Neural network adaptive control for a class of nonlinear systems based on hyperstability

This paper discussed the neural network in the application of nonlinear system adaptive control. For a typical class of nonlinear system, A BP neural network is designed to control the system based on hyperstability. The nonlinear link was eliminated by introducing its inverse model, for transforming the nonlinearity into a linear system. The neural network controllerpsilas purpose is to optimize the effect on tracing velocity and make the output of controlled object closing to the one of reference model. Satisfactory results have been achieved by using MATLAB simulation on the model for three-phase electrode regulator system of arc furnace. The simulation and analysis shows that neural network adaptive control system is more effective than the traditional PID system, and guarantees the system has a better dynamic quality.