Robust Control of Cascading Power Grid Failures using Stochastic Approximation

TL;DR

This paper introduces a stochastic approximation approach to improve the control of cascading power grid failures, accounting for noise and model errors to enhance system resilience.

Contribution

It develops a rigorous stochastic optimization methodology using Sample Average Approximation for better control of power grid failures.

Findings

Enhanced control strategies accounting for noise.

Improved robustness over deterministic methods.

Potential for reduced system failures.

Abstract

Cascading failure of a power transmission system are initiated by an exogenous event that disable a set of elements (e.g., lines) followed by a sequence of interrelated failures (or more precisely, trips) of overloaded elements caused by the combination of physics of power flows in the changed system topology, and controls. Should this sequence accelerate it can lead to a large system failure with significant loss of load. In previous work we have analyzed deterministic algorithms that in an online fashion (i.e., responding to observed data) selectively shed load so as to minimize the amount of lost load at termination of the cascade. In this work we present a rigorous methodology for incorporating noise and model errors, based on the Sample Average Approximation methodology for stochastic optimization.