MUMBAI, India, Oct. 12 -- Intellectual Property India has published a patent application (202641119412 A) filed by Dr. S. Vidya; Mr. Mohammad Anwar Siddiqui; B. Sunitha; B Venkateswara Reddy; Swathi T K; R Valli; Sabhana U; and A. Suganya on October 06, 2026, for Machine Learning-Based Electricity Theft And Abnormal Consumption Detection System For Sustainable Power Distribution.
Inventors include Dr. S. Vidya; Mr. Mohammad Anwar Siddiqui; B. Sunitha; B Venkateswara Reddy; Swathi T K; R Valli; Sabhana U; and A. Suganya.
The application for the patent was published on October 09, 2026, under issue no. 41/2026.
Abstract: The present invention relates to the development of a machine learning enabled system (100) and method (200) for identifying theft and consumption anomalies in sustainable electricity distribution grids is described. The system (100) records high-frequency temporal series information comprising active power, voltage, and current from smart meters (102a – 102c) and distribution transformers (104). A feature engineering subsystem (114) calculates rolling temporal and spectral entropy features to identify consumption anomalies (206). A hybrid machine learning classifier (116) that employs an unsupervised deep autoencoder (210) along with a gradient boosted tree algorithm (212) calculates theft probabilities in addition to applying explainable AI measures (214) to quantify the contributing risk factors. A spatial localization subsystem (118) validates the energy balance deficit of distribution transformers (104) against the theft score of individual customers to localize exact locations of direct line tapping (306) and meter bypass within secondary distribution lines (304), resulting in minimal false-positive dispatches and grid losses due to non-technical causes. FIG.1
Disclaimer: Curated by HT Syndication.
