Partial Discharge Signals From Covered Conductors Across 23 Stations

Partial discharge signals from transmission power lines collected using a contact galvanic method. The dataset contains over a terabyte of data from real environments across 23 different stations. It was created by Lukáš Klein and was last updated in April 2026.

Use Cases

• Classify fault scenarios into one of seven imbalanced classes using extracted signal features.
• Detect contacts with and without ground from processed voltage-based signals.
• Identify conductor breakdown events from the documented wide array of fault scenarios.
• Develop anomaly detection models for partial discharge behaviors using real-world signal data.

Strengths

• Over a terabyte of data from real-world environments.
• Signals collected from 23 different stations.
• Detailed labeling across seven fault classes.

Limitations

• Class distribution is described as imbalanced.
• Only 3% of the data was previously available, suggesting the full scope is newly released and may have unknown characteristics.
• Specific sample size, row count, and file formats are not provided.

Provenance

Source

Lukáš Klein, Partial Discharge Detection In Overhead Power Distribution Lines.

Collection Method

Collected via contact galvanic method from real environments.

Freshness

Last updated April 2026.

Geography

Various locations across 23 stations.