Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cascading Blackout Severity Prediction with Statistically-Augmented Graph Neural Networks

Joe Gorka, Tim Hsu, Wenting Li, Yury Maximov, Line Roald

TL;DR

This work adds to the growing literature of faster graph-neural-network (GNN)-based techniques, developing two novel techniques for the estimation of blackout magnitude from initial grid conditions and proposing a method for facilitating non-local message passing in GNN models.

Abstract

Higher variability in grid conditions, resulting from growing renewable penetration and increased incidence of extreme weather events, has increased the difficulty of screening for scenarios that may lead to catastrophic cascading failures. Traditional power-flow-based tools for assessing cascading blackout risk are too slow to properly explore the space of possible failures and load/generation patterns. We add to the growing literature of faster graph-neural-network (GNN)-based techniques, developing two novel techniques for the estimation of blackout magnitude from initial grid conditions. First we propose several methods for employing an initial classification step to filter out safe "non blackout" scenarios prior to magnitude estimation. Second, using insights from the statistical properties of cascading blackouts, we propose a method for facilitating non-local message passing in our GNN models. We validate these two approaches on a large simulated dataset, and show the potential of both to increase blackout size estimation performance.

Cascading Blackout Severity Prediction with Statistically-Augmented Graph Neural Networks

TL;DR

This work adds to the growing literature of faster graph-neural-network (GNN)-based techniques, developing two novel techniques for the estimation of blackout magnitude from initial grid conditions and proposing a method for facilitating non-local message passing in GNN models.

Abstract

Higher variability in grid conditions, resulting from growing renewable penetration and increased incidence of extreme weather events, has increased the difficulty of screening for scenarios that may lead to catastrophic cascading failures. Traditional power-flow-based tools for assessing cascading blackout risk are too slow to properly explore the space of possible failures and load/generation patterns. We add to the growing literature of faster graph-neural-network (GNN)-based techniques, developing two novel techniques for the estimation of blackout magnitude from initial grid conditions. First we propose several methods for employing an initial classification step to filter out safe "non blackout" scenarios prior to magnitude estimation. Second, using insights from the statistical properties of cascading blackouts, we propose a method for facilitating non-local message passing in our GNN models. We validate these two approaches on a large simulated dataset, and show the potential of both to increase blackout size estimation performance.
Paper Structure (33 sections, 5 equations, 6 figures, 5 tables)

This paper contains 33 sections, 5 equations, 6 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related work
  3. Contributions:
  4. Predicting Blackout Size
  5. Regression, Classification, Both?
  6. Which GNN Input Topology?
  7. Modeling Methodology
  8. GNN Regressor Architecture
  9. Classification Model
  10. Statistical Topology Augmentation
  11. Model Variants
  12. Regress (R)
  13. Regress on Augmented Topology (R+)
  14. Classify then Regress (CR)
  15. Classify then Regress on Augmented Topology (CR+)
  16. ...and 18 more sections

Figures (6)

  • Figure 1: Augmented topology for the RTS-GMLC Test System. Green Buses and Black Lines represent physical power system topology, while red lines represent additional statistical lines.
  • Figure 2: R, CR, and CVR Model Diagrams
  • Figure 3: Diagram of the DCIMSEP Cascading Blackout Simulator, adapted from Eppstein_Hines_2012.
  • Figure 4: Histogram of Blackout Sizes (Zeroes Excluded)
  • Figure 5: R+ and CR+ (Perfect Classifier) Parity Plots. The blue dots represent the predicted vs true value for all blackout samples, while the black dots and error bars represent the error statistics across a range of points (indicated by shading in blue or pink).
  • ...and 1 more figures