Risk-Aware Power Systems Operations

Abstract

Over the past decade, the soaring growth of renewable generation like wind and solar, and of distributed energy resources such as electric vehicles and rooftop solar generation, have triggered a paradigm shift in the operations of power grids. The resulting increase in operational uncertainty, both on the demand and the supply side, calls for new operational practices that explicitly quantify and manage risk. In this talk, I will present a new generation of tools that address this challenge, via a combination of optimization and machine-learning techniques, namely, probabilistic forecasting, optimization under uncertainty, principled risk assesment, and machine learning-based acceleration techniques. The presentation will focus on stochastic optimization, and ML-based optimization proxies for fast risk assessment. Computational results will be reported on a real industrial system.

Mathieu Tanneau

Research Engineer

I’m interested in mixed-integer linear and nonlinear optimization, power systems, and the integration of machine-learning techniques in optimization algorithms.