Abstract:

Electrical treeing, which causes insulation break- down in high-voltage systems, begins with partial discharges (PD) in low-density polyethylene (LDPE) nanostructured materials. Identification and study of PD features enable early detection and avoidance of deterioration, improving electrical insulation dependability. Existing PD analysis methods lack time- frequency resolution and noise interference, making transient and superimposed discharge signals difficult to interpret. This limits diagnostic precision and delays electrical tree starting.  This research uses Wavelet Transform-Time-Frequency Analysis (WT-TFA) to uncover and categorize PD patterns in LDPE nanocomposites to address such issues. Multi-resolution wavelet decomposition and high-resolution time-frequency mapping im- prove signal resolution and discharge temporal localization in the WT-TFA approach. The combined technique effectively filters noise, preserves signal integrity, and captures PD pulse dynamics. Multiple PD signal datasets from laboratory-aged LDPE nanostructured insulation samples are used to evaluate the proposed approach. Corona, surface, and interior discharges are easily distinguished by PD features as pulse form, amplitude, and frequency development. Results reveal that WT-TFA-based discrimination improves detection sensitivity and pattern discrimination over traditional techniques. This technology enhances diagnostics and illuminates electrical treeing’s early phases, allowing predictive maintenance and complex insulation system design.