MUMBAI, India, Jan. 9 -- Intellectual Property India has published a patent application (202531127989 A) filed by National Institute Of Technology, Agartala, Tripura, on Dec. 17, 2025, for 'system and method for classifying power quality disturbances in electrical power networks.'
Inventor(s) include Mr. Rajesh Debnath; and Dr. Arvind Kumar Jain.
The application for the patent was published on Jan. 9, under issue no. 02/2026.
According to the abstract released by the Intellectual Property India: "A system and method for classifying power quality disturbances (PQDs) in electrical power networks are disclosed, comprising an acquisition interface receiving voltage or current waveforms from a power distribution feeder at a sampling frequency of at least 10 kHz, a signal processing module applying continuous wavelet transform (CWT) with a complex Morlet mother wavelet to convert one-dimensional PQD signals into two-dimensional time-frequency scalogram images of 224 224 pixel resolution with three RGB channels, and a classification module implementing a masked auto-encoder based vision transformer (ViT-MAE) network featuring 16 16 pixel patch embedding into 768-dimensional tokens, a 12-layer transformer encoder with 12 attention heads, 75% masking ratio during self-supervised pre-training via mean squared error reconstruction loss, and fine-tuned classification via cross-entropy loss with a class token head; the system achieves combined classification accuracy of 97.5-98.5% across 30 PQD classes including single through quadruple simultaneous disturbances over signal-to-noise ratios of 20-40 dB, with end-to-end latency of 12-15 milliseconds (less than a half-cycle of 50 Hz grid systems) enabling online real-time detection, validated across synthetic IEEE-1159 datasets, open-source real-world datasets, and hardware-in-loop environments using Opal-RT OP4510 simulator with RT-LAB kernel demonstrating practical applicability in power quality monitoring systems."
Disclaimer: Curated by HT Syndication.
