Neural Network Predictive Control And Application In Power Plant

With the increase of large power units especially super-critical units and difference in electric network many large units need regulate load. It is of higher quality to the auto control system. In the present it is not adaptable to the control system of modern large power units for the normal control system based on PID. So it is necessary to think about a new control strategy. In this article, the Predictive Function Control strategy based on Simple Recurrent Neural Network (SRNNPFC) was developed and applied in a main temperature control system in a power plant. This design strategy was developed for a class of nonlinear systems. In this process the nonlinear objective model is identified by the simple recurrent neural network. Simulation result showed that the proposed SRNNPFC design method is superior to the conventional PID series control, which is more robust and capable of load adaptation and has wide application prospect.