Incomplete differential PID neural network temprature controller for thermal process

The system parameters of thermal process such like heating network have a feature of nonlinearity, time-variant and high time-delay. Using the intelligent control methods of combining neural network theory and incomplete differential PID controlling, adjusting the self-learning of neural network and weighting coefficients, the output of neural network is corresponding to the parameters of incomplete differential PID controller which is under the best control rate, also the system will have better controllability, robustness and adaptability. Matlab simulation shows that the presented control scheme makes the control process about second return water temperature of thermal station into the steady-state quickly, comparing with the conventional incremental PID control performance. Through online learning and auto-update parameters, the response of control system is speeding up, and it can restrain disturbance effectively.