Fuzzy Neural Network Control Method Of The Full-bridge Resonant Dc-dc Converter

This paper presents the studies of Fuzzy-Neural Network(FNN) control methods in the application of the full-bridge series type resonant DC-DC converter.There are many types Fuzzy-Neural Network at present. But in this paper, a new type of FNN structure is presented which is in form of series. The new FNN controller not only has the fuzzy controller's characters of simple structure and easily-used, but also has the self-study ability of the Elman network. Thus it is valuable for research for non-linear system .In this paper, it summarizes the development of DC-DC converter firstly, and presents a new idea which is using Fuzzy-Neural network in DC-DC converter. Then it induces the main circuit's theoretic state equation and attains all kinds of working modes. Secondly through the introduction to the fuzzy control and neural network, the new type of FNN is presented. This structure is kind of outer-coalescent. It adopts neural network's clustering function to classify the system output, then link a fuzzy controller in form of series which rules are made through the classified result.The design process and the experiment shows the well control effort to the full-bridge series type resonant DC-DC converter, at the same time the whole process avoids not only the modulating parameter process of the traditional method , but also the complexity of the traditional FNN design.Considering the application of the fuzzy controller and the neural network in practice, the structure which is presented in this paper can be divided into two parts which include the fuzzy controller design and the neural network classifier design respectively. Accordingly this research in the paper present a advantage feasibility scheme for the application of FNN in practice.