Microgrid Light-Load Efficiency Improvement Based on Online-Inverter Detection

Droop-controlled inverters are widely used in microgrids to supply the load demand. However, due to low light-load efficiency of inverters, the conventional droop control used for proportional power sharing among inverters does not guarantee an optimal system efficiency at light loads. This paper presents a decentralized online-inverter detection (OID) method based on a transient coded frequency deviation such that the OID detects the online inverters in light-load situations and determines the required minimum number of inverters with the lowest rated powers to supply the loads and improve the overall efficiency. As a result, all the selected inverters will operate at higher power levels and thus, at a higher efficiency compared with the case where all of the inverters shared the power and operated at low power levels. To study the effectiveness of the method, the OID is applied to a system with three single-phase parallel inverters, and an efficiency improvement of 8% is obtained at light loads. Moreover, experimental results are provided to evaluate the performance of the OID in practice.