MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202611070301 A) filed by Ims Engineering College, Ghaziabad on June 05, 2026, for A Predictive Grid Failure Analysis System And Method Using Machine Learning For Electrical Power Networks.
Inventors include Dr. Balwant Singh; Dr. Sourabh Maheshwary; Mr. Ankit Agarwal; and Ms. Meenu Sharma.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: The present invention relates to a predictive grid failure analysis system and method using machine learning for electrical power networks. The system receives real- time and historical grid data from smart meters, feeders, transformers, phasor measurement units, protection devices, weather sensors and maintenance databases. A preprocessing engine validates, cleans, synchronizes and normalizes the received data. A feature extraction module generates failure-indicative features including voltage deviation, frequency trend, overload duration, transformer temperature rise, phase imbalance, harmonic distortion, environmental stress and asset-health indicators. A machine learning prediction engine analyses the features to generate asset-specific failure probability scores within a selected future time window. A severity classification module classifies predicted failures based on outage risk, consumer impact, asset criticality and cascading failure possibility. An explainability module identifies contributing technical causes, and a recommendation module generates preventive actions such as load transfer, inspection, maintenance scheduling or crew dispatch. The invention improves grid reliability by enabling preventive action before outage occurrence. Accompanied Drawing [FIGS. 1-2]
Disclaimer: Curated by HT Syndication.
