JFE TECHNICAL REPORT

No. 19 (Mar. 2014)

Numerical Simulation of Combustion Behavior

in Stoker-type Incineration System†
DENDA Tomohiro*1   NAKAYAMA Takashi*2   IWASAKI Toshihiko*3

Abstract: 3-dimensional model of actual incinerator size, it is nec-

To improve the performance of the stoker-type incin- essary to include the detailed combustion reaction mech-
eration system, JFE Engineering has developed the anism, and the fact that calculations required an enor-
numerical simulator to understand the combustion mous amount of time was also an issue. As the
behavior in the incinerator. In the first step, the suitable development of a numerical simulation technology
reaction mechanism for combustion was selected. In which can easily handle actual-size large-scale 3-
next step, by using the actual-size 3-dimensional entire dimensional models would provide an extremely useful
incinerator model, temperature and CO concentration tool for research and development and design, the estab-
was calculated and compared with measured tempera- lishment of such a technology has been strongly desired.
ture and CO concentration in actual plant. As a result, With the aim of establishing a numerical simulation
calculated values agreed with measured values approxi- technology like that outlined above, JFE Engineering
mately. It showed that this numerical simulator is appro- carried out development, focusing on the JFE Hyper 21
priate for an analysis of the combustion behavior of the Stoker System, which is the company’s state-of-the-art
stoker-type incineration system and is useful for a devel- stoker-type waste incinerator. First, using a general-
opment and design tool. purpose thermal fluid code (Fluent), a simple, accurate
combustion reaction mechanism for gas combustion in
the incinerator was selected based on a study with a
1. Introduction
2-dimensional model. Next, the actual-size 3-dimensional
Since the Great East Japan Earthquake in March entire incinerator model was developed, in which the
2011, heightened expectations have been placed waste calculation region was expanded to the entire incinera-
power generation as a form of renewal energy. Accom- tor, and calculations were performed. The calculation
panying this, demands for high performance in waste results and the results of measurements of an actual
incinerators has also become stronger, increasing the incinerator were compared, and were approximately in
importance of technical development. agreement, confirming the appropriateness and useful-
Until now, development of waste incinerators had ness of this numerical simulation technology. The results
been carried out mainly by experimental techniques. of this development are described in this paper.
Although numerical simulation techniques had also been
attempted1), many of those efforts were fundamental
2. Features of JFE Hyper 21 Stoker System
studies using 1-dimensional or 2-dimensional models.
As reasons for this, because the material being inciner- First, the JFE Hyper 21 Stoker System, which is JFE
ated is waste, properties are non-uniform and quantita- Engineering’s state-of-the-art stoker-type waste inciner-
tive fluctuations are large. Moreover, in order to ator, will be explained.
make highly accurate calculations with a large-scale The JFE Hyper 21 Stoker System is based on the two

†
 Originally published in JFE GIHO No. 32 (Aug. 2013), p. 61–66 *2
 Senior Researcher,
Thermal & Combustion Engineering Gr.,
Research Center of Engineering Innovation
Technology Headquarters,
JFE Engineering

*1
 Senior Researcher, *3
 Chief Principal,
Thermal & Combustion Engineering Gr., Environmental Solutions Sector,
Research Center of Engineering Innovation JFE Engineering
Technology Headquarters,
JFE Engineering

139
 Numerical Simulation of Combustion Behavior in Stoker-type Incineration System

Photo 1  Photograph of flame in incinerator

Fig. 1  Two way gas flow incinerator pyrolysis gas evolved from the waste layer to collide,
and also forms a counterflow field. Because the counter-
flow diffusion flame stabilizes easily in the low flow
velocity region of this counterflow field, the effect of
radiant heating expands from this flame to the waste
layer. As a result, pyrolysis of the waste is accelerated
and an extremely stable combustion field is formed. This
makes it possible to realize stable combustion with a
low excess air ratio, which had been difficult to achieve
in conventional stoker-type incinerators.
Photo 1 was taken from the downstream side of the
incinerator in an actual JFE Hyper 21 Stoker System
during stable operation. A stable flame, including a lumi-
Fig. 2 F
 lame stabilization image in cross section at combustion nous flame directly above the waste layer, can be seen,
beginning area
and the counterflow field shown in Fig. 2 has formed.

way gas flow incinerator2), which was adopted indepen-

dently by this company from an early date, and is a next- 3. Selection of Simple
generation stoker-type incinerator3-5) that utilizes High- Combustion Reaction Mechanism
temperature Air Combustion Technology (HiCOT) to
3.1 Outline of
realize a further reduction in environmental pollutants
Combustion Reaction Mechanism
and substantial improvement in power generating effi-
ciency. For the plant flow, the reader may refer to a pre- The purpose of this development was to develop a
vious report6). numerical simulation technology for waste incinerators,
The structure of the two way gas flow incinerator is which enables easy handling of a large-scale 3-dimensional
shown in Fig. 1. The flow in the incinerator is divided model of actual incinerator size. From this viewpoint, it
into two parts, i.e., sub flue gas and main flue gas, by was considered desirable to select a simple combustion
providing a partition called an “intermediate ceiling” in reaction mechanism which would not increase the com-
the combustion chamber. The sub flue gas, which putational load more than necessary.
includes an uncombusted component, and the main flue Therefore, we compared and studied two simple
gas, which has a comparatively large oxygen content, combustion reaction mechanisms, which were thought
collide head-on in the gas mixing chamber, where they to be practical, using the high accuracy detailed combus-
are mixed and stirred effectively. tion reaction mechanism for CH4, GRI1.27) (32 chemical
As another technology, HiCOT is applied in the JFE species, 177 reaction formulas) as a standard. These two
Hyper 21 Stoker System. Figure 2 presents a lateral mechanisms were DRM198) (21 chemical species, 84
cross-sectional view of the incinerator, showing a sche- reaction formulas), which is a skeletal reaction mecha-
matic illustration of the combustion behavior when nism based on GRI1.2, and a reaction mechanism for
HiCOT is applied. In the combustion beginning region, CO2/H2 proposed by Yetter9) (13 chemical species, 35
where the waste in the furnace is dried and combustion reaction formulas).
begins, high temperature air is blown directly above the What is important here is the composition of the
waste layer from the right and left side walls of the waste pyrolysis gas. It is known that the main compo-
incinerator. This causes the high temperature air and the nents are H2 and CO, and the gas also contains hydro-

140  JFE TECHNICAL REPORT No. 19 (Mar. 2014)

 Numerical Simulation of Combustion Behavior in Stoker-type Incineration System

carbons such as CH4, C2H4, etc.3). Although DRM198) both cases, and in particular, the calculation time was
considers virtually all the chemical species which are reduced to about 1/4 when using the Yetter mechanism 9).
thought to be components of pyrolysis gas from waste, Thus, this study confirmed that a large reduction in the
because the reaction mechanism was originally prepared calculation load is possible by adopting the Yetter mech-
for use with CH4, it is necessary to determine the com- anism 9).
bustion characteristics when calculated with a pyrolysis Figure 4 shows a comparison of the temperature dis-
gas composition having high concentrations of CO and tribution and maximum flame temperature. As the calcu-
H2. On the other hand, CH4, C2H4, etc. can not be con-
sidered because Yetter’s mechanism9) is a reaction Table 1  Calculation conditions
mechanism for use with CO/H2. Therefore, it is neces-
Case Case A Case B Case C Case D
sary to set and perform calculations using a simulated
Reaction model GRI1.2 7) DRM19 8) DRM19 8) Yetter 9)
pyrolysis gas in which the flammable components are
Number of species 32 21 21 13
limited to CO and H2.
Number of elementary reactions 177 84 84 35

3.2 Comparison and Study Mesh size

X (m) 0.01 0.01 0.01 0.01

of Combustion Reaction Mechanisms Y (m) 0.04 0.04 0.04 0.04

by 2-Dimensional Model Number of meshes 5 054 5 054 5 054 5 054

Velocity (m/s) 0.5 0.5 0.5 0.5
In comparison and study of the combustion reaction Temperature (°C) 1 000 1 000 1 000 1 000
mechanisms, a general-purpose thermal fluid code (Flu- H2 0.003 0.003 0.009 35 0.009 35
ent) was used. The analysis domain was the 2-dimen- CH4 0.017 0.017 0 0
sional jet field shown in Fig. 3. This domain is enclosed C2H4 0.017 0.017 0 0
Pyrolysis gas
by a symmetric plane and an adiabatic wall; the inner Mass CO 0.13 0.13 0.187 04 0.187 04
side is provided with a pyrolysis gas supply part, and the fraction CHO 0 0 0 0
outer side is provided with a high temperature oxidizer CO2 0.239 0.239 0.230 61 0.230 61
gas supply part. H2O 0.20 0.20 0.192 81 0.192 81
Table 1 shows the calculation conditions. The tem- N2 0.394 0.394 0.380 18 0.380 18

perature of the pyrolysis gas was 1 000°C, and that of Velocity (m/s) 1.0 1.0 1.0 1.0

the oxidizing agent was 250°C, which is the same as in Temperature (°C) 250 250 250 250

an actual incinerator. In Case A and Case B, the pyroly- Oxidizer

O2 0.16 0.16 0.16 0.16

sis gas composition3) which was obtained experimentally Mass CO2 0.09 0.09 0.09 0.09
fraction H2O 0.08 0.08 0.08 0.08
was used, and in Case C and Case D, the simulated
N2 0.67 0.67 0.67 0.67
pyrolysis gas composition was used. Here, the simulated
pyrolysis gas composition was determined so that the
enthalpy of formation and the atomic numbers of C and Table 2  Comparison of calculation time
H were the same as those of the experimentally-obtained Reaction model CPU time/Iteration 40 000 Iteration
pyrolysis gas. 7)
GRI 1.2 1.77 s 19 h
Table 2 shows the calculation time when using the 8)
DRM 19 0.81 s  9h
respective combustion reaction mechanisms. In compar-
Yetter 9) 0.45 s  5h
ison with GRI1.27), the calculation time is shortened in
CPU: Xeon5355 (2.66 GHz)×2
Fluent: Parallel×4

Fig. 4 C
 omparison of temperature distributions and maximum
Fig. 3  Two-dimensional domain of calculation flame temperature

JFE TECHNICAL REPORT No. 19 (Mar. 2014)  141

 Numerical Simulation of Combustion Behavior in Stoker-type Incineration System

lation results of DRM198) (Case B) and GRI1.27) (Case

A) show extremely close agreement, application of
DRM198) is considered possible. Case C and Case D
(results of calculation by Yetter mechanism) are cases in
which the simulated pyrolysis gas composition is used.
Although there is some difference in the maximum
flame temperature in comparison with Case A, the tem-
perature distribution, which is most important for under-
standing combustion characteristics, shows generally
good agreement. Based on this fact, even though this
mechanism uses a simulated pyrolysis gas composition Fig. 5  The domain of the 3-dimensional entire incinerator model
in which the flammable components are replaced with
only H2 and CO, this is not considered to be a major Table 3  Calculation conditions
obstacle, and application of a simple combustion reac- FDF
FDF (Air)
tion mechanism with this composition is considered pos- (Pyrolysis gas) RDF HADF CDF
sible. Drying Comb. Drying Comb. Post
From the above, based on the characteristics of each Flow
(Nm3/h) 5 250 5 250 1 828 2 614 786 3 100 2 100 2 710
rate
combustion mechanism, and considering calculation
Temp. (°C) 600 600 60 60 60 154 220 20
time to be the most important priority in this study, the
H2 (vol%) 11.4 11.4 0 0 0 0 0 0
Yetter mechanism9) was selected as the simple combus-
CO (vol%) 16.3 16.3 0 0 0 0 0 0
tion reaction model and was used in subsequent numeri-
CO2 (vol%) 12.8 12.8 0 0 0 11.9 6.3 0
cal simulations.
H2O (vol%) 26.2 26.2 0 0 0 21.0 11.1 0
O2 (vol%) 0 0 21.0 21.0 21.0 4.0 12.0 21.0
N2 (vol%) 33.3 33.3 79.0 79.0 79.0 63.0 70.5 79.0
4. Combustion Analysis
FDF : Forced draft fan CDF : Cooling draft Fan
by 3-Dimensional Entire Incinerator Model HADF : High-temperature air draft fan RDF : Recirculated draft Fan
Drying : Drying zone Comb. : Combustion zone
4.1 Calculation Conditions Post : Post-combustion zone

Next, we examined the appropriateness and useful-

ness of this numerical simulation technology when position of waste charged to actual incinerators, and this
applied to an actual-size 3-dimensional entire incinera- gas was supplied from FDF. To give the FDF injection
tor model. nozzle a structure close to that of the fire grate in an
Figure 5(a) shows the mesh of the 3-dimensional actual incinerator, drying zone, combustion zone, and
entire incinerator model. The scale of the incinerator post-combustion zone were divided into slit shapes, and
was decided to be 13 000 mm in height, 9 000 mm in a structure which enabled alternating supply of combus-
length, and 4 000 mm in width, which is the same as that tion air and pyrolysis gas was adopted.
of an actual plant (treatment capacity: 120 t/d scale).
4.2 Analysis of Combustion Behavior in
The number of meshes is approximately 85 000.
Incinerator
The Yetter combustion reaction mechanism9) selected
in the previous section was applied to the combustion Figure 6 shows the temperature distribution at a ver-
reaction mechanism. A finite rate model10) was used for tical cross section in the incinerator longitudinal direc-
combustion, a k-e model11) was used for the turbulent
flow, and a DO model12) was used for radiation, respec-
tively. A uniform temperature (1 000°C) was given as a
wall surface condition, and heat loss through the wall
surface was considered. As in an actual incinerator, pri-
mary combustion air (FDF), secondary combustion air
(CDF), high temperature air (HADF), and recirculated
exhaust gas (RDF) were supplied from the respective
nozzles shown in Fig. 5(b).
Table 3 shows the calculation conditions. With this
model, it is necessary to give the pyrolysis gas of the
waste as the fuel. The amount and composition of the
pyrolysis gas were estimated from the amount and com- Fig. 6  Temperature distribution

142  JFE TECHNICAL REPORT No. 19 (Mar. 2014)

 Numerical Simulation of Combustion Behavior in Stoker-type Incineration System

Fig. 8  Temperature distribution (A-A' Cross section in Fig. 6)

Fig. 7  CO concentration distribution

tion. From Fig. 6, it can be understood that a high tem-

perature region exists in the sub-flue rather than in the
main flue. This is because the pyrolysis gas supplied
from FDF burns above drying zone and combustion
zone, and the larger part of the high temperature com-
bustion gas then flows downstream via the sub-flue.
This combustion gas flow can also be confirmed from
the CO concentration distribution in the vertical cross
section in the longitudinal direction shown in Fig. 7.
That is, a region of high CO concentration is distributed
in the sub-flue, and the combustion gas containing this Fig. 9 C
 O concentration distribution (B-B' cross section in Fig. 7)
high concentration of CO, which was generated from the
combustion region directly above the fire grate, passes
through the sub-flue. The CO gas that passes through the As discussed above, this study demonstrated that
sub-flue collides with the gas from the main flue side, the features of both the two way gas glow incinerator
which has comparatively high oxygen content. The CO and HiCOT can be reproduced satisfactorily in the
burns in the presence of this oxygen, and as a result, the 3-dimensional entire incinerator model in which the
temperature increases and the CO concentration combustion region is expanded to the entire incinerator.
decreases, as can be confirmed from Fig. 6 and Fig. 7. This confirmed the appropriateness of this numerical
These are distinctive features of the two way gas flow simulation technology for analysis of the combustion
incinerator, and are in agreement with the phenomena behavior in incinerators.
observed in actual incinerators.
4.3 Comparison with Measured Results of
Figures 8 and 9 show the temperature distribution
Actual Incinerator
and CO concentration distribution in a vertical cross
section in the incinerator width direction (A-A' cross Next, in order to make a quantitative evaluation, the
section in Fig. 6 and B-B' cross section in Fig. 7), results of these calculations and the results of measure-
respectively. From Fig. 8, it can be confirmed that the ments of an actual incinerator were compared. The mea-
high temperature air supplied from HADF and the pyrol- surement points in the actual plant were the combustion
ysis gas supplied from FDF collide in the center of the chamber and the gas mixing chamber. The temperature
incinerator, and form a combustion region directly above was measured at 5 points, and the CO concentration was
the incinerator bottom (supposing the top surface of the measured at 3 points. Figures 10 and 11 show a com-
waste layer). From Fig. 9, it can be understood that a parison of the calculated results and the measured results
high CO concentration region exists near the bottom of at the actual incinerator for temperature and CO concen-
the incinerator walls. This occurs because a circulating tration, respectively. The difference in the calculated and
flow is formed in the incinerator due to the effect of measured temperature was within ±50°C, and that of the
HADF supplied from the side walls, and part of the CO concentration was within ±1.5 vol. %. Thus, the
combustion gas generated in the incinerator center stag- temperature and CO concentration could be calculated
nates. This shows the distinctive feature of a stable and more accurately by incorporating the Yetter combustion
uniform combustion region of the JFE Hyper 21 Stoker reaction mechanism9) in the model, and values which
System applying HiCOT. were approximately close to the measured values of an

JFE TECHNICAL REPORT No. 19 (Mar. 2014)  143

 Numerical Simulation of Combustion Behavior in Stoker-type Incineration System

wider use of waste power generation, JFE Engineering

is grappling with the development of numerical simula-
tion technologies that provide useful tools for research
and development and design, with the aim of developing
higher performance stoker-type waste incinerators.
In this paper, first, the combustion reaction mecha-
nism proposed by Yetter9) was selected as a combustion
reaction mechanism which enables easy handling,
including large-scale 3-dimensional models of actual
incinerator size. The 3-dimensional entire incinerator
model was then developed, and the usefulness of the
numerical simulation technology as a tool for research
and development and design was clarified by a quantita-
Fig. 10  omparison between calculated temperature and
C tive evaluation.
measured temperature In the future, the authors intend to work energetically
to further improve and sophisticate this numerical simu-
lation technology.

The authors wish to take this opportunity to express

their profound appreciation to Prof. Kobayashi of Insti-
tute of Fluid Science, Tohoku University, for his gener-
ous guidance and advice in this research.

References
  1) Yasuda, Toshihiko; Shigemasa, Sachiko; Furubayashi, Michitaka;
Tomiyama, Shigeo. Numerical Prediction of Waste Bed Combus-
tion in Municipal Solid Waste Incinerator. Proceedings of the
42nd Symposium (Japanese) on Combustion. 2004, p. 417–418.
  2) Iwasaki, Toshihiko; Nakayama, Takashi; Denda, Tomohiro;
Kinoshita, Seiji; Nakagawa Tomonori; Ayukawa, Susumu.
Advance in Waste Incineration Technology. Journal of the Com-
Fig. 11 C
 omparison between calculated CO concentration bustion Society of Japan. 2011, vol. 53, no. 164, p. 85–90.
and measured CO concentration   3) The research and development project on advanced high-tempera-
ture air combustion technology. NEDO Report. 1999, p. 213–240.
  4) The research and development project on advanced high-tempera-
ture air combustion technology. NEDO Report. 2000, p. 253–292.
actual incinerator could be obtained.   5) The research and development project on advanced high-tempera-
The study described above showed that the ture air combustion technology. NEDO Report. 2001, p. 235–295.
3-dimensional entire incinerator model can also repro-   6) Denda, Tomohiro; Nakayama, Takashi; Kinoshita, Seiji; Nak-
agawa Tomonori; Ayukawa Susumu. Proceedings of the 22nd
duce the combustion behavior in the incinerator quanti- Annual Conference of Japan Society of Material Cycles and
tatively, and thus demonstrated that this numerical simu- Waste Management. 2010, p. 351–352.
lation technology is a useful tool for research and   7) http://www.me.berkeley.edu/gri-mech/
  8) http://www.me.berkeley.edu/drm/
development and design.   9) Yetter R. A.; Dryer F. L.; Rabitz H. A Comprehensive Reaction
Mechanism For Carbon Monoxide/Hydrogen/Oxygen Kinetics.
Combustion Science and Technology. 1991, vol. 79, p. 97–128.
5. Conclusion 10) ANSYS Fluent 12.0 theoretical guide book. chapter 7.
11) ANSYS Fluent 12.0 theoretical guide book. chapter 4.
In response to the increasing need to encourage 12) ANSYS Fluent 12.0 theoretical guide book. chapter 5.

144  JFE TECHNICAL REPORT No. 19 (Mar. 2014)