Deep RL-Enabled Inverters: Strengthening RES Integration in Grids with Electric Arc Furnaces

Paper in proceeding, 2023

This research work presents the application of Deep Deterministic Policy Gradient (DDPG), a Reinforcement Learning (RL) approach, for grid modeling, voltage and phase angle estimation, and control of grid-supporting inverters. The goal is to develop a grid-supporting inverter that produces virtual inertia, stabilizing the grid frequency problems from heavy loads such as electric arc furnaces and enabling seamless integration of renewable energy sources (RES). By employing DDPG, we eliminate the need for multiple estimation tools, such as Fast Fourier Transform (FFT), Synchronous Reference Frame (SRF), or lowpass filters, typically used in traditional methods for determining controller loop setpoints. Moreover, the method reduces the necessity for extensive tunings of parameters, like PID controller coefficients and lowpass filter settings. Findings indicate that the DDPG-based control system offers a fast and efficient control mechanism for maintaining the frequency stability of power systems. By improving grid stability, this innovative approach can play a pivotal role in promoting the widespread adoption of RES.