A data-driven ground fault detection and isolation method for main circuit in railway electrical traction system

Area of Science:

• Electrical Engineering
• Control Systems
• Data-Driven Analytics

Background:

• Electrical traction drive systems face harsh conditions like corrosion and static electricity.
• Ground faults in these systems generate large short-circuit currents, endangering critical components.
• Effective fault diagnosis is crucial but challenging due to system complexity.

Purpose of the Study:

• To develop an improved ground fault detection and isolation method for electrical traction drive systems.
• To leverage correlations between onboard measurements for enhanced diagnostic performance.
• To address the limitations of conventional methods relying solely on DC-link voltage measurements.

Main Methods:

• A data-driven approach utilizing canonical correlation analysis (CCA) to exploit measurement correlations.
• A fault isolation component using fault direction derived from stored operational data.
• Application and validation of the proposed method on a traction drive system.

Main Results:

• The proposed method significantly enhances fault detection rate.
• Detection delay is notably reduced compared to conventional techniques.
• Correct isolation rate is substantially improved, demonstrating superior performance.

Conclusions:

• The data-driven approach effectively utilizes correlations among onboard measurements for superior ground fault diagnosis.
• The combined fault detection and isolation strategy offers significant improvements in key performance indicators.
• This method provides a more robust and accurate solution for safeguarding electrical traction drive systems.