PI_BPNN controller for transient response improvement of LLC resonant converter

This paper proposes the proportional integral (PI) backpropagation neural network (BPNN) algorithm to improve the transient voltage control performance in an LLC resonant converter whose voltage gain is highly nonlinear. In the proposed method, the PI linear controller roughly regulates the operating frequency of the LLC converter according to the voltage gain. Moreover, the BPNN nonlinear controller compensates for the frequency in the transient and steady states. It also improves the dynamics. Compared with traditional artificial neural network algorithms, the proposed BPNN can reduce the computational burden of the digital controller and improve the accuracy of the time-series prediction by reflecting previously predicted results in the learning process. Hence, the proposed method is suitable for real-time applications, such as power electronic converters. The structure and the learning procedure of the proposed PI-BPNN controller are explained in detail, together with the modeling and the control scheme of the LLC converter. A 5.5 kW LLC converter prototype is built and tested to verify the performance of the proposed method. The load step experimental results show that compared with the traditional PI control algorithm, the proposed PI-BPNN reduces the transient response time by up to 25%.