With the increasing demand for electricity and the growing complexity of distribution networks, the identification and diagnosis of quality defects in power systems have become critical. Traditional fault diagnosis methods for distribution networks suffer from low accuracy and slow response time. In recent years, the application of data mining and artificial intelligence technologies has provided new ideas for power system fault diagnosis, especially in the diagnosis of quality defects in distribution networks. In order to improve the accuracy of fault diagnosis in distribution networks, this study proposes a quality defect diagnosis model for distribution networks based on hybrid clustering algorithm. The model first ensures the quality of data through data preprocessing, including data complementation, outlier processing, and data normalization; then, feature extraction is performed on the data through principal component analysis (PCA) dimensionality reduction to reduce the computational complexity. Finally, the clustering process is optimized by combining K-Means and hierarchical clustering algorithm to improve the accuracy of clustering results. The experimental results show that the accuracy of line loss anomaly identification of distribution network lines reaches 98.50% after using this model. In addition, by comparing with the traditional method, the optimized clustering algorithm has significant improvement in clustering time and error, the algorithm time is reduced by 26.5 seconds, and the average clustering error is reduced from 39.4832 to 7.8469. The model provides effective technical support for the intelligent operation and maintenance of the distribution network and has a better practical application value.
