IDENTIFICATION OF GENERATING PROCESS BASED ON COMBINED ERROR FEEDBACK ARTIFICIAL NEURAL NETWORK

This paper mainly researched the application of artificial neural network in real industrial systems. Because the object is a dynamic system, so we discussed the dynamic characteristics of artificial neural network used in identifying industrial process specially. With the background of power generating process, considering the characteristic of different bear, the paper proposed a novel method to identify the model that used sub network to identify the mathematics model of generating process in different loads. Considering the stability of the system, this paper introduced a new method to train neural network models with error feedback.