A multi-layer neural network controller for single-phase inverters

Single-phase inverters with varying linear or non-linear loads are widely used in critical applications such as dynamic voltage restorer (DVR) and uninterrupted power supply (UPS) system. Obtaining high performance, such as low total harmonics distortion (THD) and quick transient response is very important in such applications. This paper and its companion paper present two promising control methods based on neural networks. A direct multi-layer neural network (NN) controller is presented in this paper. Only one sensor is used in the proposed scheme. The principles, robustness, and the special features of the control scheme are discussed and the complete system is presented. The proposed NN controller with and without flexible structure was tested for comparison of performance. Simulation results are presented to verify the feasibility and expected performance of proposed approach. Simulation results also demonstrate the superior performance of the proposed controller when compared to a deadbeat controller with only one sensor operating under same conditions. The major advantages of the proposed scheme are that high performance is obtained with low sampling and switching frequency and that only sensing of output voltage is required.