Application of a neural network predictive control for the supercritical main steam

The traditional PID control is difficult in the nonlinear, delay, time-varying conditions and have a disturbance characteristics in supercritical main steam temperature control system to achieve satisfactory control effect. This paper presents a neural network predictive control method using multi-step prediction, rolling optimization and feedback correction control strategy, achieved good control results. Taking the supercritical main steam temperature as the research object, MATLAB simulation results show that, in various of the main steam temperature condition neural network dynamic model, both are well predict the dynamic characteristic, and achieved better performances than traditional PID´s.