Research and Implementation of Localization of Multiple Local Discharge Sources in Switchgear Based on Ultrasound

Area of Science:

• Electrical Engineering
• Acoustic Signal Processing
• High-Voltage Technology

Background:

• Current switchgear partial discharge detection is primarily offline, limiting simultaneous monitoring of multiple discharge sources.
• Existing methods struggle with accurate time difference acquisition in low signal-to-noise ratio environments.
• Localization of multiple partial discharge sources presents challenges for traditional time difference of arrival (TDOA) algorithms.

Purpose of the Study:

• To investigate the application of ultrasonic technology for online localization of partial discharge faults in high-voltage switchgear.
• To develop advanced algorithms for improved accuracy in time difference estimation and source separation.
• To enable simultaneous localization of multiple partial discharge sources.

Main Methods:

• Soft and hard filtering techniques were employed for background noise reduction of partial discharge signals.
• A generalized cubic correlation algorithm, based on Time Difference of Arrival (TDOA), was developed to enhance time difference acquisition accuracy.
• Single-channel signal blind source separation, singularity spectral analysis, and independent component analysis (ICA) were used for identifying and separating multiple discharge sources.
• A disordered coordinate selection method was proposed for determining the coordinates of multiple sources.

Main Results:

• The noise reduction method effectively removed interference while preserving critical partial discharge signals.
• The improved generalized cubic correlation algorithm demonstrated superior resistance to interference and reduced localization error by 17-68 mm compared to other correlation algorithms.
• The system successfully located multiple discharge sources with an accuracy of 88% in identifying the number of sources.

Conclusions:

• Ultrasonic technology, combined with advanced signal processing techniques, offers a viable solution for online, multi-source partial discharge localization in high-voltage switchgear.
• The developed generalized cubic correlation algorithm and ICA-based source separation significantly improve localization accuracy and reliability.
• This approach represents a substantial advancement over pre-improved methods, enhancing the safety and maintenance of electrical infrastructure.