Neural Network-Based Model Predictive Control of a Class of Methanol Supply System

In this paper, a neural network-based model predictive control protocol is designed to regulate the inlet pressure of main engine for a class of methanol supply system. A radial basis function neural network model is firstly trained to approximate the methanol supply system. The training data is generated by methanol supply system design institute through dynamic simulation under different working conditions by process simulation software. Then, the inlet pressure controller of the main engine of the methanol supply system is designed by combining the model predictive control theory. The effectiveness of the controller is verified by the simulation model of the supply system. The simulation results show that the designed controller can ensure the inlet pressure of main engine satisfies the requirement under working conditions.