Feature Selection for Fault Prediction in Distribution Systems

TL;DR

This paper introduces a simulation-based feature selection method for fault prediction in power distribution systems, significantly improving prediction accuracy and outperforming traditional feature sets.

Contribution

It proposes a surrogate task using simulation data for effective feature selection, addressing data scarcity and enhancing fault prediction performance.

Findings

Strong correlation (r=0.92) between surrogate task and real-world performance

Selected 374 optimal features from 1556 candidates

Achieved F1-score of 0.80 in case study, outperforming baselines

Abstract

While conventional power system protection isolates faulty components only after a fault has occurred, fault prediction approaches try to detect faults before they can cause significant damage. Although initial studies have demonstrated successful proofs of concept, development is hindered by scarce field data and ineffective feature selection. To address these limitations, this paper proposes a surrogate task that uses simulation data for feature selection. This task exhibits a strong correlation (r = 0.92) with real-world fault prediction performance. We generate a large dataset containing 20000 simulations with 34 event classes and diverse grid configurations. From 1556 candidate features, we identify 374 optimal features. A case study on three substations demonstrates the effectiveness of the selected features, achieving an F1-score of 0.80 and outperforming baseline approaches that use frequency-domain and wavelet-based features.