Decentralized Schemes for Grid-Forming Inverter System Efficiency Improvement and Online-Inverter Detection

Abstract

Parallel grid-forming (GFM) inverters are used in many modern power system applications. Therefore, improving their system efficiency is of paramount importance for energy savings. The droop control method has been conventionally used to share the power among inverters proportional to their power ratings. However, the droop strategy does not guarantee an efficient power sharing especially at light loads, where the low power demand is divided among inverters, forcing them to supply a low power at a low efficiency according to their efficiency curve. Addressing the aforementioned issue while maintaining the advantages of parallel inverters used in residential areas is the focus of this thesis. A nested-loop control system for a single-phase inverter operating in standalone mode is introduced. The control system includes current, voltage, and harmonic control loops, and is designed based on the linear quadratic tracking (LQT) optimal control method. The designed controllers exhibit a satisfactory performance in tracking the reference signals and rejecting the unwanted harmonics in case of linear and nonlinear loads. Moreover, a total harmonic distortion (THD) of 1.65% is obtained when a nonlinear load is connected to the system. A decentralized modified droop strategy is also proposed to improve the system efficiency of parallel grid-forming inverters. The main idea is to control the power sharing among the inverters such that the output power of each inverter is maintained within a proper range with respect to the inverter efficiency curves. Hence, an optimal number of inverters supply the loads efficiently, and the unnecessary inverters do not share any power. In contrast to the existing methods, the proposed method does not have a single point of failure while it features modularity, flexibility, and reliability. 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. Furthermore, a stability verification of the system within the designed operating range of the inverter no-load frequencies is conducted to ensure a stable operation. As in very-light load situations none of the inverters can operate within the proper power range with a high efficiency, an online-inverter detection (OID) method is proposed such that each inverter detects the online inverters and the unnecessary inverters stop injecting power to obtain a higher system efficiency. As the proposed OID is decentralized and modular, more inverters can be added to the system without adjusting the settings of installed inverters. Detailed derivations, simulations, and experimental results are presented to validate the effectiveness of the schemes.