A neural network-based adaptive power-sharing strategy for hybrid frame inverters in a microgrid

The capacitive-coupling inverter (CCI) is more cost-effective in reactive power conditioning and enhanced reactive power regulation ability when compared with the inductive-coupling inverter (ICI). As power conditioning capability is vital for a microgrid (MG) system, a new MG frame with hybrid parallel-connected ICIs and CCIs was proposed in this paper. With lower DC-link voltage for the CCI, an adaptive power sharing method was proposed for reducing total rated power and losses. A power-sharing control layer based on a back-propagation neural network that guarantees rapid and accurate sharing ratio computation was investigated as well. The results of simulations and experiments were used to verify the effectiveness of the proposed method.

Automated summary

A new microgrid (MG) framework with hybrid parallel-connected inductive-coupling inverters (ICI) and capacitive-coupling inverters (CCI) was proposed in this paper, which is more cost-effective in reactive power conditioning and enhanced reactive power regulation ability compared with ICI. An adaptive power sharing method was proposed to reduce total rated power and losses by lowering the DC-link voltage for CCI. A power-sharing control layer based on back-propagation neural network was investigated for rapid and accurate sharing ratio computation. The effectiveness of the proposed method was verified through simulations and experiments.