International Journal of Energy Research, Vol.43, No.9, 4580-4591, 2019
Simulation and multi-objective optimization of a fixed bed catalytic reactor to produce hydrogen using ethanol steam reforming
A mathematical model is developed for a fixed bed catalytic reactor to produce hydrogen using ethanol steam reforming in the presence of a Ni (II)-Al (III)-LDH catalyst. The simulation of the reactor has been carried out using the ode23s module of MATLAB (version 2010a) based on a mechanistic kinetic model (Langmuir-Hinshelwood approach) with proven reaction kinetics. There is a confusion regarding the values of the kinetic parameters even though earlier workers have used the same experimental data and the same model equations. The values of the model parameters obtained using two different curve-fitting techniques, the generalized reduced gradient (GRG) algorithm and a derivative-free approach based on the simplex method, were different. This leads to differences in the agreement of model predictions vs experimental results. A more powerful and recent technique, genetic algorithm (GA), has been used to resolve this problem by minimizing a sum-of-square errors (SSE). Our tuned model parameters gave good agreement with experimental data. An SSE value of the order of 10(-4) is obtained in the present study over the SSE values of the order of 10(-2) obtained from earlier studies. Using this tuned model, multi-objective optimization (MOO) of an isothermal fixed bed ESR reactor has been carried out using NSGA-II to achieve the maximum hydrogen mole fraction in the product gas while simultaneously minimizing the mole fractions of the greenhouse gases, CO + CO2. The maximum theoretical mole fraction of hydrogen obtained is 0.088 at 911.86 K vs 0.080 at 923 K, as observed experimentally. The more recent jumping gene adaptation of NSGA-II, namely, NSGA-II-JG, was also tried to check if it can give more rapid convergence to the Pareto set. It is found that the present problem is far too simple, and the advantage of the JG adaptation is small.