Methods and equipment applicable to fault type diagnosis and section location in power distribution networks
By using deep convolutional neural networks and conditional adversarial generative networks in the distribution network, a high-quality fault waveform dataset is generated, which solves the problems of accuracy and adaptability in fault type and section location in the existing technology, realizes fast and reliable fault diagnosis and location, and improves the power supply reliability of the distribution network.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- KEDA INTELLIGENT ELECTRICAL TECH
- Filing Date
- 2026-03-23
- Publication Date
- 2026-05-26
AI Technical Summary
Existing fault diagnosis methods for power distribution networks rely on manual feature extraction and identification, which are susceptible to noise and system interference, have poor adaptability, and are difficult to quickly and reliably locate fault types and sections, especially in complex interference scenarios where accuracy is insufficient.
Using single-point measurement of three-phase voltage, zero-sequence voltage, three-phase current, and zero-sequence current as the research objects, a high-quality, balanced fault waveform dataset is generated by constructing and training a deep convolutional neural network and combining it with a conditional adversarial generative network to locate the fault type and section.
It enables rapid and reliable identification of fault types and segment location under complex interference scenarios, improves the accuracy and response speed of the model, reduces power outage duration, and enhances the power supply reliability of the distribution network.
Smart Images

Figure CN121878382B_ABST