Renewable Energy 139 (2019) 611e620

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Simulation and optimization of rice husk gasification using intrinsic

reaction rate based CFD model
Xiaoyan Gao a, 1, Fei Xu b, 1, Fubing Bao a, Chengxu Tu a, Yaning Zhang c, *, Yingying Wang d,
Yang Yang e, Bingxi Li c, **
a
College of Metrology and Measurement Engineering, China Jiliang University, Hangzhou, 310018, China
b
Electronics Business Unit, Ansys, Inc., Austin, TX, 78746, USA
c
School of Energy Science and Engineering, Harbin Institute of Technology, Harbin, 150001, China
d
Taizhou Huangyan Architecture and Engineering Quality Supervision Station, Taizhou, 318020, China
e
Manufacturing Technology & Engineering Division, Corning Incorporated, Corning, NY 14831, USA

a r t i c l e i n f o a b s t r a c t

Article history: Entrained flow gasification processes of rice husk were simulated and optimized in this study using an
Received 15 December 2018 intrinsic reaction rate based CFD model. A detailed sensitivity analysis was conducted to characterize the
Received in revised form effects of operation parameters on the gas composition, gas production and cold gas efficiency. Gasifi-
4 February 2019
cation temperature, average particle diameter, ER (equivalence ratio) and CO2/biomass (mass ratio of
Accepted 20 February 2019
carbon dioxide to biomass) are important operation parameters affecting the gasification process, and
Available online 23 February 2019
they were investigated in this study. Three-objective optimization of rice husk gasification was per-
formed base on the response surface methodology (RSM) to maximize CO content, gas production, and
Keywords:
Rice husk
cold gas efficiency, and the Pareto optimal solutions were obtained from NSGA-II (non-dominated sorting
Gasification genetic algorithm) to instruct gasification operation. With standard TOPSIS (technique for order pref-
Sensitivity analysis erence by similarity to ideal situation), the optimal solutions with CO concentration of 25.15%, gas
Multi-objective optimization production of 1.96 Nm3/kg and cold gas efficiency of 65.34% were obtained.
Pareto optimal solution © 2019 Published by Elsevier Ltd.

1. Introduction structural parameters [9,10]. The experimental study conducted by

Zhou et al. [11] showed that cold gas efficiency was enhanced by
Energy crisis and environmental issues require alternative re- more than 10% with the increase in gasification temperature, since
sources of conventional fossil fuels [1e3]. Being abundantly avail- the yields of hydrogen and carbon monoxide were improved at
able and carbon neutral, biomass is deemed to have the potential to higher temperatures. Kaewpanha et al. [12] concluded from their
fulfill day to day energy demands [4e6]. Among all the thermo- steam co-gasification experiments on seaweed and cedar that the
chemical conversion methods, gasification is an attractive and alkali and alkaline earth species in seaweed promoted the gasifi-
efficient technology to convert a wide range of low energy density cation process, resulting in the growth of gas production with
materials into high value flammable gases or chemical feedstocks seaweed ratio. Kuba and Hofbauer [13] carried out an experimental
[7,8]. parametric study to investigate the effects of operating tempera-
Biomass gasification is a very complex process involving mois- ture and bed height on woody biomass gasification process in dual
ture release, devolatilization, partial combustion and gasification of fluid bed. Their results indicated that increasing either gasification
volatiles and char particles. A detailed understanding of the temperature or bed height could lead to the decreases in the total
biomass gasification mechanism requires the investigation on tar yields.
gasification performances under different operating conditions and For research on biomass gasification processes, experimental
methods are relatively difficult, expensive, and time-consuming
[14]. Therefore, CFD (computational fluid dynamics) simulation
becomes an attractive tool, which can offer the parameter varia-
* Corresponding author.
tions spatially and temporally, e.g., concentration, velocity, tem-
** Corresponding author.
E-mail addresses: [email protected] (Y. Zhang), [email protected] (B. Li). perature, etc. [15,16]. Concentrating on particulate phase with
1
These authors contributed equally to this work. blended acceleration model, Liu et al. [17] established a

https://doi.org/10.1016/j.renene.2019.02.108
0960-1481/© 2019 Published by Elsevier Ltd.
612 X. Gao et al. / Renewable Energy 139 (2019) 611e620

Nomenclature R universal gas constant (J/kmol$K); reaction rate

M molecular weight (kg/kmol)
A pre-exponential factor (1/s$Pam) Sm, SF, Sh source terms for mass, momentum and energy
De effective diffusion coefficient (m2/s) equations
Di molecular diffusion coefficient of gas component i Tg, Tp temperature of gas phase and particle (K)
(m2/s) X carbon conversion degree
Dk,i Knudsen diffusion coefficient of gas species i (m2/s) mp mass of particle (kg)
d diameter (m) u gas phase velocity (m/s)
E activation energy (kJ/mol) v velocity of particle phase (m/s)
Hrec reaction heat (J/kg) h effectiveness factor
h specific enthalpy of gas phase (J/kg) and convective qR radiation temperature (K)
heat transfer coefficient (W/m K) l thermal conductivity of gas phase (W/m K)
kc thermal conductivity of bulk gas (W/m K) mt turbulent viscosity (kg/m s)
0
kg mass transfer coefficient between vapor and bulk gas yi;r , y}i;r stoichiometric coefficients for reactant i and product i
(m/s) in r reaction
kf,r, kb,r forward rate constant and backward rate constant for rg, rp density of gas phase and particle (kg/m3)
r reaction 4 Thiele modulus
P pressure (Pa, atm)

comprehensive CFD model to study the biomass gasification in a objectives were: (a) to investigate the sensitivities of gasification
dual fluidized bed system, they concluded that syngas yield and performances varied by the changes in the operation parameters,
hydrogen content were improved by the increases in gasifier and (b) to obtain the optimal conditions to operate rice husk
temperature and steam/biomass ratio. In contrast, Couto et al. [18] entrained flow gasification.
developed and validated a 2D multiphase model for biomass
gasification in a bubbling fluidized bed, and their numerical results 2. Numerical method
indicated that the cold gas efficiency increased with oxygen con-
tent but decreased with steam/biomass ratio. Ismail et al. [19] 2.1. CFD model description
investigated gasification process of coffee husk in a fluidized bed,
and the simulated data revealed that both moisture content and An intrinsic char reaction rate based CFD model was previously
equivalence ratio had negative effects on cold gas efficiency and established and verified for predicting the biomass gasification
syngas higher heating value. According to the numerical results of performances in entrained flow reactors [25], it was therefore
CO2-biomass gasification using Eulerian method, Cheng et al. [20] adopted in the present study. The gas species inside the reactor
found CO content raised with the promotion of CO2/biomass ratio, were modeled as continuous phase according to the Euler method,
while the H2 content dropped. Commonly, the biomass gasification and the standard k-ε turbulent model was applied to close the
process is affected by operating parameters including bed tem- governing equations. The biomass particles were modeled as
perature, residence time, moisture content, particle size, etc. [9,21], discrete phase according to the Lagrange approach. The governing
and the knowledge of the effects of operating parameters on equations for both gas phase and solid phase are summarized and
gasification performances is indispensable to the optimization of they are shown in Table 1.
gasifier operations. Once entering a gasifier, the biomass particles generally go
Optimization of reactor operation is one of the critical methods through the processes of moisture release, devolatilization, ho-
to improve the performances of biomass gasification [22]. Previous mogenous reaction and heterogeneous reaction. Hence those
efforts have been made to promote gasification efficiency. Campoy chemical reactions were all considered in the CFD model, and the
et al. [23] conducted a process optimization of biomass gasification corresponding rate equations for each process are displayed in
with air-steam and reported that a maximum cold gas efficiency of Table 2. Assumed to be a diffusional limited step [26], the release of
70% could be obtained at the optimal condition. Based on the moisture stored as liquid water or water bound in biomass was
developed stoichiometric equilibrium model, Ng et al. [24] opti- modeled by wet combustion model. One-step global single reaction
mized the gasification condition to gain the maximum H2 pro- was adopted for describing the pyrolysis, thus the devolatilization
duction. They suggested that it's necessary to optimize the rate relied on the unpyrolyzed mass of biomass (as listed in Table 2).
operating parameters (such as ER, particle size, etc.) simultaneously For the partial combustion and gasification of gas phase, the reac-
in future work. With the help of CFD model and response surface tion rate was calculated from the finite-rate/eddy-dissipation
methodology, Silva and Rouboa [9] provided optimal gasification model, which chose the smaller value of the Arrhenius rate and
conditions for different objectives. Generally, maximizing the eddy-dissipation rate. Table 3 gives kinetic constants for the gas
multi-index of biomass gasification is expected in many industrial phase reactions. The heterogeneous reaction between char and
processes, therefore the multi-objective optimization combined gases is the rate controlling step inside a gasifier, thus it receives
with stable operating parameters is of great importance [9]. How- much attention. In the realistic gasification process, the combustion
ever, limited researches concerned multi-objective optimization of and gasification of char particles always take place in regime II
biomass gasification. where the actual char conversion rate depends on the combined
In order to gain a better understanding of biomass gasification effect of physical diffusion and chemical reaction. In the char re-
and offer an effective way to improve gasification performances, action submodel, the char reaction rate was expressed by a com-
sensitivity analysis and multi-objective optimization of rice husk bination of effectiveness factor h and intrinsic reaction rate Rint to
gasification in an entrained flow reactor were performed in this take the chemical reaction rate and gas diffusion rate into account,
study using an intrinsic reaction rate based CFD model. The specific and the expressions are listed in Table 2. The deduction of the
 X. Gao et al. / Renewable Energy 139 (2019) 611e620 613

Table 1
Governing equations for CFD model.

Phase Transportation equation

Gas Mass vðrg ui Þ

¼ Sm
vxi
!
Momentum vðrg ui uj Þ vPg v vu
¼  þ m i  rg u0i u0j þ rg g þ SF
vxj vxi vxj vxj
!
Energy vðrg ui hÞ v vT
¼ l g þ Sh
vxi vxj vxj
  !
Species v v m vYi DT;i vTg
ðrg uj Yi Þ ¼ rg Di;m þ t þ þ SpYi þ RfYi
vxj vxj Sct vxj Tg vxj
"  #
k-ε v v m vk
ðr ku Þ ¼ mþ t þ Gk  rg ε
vxi g i vxj sk vxj
"  #
v v m vε ε ε2
ðr εu Þ ¼ mþ t þ C1ε Gk  C2ε rg
vxi g i vxj sε vxj k k
     
Solid Mass dmp dmp dmp dmp
¼ þ þ
dt dt drying dt pyrolysis dt reaction
Momentum dv gðrp  rg Þ
mp ¼ FD ðu  vÞ þ
dt rp
 
Energy dTp dmp
mp cp ¼ hpd2p ðTg  Tp Þ þ εp pd2p sðq4R  T 4p Þ þ h
dt dt drying fg
   
dmp dmp
þ h þ Hrec
dt pyrolysis fg dt reaction

Table 2
Governing equations for CFD model.

Process Rate expression

    !
Drying dmp dmp pd p k c cp ðTg  Tp Þ
¼ pd2p Mw kg ðCs  C∞ Þ ¼ ð2 þ 0:46Re0:5 Þln 1 þ
dt v dt b cp hfg
 
Pyrolysis dmp
¼  k½mp  ð1  fv;0 Þð1  fw;0 Þmp;0 
dt pyrolysis
Homogeneous reaction ∧
Rg;r ¼ minðR i;r ; Ri;r Þ
!
∧ 0
YN 0 Y
N
}
R i;r ¼ ðy}i;r  yi;r Þ kf ;r j¼1
½Cj;r hj;r  kb;r ½Cj;r hj;r
j¼1
kf;r
kf;r ¼ Ar T ng
expðEr =RTg Þ kb;r ¼
Kr
0 ! 1
P
ε Y ε Y
Ri;r ¼ min@yi;r Mi Arg minR 0 R ; yi;r Mi ABrg PN P P A
0 0

k yR;r MR k j y} Mj
j;r
 m
Heterogeneous reaction r p d p R
Rp;j ¼ pd2p Rp;j RP;j ¼ hRint Pg;j  p;j
6 Kd
Rint ¼ A expð  E=RTÞP m j
 
3 1 1
h¼ 
4 tanh 4 4
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
dp ðm þ 1ÞA expðE=RTÞFðXÞrp vg RTg P s;j
m1
4 ¼
6 2Mc De
2Mc Di
Kd ¼
vg dp RTg

Table 3
Kinetic parameters for homogeneous phase reactions.

d a b E A Ref.

CO þ 0:5O2 /CO2 Ri ¼ A expð  E=RTg ÞC CO C O2a b

1 0.5 133 1.78  1014 [27]

H2 þ 0:5O2 /H2 O Ri ¼ A expð  E=RTg ÞC H2 C O2a b

1 1 125 1.08  1013 [28]

CH4 þ 1:5O2 /CO þ 2H2 O Ri ¼ A expð  E=RTg ÞC aCH4 C bO2
0.3 1.3 203 2.8  109 [28]
CO þ H2 O4CO2 þ H2 Rif ¼ A expð  E=RTg ÞC aCO C bH2 O
Rib ¼ A expð  E=RTg ÞC aCO2 C bH2
f 0.5 1 304.6 2.43  109 [29]
b 1 0.5 326.4 6.4  1012 [30]
 y  y
Cx Hy Oz þ x þ  z O2 /xCO þ H2 O Ri ¼ AT dg expð  E=RTg ÞC aCx Hy Oz C bO2
2 2
1 0.5 1 80.2 9.17  106 [26]

Ri in the unit of kmol/(m3∙s); E in the unit of kJ/mol; C in the unit of kmol/m3; A in the unit of (kmol/m3)m/(Kd∙s), where m ¼ 1-a-b.
f is forward reaction; b is backward reaction.
614 X. Gao et al. / Renewable Energy 139 (2019) 611e620

intrinsic reaction rate submodel was detailed in our previous work computational domain of the reactor was discretized into a finite
[25]. number of grid cells, meanwhile the discrete governing equations
were obtained through finite volume method for each grid cell. For
2.2. Computation scheme the gas phase, the governing equations of continuity, momentum,
energy and species were solved sequentially by SIMPLE algorithm.
The gasification model presented above was realized by ANSYS For the solid phase, particle streams were tracked one by one once
Fluent software according to the finite volume technique. For the biomass particles entered the computational domain through
biomass gasification in an entrained flow reactor, the injection, and the corresponding governing equations were solved
by the DPM (discrete phase model) method. Meanwhile the user
defined functions were adopted for modeling the char-gas re-
Table 4 actions using intrinsic reaction rate submodel. The iteration
Main characteristics of rice husk.
calculation was alternated between continuous phase and
Characteristics Rice husk dispersed phase until the convergence criteria were met.
Proximate analysisa (%)
Moisture content 6.86 2.3. Computation domain
Volatile mater 60.92
Fixed carbon 15.22
The gasification processes of rice husk with air or air-CO2
Ash 17.00
Ultimate analysisa (%) mixture in an entrained flow reactor were investigated through
C 37.35 using intrinsic reaction rate based CFD model. The proximate and
H 4.40 ultimate analyses of rice husk are listed in Table 4, and the intrinsic
Ob 34.05
kinetic parameters for char heterogenous reaction are given in
N 0.20
S 0.14
Table 5. Fig. 1 displays the schematic diagram of a lab-scale
Lower heating value (MJ/kg) 14.48 entrained flow gasification system, which is composed of mainly
Bulk density (kg/m3) 519 gas supplying system, sampling system and a downdraft entrained
a
Weight percentage on air dried basis. flow reactor. The internal diameter of the reactor is 10 cm and the
b
Calculated by difference. height is 190 cm. A fuel and gas injector is located at the top, and
the gas flows downward and exits the reactor from the outlet at the
bottom. The geometry of the entrained flow reactor was meshed
Table 5
with a total number of 0.37 million hexahedral elements, which
Intrinsic kinetic parameters.
met the grid independence test [15].
Reaction A (1/s$Pam) E (kJ/mol) m Ref.

Char-CO2 5.714  105 233.00 0.345 [31] 3. Sensitivity analysis

Char-H2O 2.06  103 241.00 1 [32]
Char-O2 0.744 122.62 1.586 [31]
The operating conditions, such as gasification temperature,

Fig. 1. Schematic diagram of entrained flow gasification system.

 X. Gao et al. / Renewable Energy 139 (2019) 611e620 615

Fig. 2. Effect of gasification temperature on rice husk gasification performances.

Fig. 3. Effect of average particle diameter on rice husk gasification performances.

equivalence ratio, particle size, etc., are the main factors affecting
the gasification processes inside a reactor [15,21]. Hence, a sensi-
tivity analysis was conducted to investigate the effect of four CO2 and CH4 showed slight downtrends. Although the water gas
operating parameters on the gasification performances, including shift reaction generated more CO2 amount at higher temperatures,
gas compositions, gas production and cold gas efficiency. the gasification of char consumed more CO2 at the same time,
resulting in unapparent variation in CO2 yield [15].
Additionally, the gas production increased from 1.52 Nm3/kg to
3.1. Effect of gasification temperature 1.68 Nm3/kg with the rise of gasification temperature, this is due to
the promoted devolatilization [21]. The cold gas efficiency
In this section, the average particle diameter was maintained at increased in the range of 38.52e60.61%, which can be attributed to
240 mm, the ER was kept at 0.2, and the CO2/biomass ratio was 0. the increments in combustible gases and produced gas yield at
The variations of gas compositions, gas production and cold gas elevated temperatures.
efficiency with gasification temperature are exhibited in Fig. 2.
With the gasification temperature increasing from 800  C to
1100  C, the concentrations of CO, CO2, CH4, and H2 varied in the 3.2. Effect of average particle diameter
ranges of 20.00e23.49%, 5.51e6.14%, 1.25e1.35%, and 4.34e14.43%,
respectively. As observed, the CO concentration initially increased In this part, the gasification temperature was kept at 900  C, the
and then decreased slightly to a stable value. The gasification re- ER was 0.2, and the CO2/biomass was maintained at 0. The effect of
actions of char with CO2 and H2O were favored at higher temper- average particle diameter on the entrained flow gasification per-
atures, resulting in more CO yield. However, the slightly decreased formances of rice husk is displayed in Fig. 3. As the average particle
and stable CO yield when the temperature excessed 900  C could be diameter increased from 220 mm to 350 mm, the generated CO
related to the enhanced water gas shift reaction [21]. The H2 con- concentration decreased obviously from 24.32% to 14.27%; while
centration presented a distinct uptrend, which can be explained by CO2 and H2 presented reverse trends in the ranges of 5.66e11.42%
the accelerated gasification reactions of C þ H2O and CO þ H2O and 8.17e8.82%, respectively; the amount of CH4 varied slightly
owing to the elevated temperature [32]. Both the concentrations of around 1%. According to the temperature profiles under different
616 X. Gao et al. / Renewable Energy 139 (2019) 611e620

Fig. 4. Effect of average particle diameter on temperature contour of rice husk

gasification.

average particle diameters as shown in Fig. 4, the location of high

temperature zone moved backward along the axis of the gasifier
when the average particle diameter increased. That is to say, the
distance between the reactor inlet and the high temperature zone
increased with the increasing average particle diameter. The heat
transfer resistance increased with the increasing particle diameter
[15,21], leading to time extension for the drying and pyrolysis
processes of biomass. Furthermore, the increasing mass transfer
limitation also weakened the char heterogenous reactions (both
combustion and gasification). Thus, more flammable volatiles were
consumed by oxygen, which led to the expanded high temperature
zone in the reactor (see Fig. 4). As a consequence, the water gas shift
reaction was improved due to the elevated temperature and pro-
moted H2O yield. Fig. 5. Effect of ER on rice husk gasification performances.
In Fig. 3 (b), the rise of average particle diameter from 220 mm to
350 mm had a negative impact on both gas production (1.56e1.48
Nm3/kg) and cold gas efficiency (50.03e34.29%). Because less vol- efficiency declined from 48.65% to 38.76%, and this was owing to
atiles were released during the pyrolysis process and the CO con- the lower yields of combustible gases in the produced gas.
centration dropped markedly, both gas production and cold gas
efficiency declined with larger particle diameter.
3.4. Effect of CO2/biomass mass ratio

3.3. Effect of ER In this section, the gasification temperature was kept at 900  C,
the average particle diameter was maintained at 240 mm, and the
The gasification temperature was held at 900  C, the average ER was 0. The effects of CO2/biomass mass ratio on the gasification
particle diameter was kept at 240 mm, and the CO2/biomass was performances are shown in Fig. 6. As the CO2/biomass raised from
maintained at 0. The variations of rice husk gasification perfor- 0 to 0.8, the amounts of CO, CO2, CH4, and H2 varied in the ranges of
mances with ER are demonstrated in Fig. 5. All the yields of 23.49e29.05%, 6.08e23.1%, 0.92e1.16%, and 5.88e8.26%, respec-
combustible gases displayed the downtrends in the ranges of tively. It is observed that the CO content in the produced gas
23.49e12.80% (CO), 1.16e0.43% (CH4) and 8.26e6.22% (H2) with the increased first and then declined, and the CO2 content increased
increasing ER around 0.2e0.4, whereas the yield of CO2 presented a markedly, while the yields of CH4 and H2 reduced. These were
monotone uptrend ranging from 6.08% to 11.07%. At higher ERs, because low CO2 addition in the gasifying agent accelerated the
more air amounts were introduced into the gasifier, thereby Boudouard reaction and reversed water gas shift reaction, however,
accelerating the oxidation of char and volatiles, which reduced the the high CO2 injection decreased the temperature and most of the
contents of flammable gases (CO, CH4 and H2) while increased the CO2 left the gasifier directly [35].
content of CO2 in the produced gas [33]. In Fig. 6 (b), the CO2/biomass had different effects on gas pro-
As the ER increased from 0.2 to 0.4, the gas production increased duction and cold gas efficiency. As the CO2/biomass increased in the
from 1.56 Nm3/kg to 2.17 Nm3/kg. This could be related to the facts range of 0e0.8, the gas production increased monotonously from
that (a) more gases were generated from tar combustion, and (b) 1.56 Nm3/kg to 2.08 Nm3/kg. Meanwhile the cold gas efficiency
more amount of N2 was introduced by air [34]. The cold gas increased initially from 48.65% to 65.12% and then decreased to
 X. Gao et al. / Renewable Energy 139 (2019) 611e620 617

optimization. In addition, the four operating parameters in sensi-

tivity analysis section, including gasification temperature (x1),
average particle diameter (x2), ER (x3), and mass ratio of CO2/
biomass (x4), were used as decision variables.

4.2. Optimization solution

The multi-objective optimization of rice husk gasification in the

present study can be given by a mathematical term:
   

Maximize G ¼ y1 xj ; y2 xj ; y3 xj
Suject to xLj  xj  xU
j ; i ¼ 1; 2; :::; 4

Principally, there are a set of best solutions (also known as

Pareto-optimal solutions) to a multi-objective optimization prob-
lem. As an advanced version of NSGA (non-dominated sorting ge-
netic algorithm), NSGA-II was employed to select the optimal
solutions for the rice husk gasification. The detailed procedure for
NSGA-II can be consulted in Refs. [36,37]. In the present study, the
NSGA-II based multi-objective function “gamultiobj” was adopted
to solve the optimization problem in MATLAB (R2014b) platform.

4.3. Analysis of objective responses

In order to obtain the relationship between the objective

Fig. 6. Effect of CO2/biomass on rice husk gasification performances.

60.08%, basically resulting from the variation of CO content in the

produced gas.

4. Multi-objective optimization

Multi-objective optimization of rice husk gasification was con-

ducted using the response surface methodology (RSM). The opti-
mization procedure was similar with the multi-objective
optimization procedure for sawdust gasification reported in our
previous work [15]. First CFD simulations were calculated accord-
ing to the design of experiments. And then the analysis of variance
(ANOVA) was employed to obtain the regression models for the
objective responses. Finally, multi-objective optimization algo-
rithm was adopted to select the optimal solutions for the rice husk
gasification.

4.1. Objective responses and decision variables

A multi-objective optimal study according to evolutionary al-

gorithm was carried out to determine the optimal operating con-
dition for the rice husk entrained flow gasification. Three objective
responses including CO concentration (y1), gas production (y2) and
cold gas efficiency (y3) were considered for the multi-objective Fig. 7. Response surface of CO concentration for rice husk entrained flow gasification.
618 X. Gao et al. / Renewable Energy 139 (2019) 611e620

response and the four decision variables, the fully quadratic model
was initially applied for the regression of objective function, and
then the insignificant terms were removed from the regression y2 ¼ 0:88 þ 3:72E  04x1  8:68E  04x2 þ 2:92x3 þ 0:63x4
model based on the analysis of the variance (ANOVA) to gain the (2)
revised regression model. Consequently, the relationship between
CO concentration (y1) and decision variables (xi) could be given by: The effects of the four variables on the gas production are dis-
played in Fig. 8. The gas production of rice husk gasification
y1 ¼ 75:93  1:82E  02x1  0:20x2  85:62x3 þ 12:32x4 increased with gasification temperature, ER, and CO2/biomass mass
 14:62x24 þ 8:52E  05x1 x2 þ 0:16x2 x3 ratio, while decreased with average particle diameter.
Similarly, the regression model for cold gas efficiency (y3) ob-
(1)
tained from ANOVA is given by:
Fig. 7 demonstrates the effects of decision variables on the CO
concentration in produced gas. The CO concentration increased
with gasification temperature, and the increasing amplitude y3 ¼ 52:36 þ 4:95E  02x1  0:15x2  61:54x3 þ 30:41x4
increased with the increment in average particle diameter. The CO
content decreased with average particle diameter, and the  29:96x24 (3)
decreasing amplitude reduced with gasification temperature. In Fig. 9 exhibits the effects of decision variables on the cold gas
addition, the increase in ER always had a negative effect on CO efficiency. Generally, the cold gas efficiency increased with gasifi-
production, while the increase in CO2/biomass increased CO pro- cation temperature, whereas decreased with average particle
duction initially and then decreased the CO amount. diameter and ER. However, the cold gas efficiency rose initially and
According to the ANOVA, the regression response model for gas then dropped with the increase in CO2/biomass mass ratio. This is
production, y2, can be expressed as follows: consistent with the results presented in the parametric study.

Fig. 8. Response surface of gas production for rice husk entrained flow gasification. Fig. 9. Response surface of cold gas efficiency for rice husk entrained flow gasification.
 X. Gao et al. / Renewable Energy 139 (2019) 611e620 619

4.4. Optimization results decreased it.

The Pareto optimal solutions to multi-objective optimization
Fig. 10 shows the Pareto frontier for the three-objective opti- were obtained from NSGA-II. According to the standard TOPSIS, the
mization of rice husk entrained flow gasification. The best solution optimal gasification performances of rice husk gasification were CO
to maximize the three objectives simultaneously does not exist. yield of 25.15%, gas production of 1.96 Nm3/kg and cold gas effi-
Each point on the Pareto frontier represents a potential solution for ciency of 65.34%.
maximizing CO concentration, gas production and cold gas effi-
ciency. The designers can choose the Pareto optimal solution ac- Acknowledgements
cording to the specified occasion [38]. The standard TOPSIS
(technique for order preference by similarity to ideal situation) is a This study was supported by National Natural Science Founda-
classic decision-making method [39]. With the help of a virtual tion of China (Grant No. 51606048, 11672284, 11602266), and Na-
ideal point having the highest CO content, maximal gas production tional Key R&D Program of China (Grant No. 2017YFB0603701). The
and highest cold gas efficiency, the Pareto point that being closest financial supports from the Collaborative Innovation Center of
to the ideal point is selected as the desired solution. In the present Cleaner Coal Power Plant with Poly-generation are also
study, the point A in Fig. 10 was the Pareto-optimal solution ob- acknowledged.
tained from standard TOPSIS, representing 25.15% (CO content),
1.96 Nm3/kg (gas production) and 65.34% (cold gas efficiency).
References
Obviously, it can be revealed that the Pareto-optimal solution from
TOPSIS has higher performance indexes than general situations [1] S. Liu, Q. Xie, B. Zhang, Y. Cheng, Y. Liu, P. Chen, R. Ruan, Fast microwave-
studied in 3. Sensitivity analysis section. Hence multi-objective assisted catalytic co-pyrolysis of corn stover and scum for bio-oil production
optimization can be a promising way to promote performances of with CaO and HZSM-5 as the catalyst, Bioresour. Technol. 204 (2016)
164e170.
rice husk gasification. [2] F. Guo, Y. Dong, L. Dong, C. Guo, Effect of design and operating parameters on
the gasification process of biomass in a downdraft fixed bed: an experimental
5. Conclusions study, Int. J. Hydrog. Energy 39 (2014) 5625e5633.
[3] Y. Zhang, W. Zhao, B. Li, G. Xie, Microwave-Assisted pyrolysis of biomass for
bio-oil production: a review of the operation parameters, J. Energy Resour.
A sensitivity analysis and a three-objective optimization of rice Technol. 140 (2018), 040802.
husk entrained flow gasification were carried out through using the [4] C. Wu, R. Liu, Sustainable hydrogen production from steam reforming of bio-
oil model compound based on carbon deposition/elimination, Int. J. Hydrogen
CFD model based on the intrinsic reaction rate submodel. The Energy 36 (2011) 2860e2868.
following conclusions are obtained. [5] H. Jin, B. Chen, X. Zhao, C. Cao, Molecular dynamic simulation of hydrogen
Gasification temperature, ER and CO2/biomass had positive ef- production by catalytic gasification of Key intermediates of biomass in su-
percritical water, J. Energy Resour. Technol. 140 (2017), 041801.
fects on gas production, whereas average particle diameter had a
[6] Y. Zhang, X. Gao, B. Li, H. Li, W. Zhao, Assessing the potential environmental
negative effect on gas production. Gasification temperature played impact of woody biomass using quantitative universal exergy, J. Clean. Prod.
positive role in the cold gas efficiency, while both average particle 176 (2018) 693e703.
[7] R. Yin, R. Liu, J. Wu, X. Wu, C. Sun, C. Wu, Influence of particle size on per-
diameter and ER played negative roles in the cold gas efficiency,
formance of a pilot-scale fixed-bed gasification system, Bioresour. Technol.
and CO2/biomass increased cold gas efficiency initially and then 119 (2012) 15e21.
[8] Y. Zhao, D. Feng, Z. Zhang, S. Sun, H. Che, J. Luan, Experimental study on
autothermal cyclone air gasification of biomass, J. Energy Resour. Technol. 140
(2017), 042001.
[9] V. Silva, A. Rouboa, Combining a 2-D multiphase CFD model with a Response
Surface Methodology to optimize the gasification of Portuguese biomasses,
Energy Convers. Manag. 99 (2015) 28e40.
[10] Y. Zhao, D. Feng, S. Sun, J. Luan, H. Che, Therm. Sci. 22 (2018) 439e447.
[11] J. Zhou, Q. Chen, H. Zhao, X. Cao, Q. Mei, Z. Luo, K. Cen, Biomass-oxygen
gasification in a high-temperature entrained-flow gasifier, Biotechnol. Adv. 27
(2009) 606e611.
[12] M. Kaewpanha, G. Guan, X. Hao, Z. Wang, Y. Kasai, K. Kusakabe, A. Abudula,
Steam co-gasification of brown seaweed and land-based biomass, Fuel Pro-
cess. Technol. 120 (2014) 106e112.
[13] M. Kuba, H. Hofbauer, Experimental parametric study on product gas and tar
composition in dual fluid bed gasification of woody biomass, Biomass Bio-
energy 115 (2018) 35e44.
[14] T.M. Ismail, M.A. El-Salam, Parametric studies on biomass gasification process
on updraft gasifier high temperature air gasification, Appl. Therm. Eng. 112
(2017) 1460e1473.
[15] Y. Zhang, X. Gao, F. Bao, B. Li, Y. Zhao, C. Ke, B. Jiang, CFD modeling of sawdust
gasification in a lab-scale entrained flow reactor based on char intrinsic ki-
netics. Part 2: parameter study and multi-objective optimization, Chem. Eng.
Process. Process. Intensif. 125 (2018) 290e297.
[16] H. Jin, Z. Wu, L. Guo, X. Su, Numerical investigation on the two phase flow
behaviors in supercritical water fluidized bed with swirling flow distributor,
Heat Transf. Eng. (2017) 1e12.
[17] H. Liu, R.J. Cattolica, R. Seiser, CFD studies on biomass gasification in a pilot-
scale dual fluidized-bed system, Int. J. Hydrogen Energy 41 (2016)
11974e11989.
[18] N. Couto, V. Silva, E. Monteiro, P. Brito, A. Rouboa, Using an Eulerian-granular
2-D multiphase CFD model to simulate oxygen air enriched gasification of
agroindustrial residues, Renew. Energy 77 (2015) 174e181.
[19] T.M. Ismail, M. Abd El-Salam, E. Monteiro, A. Rouboa, Eulerian e Eulerian CFD
model on fluidized bed gasifier using coffee husks as fuel, Appl. Therm. Eng.
106 (2016) 1391e1402.
[20] Y. Cheng, Z. Thow, C.-H. Wang, Biomass gasification with CO2 in a fluidized
bed, Powder Technol. 296 (2016) 87e101.
Fig. 10. Pareto frontier of three objectives optimization for rice husk entrained flow [21] K. Qin, Entrained Flow Gasification of Biomass, Technical University of
gasification. Denmark, Kongens Lyngby, 2012.
620 X. Gao et al. / Renewable Energy 139 (2019) 611e620

[22] X. Zhang, H. Li, L. Liu, C. Bai, S. Wang, Q. Song, J. Zeng, X. Liu, G. Zhang, K.S. Rothenberger, M.V. Ciocco, B.D. Morreale, S. Chattopadhyay, S. Shi, High-
Optimization analysis of a novel combined heating and power system based temperature kinetics of the homogeneous reverse water-gas shift reaction,
on biomass partial gasification and ground source heat pump, Energy Convers. AIChE J. 50 (2004) 1028e1041.
Manag. 163 (2018) 355e370. [31] X. Gao, Study on the Characteristics of Biomass Entrained Flow Gasification
[23] M. Campoy, A. Go mez-Barea, F.B. Vidal, P. Ollero, Airesteam gasification of Based on Char Intrinsic Kinetics, Harbin Institute of Technology, Harbin, 2017.
biomass in a fluidised bed: process optimisation by enriched air, Fuel Process. [32] X. Ku, T. Li, T. Løvås, Eulerian-Lagrangian simulation of biomass gasification
Technol. 90 (2009) 677e685. behavior in a high-temperature entrained-flow reactor, Energy Fuels 28
[24] R.T.L. Ng, D.H.S. Tay, W.A. Wan Ab Karim Ghani, D.K.S. Ng, Modelling and (2014) 5184e5196.
optimisation of biomass fluidised bed gasifier, Appl. Therm. Eng. 61 (2013) [33] J. Xie, W. Zhong, B. Jin, Y. Shao, H. Liu, Simulation on gasification of forestry
98e105. residues in fluidized beds by Eulerian-Lagrangian approach, Bioresour.
[25] X. Gao, Y. Zhang, B. Li, X. Yu, Model development for biomass gasification in an Technol. 121 (2012) 36e46.
entrained flow gasifier using intrinsic reaction rate submodel, Energy Convers. [34] J. Xie, Combustion/pyrolysis/gasification Processes of Combustible Solid
Manag. 108 (2016) 120e131. Waste/coal, Southeast University, Nanjing, 2015.
[26] H. Zhou, A.D. Jensen, P. Glarborg, P.A. Jensen, A. Kavaliauskas, Numerical [35] N. Couto, V. Silva, A. Rouboa, Municipal solid waste gasification in semi-
modeling of straw combustion in a fixed bed, Fuel 84 (2005) 389e403. industrial conditions using air-CO2 mixtures, Energy 104 (2016) 42e52.
[27] J. Xie, W. Zhong, B. Jin, Y. Shao, Y. Huang, EulerianeLagrangian method for [36] E. Zitzler, K. Deb, L. Thiele, Comparison of multiobjective evolutionary algo-
three-dimensional simulation of fluidized bed coal gasification, Adv. Powder rithms: empirical results, Evol. Comput. 8 (2000) 173e195.
Technol. 24 (2013) 382e392. [37] K. Deb, A. Pratap, S. Agarwal, T. Meyarivan, A fast and elitist multiobjective
[28] A. Go mez-Barea, B. Leckner, Modeling of biomass gasification in fluidized bed, genetic algorithm: NSGA-II, IEEE Trans. Evol. Comput. 6 (2002) 182e197.
Prog. Energy Combust. Sci. 36 (2010) 444e509. [38] S. Chamoli, P. Yu, S. Yu, Multi-objective shape optimization of a heat
[29] F. Bustamante, R.M. Enick, R.P. Killmeyer, B.H. Howard, K.S. Rothenberger, exchanger tube fitted with compound inserts, Appl. Therm. Eng. 117 (2017)
A.V. Cugini, B.D. Morreale, M.V. Ciocco, Uncatalyzed and wall-catalyzed for- 708e724.
ward water-gas shift reaction kinetics, AIChE J. 51 (2005) 1440e1454. [39] K.P. Yoon, W.K. Kim, The behavioral TOPSIS, Expert Syst. Appl. 89 (2017)
[30] F. Bustamante, R.M. Enick, A.V. Cugini, R.P. Killmeyer, B.H. Howard, 266e272.