat losses through
the fuel bed wall were determined from the temperatures mea-
sured with the heat loss probes and from the theoretical cal-
culations of heat loss through the fuel bed wall (see Appen-
dix B) . From Figure B.I the maximum heat loss (at a -short
time after the wall was exposed to the burning fuel) was
estimated to be of the order of 10 x 103 Btu hr~r per foot
of bed height, while at times greater than about 0.6 hr the
heat loss dropped to approximately 6 x 103 Btu hr"1 per foot
of bed heigiVc,, From the heat loss probe3data for. Run 18
the heat losses were estimated as 8 x 10 Btu hr per foot
of-jbed height. _^For calculational purposes a value of 8 x
10 Btu hr ft was chosen as representing a mean heat loss
- 254 -
-------
rate. The heat loss depended on the amount of wall exposed
to the ignited solid. From the ignition rate and the estimates
?lv®jj above, a mean-jheat loss for Run 18 was calculated to
be 10 x 1CP Btu hr~ . This value was considered to be typi-
cal for the majority of experiments conducted in this study.
For simplicity, heat losses were considered constant through-
out the duration of a run? estimates of the fuel bed heat
losses would be most markedly affected by this assumption but
the effects proved to be fairly small in relation to the
other heat losses, so this assumption was reasonable. Heat
losses were considered to take place in three places: over-
fire region, top section cooling coils, and fuel bed. Heat
losses to the grate were neglected because, for the run in
which heat release rates were calculated, the ignition front
reached the grate at a time close to the end of the experi-
ment. Heat losses from the top section (assumed to include
the refractory tray) were estimated using
2Lk (T1 - T.°)
Ql = -THTrfT^T- BtU hr (177)
where L was the length of the top section area (416 ft) ; k
the mean thermal conductivity of the refractory brickwork S
(0.18 Btu hr-ift"1 °F) ; T^ and T° the inside and outside
refractory temperature respectively; and r and r, the inside
and outside radius of the overfire section respectively (r =
0.75 ft; r1 = 1.3 ft). °
The heat loss to the top section cooling plate was calculated
from the mass flow rate of the cooling water used (450 Ib
hr ) and the temperature rise (^ 30 F max.). These heat
losses were therefore about 14 x 103 Btu hr~ and were roughly
constant over the duration of a run and between different ex-
periments.
For the specific case of Run 18, the, heat loss from the top
section, using equation (177) with T^ = 1600UF and T" = 275 F,
was about 12 x 103 Btu hr"1... Total Keat losses for this test
were therefore about 36 x 10 Btu hr"1. For this run the tem-
perature of the top section refractories remained practical-
ly constant from 1800 to 3000 sec. During this time the^aver-
age rate of enthalpy loss via the stack gas was 256 x 10
Btu hr . (This value was calculated using the molar flow
rates determined from the material balanc^p^ogram, a mean
flue gas specific heat of 8.1 Btu Ib mole F *•, the rgcord-
ed stack gas temperatures and a datum temperature of 80 F.)
From the burning rate of 68.5 Ib combustible hr"1 (41 Ib
combustible hr-lff2) the,average heating value of the com-
bustible was 4300 Btu Ib .
- 255 -
-------
The above calculation was repeated for times from 200-1600
sec. In this case, however, the temperatures of the top sec-
tion fell an average of 200 F over this time period. The
total heat generated from 200-1600 sec (82 x 10 Btu) was
computed by summing the integrated enthalpy losses to the stack,
fuel bed wall and cooling coils and subtracting the net en-
thalpy change in the refractory brickwork given by
equation (176). The total weight loss over this period was
approximately 20 lb, giving an average heating value of the
combustible of 4100 Btu Ib"1. The calculated heating values
for the two different periods of the experiment agree closely.
Using a heating value for the dry wood of 8090 Btu lb (cal-
culated from tfhe Du Long formula, AH = 14600C + 620QO (H -
0/8), Btu Ib"1) and subtracting from this the amount of ener-
gy required to vaporize any water.present (the heat of absorp-
tion term is small; ^ 30 Btu lb~ ), the heating value of
the fuel used in this experiment was calculated to be 5450
Btu lb~l. The agreement with the experimental values is
reasonable considering the approximate nature of the calcula-
tions.
Energy balances were not caluclated for Run 17 because of
the difficulties encountered in correcting for the changes
in temperature level of the refractories which occurred dur-
ing this experiment.
Overfire Air Regime
From the fuel bed gas concentrations shown in Figures 69
through 72, it is apparent that the oxygen required to com-
plete the combustion of the gases leaving the fuel bed varied
with operating conditions and with time. For a traveling
grate incinerator this means that overfire air requirements
will vary with position along the grate as well as with opera-
ting conditions.
For the experiments conducted in this study the average over-
fire (or secondary) air requirements for the burning rates
observed could be readily evaluated from Figure 61. However,
this calculation could not be performed on an a priori basis
because there were no theories available that could be used
to determine the achievable burning rate for a given under-
fire air rate. For cases where the burning rate was known as
a function of underfire air rate (i.e., from experimental
- 256 -
-------
use in
observation) a plot similar to Figure 61 could be of u=c j.,,
etermining the appropriate secondary air requirements. For
example, this technique could be used to evaluate the secon-
t-h^ air re<3uirements for a traveling grate incinerator if
p ^rra9e burning rate and the underfire air rate were known,
orK Particular example, however, some care would have
to be exercised, as the total oxygen feed rate through the
IS K 5 would need to be known. Oxygen can be supplied to
the bed both in the underfire air and in any overfire air
induced through the bed by temperature gradients; for the
latter case, the overfire air (which has a lower mass fraction
of oxygen and a lower density than the underfire air) will
tend to sink down at the "cold" walls of the furnace and to
rise up through the "hot" core of the bed. The oxygen in
this air would be expected to be rapidly consumed near the
edges of the bed. The estimation of the quantity of air being
induced into the bed in this manner is difficult but a rough
estimate of the expected magnitude of this effect can be found
as follows.
9 9
If only the significant forces are those of momentum (U pL )
and buoyancy (L3Apg), the entrainment velocity U is found
to be
(178)
where L is a characteristic length given by the distance
from the top of the bed to the roof of the incinerator;
Ap the density difference of the gas at the hot and cold
temperatures; pc the density of the cold gas; and g the
gravitational constant. The mass flux of oxygen through a
unit area of the bed by this natural draft (GNE))will be
where (M^ ) is the mass fraction of oxygen in the overfire
°2 c
air. The mass flux of oxygen through a unit of bed in the
underfire air (GFD) is given by
\
dFD = VPa (MoJa U80)
- 257 -
-------
where V is the superficial velocity of the underfire air and
p and (M_ ) the density and mass fraction of oxygen in the
a On a
underfire air, respectively.
If the effect of the natural draft is to be negligible
and using p =
C
P.T
and pH =
ri
__
where T , T and T are
aw n
the ambient cold Snd hot air and gas temperatures, gives
V (M
Q
(182)
Putting typical values into
0.5 ft/sec (- 120 Ib hr"1ft~
quation (177) as follows
>a = 0.23; (MQ ) c
V =
0.08; Ta = 520°R; TC = 2000°R; TR = 2500°R, gives
0.12 » 0.053-y/TT1
(183)
For many furnaces, L, one of the driving forces for the
buoyancy term, is of the order of 10-20 ft. Substituting
this value into equation (183) indicates that the oxygen mass
flux induced by natural draft may well be greater than that
supplied in the underfire air. The uncertainty associated
with estimating the unknown amount of air which is entrained
through the bed imposes a serious constraint upon the above
method of calculating the secondary air requirements.
The theoretical model of Niessen e_t al. (19) may also be
used to determine secondary air recfuTFemervts if the average
burning rate and fuel composition have been determined. This
model assumes that relatively small amounts of higher hydro-
carbon pyrolysis products are present in the gases at the top
of the bed and that these gases are equilibrated with res-
pect to the water gas shift reaction at the temperature of
the bed. The results of this study indicated that this assump-
tion was relatively good for the simulated refuse used; re-
sults from full scale operating units have also suggested that
this assumption holds for a real refuse (149) . Details of
the Niessen ejb al. (19) model are given on pages 106-112 .
For this study the mixing above the bed supplied by the overfire
- 258 -
-------
}ets was sufficient to provide nearly complete combustion of
all combustibles. Carbon monoxide was never detected in the
stack for any of the runs; for the later runs the limit of
detection was only about 0.1 - 0.2% but earlier runs where
the detection limit was about 0.05% also failed to show the
presence of CO. Trace quantities of soot and smoke were re-
corded, but only during periods of smoldering was any smoke
visually noticeable. Although this synthetic refuse was not
inherently a "smoky" fuel, these results are encouraging and
suggest that if, in operating units, adequate mixing is a-
chieved and high enough temperatures are maintained in the
overfire region, clean combustion of the volatiles and com-
bustible can be achieved.
As will be discussed below, for the very low degree of tur-
bulence encountered in all incinerators, and in the test in-
cinerator of this study, the rate of mixing of the combusti-
ble and oxygen is generally expected to provide the limiting
step in the combustion process. The ensuing discussion fol-
lows that of Sarofim (150). From the studies (e.g., 151,
152, 153) in well-stirred reactors the following general con
clusions can be drawn:
(a) The kinetics of oxidation for a large number of
fuels are so fast that volumetric heat release
rates of 106 to 108 Btu hr~1ft~3 are achievable
over the temperature range 2000° to 3000°F.
(b) The reactions of the hydrocarbons and ketones stu-
died are very fast relative to that of the carbon
monoxide intermediate formed in the reactions.
(c) The rate of burnout of the carbon monoxide, which
provides the rate limiting step, in the overall
combustion process, is given by
- ^£ - 1.8 x lo".xp
g/mole sec (184)
where f.^, fn , fH n are the mole fractions of CO, 02 and
i—vJ \_/ *} ti«*-/ ,~
H~0; T is the absolute temperature in K, and (P/RT) is in
g moles/liter. Using equation (184) and representative values
of temperature, oxygen and water vapor concentrations found
under most incinerator conditions, the time required to re-
duce the carbon monoxide concentration to less than .01 per-
cent is of the order of a millisecond. Thus it can be con-
cluded that if the turbulence above the fuel bed is high
enough to provide perfect mixing and if there is no equilibrium
- 259 -
-------
constraint imposed by the reaction C02 + H2 -*• CO 4- H20 no
CO would be expected in the flue gases and complete reaction
of the pyrolysis products could be achieved. Moreover, the
very rapid rate of chemical reaction suggests that if the
mixing were greatly improved above the fuel bed, furnace vol-
umes could be considerably reduced.
The rate of soot burn-out can be estimated from the expres-
sion for mass transfer and chemical reaction. The rate of
soot oxidation can be determined from the semi-empirical for-
mula proposed by Nagle and Strickland-Constable (154)to
correlate the measurements of the oxidation rate of polycry-
stalline pyrographite. Appleton (155) has summarized various
experimental evidence on surface structure to support the
assumption that the kinetics of soot oxidation are similar
to those for pyrographite. Park and Appleton (156) have pro-
vided experimental evidence that at high temperatures
(> 1500 K) the rate of soot oxidation is closely predicted
by the theory for the rate of oxidation of pyrolyzable graph-
ite. Assuming that the Nagle and Strickland-Constable re-
lationship holds for soot oxidation at the low temperatures
typical of incinerator operation, the rate of soot oxidation
can be given approximately by:
K
chem
= 12
20 exp j-15,000/TJ;
02
1 -!- 21.3 exp|2060/T[P0
'2J
-2 -1
g cm sec
(185)
where T is in K and PQ2 in atmospheres. From equation (185)
it can be concluded tha for all conceivable conditions within
an incinerator the chemical rate will provide the rate limit-
ing step for the oxidation of soot. Therefore the time to
burn a soot particle (t ) of original diameter (d cm) is
given by " °
p d d
(186)
2k
chem
''chem
On the assumption that the density of the soot particle is
about 2 g. cm"3, equation (186) shows that the time for burn-
out of the particles will be weakly dependent on the oxygen
partial pressure (P ), will increase with particle size d,
2
and will decrease with increasing temperature. Using equa-
tion (186) and typical incinerator conditions, P = 0.1 atm
and T - 1500°F {- 1090°K), the time to burn out a2l-micron
particle is 61 sec and a 200R particle is 1.2 sec.
- 260 -
-------
Jh iS ®vldent from the kinetic control of soot oxidation
^nat the temperatures in the overfire region should be kept
as nigh as possible. This result suggests that the addition
or air too high above the fuel bed may cool the gases and
•JUC2 the reaction' since the rate of reaction faU$ markedly
with decreasing temperature (at incinerator temperatures the
rate changes an order of magnitude with 200°K change in tem-
perature) . The importance of the proper design and place-
ment of the overfire air jets has been emphasized in the early
studies at Battelle under Engdahl (157).
Finally, in this study, the combustion intensities for active
burning runs were of the order of 40,000 Btu ft , assuming
that all the overfire volume and 50% of the fuel bed section
volume were needed for complete combustion. The complete
oxidation of the combustibles was probably achieved in a
volume rather less than that used in this calculation. This
combustion intensity is twice that typically used in most
incinerators. No difficulties were encountered in using a
combustion intensity of this magnitude, even for those runs
where paper was used as a constituent of the fuel? there
have been suggestions that for light material, such as paper,
high combustion intensities aggravate the problem of entrain-
ment. Extrapolating the combustion intensity results from
this study to full-scale units would be unjustified, but the
results suggest that the possibility of using higher combus-
tion intensities (which would allow furnace volumes to be
decreased) should be investigated further.
Fuel Bed Conductivity and Rate of Heat Release within Bed
An attempt was made to calculate the effective thermal con-
ductivity of the fuel bed and the net rate of heat release
within the bed following the methods proposed by Stewart and
Saville (8T) • These methods have been outlined on pages 34 -
39 and can be seen to rely upon accurate measurement of
the temperature profiles within the bed, as they require the
evaluation of the second derivative of temperature with dis-
tance —*— . The temperature measurements obtained in this
3z?
study were not of a high enough quality to permit meaningful
estimates of either the effective thermal conductivity or
the net heat release rate using the Stewart and Saville
technique.
- 261 -
-------
ACKNOWLEDGEMENTS
Financial support from the Solid Waste Research Division of
the Environmental Protection Agency is gratefully acknowledged.
The principal investigators are particularly grateful to Mr.
Louis W. Lefke and Mr. Daniel J. Keller for their guidance
and patience; during the early stages of design and construc-
tion. The authors are indebted to Mr. Robert C. Thurnau for his
encouragement and advice in the later phases of the program.
The report is based largely on the Sc.D. thesis of Joseph L.
Rogers.
The building of the experimental equipment could have been
achieved only with a great deal more trial and tribulation if
it were not for the outstanding contributions of Steven R. Le-
Mott and Charles T. Johnson. Grateful acknowledgement goes
also to Reed C. Fulton, for his many helpful suggestions and
practical assistance in constructing much of the heavy equip-
ment; and to Paul W. Bletzer, Allan H. Merrill and Arthur Clif-
ford, for the generous donation of their time in completing
many smaller but no less important tasks. Special thanks are
due Stanley R. Mitchell, who not only participated in numerous
aspects of building the equipment, but who also lent enthus-
iastic assistance with its operation. The guidance of Dr.
Anwar E.Z. Wissa of the Department of Civil Engineering with
the selection of the load cell weighing system is much appre-
ciated.
Credits for typing and assembling the final manuscript are due
to Mrs. Alice Biladeau and Mr. Wayne Wendler. We are grate-
ful to Mr. Donald C. Aldrich for his careful proofreading.
- 262 -
-------
REFERENCES
1. Wallabark, j.w.r and Taylor, A.M. "Civilization Past
and Present/' Rev. Ed., vol. 1, p. 148, Chicago, Scott
Foresman and Co. (1949).
2. American Public Works Association, Public Administration
Service, "Municipal Refuse Disposal," 2nd Edition, p. 7
(1966) .
3. Bering, R. , and Greely, s.A. "Collection and Disposal
of Municipal Refuse," McGraw-Hill, New York (1921).
4. Black, R.J., Munich, A.T., Klee, A.J., Hickman, H.L.,
Jr., and Vaughan, R.D0 "The National Solid Wastes
Survey, an Interim Report," presented at the 1968 Annual
Meeting of the Institute of Solid Wastes of the American
Public Works Association, Miami Beach, Florida (October
24, 1968) .
5. Solid Wastes Management/Refuse Removal Journal, Vol. 14
(4), No. 16, p. 24 (1971).
6. Anderson, D.R., and Colliman, J.P. "Gas Generation and
Movement in Landfills/1 Proceedings National Industrial
Solid Wastes Management Conference, Houston, Texas,
March 24-26, pp. 311-316 (1970).
7. Vaughan, R.D. "National Goals for Solid Waste Manage-
ment ," Proceedings National Industrial Solid Wastes
Management Conference, Houston, Texas, March 24-2(>,
.nagem
.. 14-
pp. 14-16 (1970).
8. Prindle, R.A. "Health Aspects of Solid Waste Disposal,"
Proceedings The Surgeon General's Conference on Solid
Waste Management, Washington, D.C., Public Health Serv-
ice, PHS-Pub-1729, pp. 15-20 (1967).
9. Kenahan, C.B., Sullivan, P.M., Ruppert, J.A., and
Spano, E.F. "Composition and Characterization of
Municipal Incinerator Residues," Report No. 7204, Bureau
of Mines, U.S. Department of Interior (1968).
10. Ostrowski, E.J. "Recycling of Ferrous Scrap from
Incinerator Residue in Iron and Steel Making,"
Proceedings National Incinerator Conference, ASME,
New York, pp. 87-96 (1972).
- 263 -
-------
11. "Municipal Refuse Disposal," loc. cit., p. 159, (1966).
12. Johnson, A.J., and Auth, G.H., Eds. "Fuel and Combus-
tion Handbook," p. 355, McGraw-Hill, New York, (1951).
13. Orning, A. A. "Principles of Combustion" in "Principles
and Practices of Incineration," Corey, R.C., ed., p. 20,
Wiley Interscience, New Yok (1969).
14. Meissner, H.G. "Central Incineration of Community Wastes,"
in "Principles and Practices of Incineration," loc. cit. ,
p. 195.
15. "Municipal Refuse Disposal," loc. cit., p. 164, (1966).
16. Hollander? H.I. "Developing Thermal Energy Release In-
tensity Parameters for Solid Waste," Proceedings Nation-
al Industrial Solid Wastes Management Conference, Hous-
ton, Texas, March 24-26, pp. 129-132, (1970).
17. Niessen, W.R., and Alsobrook, A.F. "Municipal and
Industrial Refuse: Composition and Rates," Proceedings
National Incinerator Conference, ASME, New York , pp .
319-337, (1972).
18. Niessen, W.R., and Chansky, S.H. "The Nature of Munici-
pal Solid Waste," American Society of Mechanical Engin-
eers Incinerator Division, May 15, 1969, New York, New
York.
19. Neissen, W.R., Sarofim, A.F., Mohr , C.M., and Moore, R.W.
"An Approach to Incinerator Combustible Pollutant Control,"
Proceedings National Incinerator Conference, ASME , New
York, pp. 248-259, (1972).
20. Stephenson, J.W. "Burning Wet Refuse," Proceedings Na-
tional Incinerator Conference, ASME , New York, pp. 260-
~ (1972) .
21. Thurnau, R.C., Project Manager, Thermal Degradation Pro-
ject, Solid Waste Research Laboratory, U.S. Environment-
al Protection Agency, National Environmental Research
Center, Cincinnati, Ohio, personal communication, March,
1972.
22. Koenig, L. "The Cost of Municipal Research," Proceedings
National Industrial Solid Wastes Management Conference,
Houston, Texas, March 24-26, pp. 436-442, (1970).
- 264 -
-------
*") "5
. Kreisinger, H., Ovitz, F.K., Augustine, C.E. "Combus-
tion in the Fuel Bed of Hand-Fired Furnaces," U. S.
Bureau of Mines, Technical Paper No. 137, (1916).
24. Nicholls, p. "Underfeed Combustion, Effect of Preheat
and Distribution of Ash in Fuel Beds," U.S. Bureau of
Mines, Bulletin No. 387, (1934).
25. Rose, A.H., and Crabaugh, H.R. "Incinerator Design
Standards: Research Findings," Publications of the Los
Angeles County Air Pollution Control District, No. 60.
26. Weintraub, M., Orning, A.A., and Schwartz, C.H. "Experi-
mental Studies of Incineration in a Cylindrical Combus-
tion Chamber," U. s. Dept. of the Interior, Bureau of
Mines Investigation 6908, (1967).
27. Orning, A.A., Pfeiffer, J.J., Harold, W.C., and Schultz,
J.F. "Experimental Study of Incineration," 51st Annual
Meeting of APCA, Philadelphia, Pa., May 25-29, paper 58-
12, (1958).
28. Schwartz, C.H., Orning, A.A., Snedden, R.B., Demeter,
J.J. and Bienstock, D. "Development of a Vortex Incin-
erator with Continuous Feed," Proceedings National In-
cinerator Conference, ASME, New York, pp. 145-154, TT972) ,
29. Stenburg, R.L., Hangerbrauck, R.P., VonLehunden, D.J.,
and Rose, A,H. "Effects of High Volatile Fuel on In-
cinerator Effluents," 53rd Annual Meeting of APCA,
Cincinnati, Ohio, May 22-26,(1960).
30. Essenhigh, R.H. "Incineration-A Practical and Scientific
Approach," Environmental Sci. and Tech., Vol. 2, p. 524,
(1968).
31. Stumbar, J.P., and Essenhigh, R.H. "Flame Spread Down
Cardboard Sticks and Paper Cylinders," Fire Research
Abstracts and Reviews, Vol. 10, p. 112, (1968).
32. Essenhigh, R.H. "Burning Rates in Incinerators, Parts I
and II," Proceedings National Incinerator Conference,
ASME, New York, p. 87, (1968). ~~~~~~~~~
33. Rao, S.T.R., Gelernter, G., and Essenhigh, R.H. "Scale-
up Factors for Combustion Pot Behavior in Predicting
Incinerator Performance," Proceedings Third National
Incinerator Conference, New York, p. 232,(1968).
- 265 -
-------
34. Biswas, B.K., and Essenhigh, R.H. "Heat Value of Refuse,"
Mech. Eng.f Vol. 91, p. 68, (1969).
35. Kuwata, M., Stumbar, J.P., and Essenhigh, R.H. "Combus-
tion Behavior of Suspended Paper Spheres," Proc. Twelfth
Symposium (International) on Combustion, Poitiers, France,
July, 1968, The Combustion Institute Pittsburgh, pp. 663-
675, (1969).
36. Essenhigh, R.H., and Kuo, T.J. "Combustion and Emission
Phenomena in Incinerators: Development of Physical and
Mathematical Models of Incinerators, Part I: Statement
of the Problem," Proceedings National Incinerator Con-
ference, ASME, New York, pp. 261-271,(1970).
37. Kuwata, J. , Kuo,, T.J., and Essenhigh, R.H. "Combustion
and Emission Phenomena in Incinerators: Burning Rates
and Operational Limits in a Solid Fuel Bed," Proceedings
National Incinerator Conference, ASME, New York, pp. 272-
287, (1970).
38. Stumbar, J.P., Kuwata, M., Kuo, T.J., and Essenhigh, R.H.
"Combustion and Emission Phenomena in Incinerators:
Propagation Velocity and Flame Spread Through Simulated
Fuel Beds, Part I: Flame Spread Down a Simulated Fuel
Bed, Part II, The Mechanism of Flame Spread Over a Cellu-
losic Surface," Proceedings National Incinerator Confer-
ence, ASME, New York, pp. 288-293, 293-303,(1970).
39. Biswas, B.K., Kuo, T.J., and Essenhigh, R.H. "Combus-
tion and Emission Phenomena in Incinerators: Studies on
Combustion Behavior and Extinction Limits of Smoke Flames,"
Proceedings National Incinerator Conference, ASME, New
York, pp. 304-313,(1970).
40. Rao, S.T.R., Kuo, T.J., and Essenhigh, R.H. "Combustion
and Emission Phenomena in Incinerators: Characteriza-
tion of Stirring Factors by Cold Model Simulation,"
Proceedings National Incinerator Conference, ASME, New
York, pp. 314-326, (1970).
41. Kuo, T.J., Kuwata, M., Shieh, W.S., and Essenhigh, R.H.
"Combustion and Emission Phenomena in Incinerators:
Physical and Mathematical Model of a Simple Incinerator:
Part II, The Mechanistic Analysis," Proceedings Nation-
al Incinerator Conference, ASME, New York, pp. 327-330,
(1970) .
- 266 -
-------
• Biswas, B.K., and Essenhigh, R.H. "Some Characteristics
or stabilized Smoke Flames," Combustion and Flame,
vol. 15, pp. 93-96, (1970).
43. Essenhigh, R.H., and Kuo, T.J. "Development of Funda-
mental Basis for Incinerator Design Equations and
Standards," Proceedings Third Mid-Atlantic Industrial
Waste Conference. University nf Marylandr MOV. lQfiQr
PP. lOb-146, (1970).
44. Rao, S^T.R, and Essenhigh, R.H. "Experimental Determina-
tion of Stirring Factors Generated by Straight and Swirl-
ing Jets in Isothermal Combustion Chamber Models," Pro-
ceedings Thirteenth Symposium (International) on Combus-
tion, University of Utah, August iQ7n, Th^ rornhng-Hon
Institute, pp. 603-614, (1971).
45. Biswas, B.K., and Essenhigh, R.H. "The Problem of
Smoke Formation and Control," Paper No. 39f presented
at A.I.Ch.E. 70th National Meeting, Atlantic City,
Aug/Sept. 1971.
46. Shieh, W.S., and Essenhigh, R.H. "Experimental Behavior
of Waste Combustion in a Test Incinerator," Fourth Mid-
Atlantic Industrial Waste Conference, University of Dela-
ware , November,(1970).
47. Shieh, W.S., and Essenhigh, R.H. "Combustion of Computer
Cards in a Continuous Test Incinerator: A Comparison of
Theory and Experiment," Proceedings National Incinerator
Conference, ASME, New York, pp. 120-134, (1972).
48. "Municipal Refuse Disposal," loc. cit., pp. 45-54, (1966)
49. Kaiser, E.R. "Chemical Analyses of Refuse Components,"
publication 65-WA/PID-9, ASME, New York, 1965.
50. "Municipal Refuse Disposal," loc. cit., p. 12, (1966).
51. Kaiser, E.R. "Refuse Composition and Flue Gas Analyses
from Municipal Incinerators," Proceedings National In-
cinerator Conference, ASME, New York, p. 180, (1964) .
52. "Municipal Refuse Disposal," loc. cit., p. 47, (1966).
53. "Control Techniques for Particulate Air Pollutants,"
U.S. Dept. of HEW, PHS, Consumer Protection and Environ-
mental Health Service, National Air Pollution Control
- 267 -
-------
administration Publication No. AP-51.
54. "Proposals for a Refuse Disposal System in Oakland Coun-
ty, Michigan/' Report SW-7d prepared by Jones and Henry
Engineers Limited under Grant D01-U1-00068 from the Bu-
reau of Solid Waste Management, Public Health Service
Publication No. 1960 (1970).
55. Bortz, S.A., and Pincus, A.G. "High Temperature Inciner-
ation," Proceedings National Industrial Solid Wastes
Management Conference, University of Houston, March 24-
26, (1970).
56. Bernard, R.A., and Wilhelm, R.H. "Turbulent Diffusion
in Fixed Beds of Packed Solids," Chem. Eng. Prog., Vol.
46, No. 5, pp. 233-244, (1950).
57. Schotte, W. "Thermal Conductivity of Packed Beds,"
A.I.Ch.E. Journal, Vol. 6, No. 1, p. 63 (1960).
58. Yagi, S., Kunii, D., and Wakao, N. "Internal Develop-
ments in Heat Transfer," Part IV, p. 742, ASME, New York,
(1961).
59. Luikov, A.V., Shashkov, A.G., Vasiliev, L.L., and Fraiman,
Yu.E. "Thermal Conductivity of Porous Systems," Int. Jour-
nal of Heat and Mass Transfer, Vol. 11, pp. 117-137^ TT9~6"8)
60. Barker, J.J. "Heat Transfer in Packed Beds," Ind. and
Eng. Chem., Vol. 57, No. 4, p. 43 (1965).
61. Heggs, P.J., and Handley, D. "Momentum and Heat Trans-
fer Mechanisms in Regular Shaped Packings," Trans. Instn.
Chem. Engrs, Vol. 46, T251 (1968).
62. Furnas, C.C. "Heat Transfer from a Gas Stream to a Bed
of Broken Solids," U.S. Bureau of Mines Bulletin 261
(1932) .
63. Gillespie, B.M., Crandall, E.D., and Carberry, J.J. "Lo-
cal and Average Interface Heat Transfer Coefficients in
a Randomly Packed Bed of Spheres," A.I.Ch.E. Journal,
Vol. 14, pp. 483-490, (1968).
64. Bradshaw, A.V., Johnson, A., McLachlan, N.H. and Chin,
Y-T. "Heat Transfer Between Air and Nitrogen and Packed
Beds of Non-reacting Solids," Trans. Instr. Chem. Engrs.,
Vol. 48, pp. T77-T84, (1970).
65. Kitaev, B.I. "Heat Exchange in Shaft Furnaces, Metallur-
gizdat, (1945).
- 268 -
-------
° lda" F- i Ramaswami, D. , and Hougen, D.A. "Temper-
atures and Partial Pressures of the Surfaces of Catalyst
^articles," A.I.Ch.E. Journal, Vol. 8, pp. 5-11, (1962).
"
' B-z-» Yaroshenko, Yu. G. , and Suchkov, V.I.
Exchange in Shaft Furnaces," p. 90, Pergammon Press,
Oxford, (1967).
68. Baumeister, E.B., and Bennett, C.O. "Fluid-Particle Heat
Transfer m Packed Beds," A.I.Ch.E. Journal, Vol. 4, No.
1, pp. 69-74, (1958). — -
69. Denton, w.H. "Proceedings of the General Discussion on
>" Inst° °f Mech. Eng. (London), p. 370
70. Yagi, s.f and Kunii, D. "Studies on Effective Thermal
Conductivities in Packed Beds," A.I.Ch.E. Journal, Vol.
3, No. 3, pp. 373-381 (1957). -
71. Deissler, L.G.f and Eian, G.S. Natl. Advisory Comm. ,
Aeronaut . , RME. 52005 (1952). - -
72. Yagi, S. , Kunii, D. , and Wakao, N. "Studies on Axial Ef-
fective Thermal Conductivities in Packed Beds," A.I.Ch.E.
Journal , Vol. 6, No. 4, pp. 543-546 (1960).
73. Spalding,, D.B. "Some Fundamentals of Combustion," pp.
53-81, Gas Turbine Series, Vol. 2, Butterworths Scien-
tific Publications, London (1955) .
74. Kays, W.M. "Convective Heat and Mass Transfer," McGraw-
Hill, New York, pp. 327-329 (1966).
75. Amundson, N.R. "Solid-Fluid Interactions in Fixed and Mo-
ving Beds: Part 2, Fixed Beds with Large Particles,"
Ind. and Eng. Chem. , Vol. 48, p. 35 (1956).
76. Sauders, O.K., and Ford, H. "Heat Transfer in the Flow
of Gas Through a Bed of Solid Particles," J. Iron and
Steel Inst. , Vol. 1, pp. 292-328 (1940).
77. Chukhanov, Z.F. "Thermal Regime of Semi Coking, Burning
and Gasification of Particles in a Solid Fuel," Dokl.
Akad. Nauk. SSSR, Vol. 47, No. 3, 1945. Clearing House
for Scientific and Technical Information, Washington,
D.C. 20230 Translation No. 62-10785.
78. Handley, D. , and Heggs, P.J. "The Effect of Thermal Con-
ductivity of the Packing Material on Transient Heat
- 269 -
-------
Transfer in a Fixed Bed," Int. J. Heat Mass Transfer,
Vol. 12, pp. 549-570, Pergammon Press (1969).
79. Schumann, T.E.W., "Heat Transfer: A Liquid Flowing Through
a Porous Prism," J. Franklin Inst., Vol. 208, No. 3,
p. 405 (1929).
80. Klinkenberg, A. "Numerical Evaluation of Equations Des-
cribing Transient Heat and Mass Transfer in Packed Solids,"
Ind. and Eng. Chem., Vol. 40, Pt. 2, No. 10, p. 1992
(1948)"
81. Singer, E., and Wilhelm, R.H. "Heat Transfer in Packed
Beds: Analytical Solution and Design Method; Fluid Flow,
Solids Flow and Chemical Reaction," Chem. Eng. Prog.,
Vol. 46, No. 7, p. 343, (1950).
82. Brinkley, S.R. "Heat Transfer Between Fluid and Porous
Solid Generating Heat," J. App. Physics, Vol. 18, No. 6,
p. 582 (1947).
83. Wilhelm, R.H., Johnson, W.C., and Acton, F.S. "Conduc-
tion, Convection and Heat Release in Catalytic Converters,"
Ind. and Eng. Chem., Vol. 35, No. 5, p. 562 (1963).
84. Kasten, P.R., Lapidus, L., and Amundson, N.R. "Mathema-
tics of Absorption in Beds: Part V, Effect of Intrapar-
ticle Diffusion in Flow Systems in Fixed Beds," J. Phy s.
Chem., Vol. 56, No. 6, P. 683, (1952).
85. Munro, W.D., and Amundson, N.R. "Solid-Fluid Exchange
in Moving Beds," Ind. and Eng. Chem., Vol. 42, No. 8,
p. 1481 (1950).
86. Amundson, N.R. "Solid-Fluid Interactions in Fixed and
Moving Beds: Part 1, Fixed Beds with Small Particles,"
Ind, and Eng. Chem,. , Vol. 48, No. 1, P. 26 (1956).
87. Stewart, M.C., and Saville, J. "A Simplified Heat Trans-
fer Equation for Sinter Beds," Trans. Australian Mech.
and Chem. Eng., p. 135 (1968).
88. Woods, S.E., and Harris, C.F. "Heat Transfer in Sinter
Roasting," Proceedings Symposium on Chemical Engineer-
ing in the Metallurgical Industries, Edinburgh 25-26,
Sept. 1963, London I. Chem. E., p. 77 (1963) .
89. Walker, P.L., Rusinko, R., and Austin, L.G. "Gas Reac-
tions of Carbon," in "Advances in Catalysis," Vol. XI,
- 270 -
-------
P- 135, Academic Press (1959).
. Mayers, M.A. "Temperatures and Combustion Rates in a
nQ^7?ed'" Trans° ASME, Vol. 59, Paper No. FSP-59-8
\ ^- -? J / / .
91. Kolodtsev, K.I. "Dynamics of Gas Formation in Carbon
Beds' J- phy. Chem. U.S.S.R., Vol. 19, p. 417 (1945).
92. Yagi, s., and Kunii, D. "Studies on Combustion of Car-
bon Particles in Flames and Fluidized Beds," 5th Sym-
posium (International) on Combustion, Reinhold, New York,
p. 231 (1955)": - -
93. Parker, A.S., and Hottel, H.C. "Combustion Rate of Car-
bon: Study of Gas-Film Structure by Micro sampl ing, "
^;^fnd Eng° Cheiru° v°l- 28, No. 11, pp. 1334-1341
(1936) .
94. Marskell, W.C0,and Miller, J.M. "Mode of Combustion of
Coal on a Chain Grate Stoker," Fuel, Vol. 25, No. 1,
pp. 4-12 (1946). -
95. Kreisinger, H., Augustine, C.E., and Harpster, W.C.
"Combustion Experiments with North Dakota Lignite,"
U.S. Bureau of Mines, Technical Paper 207 (1919).
96. Mayers, M.A. "Combustion Characteristics of Coke,"
Fuel in Science and Practice, Vol. 21, No. 1, p. 4 (1941),
97. Mayers, M.A. "Some Factors Affecting Combustion in Fuel
Beds," ASME Tech. Publication, No. 77 (1937).
98. Thring, M.W. "Physics of Fuel Bed Combustion," Fuel,
London, Vol. 31, p. 355 (1952).
99. Thring, M.W. "Physical Processes in a Bed of Fuel,"
Inst. of Fuel Journal, Vol. 24, No. 75, pp. 47-65 (1940).
100. Silver, R.S. "Theoretical Treatment of Combustion in
Fuel Beds I, Gas Compositions and Heat Release," Fuel,
Vol. 32, p. 121, (1953) .
101. Silver, R.S. "Theoretical Treatment of Combustion in
Fuel Beds II - Temperature Attained in Combustion,"
Fuel, Vol. 32, p. 138 (1953).
102. De Acetis, j. , and Thodos, G. "Flow of Gases Through
Spherical Packings," Ind. and Eng. Chem. , Vol. 52, p.
1003 (1960) . -
- 271 -
-------
103. Terres, E., Tjia, Hie A., Herrmann, W., Johswich, F.,
Patscheke, G., Pfeiffer, J., and Schwarzmann, H.
Beitrage zur Kenntris der thermischen grundlagen der
Verschwelung und Verkokung von Braunkohlen (Ausang),"
Angew. Chm., Vol. 48, pp. 17-21 (1935).
104. Bangham, D.H., and Thring, M.W. Coal Research, p. 48
(Sept., 1944).
105. Gamson, B.W. "Heat and Mass Transfer: Fluid Solid Sys-
tems," Chem. Eng. Prog., Vol. 47f No. 1, pp. 19-28
(1951).
106. Nichols, L.H.F. "Gas Sampling and Temperature Measure-
ment in Fuel Beds," Inst. of Fuel Journal, Vol. 14,
No. 75, pp. 71-73, (1940).
107. Mott, R.Ao, and Wheeler, R.V. "Coke for Blast Furnaces,"
Iron and Steel Industry Research Council Report No. 1,
Midland Coke Research Committee (1930).
108. Lof, G.O.G., and Hawley, R,W. "Unsteady State Heat Trans-
fer Between Air and Loose Solids," Ind. and Eng. Chem.,
Vol. 40, p. 1061 (1948).
109. Thring, M.W. Coal Research, p. 70 (Sept. 1944).
110. Essenhigh, R.H. "The Speed of Combustion of Solid Fuel
on Beds," in "The Science of Flames and Furnaces,"
Thring, M.W., p. 184, 2nd Edition, John Wiley & Sons,
Inc., New York (1962) .
111. Wu, P.C. "The Kinetics of the Reaction of Carbon with
Carbon Dioxide," Sc.D. Thesis, Dept. of Chem. Eng.,
M.I.T. (1949).
112. Arthur, J.R. "Reactions Between Carbon and Oxygen,"
Trans. Faraday Soc., Vol. 47, pp. 164-178 (1951).
113. Rice, C.W., "Forced Convection of Heat in Gases and Li-
quids - II," Ind. and Eng. Chem., Vol. 16, No. 5, pp.
460-467 (1924).
114. Kaiser, E.L., personal communication to Niessen, W.R.,
Mohr, C.M., and Sarofim, A.F. (1970).
115. von Fredersdorff, C.G., and Elliott, M.A. "Coal Gasi-
fication," Chapter 20 in "Chemistry of Coal Utilization,"
Lowry, H.H., ed„, Supplemental volume, John Wiley and
Sons, New York (1945).
- 272 -
-------
116. Lewis W.K., Gilliland, E.R., McBride, G.T., Sr. "Gas-
irication of Carbon by Carbon Dioxide in Fluidized Pow-
1226 (19" Ind' and En?- Chem., Vol. 41, No. 6, pp. 1213-
17. Simms, D.L., and Law, M. "The Ignition of Wet and Dry
wood Ignited by Radiation," Combustion and Flame, Vol.
11, p. 377-388 (1967).
118. Spalding, D.B. "Conditions for the Stable Burning of
Carbon in an Air Stream," Inst. of Fuel Journal, Vol.
26, No. 155, pp. 289-294 (1953) .
119. Vortmeyer, D. "Das Verhalten von Zvindzonen in exotherm
reagierenden gasdurchstromten Kornerschuttungen," Zeit-
schrift fur Elektrochemie. Vol. 65, pp. 282-289 (196TT7
120. Eigenberger, G. "On the dynamic behavior of the cataly-
tic fixed-bed reactor in the region of multiple steady
states - I. The influence of heat conduction in two
phase models," Chem., Eng. Sci., Vol. 27, pp. 1909-1915
(1972). ~
121. Vortmeyer, D., and Jahnel, W. "Moving reaction zones in
fixed bed reactors under the influence of various para-
meters," Chem. Eng. Sci., Vol. 27, pp. 1485-1496 (1972).
122. Wicke, E. "Einfluss des Stofftransportes auf den Verlauf
heterogener Gasreaktionen," Chem. Eng. Sci., Vol. 8,
p. 61 (1958).
123. Wekmeister, H. "Verbrennungsverlauf bei Steinkolen mitt-
lerer Korngrossen," Arch, fur Warmewirtschaft, Vol. 12,
pp. 225-232 (1931).
124. Maughan, J.D., Spalding, H.B., and Thornton, B.M. "A
Method of Controlling the Temperatures of Travelling
Grate Stokers," Engineering, Vol. 137, pp. 587-589,
656-658, 669-671 (1934).
125. Williams, C.C. "Damage Initiation in Organic Materials
Exposed to High Intensity Thermal Radiation," Technical
Report No. 2, Fuels Research Laboratory, Massachusetts
Institute of Technology, Cambridge, Mass. (1953).
126. Garden, R. "Temperature Obtained in Wood Exposed to High
Intensity Thermal Radiation," Technical Report No. 3,
Fuels Research Laboratory, Massachusetts Institute of
Technology, Cambridge, Mass. (1953).
127. Kreith, F., and Romie, F.E. "A Study of the Thermal
- 273 -
-------
Diffusion Equation with Boundary Conditions Correspon-
ding to Solidification or Melting of Materials Initially
at the Fusion Temperature," Proc. Phys. Soc. B., Vol.
68, pp. 277-291 (1955).
128. Goodman, T.R. "The Heat-Balance Integral and Its Appli-
cation to Problems Involving a Change of Phase," Trans.
ASME, Vol. 8, No. 2, pp. 335-342 (1958).
129. Rosseland, S. "Astrophysik und Atom-Theoretisches
Grundlage," Springer Verlag, Berlin, 1931, pp. 41-44.
130. Bamford, C.H., Crank, J., and Malan, D.H. "The Combus-
tion of Wood. Part I," Proc. Camb. Phil. Soc., Vol.
42, pp. 146-181, (1946).
131. Tao, C.H.j "Generalized Numerical Solutions of Freezing
a Saturated Liquid in Cylinders and Spheres," A.I.Ch.E.
Journal, Vol. 13, No. 1, pp. 165-169 (1967).
132. Shih, Yen-Ping, and Chou, Tse-Chuan, "Analytical Solu-
tions for Freezing a Saturated Liquid Inside or Outside
Spheres," Chem. Eng. Sci., Vol. 26, pp. 1787-1793 (1971).
133. Siegel, R., and Savino, J.M. "An Analytical Solution for
Solidification of a Moving Warm Liquid onto an Isother-
mal Cold Wall," Int. J. Heat Mass Transfer, Vol. 12,
p. 803 (1969).
134. Goodman, T.R. "Integral Methods for Non-Linear Heat
Transfer," Advan. Heat Transfer, Vol. 1, pp. 51-122
(1964).
135. Rossini, F.D., Pitzer, K.S., Arnett, R.L., Braun, R.M.,
and Pimental, G.C. "Selected Values of Physical and Ther-
modynamic Properties of Hydrocarbons and Related Compounds,"
Carnegie Press, Pittsburgh (1953) .
136. Hougen, B.A., and Watson, K.M. "Chemical Process Prin-
ciples," Part I. p. 254, Wiley and Sons, New York, 1950.
137- Dunningham, A.C., Grummell, E.S. "Contribution to the
Study of the Ignition of Fuel Beds," Inst. of Fuel Jour-
nal, Vol. 11, pp. 117-122, 129-133 (19T7Ti
138. Johnstone, R.E., and Thring, M.W. "Pilot Plants, Models
and Scale-up Methods in Chemical Engineering," McGraw-
Hill, New YOrk, 1957, p. 214.
139. Spink, L.K. "Principles and Practices of Flow Meter
- 274 -
-------
Engineering," The Foxboro Company, March 1967, Chapter
i'E-L" Sarofim, A.F., and Howard, J.B. "The
n 2 Underfire Air Rate on a Burning Simulated Re-
« ' Proceedings National Incinerator Conference,
, New York, pp. 135-144 (1972). - - : -
141. Wheeler, A. "Reaction Rates and Selectivity of Catalyst
n ?Jo ChaPter in "Advances in Catalysis," Vol. 3,
n o
PP. <249-327, Academic Press, Inc., New York, 1951.
142. Weisz, _P.B., and Prater, C.D. "Interpretation of Measure-
ments in Experimental Catalysis," Chapter in "Advances
in Catalysis," Vol. 6, pp. 143-196, Academic Press, Inc.,
New York, 1954.
143. Van Heerden, C. "The Character of the Stationary State of
Exothermic Processes," Chem. Eng. Sci". , Vol. 8, p. 133
144. Liu, S.f and Amundson, N.R. "Stability of Adiabatic Packed
Bed Reactors: An Elementary Treatment," Ind. Eng. Chem.
Fundamentals, Vol. 1, p. 200 (1962). -
145. Aris, R. "Some Problems in the Analysis of Transient Be-
havior and Stability of Chemical Reactors," Advances
in Chemistry, No. 109, p. 578, American Chemical Society
(1972) .
146. Aris, R. and Schruben, D.L. "Transients in Distributed
Chemical Reactors, Part I: A Simplified" Model* "' Chem.
Eng. J., Vol. 2, p. 179 (1971). - -~-^**^^— -• -
147. Aris, R. , and Fatina, I.H. "Transients .in Distributed
Chemical Reactors, Part 2: Influence of Diffusion in the
Simplified Model," Chem. Eng. J., Vol. 4, p. 149 (1972).
148. Rhee, Hyun-Ka, Foley, D., and Amundson, N.R. "Creeping
Reaction zone in a Catalytic, Fixed-Bed Reactor: A
Cell Model Approach," Chem. Eng. Sci., Vol. 28, pp.
607-615 (1973).
149. Sarofim, A.F., personal communication, March 5, 1973.
150. Sarofim, A.S. "Combustion Reactions in Overfire Volumes,"
Presented to the Grates and Combustion Committee ASME
Incinerator Division, Thursday, Jan. 15, 1970.
151. Longwell, J.P. and Weiss, M.A. , "High Temperature Reac-
tion Rates in Hydrocarbon Combustion," Ind. Eng. Chem. ,
- 275 -
-------
Vol. 47, p. 1534 (1955).
152. Hottel, B.C., Williams, G.C., Nerheim, M.N. , and Schneider
G.R. , "Kinetic Studies in Stirred Reactors: Combustion
of Carbon Monoxide and Propane," Tenth Symposium (In-
ternational) on Combustion, pp. 111-121, The Combustion
Institute (1965) .
153. Williams, G.C. , Hottel, H.C., and Morgan, A.C. "The Com-
bustion of Methane in a Jet Mixed Reactor , " Twelfth Sym-
posium ( International ) on Combustion , pp. 913-925, The
Combustion Institute
154. Nagle, J. and Strickland-Constable, R.F. "Oxidation of
Carbon Between 1000-2000 C," Proceedings of Fifth Car-
bon Conference, Vol. 1, p. 154 (1962) .
155. Appleton, J.P. "Soot Oxidation Kinetics at Combustion
Temperatures," Presented at AGARD Propulsion and Ener-
getics Panel, 41st Meeting on "Atmospheric Pollution by
Aircraft Engines," London, England, 9-13 April (1973).
156. Park, C., and Appleton, J.P. "Shock-Tube Measurements
of Soot Oxidation Rates," Combustion and Flame, (in
Press), (1973).
157. Engdahl, R.B., and Major, W.S. Bituminous Coal Research,
Inc., Tech. Paper No. VII, 1957.
158. American Gas Association Report No. 3 (1954).
- 276 -
-------
Biographical Note
n SaVx Edward Lloyd Rogers was born on January 30. 1945,
Rhodesia. and is the son of Dr. and Mrs. Edward
o . . .
no • Salisbury, Rhodesia. He attended Ruzawi Preparatory
Sali*£ in Marandellas, Rhodesia, and Prince Edward High School,
1962 hY' Rhodesia- After completing high school in December
h WaS awarded a four year Student Engineer Scholarship
?ne.Angl0 American Corporation of South Africa and worked
Anglo American at the Rhokana mine in Zambia until enter-
University of Edinburgh, Scotland, in the fall of
. He graduated with honors from the University of Edin-
burgh in June 1967. m the summer of 1967 he was awarded a
Fulbright Travel Grant and he entered M.I.T. in the fall of
that year as a teaching assistant. He attended the School of
Chemical Engineering Practice in the fall of 1968 and received
the degree of Master of Science in Chemical Engineering Prac-
tice in February 1969. He then accepted a one-semester appoint-
ment as the Assistant Director of the Bound Brook Practice
School in the spring of 1969. In the summer of 1969 he returned
to M.I.T. as a Research Assistant and commenced work on his doc-
torate.
Joseph Rogers is a member of the honorary society of the
Sigma Xi, the American Society of. Chemical Engineers, the In-
stitution of Chemical Engineers (London) , the Combustion Insti-
tute, and is an honorary citizen of the State of Tennesse. He
is co-author of
"The Effect of Underfire Air Rate on a Burning Simulated
Refuse Bed," Proceedings National Incinerator Conference,
ASME, New York, pp. 135-144, (1972).
"Combustion Characteristics of a Simulated Refuse Bed,"
presented at the 1973 Technical Session of the Central
States Section of the Combustion Institute, March 27,
28, Champaign, Illinois.
He was married to the former Sally Ann Doonan in August
1972 and has accepted a position with Halcon International,
Inc., in New York City.
- 277 -
-------
NOMENCLATURE
Dimensional units are expressed in terms of length (L) , mass
(M) , quantity of heat (H) , time (t) and temperature (T) .
Dimensional quantities are given in terms of Btu, hr, ft, °F
unless otherwise specified.
Symbol
English
A
B
D
D
E
E
Meaning
Frequency factor in kinetic
expression
Ash fraction of fuel, equation
(95)
Specific heat ratio, C /C
ps pg
Interfacial area per unit
volume of bed
Mass transfer driving force
Concentration of specie j
Heat capacity of gas
Heat capacity of solid
Coefficient of molecular
diffusion
Effective coefficient of
diffusion
Particle diameter
Mean radiating beam length,
equation (65)
Turbulent eddy diffusion
Activation energy
Dimensions
Dimensions of
reaction rate
constant
L2lT3
ML
~3
HM"1!'1
LV1
L
L
H lb mole
- 278 -
-------
A Area firing rate, equation
(95)
p
RCS Relative Carbon Saturation
factor, defined by equation
(82)
•
G Mass velocity of fluid ML"2t"1
9 Mass transfer conductance Mt"1!,"2
A 1 O
9 Mass transfer conductance Mt L
for very small mass transfer
rates
Hv Effective latent heat of vapor- HM"1
ization,
AH ps
v —. 1
HV Latent heat of vaporization HM
c -1
H Overall heat of combustion HM
H. Heat of reaction per mole of H mole
-1 j - endothermic negative,
exothermic positive
-1 -2 -1
h Radiation contribution to Ht L T
r effective thermal conductivity
of porous solid - defined by
equation (14)
-1 -2 -1
h Heat transfer coefficient Ht L T
S based on unit area
-1 -2 -1
h Effective heat transfer coef- Ht L T
S ficient based on unit area,
defined by equation (6)
hv Heat transfer coefficient Ht L T
based on unit volume
•j_ k p (N,,
JD eg Sc
- 279 -
-------
-------
NBi Biot Number /h D
NLe Lewis Number /BL, \
l-^£\
VW
NPr Prandtl Number /C V
\
'
g
NRe Modified Reynolds Number /D G'
NSC Schmidt Number / y
Nc. Sherwood Number /k D
Sh let
Qne New total heat release from Ht L
z=0 to z=°°
Q(z) Net heat release Ht"1!,""3
— -2
Q(z) Conductivity normalized net TL
heat release
2 -2 -1 -1
R Gas constant ML t T mole
R Radius of particle L
r. Rate of reaction i ML t
r. Effective rate of reaction i ML t
-3 -
R* Rate of reaction
T Ambient temperature T
T d Adiabatic temperature T
T Gas temperature T
- 281 -
-------
t
U
V
V
x
Y
z
Symbol
Greek
B
Y
<5
e
Solid temperature
Time
Mass velocity of solid
Volatile fraction of fuel,
equation (95)
UC
GC
•8
T
t
L
L
Weight fraction of pyrolyzable
material
Defined by equation (35)
Defined by equation (34)
Vertical position
Oxygen consumption distance
Meaning
Defined by equation (45)
V/k", equation (51)
GC
pg, equation (41)
Void fraction
Emissivity
Stoichiometric coefficient of com-
ponent i in reaction j
Dimensions
Gas density
Solid density
ML
ML
-3
- 282 -
-------
*
Shape factor (unity for
spheres), equation (10)
Gas viscosity
Stefan-Boltzman constant
Net heat absorbed in endo-
thermic reactions, equa-
tion (4)
Mass rate of appearance of
specie j, equations (3)
and (5)
Symbol
Superscript
b
d
i
o
v
w
Subscript
j
Meaning
Bulk
Dry
Interface
Initial
Vaporization
Wet
Specie j
ML'V1
- 283 -
-------
THE FOLLOWING PAGES ARE DUPLICATES OF
ILLUSTRATIONS APPEARING ELSEWHERE IN THIS
REPORT. THEY HAVE BEEN REPRODUCED HERE BY
A DIFFERENT METHOD TO PROVIDE BETTER DETAIL
-------
I
u" 3 cr H
£-6 £ 2?
* 3 *"
III
0-0
2 -PS
Ib
Fig. 44. Experimental Incinerator, Gas Analysis
Train and Peripheral Equipment.
-------
Fig. 45. Interior of Fuel Bed Section with Thermocouple
Probes Installed.
This page is reproduced at the
hack of the report by a different
reproduction method to provide
- 1 6 0 - better detail.
-------