MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641074783 A) filed by Vellore Institute Of Technology on June 17, 2026, for “a Method For Analyzing Vibration And Strain Data Using Deep Learning Models To Forecast Structural Failure In A Structure”.
Inventors include Dr. Sudha S; and Pydi Meghana.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: The present invention relates to a method and system for forecasting structural failure using deep learning-based analysis of multimodal structural response data are disclosed. A distributed sensing arrangement, including a smart sensing coating and/or embedded sensing nodes, is deployed on a structure to capture vibration, strain, and electrical response signals, which are conditioned, synchronized, and preprocessed into spatially and temporally aligned data. The processed data is analyzed using a deep learning model comprising convolutional and recurrent neural network layers to identify spatial features and temporal degradation patterns, and to predict a probability of structural damage progression within a future time window. A structural risk indicator and/or damage index is computed and output to a monitoring interface for real-time risk assessment and failure forecasting. The system further includes a polymer-based sensing coating with piezoelectric and conductive nanomaterials, a data acquisition unit with embedded control, an AI processing module, and a monitoring dashboard for visualization of structural health and alerts.
Disclaimer: Curated by HT Syndication.
