Artificial Neural Networks for Volt/VAR Control of DER Inverters at the Grid Edge

Abstract

With increased grid penetration of distributed energy resources (DERs), volt/VAR control (VVC) is becoming increasingly complex and challenging in distribution systems. Conventional VVC approaches in distribution networks reside typically on the primary feeder. However, these approaches are ineffective for handling the voltage variability caused by DERs and load dynamics on the secondary side, termed as the grid edge. With more and more deployments of DERs at the grid edge, advanced VVC technologies through controlling DER inverters are perceived to be the key enabling technologies for distributed voltage control to meet the future smart grid requirements. Nevertheless, a great challenge is that DER inverters would fight each other when operated autonomously in participating grid voltage control using the conventional control methods. This paper proposes a novel VVC strategy based on an artificial neural network (ANN) control. The proposed approach is able to properly handle DER inverter constraints in achieving VVC objectives at the grid edge and overcomes the challenges of conventional DER inverter control techniques. The behaviors of the proposed method are evaluated via simulation and hardware experiments, which demonstrate great advantages of the ANN-based VVC for controlling DER inverters at the grid edge.