Energy storage inverter control based on neural network inverse model

Abstract

The output voltage waveform quality of single-phase energy storage inverter is an important measurement index of its performance. In this paper, the mathematical model of single-phase energy storage inverter is analyzed, and its inverse model is established using BP neural network. Combined with a single loop PI controller, two different control methods are proposed based on the inverse model. One method is to take the output of the neural network inverse model as a feed forward, and superimpose it to the output of a single loop PI controller; another method is to cascade the neural network inverse model and its original model to form a pseudo linear system, and then adopt PI controller to perform single loop control. Simulation results show that, comparing with simple single loop PI controller, the two control methods proposed in this paper could effectively improve the dynamic response speed of the inverter output voltage and reduce the harmonic content.