Distributed Risk-Constrained Learning-Based Controls: Designs Focusing on Inverter-Dominated Grid Operations
Recent advances in data-driven methods have provided significant advantages for solving complex optimal control problems. However, physical and engineering complex dynamical systems involve unforeseen uncertainties and constraints on control implementations, leading to predominantly distributed requirements. We will describe some recent developments in learning risk-constrained linear quadratic regulator-based optimal controllers that also encompass structural constraints on the controller gains targeting networked dynamical behaviors. The algorithms utilize zero-order policy gradients and provide guarantees to convergence to a stationary point with high probability. The designs are adapted for power electronics-dominated grid operations with high variability to improve the oscillatory performances. The distributed aspects of the inverter-dominated power system control stem from multiple factors, including constraints on the information exchange graph and underlying dynamic structural impacts. The talk will highlight examples including risk-aware controllers for grid-forming inverters (GFMs) aimed at minimizing large frequency oscillations, as well as clustering structure–aware, distributed learning–based approaches designed to damp sustained oscillations across wide-area inverter-dominated power systems.
Dr. Sayak Mukherjee
Senior Staff Scientist, Pacific Northwest National Laboratory (PNNL) on February 28, 2025 at 10:15 AM in EB2 1231
Join Zoom Webinar
Sayak Mukherjee is a Senior Staff Scientist in the Optimization and Control group at the Pacific Northwest National Laboratory (PNNL). He received Ph.D. in Electrical Engineering from North Carolina State University, USA, in 2020 and B.E.E. from Jadavpur University, India, in 2015. He joined PNNL as a Post-doctoral Research Associate. He has received outstanding performance awards at PNNL. His research interests and areas of expertise include system theory, control, reinforcement learning, adaptive dynamic programming, AI/ML for dynamics, resilient control designs, large-scale power system stability and control, grid operation with distributed energy resources with power electronics, and wind energy systems.
This lecture series features exciting and dynamic visiting and virtual speakers from across the range of ECE disciplines. Take some time every Friday morning to be inspired by these great scientists and engineers before heading into the weekend!