Theoretical Simulation of Entrained Flow IGCC Gasifiers: Effect of

However, the fluctuations of mixture fractions concerning all the reactions have little effect on the distribution of particle concentration. For gasi...
0 downloads 0 Views 209KB Size
1280

Energy & Fuels 2002, 16, 1280-1286

Theoretical Simulation of Entrained Flow IGCC Gasifiers: Effect of Mixture Fraction Fluctuation on Reaction Owing to Turbulent Flow Hao Liu,† Caixia Chen,† and Toshinori Kojima* Department of Applied Chemistry, Seikei University, 3-3-1 Kichijojikita-machi, Musashino-shi, Tokyo 180-8633, Japan Received March 11, 2002

A three-dimensional numerical model was developed for a two stage entrained flow coal gasifier. In the model, the coal devolatilization, volatile and char combustion, char-H2O and char-CO2 gasification, and gas-phase shift reaction were taken into account in accordance with the actual process. Four mixture fractions were used to simulate the coal gasification. The influence of turbulence on gas composition was taken into account by a pdf model with a clipped Gaussian distribution function. The model uses conventional numerical methods and submodels to calculate gas and particle temperatures, gas and particle velocities, mean turbulent kinetic energy and turbulent energy of dissipation, gas species composition, particle trajectory, extent of reaction, and radiant heat flux. The effect of mixture fraction fluctuations on the overall gasification characteristics due to turbulent flow was investigated for gasification in an IGCC gasifier. It was revealed that the fluctuations of mixture fractions of gases produced from devolatilization and char-O2 reactions have important influences on not only distributions of temperature and gas composition, but also cold gas efficiency and heating value of product gas. However, the fluctuations of mixture fractions of gases produced from char-CO2 and char-H2O reactions have only a limited effect. Including the fluctuations of mixture fractions smoothens the distribution of temperature. However, the fluctuations of mixture fractions concerning all the reactions have little effect on the distribution of particle concentration. For gasification in an entrained flow IGCC gasifier, with little influence on the validity of the simulated results, the neglecting of the fluctuations of mixture fractions of gases produced from char-CO2 and char-H2O reactions can significantly shorten the calculation time. However, the fluctuations of mixture fractions of gases produced from devolatilization and char-O2 reactions cannot be neglected. Comparison between prediction and measurement suggests that including fluctuations of mixture fractions gives a better prediction.

Introduction As one of the key components of IGCC, an entrained flow coal gasifier has been highlighted in recent years. In Japan, a 50 ton/day oxygen blown HYCOL (Hydrogen from Coal) gasifier1 and a 200 ton/day air blown IGCC gasifier have been developed.2 A 1500-2000 ton/day commercial scale plant is also being developed for IGCC process. There are several mathematical models developed for entrained flow gasifiers in the literature.3-7 Those models have common features in solving the mass, * Corresponding author. † Research fellow of NEDO (New Energy and Industrial Development Organization), Japan. Phone: +81-422-37-3758. Fax: +81-42237-3750. E-mail: [email protected]. (1) Miyadera, H.; Koyama, S.; Morimoto, T.; Eida, E.; Ueda, A.; Ueda, F.; Yoshida, N. Development of HYCOL coal gasification technology. J. Jpn Inst. Energy 1995, 74, 691-698. (2) Araki, S.; Harai, Y. J. Jpn. Inst. Energy. 1996, 75, 839-850. (3) Wen, C. Y.; Chaung, T. Z. Ind. Eng. Chem. Process Des. Dev. 1979, 18, 684-694. (4) Govind, R.; Shah, J. AIChE J. 1984, 30, 79-91. (5) Ni, Q.; Williams, A. Fuel 1995, 74, 102-110. (6) Smoot, L. D.; Smith, P. J. Coal Combustion and Gasification; The Plenum Chemical Engineering Series; Plenum Press: New York, 1985. (7) Hill, S. C.; Smoot, L. D. Energy Fuels 1993, 7, 874-883.

momentum and energy conservation equations with similar submodels by nearly the same mathematical methods.8 Some models have been applied to benchscale entrained flow gasifiers,9 but the simulation capability for gasification was quite limited because of their oversimplification on the chemistry. To simulate gasification reaction and reactant mixing process in practical entrained flow gasification facilities, a comprehensive coal gasification model with a multisolid progress variables (MSPV) method has been developed by us. The gas flow fields, gas temperature distributions, gas composition, extent of burnout, particle trajectories (particle concentration), etc., within the gasifiers were predicted by the model in our previous work.10-13 The predicted results are in good agreement with experimental data. (8) Smoot, L. D., Ed. Fundamentals of Coal Combustion for Clean and Efficient Use; Coal Science and Technology 20; Elsevier: New York, 1993; pp 567-630. (9) Smoot, L. D.; Brown, B. W. Fuel 1987, 66, 1249-1256. (10) Chen, C.; Horio, M.; Kojima, T. Chem. Eng. Sci. 2000, 55, 38613874. (11) Chen, C.; Miyoshi, T.; Kamiya, H.; Horio, M.; Kojima, T. Can. J. Chem. Eng. 1999, 77, 745-750. (12) Chen, C.; Horio, M.; Kojima, T. Chem. Eng. Sci. 2000, 55, 38753883. (13) Chen, C.; Horio, M.; Kojima, T. Fuel 2001, 80, 1513-1523.

10.1021/ef0200626 CCC: $22.00 © 2002 American Chemical Society Published on Web 07/16/2002

Entrained Flow IGCC Gasifiers

Energy & Fuels, Vol. 16, No. 5, 2002 1281

Although a lot of research work was conducted concerning the simulation of coal gasifiers, many topics remain to be clarified. In an entrained flow gasifier, the flow field is highly turbulent and leads to the fluctuations of mixture fractions of gases. Accordingly, these fluctuations of mixture fractions due to turbulent flow may affect the overall gasification characteristics of a gasifier, which can only be clarified through simulation. The main objective of the present paper was to examine the effect of fluctuations of mixture fractions on the overall gasification characteristics, i.e., temperature and gas composition distributions, cold gas efficiency, so as to clarify the interaction between turbulent flow and reaction as well as to what extent the fluctuations of mixture fractions can be neglected. The research method is comparing the calculated results when the fluctuations of mixture fractions are included with those when the fluctuations of mixture fractions are not included. This study is essentially different from the work of Smoot and Smith6 in that we studied entrained flow coal gasification, but they studied the effect of fluctuations on the temperature and CO2 concentration profiles in a lifted coal combustion flame. Model Description and Simulation Method General Description of the Model. The following physical and chemical processes are included in the present model: (1) turbulent flow of gas and mixing of gaseous reactants, (2) entrainment of coal particles and their turbulent dispersion, (3) coal devolatilization and volatile combustion, (4) heterogeneous char reactions including combustion and gasification, (5) convective and radiative heat transfer among coal, char and ash particles. The numerical methods and the submodels recommended for the entrained flow coal combustion and gasification process6 were used in the present model. The gas phase is assumed to be a steady state, reacting, continuum field that can be described by general conservation equation as follows:

instantaneous gas velocity is decomposed into time mean and fluctuating component to account for the effects of turbulence on the particle dispersion. The values of the fluctuating velocities are assumed to follow a Gaussian distribution and determined in a stochastic manner within a turbulent eddy. Coal devolatilization is modeled by a simple, two-step mechanism. The particle surface reactions are characterized by the 0.5 order, multiple, parallel reaction rate formulation of Smith.18 The gas-phase shift reaction is assumed to reach equilibrium rapidly. The details of the particle reaction model have been described elsewhere.10 Turbulent Reaction Model. An extended version of the statistical coal gas mixture fraction model with the multiple solid progress variables (MSPV) method19 is used. Four components of coal off-gas, i.e., four coal gas mixture fractions are used to track the reaction products. These mixture fractions at a point are defined as the mass ratio of coal offgas to the total gas product and written as19

) (

) (

(14) Crowe, C. T.; Sharma, M. P.; Stock, D. E. J. Fluids Eng. 1977, 99, 325-332. (15) Launder, B. E.; Spalding, B. Mathematical Models of Turbulence; Academic Press: New York, 1972; pp 90-110. (16) Launder, B. E.; Sharma, B. I. Lett. Heat Mass Transfer 1974, 1, 131-138.

∑m

j

j)1

where mi (i ) 1-4) represents the mass of gas originating from devolatilization, char-O2, char-H2O and char-CO2 reactions respectively, mair is the mass of inlet air, and fi (i ) 1-4) is the conserved mixture fraction calculated by transport equations. From the general theory of MSPV method, the other four mixture fractions are defined as

mi

Fi )

(3)

i

mair +

∑m

j

j)1

By these definitions, each mixture fraction Fi (i ) 1-4) varies independently between zero and unity and, to a first approximation, is assumed to be statistically independent of the other mixture fractions. Thus Fi is related to fi as

fi

Fi )

)

Φ refers to any quantity of mass, velocity components (u, v, and w), turbulent kinetic energy (k) and turbulent kinetic energy dissipation rate (), gas enthalpy (h), mixture fractions (fi) and their variances (gi), SΦ is the source term, and SΦp is an additional source term representing the interaction between gas and particle phases.6 With a Lagrangian method the pulverized coal particles are tracked. The particle-source-in-cell model14 is used to deal with the interaction between gas and particle phases through various particle source terms. The net difference in the particle properties between leaving and entering any given cell provides the particle source term SΦp for the gas-flow equations. Turbulence in gas phase is modeled by Favre-averaging of the gradient diffusion processes with the two-equation k- model15 for closure. The standard constants of the k- model16 were used in the simulation. The particle stochastic trajectory model17 based on instantaneous gas velocities is used to simulate the particle motion and turbulent dispersion. The

(2)

4

mair +

∂ ∂ ∂ (FuΦ) + (FvΦ) + (FwΦ) ) ∂x ∂y ∂z ∂ µ ∂Φ ∂ µ ∂Φ ∂ µ ∂Φ + + + SΦ + SΦp (1) ∂x σΦ ∂x ∂y σΦ ∂y ∂z σΦ ∂z

(

mi

fi )

(4)

4

1-

∑f

j

j)i+1

The average value of any dependent fluctuating gas property β (gas species, temperature, density, or viscosity) is a function of F1, F2, F3, and F4, and can be calculated by convoluting the instantaneous value over the probability density functions (pdf) of the independent mixture fractions.

∫ P˜ (F )∫ P˜ (F )∫ P˜ (F ) × ∫ P˜ (F )β(F ,F ,F ,F ) dF dF dF

β˜ (F1,F2,F3,F4) )

1-

1-

4

0+ 1-

0+

1

1-

3

0+ 1

2

3

2

0+

4

1

2

3 dF4

(5)

The local variances of the mixture fractions are calculated from gi transport equations similar to those of Launder and Spalding.15 The pdfs are assumed to have the form of a clippedGaussian distribution, adjusted to account for turbulent intermittency.6 Numerical Solution Procedure. By integrating over the computational cells, the governing partial differential equa(17) Shuen, J. S.; Chen, L. D.; Feath, G. M. AIChE J. 1983, 29, 167170. (18) Smith, I. W. The Combustion Rate of Coal Chars: A Review. In Nineteenth Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, 1982; pp 1045-1065. (19) Brewster, B. S.; Baxter, L. L.; Smoot, L. D. Energy Fuels 1988, 2, 362-370.

1282

Energy & Fuels, Vol. 16, No. 5, 2002

Liu et al.

Figure 1. Effect of fluctuations of mixture fractions on the distribution of temperature. Table 1. Selected Model Parameters and Operating Conditions parameters

values

Proximate Analysis (wt %) moisture fixed carbon volatile matter ash heating value (kJ kg-1)

values

Particle Size Distribution (µm) (wt % for Each Size Classification) 5.3 35.8 46.7 12.1 27 399.88

Ultimate Analysis (wt %) carbon hydrogen nitrogen oxygen sulfur

parameters

77.6 6.5 1.13 13.9 0.22

10% 10% 20% 20% 20% 20% mass mean diameter (µm)

150 100 40 20 10 4 39.8

Gas Flow Rate (kg s-1) lower combustor coal burner higher char/coal burner reductor coal burner

4.708 4.708 1.832

combustor diffuser reductor

1897 1073 873

Particle Loading (kg s-1) lower combustor coal burner higher char/coal burner reductor coal burner

0.472 1.112 1.832

pressure (MPa)

2.7

oxygen vol % in inlet gas

21%

Wall Temperature (K)

tions for all the quantities were reduced to their finite difference analogues. The conservation equations are Favreaveraged and solved by the SIMPLER method.20 A line-byline iteration technique was used to solve the finite difference equations. Solution of the particle phase conservation equations is coupled with the gas phase. An overall convergence of the two phases is achieved using an iteration procedure. Simulation is based on the Nakoso 200 ton/day two stage air blown entrained flow gasifier.2,10 The operating conditions used in the simulations are listed in Table 1, and the coal property is based on that of Taiheiyo bituminite. The gas flow rate and particle loading for each set of burners are the total flow rate into four nozzles. The wall temperatures are determined from the information of measurements. The chemical reaction rate constants from the literature21 for similar coal were used in the present calculation. (20) Patankar, S. V. Numerical Heat Transfer and Fluid Flow; Hemisphere Publishing Corp.: Bristol, PA, 1980; pp 113-135, (21) Mann, A. P.; Kent, J. H. Combust. Flame 1994, 99, 147-156.

Coal particles from each nozzle were tracked. To obtain a steady convergence in the calculations, the model tracked 11500 particle trajectories. A 21 × 21 × 62 grid mesh was adopted. A converged solution is defined when the global energy balances to within 0.3% deviation of the total combustion energy, and the normal residual (mean residual/mean inlet velocity) for each velocity component is less than 0.3%.

Results and Discussions Effect on Temperature Distribution. The effect of fluctuations of mixture fractions on the distributions of temperature, gas composition, etc. due to a turbulent flow, was investigated for gasification in an IGCC gasifier. Figure 1 shows the effect of fluctuations of mixture fractions on temperature distribution. The height of the throat above the bottom of the gasifier is about 1.5 m. In case a, no fluctuation of mixture fraction was included. In case b, the fluctuations of mixture fractions F1 and F2 were included in the calculation, i.e.,

Entrained Flow IGCC Gasifiers

Energy & Fuels, Vol. 16, No. 5, 2002 1283

Figure 2. Effect of fluctuations of mixture fractions (F1-F4) on the distribution of overall mixture fraction f.

Figure 3. Effect of fluctuations of mixture fractions on the distribution of H2 concentration.

the fluctuations of mixture fractions concerning devolatization and char-O2 combustion reactions were included. On the other hand, in case c, the fluctuations of all mixture fractions, i.e., F1 through F4, were included in the calculation. In other words, the fluctuations concerning gasification reactions were also considered in case c. From Figure 1 it can be seen that the results of cases b and c are nearly identical, with only the result of case a very different from the results of cases b and c. These results suggest that the effect of fluctuations of mixture fractions of gases produced from devolatilization and char-O2 reaction on temperature distribution is important. On the other hand, the fluctuations

of mixture fractions of gases produced from char-CO2 and char-H2O reactions do not have significant effect on the temperature distribution. Moreover, the peak temperature of case a is higher than the peak temperatures of cases b and c, which suggests that including the fluctuations of mixture fractions smoothens the distribution of temperature. This result was caused by the fact that the turbulent fluctuations spread the reaction zone over a wider area and thus moderate hightemperature. A similar conclusion was also obtained by Smoot and Smith.6 Effect on Gas Composition Distribution. The effect of fluctuations of mixture fractions on the distri-

1284

Energy & Fuels, Vol. 16, No. 5, 2002

Liu et al.

Figure 4. Effect of fluctuations of mixture fractions on the distribution of H2O concentration.

Figure 5. Effect of fluctuations of mixture fractions on the overall gasification characteristics of the gasifier.

bution of gas composition was also investigated. Figure 2 shows the distribution of the overall mixture fraction, f, for various cases. Figure 2 indicates that the results of cases b and c are nearly identical, with only the result of case a different from the others. These results suggest that whether including the fluctuations concerning charCO2 and char-H2O reactions or not makes little difference. Figure 3 shows the effect of fluctuations of mixture fractions of gases produced from various reactions on

H2 concentration distribution. The fluctuations of mixture fractions concerning devolatilization and char-O2 reactions significantly influence the distribution of H2 concentration, but the effect of the fluctuations of mixture fractions concerning gasification reactions (charCO2 and char-H2O reactions) is limited. This is very likely due to the high reaction rates of devolatilization and char-O2 reactions, and the low reaction rates of gasification reactions (char-CO2 and char-H2O reactions). Figure 4 shows the effect of the fluctuations of

Entrained Flow IGCC Gasifiers

Energy & Fuels, Vol. 16, No. 5, 2002 1285

Figure 6. Effect of fluctuations of mixture fractions on the distribution of particle concentration (log pc, kg/m3)

Figure 7. Predicted and measured temperature profiles along the height of a 200 ton/day IGCC gasifier. Line: prediction. Circle symbol: experiment.

mixture fractions on H2O concentration distributions. Similarly, the results of cases b and c are nearly identical, with only case a very different.

Effect on Cold Gas Efficiency. Figure 5 shows the cold gas efficiency and heating value of product gas for cases a, b, and c, respectively. It can be seen that the

1286

Energy & Fuels, Vol. 16, No. 5, 2002

Liu et al.

Table 2. Predicted and Measured Heating Values of Product Gas case

heating value of product gas (kcal/m3)

(a) no fluctuation (b) flluctuations with F1 and F2 (c) fluctuations with F1, F2, F3, and F4

1014.1 1046.7 1056.6

experimental measurement

1055.8

cold gas efficiency and heating value of product gas for case b are nearly identical to those for case c. However, the cold gas efficiency and heating value of product gas for case a are very different from cases b or case c. These results identify that the fluctuations of mixture fractions of gases produced from devolatilization and char-O2 reactions significantly influence not only the distributions of temperature and gas composition, but also the cold gas efficiency and heating value of product gas, during gasification in an IGCC gasifier. On the other hand, the fluctuations of mixture fractions of gases produced from char-CO2 and char-H2O reactions have little effect on cold gas efficiency and heating value of product gas. Effect on Particle Concentration. The effect of fluctuations of mixture fractions on the distribution of particle concentration was also investigated (Figure 6). Nearly the same distributions are obtained for all the three cases. From this result it is known that the fluctuations of mixture fractions concerning all the reactions hardly influence the distribution of particle concentration. This independence of particle concentration is probably due to the inertia of particles. Comparison between Prediction and Experiment. Figure 7 compares the predicted and measured temperature profiles along the height of a 200 ton/day IGCC gasifier. The prediction including fluctuations of mixture fractions (F1 and F2) much better agrees with the experimentally measured temperature profile than does the prediction neglecting fluctuation. Table 2 lists the predicted and measured heating values of product gas. The predicted heating values of product gas of cases b and c are closer to the measured one than the prediction of case a, i.e., the inclusion of fluctuations gives a better prediction. Conclusions A comprehensive computer simulation model was developed for an entrained flow coal gasifier. The effect of mixture fraction fluctuations on the overall gasification characteristics due to turbulent flow was investigated for gasification in an IGCC gasifier. It was revealed that the fluctuations of mixture fractions of

gases produced from devolatilization and char-O2 reactions have important influences on not only distributions of temperature and gas composition, but also cold gas efficiency and heating value of product gas. However, the fluctuations of mixture fractions of gases produced from char-CO2 and char-H2O reactions have only a limited effect. Including the fluctuations of mixture fractions smoothens the distribution of temperature. However, the fluctuations of mixture fractions concerning all the reactions have little effect on the distribution of particle concentration. These results are believed to be due to the high reaction rates of devolatilization and char-O2 reactions, and the low reaction rates of gasification reactions (char-CO2 and char-H2O reactions). For gasification in an entrained flow IGCC gasifier, with little influence on the validity of the simulated results, the neglecting of the fluctuations of mixture fractions of gases produced from char-CO2 and char-H2O reactions can significantly shorten the calculation time. On the other hand, the fluctuations of mixture fractions of gases produced from devolatilization and char-O2 reactions cannot be neglected. Comparison between prediction and measurement suggests that including fluctuations of mixture fractions gives a better prediction. Nomenclature f: overall mixture fraction, the mass ratio of total coal off-gas to the total gas product fi, Fi: mixture fractions defined in eqs 2 and 3 gi: variance of fi h: thermal enthalpy, J kg-1 k: turbulent kinetic energy, m2 s-2 m1...m4: mass of gas originating from gas solid reactions, kg mair: mass of air, kg mp: mass of particle, kg SΦ: generalized source term SΦp: source term arising from particles u, v, w: gas velocity components, m s-1 x, y, z: coordinate of three directions, m Greek letters β˜ : average value of any gas property : dissipation rate of turbulent kinetic energy, m2 s-3 σφ: turbulent model constants Φ: generalized variable F: gas density, kg m-3 µ: turbulent viscosity, kg m-1 s-1

Acknowledgment. The authors would like to thank NEDO/CCUJ for financial support of this work under BRAIN-C program. EF0200626