CFD Simulations of Allothermal Steam Gasification Process for Hydrogen Production

: The article presents an experimental laboratory setup used for the empirical determination of the gasification of coal samples in the form of solid rock, cut out in the form of a cylinder. An experimental laboratory set enabled a series of experiments carried out at 700 °C with steam as the gasification agent. The samples were prepared from the coal seam, the use of which can be planned in future underground and ground gasification experiments. The result of the conducted coal gasification process, using steam as the gasification agent, was the syngas, including hydrogen (H 2 ) with a concentration between 46% and 58%, carbon dioxide (CO 2 ) with a concentration between 13% and 17%, carbon monoxide (CO) with a concentration between 7% and 11.5%, and methane(CH 4 ) with a concentration between 9.6% and 20.1%.The results from the ex ‐ situ experiments were compared with the results of numerical simulations using computational fluid dynamics (CFD) methods. A three ‐ dimensional numerical model for the coal gasification process was developed using Ansys ‐ Fluent software to simulate an ex ‐ situ allothermal coal gasification experiment using low ‐ moisture content hard coal under atmospheric conditions. In the numerical model, the mass exchange (flow of the gasification agent), the turbulence description model, heat exchange, the method of simulating the chemical reactions, and the method of mapping the porosity medium were included. Using the construction data of an experimental laboratory set, a numerical model was developed and its discretization (development of a numerical grid, based on which calculations are made) was carried out. Tip on the reactor, supply method, and parameters maintained during the gasification process were used to define the numerical model in the Ansys ‐ Fluent code. A part of the data were supplemented on the basis of literature sources. Where necessary, the literature parameters were converted to the conditions corresponding to the experiment, which were carried out. After performing the calculations, the obtained results were compared with the available experimental data. The experimental and the simulated results were in good agreement, showing a similar tendency.


Introduction
The development of civilization is undoubtedly associated with the acquisition of mineral resources, including fossil energy resources. Energy resources are the basis for the production of world energy, initially as heat and then in the form of electricity. Due to the scarcity of energy resources and the accompanying release of greenhouse gas emissions to the natural environment, it has become necessary to manage energy economically, which is one of the fundamental problems facing the modern world. In particular, this problem is felt in European Union countries, including Poland. In addition, the situation is complicated by the awareness of the limited supply of natural resources of oil and natural gas, and the uncertain situation regarding changes in prices of these energy resources, which inevitably has an impact on the energy security of countries [1][2][3][4]. different operating parameters. The numerical model includes reaction kinetics for the solid-phase reactions, together with gas diffusion and equilibrium for the gas-phase reactions. Lee at al. [13] developed a two-dimensional numerical model of the coal gasification process in order to predict product gas yield, syngas composition, and efficiency, calorific value of the syngas, and carbon conversion. Shirazi, in [25], developed a two-dimensional model in COMSOL Multiphysics as well as in Ansys-Fluent to predict the effect of different parameters on the UCG process performance with temperature. Sreedharan in [22] developed a three-dimensional numerical model of the coal gasification and heavy oil in order to predict syngas production using a discrete phase model. Mota et al., in [26], developed a three-dimensional model of a bubbling bed gasifier using CFD method to predict and optimize hydrogen-rich syngas during the thermochemical conversion of a lignite coal using steam and oxygen as oxidants.
The numerical modeling available in the literature concentrates mainly on the aspects of key influential parameters for coal gasification process, value, and composition of product syngas, optimal efficiency, and process factors, in which coal is mostly simulated in form of slurry or particles.
In this paper, the task of developing and constructing an experimental laboratory setup with the intention of conducting a series of coal gasification experiments, occurring in a homogeneous material block, with a given geometry in the form of a cylinder, was undertaken. The obtained experimental data referred to the developed numerical model of the coal gasification process, validating the solution obtained by comparing them to the results obtained from the real system from the gas-flow reactor.
An experimental laboratory setup will lead to the gaining of more knowledge concerning the gasification of coal in the gasification reactor in the phenomenological sense, through the possibility of studying the volume composition of CO2, CO, H2, and CH4 gas, which is the desired end result of the whole process. Moreover, it will offer the possibility of forecasting based on the developed mathematical and numerical model using the computational fluid mechanics (CFD) method.

Materials and Methods
The implementation of the work required access to relevant experimental data. Due to the complexity of the work, data from an experiment conducted on a laboratory scale were used. Based on this, the model implemented in the Ansys-Fluent software was prepared and verified. The data from the experiments carried out on a laboratory scale also served to validate the mathematical model developed.

Physical Model
The experiment carried out on a laboratory scale consisted of the gasification of coal samples with steam in the form of a solid rock in the shape of a cylinder. Laboratory tests of the coal gasification process were carried out using an experimental laboratory setup consisting of (  (Figure 1b), whose operation was controlled using a digital proportional-integral-derivative (PID) controller. Steam as a gasification agent was obtained as a result of the rapid evaporation of water from the surface of a ceramic bed (aluminum oxide Al2O3), in the shape of spheres with a diameter of 6 mm, filling the cylindrical vessel (Figure 2k), located directly in front of the test coal sample. In order to eliminate the possibility of the uncontrolled movement of the ceramic bed (dialuminium trioxide-Al2O3) in the gasification reactor, the tank (Figure 1e) was closed from the top with a tight steel ring together with a steel mesh with a 0.2 mm diameter. The synthesis gas emerging in the reactor, flowing through a heat exchanger ( Figure 1d) fed with liquid pumped into the system by a piston pump (Figure 1f), condensates to the tank ( Figure 1e) and then through the outlet valve ( Figure 1j) it was directed to high density polyethylene (HDPE) bags with a capacity of 2 dm 3 . The chemical composition of the stored process gas was measured using a gas chromatograph. Figure 1. An experimental laboratory setup developed for the coal gasification: a-reactor, b-furnace, c-water pump piston, d-syngas cooler (heat exchanger), e-condensate tank, f-pump supplying the heat exchanger, g-manometer at inlet, h-manometer at outlet, i,j-inlet and outlet valves.
The coal gasification experiments were carried out on an experimental laboratory setup in a gasification reactor (Figure 1a) designed and constructed with the aim of reproducing, in an approximate way, the conditions prevailing in its interior during gasification occurring under conditions of underground coal gasification. The behavior of the coal sample was to be retraced under the conditions of the gasification process. The reactor, designed as a pipe with an inner diameter of 30 mm, length of 1800 mm, and wall thickness of 2 mm, allowed these requirements to be met, and for both samples of coal simulating the bed, and the walls of the pipe as surrounding rocks, to be place in its interior. Coal was delivered in the form of blocks with dimensions of 400 mm × 400 mm × 250 mm (Figure 3), from which cylindrical coal samples were taken using traditional methods of borehole drilling. A raw coal sample was placed in the reactor (Figures 2a and 4) and then heated to 700 °C in a nitrogen atmosphere. Nitrogen was supplied from the gasification reactor bottle with an inlet valve (Figure 1i). After the temperature stabilized, the nitrogen flow was turned off, and then water was fed using a piston pump with a mass flow rate of the supplied agent of 0.1 mL per min −1 , determined experimentally. The construction of the reactor enables the monitoring of the pressure, the temperature of the furnace, and the value of the mass flow of the water at the reactor inlet. The reactor allows the coal gasification process to be conducted at a maximum temperature of 1200 °C. The maximum size of the coal sample was determined experimentally as a solid cylinder with a diameter of 30 mm and a length of 85 mm ( Figure 5). In the coal sample, each time before starting the experiment, two gasification channels measuring 2 mm × 2 mm × 85 mm were made symmetrically to the coal sample axis ( Figure 6). Then the mass measurement of the sample was measured using an electronic Precisa scale, allowing mass measurement with an accuracy of ±0.01 g. The reactor together with the test coal sample placed inside was sealed with a flange ( Figure 2l) and was secured against the opening with clamps, as shown in Figure 7. At the end of each of the gasification experiments, the coal sample was obtained in the form of a degassed coal mass, as shown in Figure 8.

Coal Samples
The quality of the solid fuel depends on many factors, but primarily on its chemical composition. Based on the chemical composition of coal, the suitability of the fuel is determined. The content of the carbon and hydrogen in the coal determines the amount of heat evolved during gasification processes. Sulphur is an undesirable component, because its compounds cause the release of harmful odors (including H2S). They adversely affect the natural environment and contribute significantly to the formation of corrosion. Moisture and ash in the coal mass are the passive part, i.e., the so-called ballast, which does not take part in energy processes.
The basic quality parameters of hard coals are:  Total moisture content;  Ash content;  Total sulfur content;  Heating value.
Three coal samples were taken from hard coal mines operated by the Polish Mining Group (PGG S.A.). The coal samples were characterized by the following parameters (average values) from the ultimate and proximate analysis, as shown in Table 1. Through analysis of the results of the chemical composition tests, selected samples of coal intended for gasification in a laboratory installation can be classified as typical steam coal due to the volatiles content being above 28%.

Laboratory Tests
The measurement of the chemical composition of the process gas was carried out using the Agilent 3000A gas chromatograph, allowing measurement with an accuracy of ±0.01%.A typical single-acting piston pump consisting of a chamber (cylinder) (Figure 9a) was used to supply the gasification reactor, in which the eccentrically driven piston moves with the help of a shackle (Figure 9b), where the piston stroke was regulated by a micrometer screw (Figure 9c). The pump chamber is separated from the suction part by a self-acting valve (Figure 9d) enabling the fluid to flow from the suction part to the pump chamber and blocking the flow in the reverse direction. The piston makes a typical reciprocating movement in the cylinder ( Figure 9). The pump's working range and the value of the supplied mass stream at the inlet of the reactor had an unsettled character and resembled a sine wave (due to the step nature of pump operation every 2.5 s); hence, it was necessary to develop a proper mathematical model that would reflect the cycle of the numerical operation of the reactor in the Ansys-Fluent software.
The approximate conditions prevailing at the inlet to the reactor were expressed in the following form: Figure 10 shows the effects of the developed mathematical model (1) taking into account the following input data:


Mass stream value for t = 0 Amplitude of the reactor's working cycle A = 0.001;  The period of the reactor's work cycle ω = 1 (s −1 ). The User Defined Function (UDF) programming tool enabled the implementation of the mathematical model (1) in the numerical model as a boundary condition of the supplied gasification agent at the inlet to the gasification reactor [27].
In addition to testing the volume fraction of individual process gas components, the following were also measured: the duration of filling HDPE bags with process gas (texp), the volume of synthesis gas (Vsyngas), the coal sample mass before gasification (mC1) and after gasification (mC2), and determining their difference (mC3), as well as the mass (mH2O(l))and volume (VH2O(l)) of the pure water supplied into the reactor with a piston pump, in order to generate water vapor for the needs of the experiment.
An important aspect during the implementation of each coal gasification process is the well-defined porosity of hard coal, ensuring the correct access of the gasification agent to the layer consisting of micro and submicropores. As experience shows [2,3,20], in the conditions of UCG, low permeability of coal beds is a frequent reason why a reactor may stop working, due to high resistance preventing the normal flow of gasification and the outflow of reaction products. Therefore, an attempt was made to capture the changes in the porosity of the tested coal samples and the impact on the quality of the process gas obtained by examining the value of pressure changes at the inlet to the gasification reactor (the results are given in the tables). The values of pressure change read from the measurements made it possible to estimate the flow resistance parameter in the Ansys-Fluent software.
Each experimental test was repeated three times, and the results are presented in Table 2   Exp.-ex-situ experiment.      Figure 15 presents the syngas composition obtained during the gasification of the cola sample from the hard coal mine C, where the average concentration of H2 in the syngas was approximately 56.30%, the average concentration of CO2 was approximately 7.30%, the average concentration of CH4 in the syngas was approximately 11.10%, and the average concentration of CO in the syngas was approximately 7.30%. However, the average concentration of the other syngas components, such as O2, C2H6, N2, and H2S, was approximately 18%.  Figure 17 presents the geometric model of the reactor, which corresponds to the real dimensions of the gasification reactor designed and constructed for the needs of the tests, meeting the requirements in terms of geometric dimensions and flow characteristics. When developing the spatial model of a gasification reactor, the following information was taken into consideration:  Gas-field geometry;  Geometry of the gasification channels;  Geometry of the tested coal sample.   Figure 18 shows the solid model, which makes it possible to carry out the numerical calculations presented in this paper, which are aimed at determining the volume composition of CH4, CO, CO2, and H2 in process gas. Figure 19 shows the results of the developed numerical grid of the entire area of the medium flow in the gas and coal sample.

Numerical Grid
In the case of flows with a simultaneous chemical reaction, it is recommended that very dense numerical grids of the area occupied by the fluid should be aimed for in order to capture the processes occurring in the modeled flow more accurately. These requirements were decisive and resulted in the development of two types of numerical grids, namely:  The numerical grid should be characterized by having computational cells of regular shape and identical size, which in practice proves to be too difficult to achieve. This necessitates the need for an appropriate criterion, which, by its description, will provide a measure of deviation from this optimum. In Ansys-Fluent, this criterion is the aspect ratio (aR) defined by the relation [27]: These criteria are satisfied if the obtained value from Equation (2) is strongly less than the value of 100. This criterion ensures obtaining a stable numerical solution as well as a faithfully reproduced volume occupied by the fluid [27]. Figure 20 illustrates the idea of the mesh convergence adapted in the Ansys-Fluent code based on the aspect ratio. For the developed numerical grid, the minimum value of the aspect ratio was aR = 21.07, which means that the numerical grid was developed properly, according to [27] recommendations.
Moreover, in order to ensure that the numerical grid was correct, the influence of the mesh density was considered. Figure 21 illustrates the influence of the mesh refinement study on the temperature results in the numerical model. The fluid temperature was measured at the outlet of the reactor model. The results of the four mesh densities were compared in Table 3. It can be observed in Table 3 that the coarse and normal mesh forecast less accurate temperatures, but the fine and very fine mesh forecast similar results. In this case, it was decided that the mesh in the numerical model will contain above the 1,000,000 computational cells, according to the convergence study results. 500,000 1,000,000 1,500,000 2,000,000

CFD Method Assumptions
The quantitative description of the reactive flow depends on the nature of the flow as well as the kinetics of the reaction. Therefore, an essential condition for obtaining an unambiguous solution is to precisely specify the conditions of uniqueness, i.e., the characteristic features specific to a given process and object. The basic conditions of uniqueness, which is characteristic for the gasification process, is to consider four pieces of information: the scope of the interpretation of the transport equations and the accompanying models of chemical reactions, the initial and boundary conditions, and the state equation. By approaching the modeling of the coal gasification process, in terms of computational fluid dynamics (CFD) in the description of the process gas flowing, it is very important to show its properties in the form of the distribution of changes in parameters, such as temperature, pressure, and chemical composition in the reaction space with a given geometry and time interval. The modeling of the transfer process using CFD methods consists of solving a system of differential equations, describing the principle of the conservation of mass, momentum, and energy, as well as the equations of transporting fluid components, together with chemical reactions and state equations [28].
The basic equations describing the behavior of the flowing fluid along the gasification reactor in the Ansys-Fluent software are expressed by the following relationships [2,25,27,28]:  Mass conservation equation:  The source term Sp in Equation (2) is the additional mass due to the devolatilization of coal momentum conservation equation: :  Energy conservation equation: : The source term Sh in Equation (5) is the source for reaction heat.  Chemical reaction conservation equation: : The source terms such as Sh, Ri, and ⃗ in Equation (6) appear because of the gradients in temperature and concentration. In Table 4, the source terms are listed.  Table 4, , is the mass flow rate of volatile mater and pure steam, ∆ ℃ is the heat of reaction of coal gasification, M , is the molecular weight of species, R , is the Arrhenius molar rate, s is the turbulent Schmidt number (sct = 0.7 [27]).
For the purpose of modeling the coal gasification process in the Ansys-Fluent software related to the gasification process, the most common CFD turbulence model was used to determine the viscosity of μt turbulence using the kinetic energy of vortices and dissipation rate (loss of fluid energy over time as a result of friction or turbulence) referred to as the k-ε turbulence model.
The model of viscosity of turbulence μt is expressed in the following equation [27]: The most common forms of transport equation for the kinetic energy of turbulence k and dissipation ε have the form [27]:  For kinetic energy of turbulence The k-ε turbulence model has been adopted in many previous literatures for numerical modeling of the coal gasification process [5,6,11,12]. In additional, the pressure-based solver was used, where the flow equations and species were solved using the second-order scheme as a spatial discretization. The SIMPLE algorithm procedure was used to solve the Navier-Stokes equations. The second order implicit was used as a transient formulation.

Gasification Reaction
The following set of solid and gas phase reactions were adopted [2,[4][5][6]8,17,21,25,29]: where: v-volatile matter. Equation (10) interprets the coal-drying process, while expressions (11) and (12) describe the pyrolysis process. Equation (13) ÷ (16) describe the solid phase reaction. Equation (17) ÷ (18) describes the gas phase reaction. The kinetics of the gasification process was defined by the following equation [27]: where A is a constant for each chemical reaction, T is the absolute temperature of the reaction, E is the activation energy, and R is the universal gas constant. The parameter values used in the calculation of the gasification reaction kinetics are listed in Table 5. system for the CH4, CO2, CO, and H2 gas composition was located at the reactor outlet, as shown in Figure 22. The value of the thermal conduction coefficient-0.9 (W m −1 K −1 );  The chemical composition of coal samples adopted in the simulation are shown in Table 6. Other physicochemical parameters for gasification reaction products were taken from the implemented internal database in the Ansys-Fluent software.
The following system settings have been considered in the Ansys-Fluent software, namely:  Transient state was considered;  Gasification pressure-101,325 (Pa);  Turbulence model-k-ε;  The fluid used in the calculation is water vapor;  The heat transfer coefficient of the gas mixture-0.0454 (W m −1 K −1 );  The timescale of the phenomenon-1040 s (hard coal mine A), 704 s (hard coal mine B) and 545 s (hard coal mine C);  The roughness of the gasification channel-0.1 (m);  Time step-0.25 s;  The convergence of calculations-1 ⸱ 10 −4 .

Results
In Table 7, and the corresponding figures, the values of syngas compositions including CH4, CO2, CO and H2 obtained from the numerical calculations. Figure 22 presents the composition of synthetic gas including CH4, H2, CO2, and CO obtained from the CFD simulation and the ex-situ experiment of the base case.  Table 7 and Figure 23, it can be observed that the experimental and the simulated results were in good agreement as well as showed a similar tendency. Although, it can be observed that there were small deviations. In the case of the hard coal mine A, H2, CH4, CO were the three gases, whose percentages were less than 20%, but CO2 was the gas, which percentages was more than 17% for the numerical simulation and experiment. In the case of the coal sample from the hard coal mine B, CH4, and H2 were the two gases, whose percentages were less than 50% (CH4) and 10% (H2), but CO2 and CO were the two gases whose percentages were more than 20%for the numerical simulation and experiment. In the case of the hard coal mine C, CH4 and H2 were the two gases whose percentages were less than 40% (CH4) and 30% (H2), but CO2 and CO were the two gases whose the percentages were more than 20% (CO2) and 50% (CO) for the numerical simulation and experiment.     The tables below present the results of relative error analysis between the average process gas concentration, including CH4, CO2, CO and H2 obtained during the in-situ experiment and the modeled values obtained from the numerical simulations (listed in Table 7).
It can be observed that a satisfactory relative error was obtained for the results summarized in Table 8 for the coal sample from the hard coal mine A. In the case of the results summarized in Tables 9 and 10, the significant discrepancies for the concentration of H2, CO, and CH4 were observed, which prove that the constants in the kinetic of chemical reactions listed in Table 5 should be further examined.

Modeling Discussion
The discrepancies between the numerical simulated and the ex-situ experimental results are caused by many parameters. The coal gasification process is a very complicated technology. For example, the coal, it is a heterogeneous material with potential minerals and rock fractions impurities. Moreover, the permeability and porosity variations of the coal, which cannot be monitored experimentally, were described by a semi-empirical relationship in order to minimize the complexity in the real ex-situ situation. The phenomena such as the ash accumulation, cracking, and spalling of coal as well as existed for the hard coal, drying and evaporation of moisture, have been the beyond the ability of the numerical model. Due to a wide range of parameters reported by various scholars [4,17,21,25], the kinetic parameters of the reactions might have a huge impact on the overall numerical calculation results.
However, the main mechanism and phenomena of the allothermal gasification process for coal were well captured by the numerical model using the CFD method.

Conclusions
The paper presents a laboratory rig that enables the testing of the allothermal coal gasification process on samples in the form of a cylinder, as well as the development of a numerical model using computational fluid dynamics (CFD) methods. The Ansys-Fluent software was selected for the analysis of the technological parameters of the coal gasification process. The parameter values obtained were referred to the results obtained, based on the mathematical model, and the results of measurements in the experimental reactor. The linear correlation coefficient was used as an indicator of the model fitting to the experimental results. Based on the determined correlation coefficient, satisfactory (in most cases) compliance of the developed models with the experimental result was found.
The experimental results and the numerical simulations allowed to formulate the following conclusions:  The simulated composition of syngas were in good agreement with the experimental results showing a similar tendency;  The fundamental goal of the work was to develop a laboratory rig for the gasification of coal in the form of a solid rock in order to verify the numerical model;  The use of the CFD method allowed presenting the process of coal gasification in a quantitative manner;  The exact purposes of using the proposed model was for the possibility of predicting and investigating the process of UCG under the atmospheric pressure;  The advantages of the work involved applying a thermal-hydraulic-chemical model using the CFD method to simulate the effectiveness of the coal gasification process for hydrogen production under the atmospheric condition. Moreover, parameters, such as permeability, porosity, and solid loss, etc., which cannot be achieved by experiments, is able to predict by the numerical model;  The CFD model can be further developed to get closer to the actual situation aided by the advance of experimental monitoring;  The demonstration of significant interdependencies between the analyzed variables obtained from the numerical simulations and the experimental results confirm the adequacy of the assumptions made for the numerical model of the coal gasification process;  The CFD method proved to be a useful numerical method for enabling the development of numerical models of the coal gasification process with the use of steam as the gasification agent;  The numerical results provide opportunities for easy interpretation of phenomena during the UCG process;  The design of the coal gasification process using the CFD method can shorten the time and reduce the costs needed to develop a potential scenario for the underground coal gasification process to take place under specific geological and mining conditions.  The rate of mass creation by addition from the dispersed phase sources, kg m −3 s −1 .