A Novel Wide-Band Directional MUSIC Algorithm Using the Strength Proportion.

Area of Science:

• Electrical Engineering
• Signal Processing
• Electromagnetics

Background:

• Partial discharge (PD) detection in substations is crucial for grid reliability.
• Existing directional multiple signal classification (Dir-MUSIC) algorithms struggle with the wideband nature of PD signals due to frequency-dependent antenna gain.
• Improved accuracy in PD source localization is needed for effective fault diagnosis.

Purpose of the Study:

• To discuss wideband Dir-MUSIC algorithms for PD source localization.
• To propose a novel wideband Dir-MUSIC algorithm, DirSP (Directional Signal Proportion), for enhanced accuracy.
• To evaluate the performance of DirSP against conventional methods in simulations and experiments.

Main Methods:

• The study discusses existing wideband Dir-MUSIC algorithms.
• A novel DirSP algorithm is proposed, which focuses PD signals at a specific frequency using strength proportion before applying Dir-MUSIC.
• Simulations and experimental validation using six sets of PD signal samples were conducted.

Main Results:

• The conventional Dir-MUSIC algorithm loses accuracy when antenna gain functions vary significantly across frequencies.
• The proposed DirSP algorithm demonstrates robust performance even with differing antenna gain functions.
• Experimental results show DirSP achieves a mean error and standard deviation of less than 4°, outperforming other methods.

Conclusions:

• The DirSP algorithm offers a significant improvement in the accuracy of wideband PD source localization compared to traditional methods.
• DirSP effectively addresses the challenges posed by frequency-dependent antenna gain in PD signal analysis.
• The proposed method is a valuable advancement for monitoring and diagnosing partial discharge activity in electrical substations.