Neural networks internal model control for water level of boiler drum in power station

Abstract

Water level of boiler drum in power station is one of the main control parameters for turbine-generator unit, which reflects the balance of boiler load and feed-water indirectly. Aiming at the water level control in power station, an internal model control based on neural networks is presented in this paper. It adopts the steam flux signal to the internal model controller considering the influence of load changing, which has the ability of feed-forward compensation for steam flux disturbance and can avoid "false water level" phenomenon. At the same time, the whole input signal of controlled object is combined by NNC and a robust controller (RC), which make up of a changing robust controller to control the object. Thus, good regulating performance is guaranteed in the initial control stage, even when the controlled object varies later. Simulation results show that this strategy is more effective and practicable compared with general cascade PID control.