MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202611027254 A) filed by Ashoka Institute Of Technology And Management; Dr. Sarika Shrivastava; Shambhu Kumar; Jitendra Nalwaya; Santosh Kumar; Mr. Utkrisht Verma; Sandeep Kr. Singh; Somendra Bannerjee; Divesh Bhaskar; and Mr. Priyanshu Upadhyay, Varanasi, Uttar Pradesh, on March 9, for 'multi-stage memory-integrated gradient boosting system for distribution transformer failure prediction.'
Inventor(s) include Dr. Sarika Shrivastava; Shambhu Kumar; Jitendra Nalwaya; Santosh Kumar; Mr. Utkrisht Verma; Sandeep Kr. Singh; Somendra Bannerjee; Divesh Bhaskar; and Mr. Priyanshu Upadhyay.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to a predictive analytics system for forecasting failures of distribution transformers using a Multi-Stage Feature Engineering Framework (10) integrated with a Sequential Gradient Boosting Model (110). The system derives identity features through an Identity Feature Module (20), contextual interaction indicators through a Load Density Index Calculator (60), Seasonal Stress Factor Generator (70), and Critical Node Flag Module (80), and temporal degradation indicators through a Dynamic Memory Feature Module (40) including a Cumulative Fatigue Index Engine (50) and Capacity-Density Ratio Engine (90). Environmental exposure is quantified using a Geo-Vulnerability Score Module (100). The computed features are processed by the Sequential Gradient Boosting Model (110) to generate daily failure predictions and ranked outputs through a Risk Ranking and Alert Engine (120). The architecture enables reproducible predictive modeling using defined numerical ranges and rolling 30-day stress accumulation."
Disclaimer: Curated by HT Syndication.
