Low-Current Feeder Fault Selection With KPCA-BIRCH
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fault line selection devices struggle with low detection accuracy in low-current grounded distribution networks due to single-phase grounding faults, which can lead to equipment damage and potential fires.
Innovation Solution
A fault line selection method using the KPCA-BIRCH clustering algorithm for unsupervised clustering of zero-sequence current and voltage data to identify faulted feeders, employing KPCA for dimensionality reduction and BIRCH for precise fault detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional fault line selection devices are used in low-current grounded distribution networks, then the system maintains fault-tolerant operation capability, but the detection accuracy of faulted feeders is insufficient
Solution Approach 1:
The patent transforms the fault detection problem from traditional current magnitude analysis to a clustering-based dimensional analysis in feature space. By extracting multiple feature dimensions (current, voltage, impedance characteristics) and applying clustering algorithms, the system creates a new diagnostic dimension that enables accurate fault line identification while preserving the low-current operational characteristics of the distribution network
Solution Approach 2:
The patent replaces traditional mechanical/electrical fault detection mechanisms with data-driven clustering algorithms. Instead of relying on hardware-based protective relays and current transformers alone, the system uses software-based unsupervised clustering (KPCA-BIRCH) to analyze electrical characteristics and identify faulted feeders, achieving higher detection accuracy without altering the physical low-current grounding system
2Reliability
If single-phase grounding faults are allowed to operate for a period, then power supply reliability is enhanced, but arc faults may develop causing equipment damage and fires
Solution Approach 1:
The patent implements continuous monitoring and clustering-based feedback analysis of electrical characteristics during single-phase grounding faults. The system repeatedly clusters updated feature data to track fault evolution in real-time, providing feedback that enables timely detection of arc fault development while maintaining fault-tolerant operation, thus preventing equipment damage and fires
Solution Approach 2:
The patent applies preliminary clustering analysis to detect early signs of arc fault development before they cause damage. By continuously analyzing clustered features during the fault-tolerant operation period, the system takes preliminary protective action to identify transitioning faults, enabling preventive measures to be taken before arc faults cause equipment damage or fires
Data Source
AI summary
A method, system, and readable storage medium for fault line selection in distribution networks is provided. The method includes: obtaining the zero-sequence current of each feeder and the zero-sequence voltage of the busbar within a preset time window after a fault occurs; using these to process the feeder's short-time window zero-sequence instantaneous power curve cluster in the distribution network through KPCA (Kernel Principal Component Analysis) for dimensionality reduction, determining the principal component scores; and performing BIRCH (Balanced Iterative Reducing and Clustering using Hierarchies) clustering based on these scores to identify whether a feeder is faulted. This clustering process allows for precise and rapid identification of the faulted feeder, even when the current is small, improving detection accuracy. This solves the problem of quickly identifying the faulted feeder in a small current grounding distribution network during single-phase grounding faults.