%0 Journal Article
%@ 00162361
%A Chai, B.W.L.
%A Foo, H.C.Y.
%A Tan, I.S.
%A Lam, M.K.
%A Lau, L.C.
%D 2021
%F scholars:15190
%I Elsevier Ltd
%J Fuel
%K Analysis of variance (ANOVA); Catalytic oxidation; Chemical analysis; Desulfurization; Errors; Genetic algorithms; Hydrogen sulfide; Kinetics; Least squares approximations; MATLAB; Mean square error; Sensitivity analysis; Sulfur compounds, Advanced adsorption; Detailed chemical kinetic; Normalized sensitivity; Operating variables; Prediction uncertainty; Response surface method; Root mean square errors; Sensitivity coefficient, Data reduction
%R 10.1016/j.fuel.2020.119406
%T Complex chemical kinetic mechanism reduction for simultaneous catalytic oxidation and desulphurization of hydrogen sulphide
%U https://khub.utp.edu.my/scholars/15190/
%V 286
%X Complex chemical kinetic mechanism complicates simulation of advanced adsorption column, particularly in the coupling of simultaneous catalytic oxidation and desulphurization of hydrogen sulphide. In the present study, a new reduction method of the detailed chemical kinetic mechanism was postulated based on genetic algorithm and least square error to solve for both reactions class-based global sensitivity and path sensitivity analyses. During the reduction process, the influence of the species and reactions was determined according to the contribution of their corresponding reaction classes to the prediction uncertainties by calculating the normalized sensitivity index (NSI) and the path sensitivity coefficient (PSC) of each reaction class from the detailed mechanism. Through solving both NSI and PSC with MATLAB coding incorporating genetic algorithm and least square error, a reduced hydrogen sulphide mechanism with 15 species and 7 reactions is obtained. By comparing the calculated value from reduced mechanism with experimental data, 16 useful kinetic parameters were estimated and validated through the root mean square error (RMSE). Good agreements for the predicted data between the reduced and experimental data indicate the advantages of the present reduction method. Furthermore, the validated reduced mechanism also suggests that the exponential-like diffusion coefficient is suitable to explain the behaviour of pore blockage due to solid product formation. Lastly, the interaction between significant operating variables such as temperature, concentration, flow rate, and mass of adsorbent were also studied and presented under the umbrella of response surface method and analysis of variance (ANOVA). © 2020 Elsevier Ltd
%Z cited By 1