Coordinated Control System (CCS) is an important topic on the power plantautomatic control. It is a highly complicated multi-variable control system. It has somecharacters such as nonlinear, parameters time varying, large delay and severe coupling.The control strategy of traditional coordinated control system is far from satisfying withthe request of the electric network. We design a method on decoupling when thedynamic characteristic of the unit power set have been analyzed. Based on neuralnetwork theory, a neural network PID control algorithm with model reference using BPneural network has been brought forward. With the case of the mathematic-model of thehomemade 300MW unit power set, its good control quality has been validated by a lot ofsimulation experiments and comparing it with traditional PID control algorithm. Thepaper supplied a new thinking to research the nonlinear, time-variable andmulti-variables control system and have certain theory meaning and project using value. |