Article pubs.acs.org/EF
Parametric Analysis Using a Reactor Network Model for Petroleum Coke Gasification M. Hossein Sahraei,† Robert Yandon,‡ Marc A. Duchesne,‡ Robin W. Hughes,‡ and Luis A. Ricardez-Sandoval*,† †
Department of Chemical Engineering, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada CanmetENERGY, Natural Resources Canada, Ottawa, Ontario K1A 1M1, Canada
‡
ABSTRACT: Usage of reduced order models (ROMs) and reactor networks are becoming widely accepted tools for the modeling of complex reactors, such as entrained-flow gasifiers. The approximations made in a ROM reduce the required computational costs compared to computational fluid dynamic (CFD) models; however; the capabilities of the model in predicting the outputs for a range of operating conditions in the gasification unit face challenges. The following contribution presents a comparison between a ROM and the corresponding CFD model of a short-residence-time gasifier under different operating conditions and kinetic parameters. Although the framework of the proposed ROM was fixed and developed on the basis of CFD simulations generated at a base-case condition, the results showed reasonable agreement between the two models in predicting syngas composition, carbon conversion, and temperature profile in the gasification system. Sensitivity analysis of the inputs of the ROM (including test condition and reactor network parameters) has also been performed. This analysis has shown that the recirculation ratio and oxygen flow rate have a greater effect on the outputs compared to model geometry and kinetic parameters.
1. INTRODUCTION Coal-fired power plants have been the largest source of power generation for many years. Market concerns regarding greenhouse gas emissions have slowed the development of these power generation systems. It is predicted that clean natural gas will surpass coal as the fuel to fulfill the electricity demands of the world by 2040; however, coal would still be used to produce one-third of the power demands of the world.1 Therefore, sustainable coal-fired power plants that reduce the environmental impacts, such as carbon dioxide (CO2) emissions, are needed. Usage of CO2 capture technologies in power plants decreases their net efficiency and significantly affects the process economics. Among the choices for generating electricity from coal, integrated gasification combined cycle (IGCC) has been accepted as the most efficient power plant when a CO2 capture technology is considered in the layout of the plant.2 To become competitive, IGCC still requires improvement in terms of process economics, particularly in the capital expenses associated with the gasification unit. Gasification involves the transformation of carbonaceous fuels (through thermal treatment) into a viable gaseous form typically referred to as syngas. The most important reactions taking place in this process are depicted in Figure 1. During gasification, the fuel is often mixed with steam and oxygen. Upon this mixing, the solid fuel is dried and decomposed to volatiles, ash, and char. This process is referred to as coal pyrolysis. The amount of pyrolysis products are determined on the basis of the properties of the fuel and the operating conditions of the system. Among the volatile products, combustible gases, such as hydrogen, methane, ethane and carbon monoxide, are produced. These species react with oxygen to produce heat. The heat emanated from these © 2015 American Chemical Society
Figure 1. Reaction systems of the gasification process.
reactions provides the energy required by the heterogeneous char gasification reactions to produce the syngas. During the gasification reactions, the sulfur content within the fuel appears in the gas phase as hydrogen sulfide (H2S).3 Note that H2S and CO2 in the product gases can be separated in the downstream gas cleaning unit of an IGCC power plant.4,5 During the design of a gasification unit, many parameters, such as the geometry of the reactor, number of injectors, feed flow rates, and operating conditions, must be specified such that they accomplish an optimal performance for a given capital Received: July 28, 2015 Revised: October 7, 2015 Published: October 8, 2015 7681
DOI: 10.1021/acs.energyfuels.5b01731 Energy Fuels 2015, 29, 7681−7688
Article
Energy & Fuels
Figure 2. Entrained-flow gasifier of CanmetENERGY.
gasifiers.13 Yang el al. used the ROM to investigate the performance of a membrane-walled entrained-flow gasifier with two-stage oxygen supply.14 A sensitivity analysis of the inputs on the slag layer thickness was also performed in that study. Kong et al. performed a steady-state simulation for an industrial-scale Texaco gasifier by implementing a compartment reactor network.15 Recently, Sahraei et al. proposed a reactor network to study the performance of a short-residence-time gasifier.16 Despite these efforts, the above studies have only used one set of CFD simulation results for validation of their ROM. That is, the ability of the ROM to predict the performance of the gasifier over a range of test conditions still needs to be addressed. The simplifying assumptions considered in model order reduction techniques might affect the results beyond the nominal operating conditions of gasifier at which the framework of the model was developed. To the knowledge of the authors, a study that investigates the capabilities of the ROM in predicting temperature distribution and syngas composition under different operating conditions and kinetic parameters is not available in the literature. In the present work, comparisons between the predictions obtained by a ROM previously developed by our group16 and the corresponding CFD model for different test conditions of an entrained-flow gasifier are performed. The outcomes of the present study improve the confidence of the ROM and its capabilities to describe the operation of highly complex systems, such as a short-residence-time gasifier. Furthermore, a set of sensitivity analyses are performed to investigate the effects of various key inputs and model parameters on the results obtained from the ROM. These analyses have provided insights for modifying the proposed modeling framework to predict the behavior of the gasifier at different operating conditions. Most of the existing modeling studies have focused on the gasification of coals and biomasses; less work has been reported on petroleum coke gasification.17 It is predicted that almost 15% of planned gasification plants will be using petroleum coke as the feed.17 Thus, the present study considers the gasification of petroleum coke in a short-residence-time entrained-flow gasifier. The structure of this paper is as follows: Section 2 describes the gasification system used in this study. The details of the ROM and the CFD simulation are described in section 3.
cost. Conducting experimental tests and having access to sample measurements are costly and sometimes prohibitively so. Therefore, computational models can be developed and used to assess the performance and economics of the gasification units at different operating conditions. Two wellestablished approaches can be employed to model the behavior of the gasification process: computational fluid dynamic (CFD) models and reduced order models (ROMs). In a CFD model, detailed multi-phase sub-models are often used to simulate the flow patterns and complex processes taking place inside the gasifier. Validated CFD models can provide insights on the locations of injectors, design of the gasification unit, optimization of experimental test matrices, and investigation of commercial-scale gasifiers.6 However, CFD simulations are usually computationally intensive; therefore, their implementation to perform sensitivity analysis, co-simulation (integration of the CFD model to process simulators), controllability analysis, and process optimization may be limited or even prohibitive. One of the alternatives to CFD is the implementation of lower order models to approximate comprehensive CFD models. In a ROM, a reactor network describing the structure of the flow pattern is implemented to consider the most important characteristics of the gasification process, such as conversion, reaction systems, compositions, and temperature profiles. Such models require lower computational effort compared to CFD models while enabling the integration of gasifier models with process simulators. Furthermore, ROMs can be extended to study the transient performance of the gasifier during operational changes where the CFD models cannot be used as a result of unattractive computational costs. CFD models have been used to investigate biomass gasification,7 effects of coal particle density and size fraction in the process,8 feasibility of industrial-scale gasification systems,9 slag layer behavior,10 and optimization.11 On the other hand, the implementation of ROMs to model gasifiers has gained attention in recent years. Gazzani et al. presented a ROM to study the different operating conditions of a Shell− Prenflo gasifier.12 Monaghan et al. presented a ROM framework that simulates different entrained-flow gasifiers.3 Those authors also performed a sensitivity analysis of the key inputs on the char conversion and outlet temperature of the 7682
DOI: 10.1021/acs.energyfuels.5b01731 Energy Fuels 2015, 29, 7681−7688
Article
Energy & Fuels
data for model validation. Identification of reduced low-order models is a widely used technique in academia and industry to describe the behavior of large-scale, highly complex processes. Ricardez-Sandoval et al. reduced computationally intensive dynamic optimizations by representing the closed-loop nonlinear behavior of a system as a linear state-space model consisting of uncertain model parameters.20 Sanchez-Sanchez and Ricardez-Sandoval considered simultaneous evaluation of dynamic flexibility and feasibility to obtain optimal process synthesis and control structures based on worst case performance evaluation.21 This method reduced the computational costs of dynamic optimizations by a factor of 24 compared to formal dynamic optimization formulation. A review on approximation of large-scale systems for linear and nonlinear systems is provided by Antoulas and Sorenson.22 In the present study, the order of the mathematical model is reduced to a set of one-dimensional differential equations describing the behavior of the system along the axial domain of the gasifier. To develop a low-order model for a gasification system, a reactor network that captures the internal flow patterns of the system is needed. The reactor network describes the distribution of heat and mass transfer in different flow fields with the aim to simulate the physical structure of the actual gasifier. In the present study, the reactor network proposed by Sahraei et al.16 for the gasifier of CanmetENERGY is employed. The gasifier under consideration is divided into three specific flow regions: the jet expansion zone (JEZ), the external recirculation zone (ERZ), and the downstream zone (DSZ). However, the mixing of steam and hot recirculated gases divided the JEZ and ERZ into two sections. The reactor network is presented in Figure 3. As shown in this figure, the mixing zones (ERZ1 and ERZ2) are modeled using continuous stirred-tank reactors (CSTRs), while the JEZ1, JEZ2, and DSZ are modeled using onedimensional plug flow reactors (PFRs). The sampling probe of the gasifier (shown in Figure 2) is located at the outlet of the DSZ in the reactor network. The model parameters, volume of each reactor, and split ratio of each stream are provided in ref 16. A brief description of the framework of the ROM employed in this study is presented next. The ROM employed in this study used a multi-component approach to deal with volatile components, which were implemented through the simulation of fuel pyrolysis using PC Coal Lab software. This software takes into account the effect of the pressure, temperature, and heating rate to compute the kinetic parameters of the reactions.23,24 The pyrolysis products from PC Coal Lab consist of materials such as CO2, H2O, CO, CH4, H2, H2S, and char.24 The ROM considers the formation of sulfurous and nitrogenous pollutants. In the case of sulfur, SO2 takes part in further reactions to initially form COS, which continuously reacts with H2 to produce H2S. Moreover, the chemically bound nitrogen within the volatiles can form HCN, which reacts with NO to produce N2 in the product gases. The relevant kinetic reactions of pollutant formation as well as the reactions employed for petroleum coke gasification and partial combustion of volatiles are presented in ref 16. Momentum, energy, and mass conservation equations were developed for each zone of the reactor network shown in Figure 3. Momentum conservation equations for gas and solids consider advection momentum transfer, pressure gradient, gravity, gas−particle viscous forces, and gas−wall friction forces. The model assumes that the gas−wall friction only occurs in the external layers of the reactor network, i.e., ERZ2 and DSZ. Energy conservation equations for gas and solid phases take into account advection, convective heat transfer, heat as a result of the homogeneous and heterogeneous reactions, particle−gas radiation, and convection sources for gas to particle, gas to wall, and wall to surroundings. Mass conservation equations for gas and solid phases account for molecular diffusion, advection, convective heat transfer, reactions, and particle mitigation to the walls of the gasifier. The resulting set of nonlinear differential equations describing the operation of the reactor network was simultaneously solved using finite difference approximations on a discretized domain. The modeling details of the ROM are presented in ref 16. 3.2. CFD Development. A CFD model describing the behavior of the entrained-flow gasifier of CanmetENERGY was developed using
The results of the two models are compared in section 4. The insights gained from sensitivity analyses are presented in section 5. Concluding remarks are stated in section 6.
2. GASIFICATION SYSTEM To overcome the current limitations associated with the capital expenses in gasification units, more compact and economically attractive gasifiers need to be developed. A new class of compact gasifiers are widely known as short-residence-time gasifiers.4 The volume of such gasifiers is approximately 90% lower than that required by conventional entrained-flow gasifiers,18 and therefore, the associated capital costs are reduced. This technology is capable of achieving high carbon conversions within a residence time of less than 0.5 s.4 Testing of the components for a short-residence-time gasifier at the pilot scale has been performed by CanmetENERGY in Ottawa, Canada. As shown in Figure 2, the pilot-scale gasifier of CanmetENERGY is a single-stage, pressurized, downward-feed gasification facility capable of converting 1 tonne/day of various solid fuels, such as petroleum coke, to syngas. Petroleum coke is a carbonaceous byproduct of crude-oil refineries for heavy oil upgrading with thermal cracking processes. It is often more economical than coal and typically used as a source of energy in steam and power plants. On the other hand, it produces more carbon emissions than coal; i.e., on a per unit of energy basis, petroleum coke emits 5−10% more CO2 than coal.19 In Canada, petroleum coke is primarily produced by oil sands upgraders. The proven oil sands reserves of Canada will yield roughly 5 billion tons of petroleum coke, which is adequate to provide fuel for 111 U.S. coal power plants until 2050.19 In this gasifier, dry fuel particles are carried by nitrogen and injected into oxygen jets created by eight oxygen nozzles. Superheated steam is fed to the gasifier to moderate the reactor temperature and enhance hydrogen production. The average particle residence time reported for the gasifier of CanmetENERGY is 0.3−0.5 s.16
3. GASIFIER MODELING The ROM and CFD simulations of the gasifier of CanmetENERGY are described in this section. Base-case conditions used in this study are listed in Table 1. 3.1. ROM Development. Model order reduction techniques comprise the following aspects: appropriate approximation schemes to describe the process under consideration, preservation of key decisive parameters and variables of the original system, computational stability, i.e., avoid numerical errors, and having access to experimental
Table 1. Base-Case Conditions for Petroleum Coke Gasification test condition petroleum coke (kg/h) carrier gas (nitrogen) (kg/h) oxygen (kg/h) steam (kg/h) operating pressure (bar) petroleum coke composition proximate analysis (as received)
ultimate analysis (dry and ash-free)
particle diameter (μm) recirculation ratio jet angle (deg)
ash volatile matter moisture carbon hydrogen sulfur nitrogen oxygen carbon
50 13 67.9 10.5 16 mass fraction 0.046 0.127 0.005 0.822 0.042 0.061 0.018 0.015 0.864 79 1.1 17 7683
DOI: 10.1021/acs.energyfuels.5b01731 Energy Fuels 2015, 29, 7681−7688
Article
Energy & Fuels
Figure 3. Proposed reactor network for the gasifier of CanmetENERGY.16
Figure 4. Streamlines and normalized velocity vectors of CFD simulation: (a) −10% O2 flow rate, (b) base case, and (c) +10% O2 flow rate. and the parameters are from Kovacik et al.26 Particles impacting the wall will either stick or reflect depending upon the following conditions: particles with a high char content (less than 88% conversion) are assumed to reflect, which is based on the observations of Li et al.27 for bituminous coal. For particles with higher conversion, a probability of sticking is calculated from the ash viscosity, following an approach similar to the work by Richter et al.,28 with an assumed critical viscosity of 100 Pa s. If a particle sticks to the wall, it is removed from the simulation.
the ANSYS CFX software (version 16). It was built based on the foundational work reported by Chui et al.6 The model geometry consists of a half-section of the gasifier, includes the sampling probe and a portion of the nozzle, and ends at the entrance to the quench section. Conjugate heat transfer is modeled through the innermost refractory layer and select parts of the nozzle. Beyond the inner layer, an estimated heat-transfer coefficient is applied to represent the remaining insulating layers of the gasifier. The same coefficients have been implemented in the ROM to model heat transfer through the walls. This estimate takes into account temperature-dependent material properties and includes radiation from the steel shell to the surroundings. The model uses the Reynolds-average Navier−Stokes equation with k−ε turbulence closure, discrete transfer radiation model, composition-dependent gas properties, and Lagrangian particle tracking. The volatiles have been modeled in the CFD model as a single component, and the rate of devolatilization is modeled using a first-order Arrhenius rate equation. The char oxidation and gasification reactions are the same as those used in the ROM. For simplicity and numerical stability, the reaction of char with H2 has been omitted and the inhibition of the char gasification reactions by CO and H2 has been neglected. The formation of NO and SO2 is based only on the composition of char to balance the mass of the elements. In the reducing environment of the gasifier, nitrogenous and sulfurous species are expected to be produced in other forms; however, the relevant gas-phase reduction reactions were not implemented in the CFD model because of computational limitations. Because the amount of sulfur is 6.1% of the dry and ash-free fuel, the changes in energy in converting SO2 to form H2S may explain some differences between CFD and ROM. The stoichiometry of the volatile oxidation reaction was obtained from PC Coal Lab, but no reaction rate was available for this matter. Thus, it is assumed that the volatile reaction was limited by turbulent mixing, using the eddy dissipation model (EDM). The rate of the hydrogen oxidation reaction is the slower of the EDM and the finite rate chemistry (FRC) model.25 The water-gas shift reaction uses only FRC,
4. ROM AND CFD RESULTS The ROM framework was designed according to the flow patterns obtained from CFD simulations generated at the basecase operating conditions presented in Table 1. To assess the accuracy of the proposed ROM, the capability of the framework of the ROM to predict the behavior of the gasifier needs to be investigated over a range of operating conditions. In the present work, the accuracy of the ROM in predicting the carbon conversion, temperature, and composition profiles was examined by performing a sensitivity analysis on a test condition (oxygen flow rate) and a reactor network parameter (heterogeneous reaction kinetics). To obtain the results of the ROM, the momentum equations were primarily solved on the basis of an average gas density in each zone of the reactor network. The resulting velocity profiles were provided as inputs to solve for the mass- and energy-transfer equations simultaneously at each grid point inside the reactor network. Note that, in the corresponding CFD model, all of the governing equations (velocity, mass, and heat) were simultaneously solved for the two-dimensional mathematical system. As indicated in Figure 4, no distinguishable changes were observed in the streamlines of CFD simulations when the 7684
DOI: 10.1021/acs.energyfuels.5b01731 Energy Fuels 2015, 29, 7681−7688
Article
Energy & Fuels oxygen flow rate was varied by ±10%. Also, the recirculation zone ends at the same axial distance in all cases. Therefore, the framework of the ROM was not adjusted for the changes in the oxygen flow rate considered in this work. The temperature distribution predicted by the ROM and CFD simulation for three different oxygen flow rates are presented in Figure 5. As
H2 are presented in Figures 6 and 7, respectively. As shown in Figure 6 and Table 2, the composition of CO obtained from
Figure 6. Carbon monoxide molar fraction: (a) −10% O2 flow rate, (b) base case, and (c) +10% O2 flow rate. Figure 5. Temperature distribution in kelvins: (a) −10% O2 flow rate, (b) base case, and (c) +10% O2 flow rate.
shown in this figure, the gas temperature is highest at the beginning of the reactor as a result of the combustion of volatiles for the three oxygen flow rates considered. The released heat provides the required energy to ignite the solid particles and run the gasification reactions. By depletion of oxygen, endothermic reactions, which are dominant in the DSZ, cause a drop in the temperature. When the oxygen flow rate increases, more carbon particles are expected to react with oxygen; thus, more energy is released, causing a higher temperature distribution within the gasifier, as shown in Figure 5. The modeling results at the sampling probe are presented in Table 2. According to Figure 5 and Table 2, the ROM predicted a similar temperature distribution to that observed from CFD simulations; however, there is a temperature deviation of 70−120 K for the test conditions. The different approaches implemented in modeling of volatile species, homogeneous reactions, and pollutant formation in the ROM are the main expected causes for these deviations. As shown in Table 2, the proposed ROM predicted similar carbon conversions compared to CFD results. According to the results, high temperatures were attained in the gasifier in the gasifier. This increased the gasification reaction rates and resulted in high carbon conversion within a short residence time. The distributions of the molar compositions for CO and
Figure 7. Hydrogen molar fraction: (a) −10% O2 flow rate, (b) base case, and (c) +10% O2 flow rate.
the CFD model decreased when more oxygen was injected into the system. As shown in Figure 5, the temperature profile increased proportionally with the oxygen flow rate. This expected behavior shifted the water-gas shift reaction to the right-hand side, which led to more CO converted to CO2. Both the ROM and CFD simulations predicted this trend for the CO
Table 2. Molar Compositions, Temperature, and Char Conversion at the Sampling Probe −10% oxygen flow rate
base case CO H2 CO2 H2O T (K) Xchar (%)
+10% oxygen flow rate
/10× kinetic parameters
1
10× kinetic parameters
CFD
ROM
CFD
ROM
CFD
ROM
CFD
ROM
CFD
ROM
0.520 0.148 0.084 0.143 2299 99.2
0.552 0.141 0.074 0.112 2172 98.5
0.535 0.173 0.063 0.111 2099 95.8
0.558 0.165 0.064 0.097 1990 94.9
0.485 0.104 0.12 0.188 2395 99.7
0.541 0.097 0.109 0.152 2469 99.2
0.473 0.117 0.113 0.182 2329 89.5
0.503 0.111 0.109 0.160 2299 91.4
0.522 0.137 0.083 0.140 2301 99.9
0.565 0.139 0.073 0.111 2191 98.7
7685
DOI: 10.1021/acs.energyfuels.5b01731 Energy Fuels 2015, 29, 7681−7688
Article
Energy & Fuels molar composition when the oxygen flow rate was decreased. According to Figure 7, the ROM predictions in terms of the H2 molar compositions are in agreement with those obtained from CFD simulations. At higher concentrations of oxygen, more H2 is reacted to form H2O; thus, in both models, H2 molar composition at the outlet of the gasifier is decreased when more oxygen reacts with the fuel. On the basis of the results shown in Table 2, the proposed ROM was able to predict satisfactory results for CO2 and H2O molar fractions when compared to CFD simulation results. The major difference between the two models in predicting syngas composition was observed in the ERZ (the wall regions of the upper section of the gasifier). As shown in Figures 6 and 7, the molar compositions of CO and H2 are higher in this zone of the ROM compared to CFD simulations. The ROM considers the recycle streams with the molar composition calculated from JEZ2 to enter ERZ, where more CO and H2 are already formed. In the CFD model, however, these streams recirculate gradually into ERZ2. As a result, higher molar compositions are observed in ERZ2 sections of the ROM. To examine further the capabilities of the ROM, changes in the reactor network parameters were performed by modifying the constant factor of the Arrhenius rate equations for the heterogeneous reactions. Two case studies were considered: increasing and decreasing the Arrhenius constants by a factor of 10. As the kinetic parameters are reduced, the heterogeneous reactions become slower and less carbon conversion is expected within the gasification unit. According to Table 2, the carbon conversion is reduced to 91.4% when the kinetic parameters are decreased; however, the carbon conversion did not change considerably for higher reaction rates because nearly complete carbon conversion was already achieved for the base-case condition. Thus, the results for the case study when the kinetic parameters are increased are expected to be similar to those obtained for the base-case condition. According to Table 2, the ROM outputs for conversion, temperature, and compositions of these case studies are in good agreement with the results of CFD simulation. The molar composition profiles of CO and H2 are presented in Figures 8 and 9, respectively. The trend in the composition profiles is similar for the two models, with the ROM predicting slightly higher CO composition for each case,
Figure 8. Carbon monoxide molar fraction: (a) parameters and (b) 10× kinetic parameters.
1
Figure 9. Hydrogen molar fraction: (a) 1/10× kinetic parameters and (b) 10× kinetic parameters.
i.e., 3−4%. The dominant reactions within the gasifier are the endothermic char gasification reactions rather than partial combustion of carbon. Therefore, the outlet temperature of the gasifier must increase for the case when the kinetic parameters are decreased. As shown in Table 2, both models predicted higher outlet temperatures for this case compared to the basecase operating condition. The required computational time to obtain the results using the proposed ROM was on average 200 s per case study (Core i7, 3.4 GHz with 8 GB of RAM) compared to the computational time of 7−10 days for a CFD simulation (using a server cluster of 32 processing cores). The significant reduction in the computational time gives the ROM an advantage over CFD simulations to perform further modeling analysis. Moreover, the ROM was capable of predicting the base-case results with more accuracy compared to the other four cases presented above. This is due to the fact that the geometry and framework of the ROM were developed using the streamlines of the base-case CFD simulation. Although the ROM is only an approximation to the CFD model, the overall results (obtained over a range of operating conditions and kinetic parameters) have demonstrated the prediction capabilities of the ROM as a computationally efficient alternative to describe the performance and operation of the gasification unit of CanmetENERGY.
5. SENSITIVITY ANALYSIS A sensitivity analysis can provide insights into the effects of the design and reactor network parameters on the gasification behavior and can therefore be used to increase the quality and prediction capabilities of the proposed ROM. Accordingly, a univariate sensitivity analysis of the inputs of the ROM was conducted to identify the effect of key parameters on the simulation results. These parameters include reactor network parameters, i.e., the recirculation ratio, jet angle, steam gasification, and Boudouard reaction rates, and design parameters, i.e., particle size, oxygen flow rate, and steam flow rate. To perform the sensitivity analysis, all of the input parameters were individually varied over a range of ±10% around their base-case values, which are listed in Table 1. The carbon conversion, H2/CO ratio, and temperature profile obtained for each input change were recorded. The results of
/10× kinetic 7686
DOI: 10.1021/acs.energyfuels.5b01731 Energy Fuels 2015, 29, 7681−7688
Article
Energy & Fuels
Figure 10. Sensitivity analysis of (a) carbon conversion, (b) H2/CO ratio, (c) outlet temperature, and (d) peak temperature.
these temperatures. During gasification, endothermic reactions were dominant; thus, larger particles decreased the peak temperature. However, the size of the particles has no significant effect on the outlet temperature. Other parameters, such as the diameter of ERZ1, particle-carrier flow rate, average gas density used for momentum equations, and number of particles, were also examined, but they have no significant effects on the performance of the gasification unit.
the sensitivity analysis are presented in Figure 10. As indicated in Figure 10a, carbon conversion is mostly influenced by the oxygen flow rate. In the gasifier design of CanmetENERGY, the steam is injected at a higher temperature than the other feeds, i.e., oxygen and fuel. Therefore, increasing the steam flow rate has a direct effect on the carbon conversion, unlike slurry gasifier designs discussed by Monaghan and Ghoniem.13 Among the reactor network parameters, the recirculation ratio has the highest effect on carbon conversion. As this parameter is increased, more particles are recirculated within the ERZ2, leading to higher residence time and carbon conversion. In addition, the jet angle and kinetic parameters have a limited effect on carbon conversion. The input parameters considered in this analysis have almost the same effects on the H2/CO ratio as they have on carbon conversion, except for the C + CO2 kinetic rate, oxygen flow rate, and jet angle. As shown in Figure 10b, more oxygen and an increased degree of the jet angle (which contributes to a larger expansion zone) decreased the H2/CO ratio as a result of higher conversion of carbon to CO and H2 to H2O. The effects of changes on each individual parameter over the outlet temperature and peak temperature are presented in panels c and d of Figure 10, respectively. The oxygen flow rate and recirculation ratio are the most important parameters affecting
6. CONCLUSION A comparison between a ROM and a CFD simulation for the short-residence-time gasifier of CanmetENERGY was presented. The ROM includes a reactor network and various submodels to account for laminar and mixing flows, devolatilization, homogeneous and heterogeneous reactions, fluid dynamics, pollutant formation, and heat transfer through the walls of the gasifier. Apart from the pollutant formation, the CFD simulation used more comprehensive sub-models to account for these features. Validation of the ROM was performed over a range of operating conditions using CFD simulations. Although the framework of the ROM was developed for the base-case test condition, the results showed satisfactory agreement with the CFD simulations for conversion, temperature, and composition profiles under different operating conditions and 7687
DOI: 10.1021/acs.energyfuels.5b01731 Energy Fuels 2015, 29, 7681−7688
Article
Energy & Fuels
(12) Gazzani, M.; Manzolini, G.; Macchi, E.; Ghoniem, A. F. Reduced order modeling of the Shell−Prenflo entrained flow gasifier. Fuel 2013, 104, 822−837. (13) Monaghan, R. F. D.; Ghoniem, A. F. A dynamic reduced order model for simulating entrained flow gasifiers. Part II: Model validation and sensitivity analysis. Fuel 2012, 94, 280−297. (14) Yang, Z.; Wang, Z.; Wu, Y.; Li, Z.; Ni, W. Use of a reactor network model in the design and operation of a new type of membrane wall entrained flow gasifier. Energy Fuels 2013, 27, 6322− 6332. (15) Kong, X.; Zhong, W.; Du, W.; Qian, F. Compartment modeling of coal gasification in an entrained flow gasifier: A study on the influence of operating conditions. Energy Convers. Manage. 2014, 82, 202−211. (16) Hossein Sahraei, M.; Duchesne, M. A.; Yandon, R.; Majeski, A.; Hughes, R. W.; Ricardez-Sandoval, L. A. Reduced order modeling of a short-residence time gasifier. Fuel 2015, 161, 222−232. (17) Murthy, B. N.; Sawarkar, A. N.; Deshmukh, N. A.; Mathew, T.; Joshi, J. B. Petroleum coke gasification: A review. Can. J. Chem. Eng. 2014, 92, 441−468. (18) Fusselman, S. P.; Sprouse, K. M.; Darby, A. K.; Tennant, J.; Stiegel, G. J. Pratt & Whitney Rocketdyne/DOE Advanced Single-Stage Gasification Development Program; National Energy Technology Laboratory (NETL): Pittsburgh, PA, 2007. (19) Stockman, L.; Turnbull, D.; Kretzmann, S. Petroleum Coke: The Coal Hiding in the Tar Sands; Oil Change International: Washington, D.C., 2013. (20) Ricardez Sandoval, L. A.; Budman, H. M.; Douglas, P. L. Simultaneous design and control of processes under uncertainty: A robust modelling approach. J. Process Control 2008, 18, 735−752. (21) Sánchez-Sánchez, K.; Ricardez-Sandoval, L. Simultaneous process synthesis and control design under uncertainty: A worst-case performance approach. AIChE J. 2013, 59 (7), 2497−2514. (22) Antoulas, A. C.; Sorensen, D. C. Approximation of large-scale dynamical systems: An overview. Int. J. Appl. Math. Comput. Sci. 2001, 11 (5), 1093−1121. (23) Merrick, D. Mathematical models of the thermal decomposition of coal: 1. The evolution of volatile matter. Fuel 1983, 62, 534−9. (24) Niksa Energy Associates LLC. http://www.niksaenergy.com (accessed March 28, 2015). (25) Jones, W. P.; Lindstedt, R. P. Global reaction schemes for hydrocarbon combustion. Combust. Flame 1988, 73, 233−249. (26) Kovacik, G. J.; Knill, K. J. Numerical simulation of coal gasification reactors. Proceedings of the International Joint Power Generation Conference; Phoenix, AZ, Oct 2−6, 1994; pp 101−106. (27) Li, S.; Wu, Y.; Whitty, K. J. Ash deposition behavior during charslag transition under simulated gasification conditions. Energy Fuels 2010, 24 (3), 1868−1876. (28) Richter, S.; Ströhle, J.; Schnell, U.; Hein, K. R. G. Application of the 3D combustion simulation code AIOLOS to the prediction of ash deposition in a pulverised coal-fired utility boiler. Proceedings of the 4th International Symposium on Coal Combustion; Beijing, China, Aug 18− 21, 1999.
kinetic parameters. The ROM requires much lower computational time than that required by the CFD model. This makes the ROM a computationally attractive tool for performing further modeling analysis. A sensitivity analysis was performed on the ROM to obtain insights regarding the impacts of different inputs, such as operational variables and reactor network parameters, on the results of the ROM. This analysis has shown that the major input parameters affecting the carbon conversion, H2/CO ratio, and temperature profile are the oxygen flow rate and the recirculation ratio. The addition of oxygen increases the operating temperature of the gasifier, whereas a higher recirculation ratio increases the residence time of the particles. These two effects lead to higher conversions. Other input or design parameters, such as the jet angle, particle size, and kinetic rates, have been shown to have a secondary effect on the operation of the gasification unit studied in this work.
■
AUTHOR INFORMATION
Corresponding Author
*Telephone: +1-519-888-4567, ext. 38667. Fax: +1-519-8884347. E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the Program of Energy Research and Development of the Government of Canada and the ecoENERGY Innovation Initiative.
■
REFERENCES
(1) Exxon Mobil Corporation. The Outlook for Energy: A View to 2040; Exxon Mobil Corporation: Irving, TX, 2014; http://exxonmobil. com/energyoutlook (accessed June 16, 2015). (2) Rubin, E. S. The outlook for power plant CO2 capture. Erice International Seminar; Aug 2009; http://repository.cmu.edu/epp (accessed June 16, 2015). (3) Monaghan, R. F. D.; Ghoniem, A. F. A dynamic reduced order model for simulating entrained flow gasifiers: Part I: Model development and description. Fuel 2012, 91, 61−80. (4) Hossein Sahraei, M.; McCalden, D.; Hughes, R.; RicardezSandoval, L. A. A survey on current advanced IGCC power plant technologies, sensors and control systems. Fuel 2014, 137, 245−259. (5) Hossein Sahraei, M.; Ricardez-Sandoval, L. A. Controllability and optimal scheduling of a CO2 capture plant using model predictive control. Int. J. Greenhouse Gas Control 2014, 30, 58−71. (6) Chui, E. H.; Majeski, A. J.; Lu, D. Y.; Hughes, R.; Gao, H.; McCalden, D. J.; Anthony, E. J. Simulation of entrained flow coal gasification. Energy Procedia 2009, 1, 503−509. (7) Xue, Q.; Fox, R. O. Multi-fluid CFD modeling of biomass gasification in polydisperse fluidized-bed gasifiers. Powder Technol. 2014, 254, 187−198. (8) Slezak, A.; Kuhlman, J. M.; Shadle, L. J.; Spenik, J.; Shi, S. CFD simulation of entrained-flow coal gasification: Coal particle density/ sizefraction effects. Powder Technol. 2010, 203, 98−108. (9) Vascellari, M.; Roberts, D. G.; Hla, S. S.; Harris, D. J.; Hasse, C. From laboratory-scale experiments to industrial-scale CFD simulations of entrained flow coal gasification. Fuel 2015, 152 (15), 58−73. (10) Yong, S. Z.; Ghoniem, A. Modeling the slag layer in solid fuel gasification and combustionTwo-way coupling with CFD. Fuel 2012, 97, 457−466. (11) Tremel, A.; Spliethoff, H. Gasification kinetics during entrained flow gasificationPart III: Modelling and optimization of entrained flow gasifiers. Fuel 2013, 107, 170−182. 7688
DOI: 10.1021/acs.energyfuels.5b01731 Energy Fuels 2015, 29, 7681−7688
