The parallel control of single-phase inverters based on BP neural network

In order to improve the output waveform quality and reduce the circulation of current in parallel inverter system, droop control that is combined with double loop control is used, and two BP neural network controllers are proposed to control the designed inverter in this paper. One BP neural network control the outer-loop part of voltage to obtain stable current reference value, and the other BP neural network control the inner-loop part of current to achieve the output voltage with sinusoidal wave. A simulation studies based on Matlab is done, and by comparing the results of different capacities and load conditions based on BP neural network controller. The result shows that the designed parallel inverter system with the BP neural network controller can have a small circulation and quick adjustable speed. So the suggested control strategy can satisfy the requirement of performance and reliability for the parallel inverter system.