Load Pattern Mining and Label Recognition for Farmland Irrigation and Drainage Users Based on GMM Clustering

To address the challenges of low efficiency in identifying the electricity consumption characteristics of farmland irrigation and drainage users and inadequate assessment of power supply service risks, this paper proposes a technical framework for load pattern mining and precise label recognition based on Gaussian Mixture Model (GMM) clustering. Using the 2024–2025 electricity consumption data of 1023 farmland irrigation and drainage users as a sample, a 10-dimensional load feature system is constructed, encompassing four key dimensions: consumption volume, fluctuation, time-of-use, and continuity. Following feature extraction, typical electricity consumption patterns are mined using unsupervised GMM clustering. Finally, association rules linking these features to business labels are established to enable accurate recognition. The research results demonstrate that the framework achieves a label recognition accuracy of 98%, providing technical support for the refined management of farmland irrigation and drainage users.