A Full-ANN Control Scheme of Single-Phase Grid-Connected Inverter

This paper presents a full artificial neural network (ANN) control method of single-phase grid-connected inverter, which directly obtains the modulation wave required by PWM for inverter control. Firstly, the principle of inverter circuit is analyzed to select auxiliary variables that affect modulated waves. Secondly, auxiliary variables and echo state network based on gradient descent algorithm (ESN-GD) are used to complete the accurate fitting of the reference modulated wave. At the same time, echo state network based on the least squares algorithm (ESN-LS) is used to obtain the amplitude and phase relationship between the reference modulation wave and the modulation wave corresponding to the reference current. Finally, the modulating wave is obtained by the amplitude and phase transformation of the reference modulating wave. The proposed method is verified by Matlab/Simulink platform with ideal steady-state performance and transient performance.