MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202631112969 A) filed by Brainware University on September 21, 2026, for Ai-Based Predictive Fault Detection And Self-Healing System For Smart Energy Grids.
Inventors include Amitava Podder; Subrata Paul; Piyal Roy; Shyamalendu Paul; and Sanchita Ghosh.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: An AI-based predictive fault detection and self-healing system for an electrical distribution grid (100) includes a sensor and IoT network (102), signal-conditioning and edge-processing units (104, 116), a fault prediction engine (106), a central control unit (108), a self-healing switching mechanism (110), a communication module (112), and a monitoring dashboard (114). Multi-parameter voltage, current, temperature and vibration or partial-discharge data are pre-processed and supplied to a trained model comprising an LSTM temporal layer (122), an attention-based feature-weighting layer (124), and an ensemble classifier layer (126). The model produces a fault-probability output (128) and predicted fault location. When the probability exceeds a predefined threshold, the control unit commands isolation of the predicted fault segment and reconfiguration through an alternate feeder path. Events may be logged for periodic model retraining. Figure 1 is representative of the invention.
Disclaimer: Curated by HT Syndication.
