Adaptive Neuro-control System for Superheated Steam Temperature of Power Plant over Wide Range Operation

In this paper, the dynamics of a superheated steam temperature process in a 600MW supercritical once-through boiler is first analyzed. The dynamics is influenced by the operating conditions represented by steam flow and steam pressure mainly. An intelligent cascade control system is then designed for controlling superheated steam temperature over wide range conditions. The parameter of the proportional controller in the inner loop is fixed all the time. The controller in the main loop is a self-tuning PID neuro-controller whose parameters are adjusted by gradient algorithm, where the Jacobian information of the plant is obtained by RBF neural networks. Finally, the designed control system is applied to control the superheated steam temperature in a 600MW supercritical once-through boiler, and the simulation indicates it has satisfactory performances over wide range