Research On Fuzzy Neuro Network Control Technology Of Three-phase Grid-connected Inverter

The research and development of new energy has led to the rapid development ofdistributed generation technologies, the control technology of the stand-alone mode andgrid-connected mode of three-phase inverter is becoming more and more important. In thedesign process of traditional controller, it’s necessary to know the mathematical model ofthe controlled object, but it’s difficult to obtain the accurate mathematical model, besides,once the parameters are determined, it’s difficult to reach the performance requirementsunder different working environments, In this paper, a novel control strategy combinedFNN (fuzzy neural network) control and PI control is proposed, and the performances ofthree-phase grid-connected inverter have been improved by using this method.In this article, the models of stand-alone and grid-connected mode of three-phaseinverter have been analyzed, on this basis, PI controller parameters have been designed.Due to the switching state which inverter is based on, it a nonlinear system, besides, theimpact of nonlinear component make it more difficult to establish the accurate model ofthe inverter. Therefore, an intelligent control method unrelated to the objected model isintroduced in this paper. By using fuzzy self-adjust PI control and fuzzy neural networkmethod, the T-S model-based adaptive neural network fuzzy inference system (ANFIS) isselected. Finally, simulation analysis is taken after appling the proposed method inthree-phase grid-connected inverter. Compared the simulation results of PI control, fuzzyPI control FNN control, it’s proved that FNN controller is combined with advantages offuzzy control and neural network control, it can improve the dynamic and staticperformance of the inverter system.To verify the feasibility and correctness of the proposed theory, the experiment hasbeen operated based on the output voltage 110V, output power 2kW three-phase gridinverter platform.