Modified Droop Strategy for Wide Load Range Efficiency Improvement of Parallel Inverter Systems

Parallel inverters are used in many modern applications, and thus, improving the inverter system efficiency plays a key role in energy savings. The conventional droop strategy used for power sharing among inverters, however, leads to a low efficiency especially at light loads, as the low power demand is divided among inverters, forcing them to process a fraction of the low power at a low efficiency according to their efficiency curve. To avoid such operating conditions, a communicationless modified droop strategy is proposed in this article to select an optimal number of inverters to process fractions of the power demand that leads to a higher system efficiency considering the efficiency curve of the inverters. To achieve this objective at very light load situations, an online-inverter detection method is developed so that each inverter detects the online inverters and the unnecessary inverters turn off. The proposed method is employed in a system with three single-phase parallel inverters to evaluate the effectiveness of the method. It is observed that the proposed strategy can improve the system efficiency by up to 14% at light loads compared with the conventional droop. Additionally, the reliability of the system is enhanced by extending the lifetime of inverters with higher power ratings, which are considered as valuable assets of the system. Detailed derivations, simulations, and experimental results are presented to validate the proposed method.

Automated summary

This article proposes an improved droop strategy to optimize the efficiency of parallel inverters. The strategy selects an optimal number of inverters based on their efficiency curve to process fractions of the power demand, and introduces an online inverter detection method to deal with light load situations. Experimental results demonstrate that this method can significantly improve system efficiency and enhance system reliability under light load conditions.