Early Anomaly-Onset Detection based on Wigner--Ville Distribution Slice Spectra: A Transmission-Grid Test Case

2026-06-14 · Source: Machine Learning · Field: Energy & Utilities — Utilities & Infrastructure, Energy Storage & Grid Technology, Artificial Intelligence & Machine Learning · Depth: Expert, quick

Summary

A study evaluates full-vector Wigner--Ville Distribution Slice (WVDS) spectra for early anomaly-onset detection in high-voltage grid-voltage waveforms, addressing the need for real-time operational disturbance monitoring. This approach represents each 128-sample voltage window as a 128-dimensional slice spectrum, preserving the Wigner--Ville distribution's bilinear midpoint interaction structure and eliminating the need for manual fault-frequency markers. The WVDS method, combined with a baseline-normalized deviation (BND) score, was compared against FFT-BND and various autoencoder configurations (raw-window, FFT, WVDS) using a synthetic autoencoder-clustering teacher on RTE fault records. While FFT-BND showed higher sensitivity, WVDS-BND achieved the lowest false-alarm operating point, reducing record-level pre-onset false alarms to 0.69%. The findings suggest that preserved WVD cross-term information offers a selective representation for online grid-waveform anomaly monitoring, particularly where false alarms are costly.

Key takeaway

For Machine Learning Engineers developing real-time power grid anomaly detection systems, consider Wigner--Ville Distribution Slice (WVDS) spectra. If minimizing false alarms in high-voltage waveform monitoring is your priority, WVDS-BND significantly reduces pre-onset false alarms to 0.69%. This outperforms FFT-BND in selectivity. This approach is crucial where false positive costs are high. It guides your choice towards robust, potentially less sensitive, detection mechanisms.

Key insights

WVDS spectra offer a selective, low false-alarm method for early power grid anomaly detection, leveraging WVD cross-term information.

Principles

• WVD cross-term information enhances anomaly selectivity.
• Real-time monitoring benefits from sequential window analysis.
• Reducing false alarms is critical when costs are high.

Method

Apply full-vector WVDS to 128-sample voltage windows, generating 128-dimensional slice spectra. Calculate a baseline-normalized deviation (BND) score for sequential anomaly-onset detection.

In practice

• Implement WVDS-BND for high-voltage grid monitoring.
• Prioritize WVDS where false alarm costs are significant.
• Use WVDS autoencoders for reduced reconstruction false alarms.

Topics

• Wigner-Ville Distribution
• Anomaly Detection
• Power Grid Monitoring
• High-Voltage Waveforms
• False Alarm Reduction

Best for: AI Scientist, Research Scientist, Machine Learning Engineer

Related on AIssential

• Distribution-Free Stochastic Analysis and Robust Multilevel Vector Field Anomaly Detection — A distribution-free stochastic functional analysis method detects anomalies in vector fields using covariance and multilevel orthogonal subspaces.
• Navigating the Safety-Fidelity Trade-off: Massive-Variate Time Series Forecasting for Power Systems via Probabilistic Scenarios — Large-scale power system forecasting requires new benchmarks and models to balance safety and fidelity.
• I Predicted the T20 World Cup 2026 Final With an Anomaly Detection Pipeline. — Anomaly detection system failures reveal more about production challenges than successes, especially regarding evidence thresholds and cold-start problems.
• Physics-Guided Dual Decoding and Spectral Supervision for Global 3D Hydrometeor Prediction — PredHydro-Net uses physics-guided dual decoding and spectral supervision to improve global 3D hydrometeor prediction, overcoming smoothing issues.
• CRAFTIIF: Cross-Resolution Analytic Four-Type Interpretable Isolation Forest for Multivariate Time Series Anomaly Detection — CRAFTIIF addresses four distinct anomaly types in multivariate time series using specialized wavelet features and Isolation Forests for direct attribution.
• PhaseNet++: Phase-Aware Frequency-Domain Anomaly Detection for Industrial Control Systems via Phase Coherence Graphs — Phase information, often discarded, offers a complementary and powerful modality for ICS anomaly detection.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning.