Generating adversarial inputs for a graph neural network model of AC power flow

TL;DR

This paper develops optimization techniques to generate adversarial inputs for a graph neural network modeling AC power flow, revealing vulnerabilities and motivating robust training and verification methods.

Contribution

It introduces a method to generate high-error adversarial inputs for GNN-based AC power flow models, highlighting the need for improved robustness.

Findings

Adversarial points cause errors up to 3.4 per-unit in reactive power.

Minimal perturbations of 0.04 per-unit can satisfy adversarial constraints.

Generated adversarial inputs demonstrate vulnerabilities in the neural network model.

Abstract

This work formulates and solves optimization problems to generate input points that yield high errors between a neural network's predicted AC power flow solution and solutions to the AC power flow equations. We demonstrate this capability on an instance of the CANOS-PF graph neural network model, as implemented by the PF$\Delta$ benchmark library, operating on a 14-bus test grid. Generated adversarial points yield errors as large as 3.4 per-unit in reactive power and 0.08 per-unit in voltage magnitude. When minimizing the perturbation from a training point necessary to satisfy adversarial constraints, we find that the constraints can be met with as little as an 0.04 per-unit perturbation in voltage magnitude on a single bus. This work motivates the development of rigorous verification and robust training methods for neural network surrogate models of AC power flow.