Integrated advanced nonlinear neural network-simulink control system for production of bio-methanol from sugar cane bagasse via pyrolysis

Abstract

This study focuses on development of control system for the production of bio-methanol via thermochemical pyrolysis process. The dynamic control system was developed and simulated using Matlab®. The sustainability concern of bio-methanol involved the integration of NN-Simulink control strategy applied with an advanced controller in NN control toolbox. Model reference control in dynamic study of nonlinear bio-methanol production regarding the process design and control parameters was employed. The significant parameters that influence the process are namely the reaction time, reaction temperature and nitrogen flow operated in pyrolysis batch reactor. The objective is emphasised for maintaining high yield of bio-methanol production in pyrolysis with set point constant values at optimum conditions from experimental studies. Thus, the development of model reference control with neural controller has attained a better control in manipulating of reaction temperature for pyrolysis batch reactor. The predicted data from ANN fitting model in Matlab® was determined with validation, R2 achieved at 0.98 for nonlinear quadratic model. The bio-methanol production yield attained 3.09 wt.% in prediction value. MSE error was indicated at small difference of 0.2617 in validation result. Therefore, the non-linearity of regulating input temperature to linearity of measured output of bio-methanol yield was achieved in this study.