Neural network prediction of the Fischer-Tropsch synthesis of natural gas with Co (III)/Al2O3 catalyst

Abstract

Application of Co (III)/Al₂O₃ catalyst in Fischer-Tropsch synthesis (FTS) was studied in a wide range of synthesis gas conversions and compared with ANN Simulation results. Present study applies Neural Network model to predict composition of CH₄, CO₂ and CO of the FischerTropsch Process of Natural Gas, while the input vector was 4-dimension vector including four variables from operating pressure, operating temperature, time and ratio of CO/H₂ of 70 different experiments and the output were composition of CO₂, CO and CH₄. The MLP algorithm has been applied for the training and the test set was applied to evaluate the performance of the system including R2, MAE, MSE and RMSE. The results exposed that the predicted values from the model were in good agreement with the experimental data. The paper indicates how Neural Network, as a promising predicting technique, would be effectively used for FTS.

DOI: http://dx.doi.org/10.3329/cerb.v17i1.22915

Chemical Engineering Research Bulletin 17(2015) 25-33