Power Quality Disturbance Detection and Classification Techniques

Power quality disturbances encompass a range of deviations from ideal voltage or current waveforms in electrical networks. Detection and classification techniques aim to identify events such as sags, swells, transients, harmonics and interruptions, and to assign each event to its proper category for remedial action. Traditional methods rely on time-domain analysis and Fourier-based signal processing, while more recent approaches incorporate time–frequency tools such as wavelet and Stockwell transforms to capture non-stationary characteristics. Machine learning techniques have become prevalent, using feature extraction and selection to train classifiers—from support vector machines to deep convolutional neural networks—that automate recognition with high accuracy. Synchronized measurement devices and phasor measurement units offer enhanced data resolution, enabling real-time monitoring in smart grids and renewable-integrated systems. Emerging trends include hybrid models combining expert rules with data-driven algorithms, semi-supervised learning for scarce labelled data, and explainable artificial intelligence to improve transparency. Together, these advances support robust, real-time disturbance management across evolving power systems worldwide.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Power Quality Disturbance Detection and Classification Techniques publication trend

The graph below shows the total number of articles in power quality disturbance detection and classification techniques across all publications each year (not limited to Nature Index journals).

Technical terms

Power Quality Disturbance (PQD): Any deviation from ideal sinusoidal voltage or current that may impair equipment performance.

Voltage Sag: A brief reduction in RMS voltage, often caused by short-circuit faults or motor startups.

Wavelet Transform: A time–frequency analysis technique that decomposes signals into scaled and shifted wavelets to detect transient features.

Convolutional Neural Network (CNN): A deep learning architecture that automatically extracts hierarchical features from input data, often used for image-based representations of signals.

Feature Extraction: The process of transforming raw measurement data into metrics (such as energy, entropy or statistical moments) that are informative for classification.

References

1. Power quality monitoring in electric grid integrating offshore wind energy: A review. Renewable and Sustainable Energy Reviews (2024).
2. A systematic review of real-time detection and classification of power quality disturbances. Protection and Control of Modern Power Systems (2023).
3. Deep learning for power quality. Electric Power Systems Research (2023).