Neural network approach for linearizing control of nonlinear process plants

Abstract

The application of a feedback linearization technique using artificial neural networks (ANNs) for a nonlinear industrial process plant is considered in this paper. The process plant is modeled first using an ANN, and then the dynamic neural network model acting as a process plant emulator is feedback linearized. A novel configuration for linearization of an ANN emulator using only backpropagation is used. Effective control of the linearized emulator is then exhibited using a linear controller. Experimentation and simulation results on the linearized emulator are presented to demonstrate the effectiveness of the feedback linearization technique.