A Novel Pix2Pix Enabled Traveling Wave-Based Fault Location Method

Area of Science:

• Electrical Engineering
• Power Systems
• Artificial Intelligence

Background:

• Traveling wave-based fault location is crucial for power system stability.
• Existing methods often require high-frequency data from Phasor Measurement Units (PMUs), limiting their applicability.
• Low-frequency PMU data presents challenges for accurate fault localization.

Purpose of the Study:

• To develop a novel deep learning approach for fault location using Image-to-Image Translation (Pix2Pix).
• To overcome the dependency on high-frequency PMU data for accurate traveling wave analysis.
• To enhance the precision of fault location estimation in power systems.

Main Methods:

• Utilized a Pix2Pix deep learning model for Image-to-Image Translation.
• Translated low-frequency detail component images from PMU data to higher frequency representations.
• Employed the YOLO v3 object recognition algorithm to validate generated images and estimate traveling wave arrival times.

Main Results:

• The Pix2Pix model successfully translated low-frequency PMU data to higher frequency images.
• Generated images were accurately identified by the YOLO v3 algorithm.
• Improved accuracy in estimating the arrival time of the traveling wave head was achieved.
• Enhanced overall fault location accuracy was demonstrated.

Conclusions:

• The proposed Pix2Pix-enabled deep learning method offers a viable solution for traveling wave-based fault location with low-frequency data.
• This approach significantly improves fault location accuracy compared to traditional methods.
• The integration of Pix2Pix and YOLO v3 provides a robust framework for advanced power system fault analysis.