Deep Reinforcement Learning-Based Robust Protection in DER-Rich Distribution Grids
Dongqi Wu, Dileep Kalathil, Miroslav M. Begovic, Kevin Ding, Le Xie
Abstract
This paper introduces a new framework of deep reinforcement learning based protective relay design in power distribution systems with many distributed energy resources (DERs). With increasing penetration of power electronically-interfaced resources, conventional overcurrent relays’ performance is rendered less effective due to the two-way uncertainties in power flow patterns. In this paper, a machine learning-based protective relay that is designed for adaptively deciding the threshold for relay action is proposed. The particular algorithm used is an Long Short-Term Memory (LSTM) enhanced deep neural network that is highly accurate, communication-free and easy to implement. The proposed relay design is tested in OpenDSS simulation on the IEEE 34-node test feeder and a collection of large synthetic feeders in Austin, Texas area. By designing adaptability upfront, the proposed relay is shown to substantially improve the performance of relay in terms of failure rate, robustness, and response speed, in particular in scenarios with high level of distributed energy resources.