MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202621065132 A) filed by Sayali Prakash Shinde; Vishwakarma Institute Of Technology; Prof. Prajakta Dandavate; Sanika Mahesh Kulkarni; Manjiri Santosh Kulkarni; Akshada Sanjay Lohakre; and Anuja Dhiraj Kumbhar on May 23, 2026, for Adaptive Battery Thermal State Prediction And Early Overheat Warning System Based On Lstm Networks.
Inventors include Prof. Prajakta Dandavate; Sanika Mahesh Kulkarni; Manjiri Santosh Kulkarni; Akshada Sanjay Lohakre; Anuja Dhiraj Kumbhar; and Prof Sayali Shinde.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: Lithium-ion batteries are the dominant energy storage technology in electric vehicles, renewable systems, and consumer electronics. Ensuring thermal stability and detecting early overheating trends remain critical challenges, as temperature strongly influences capacity fade, internal resistance, safety, and overall battery lifespan. This paper presents a data-driven battery temperature forecasting model developed using a multivariate Long Short-Term Memory (LSTM) neural network. The model is trained on real-world recommissioned battery datasets and predicts short-term temperature evolution using a sequence history of voltage, current, battery mode, and temperature signals. A novel adaptive thresholding mechanism is proposed that automatically computes risk thresholds for voltage, temperature, and load conditions based on statistical patterns of high-risk samples. The system outputs an interpretable JSON risk report containing predicted temperatures, probability of overheating, data-driven risk thresholds, and feature influence analysis. Experimental evaluation demonstrates stable short-term forecasting performance several steps ahead, enabling proactive risk mitigation in lithium-ion battery systems across diverse lithium-ion battery applications.
Disclaimer: Curated by HT Syndication.
