Dynamic Droop Setting of Smart Inverters in Active Distribution Grid Using Neural Networks

Increasing penetration of Smart Inverters (SIs) will shift the voltage control functionalities to the grid edge. With the advanced control mechanism and capabilities of SIs to dynamically adjust Volt-Watt and Volt-VAr droops, better and granular voltage regulation can be achieved on distribution feeders. In this context, this paper discusses the application of Neural Networks (NNs) through which effective setting of SIs' droops could be obtained leading to better voltage regulation on distribution feeders. Case studies on the IEEE 123-node feeder shows effective voltage regulation from SIs with droop settings obtained from the NNs.