MUMBAI, India, Oct. 12 -- Intellectual Property India has published a patent application (202641118957 A) filed by Sreyas Institute Of Engineering And Technology on October 05, 2026, for Explainable Ai Framework For Predictive Maintenance And Anomaly Detection In Large-Scale Industrial Cyber-Physical Systems.
Inventors include Dr T V Divya; E Keerthi; and R Sailaja.
The application for the patent was published on October 09, 2026, under issue no. 41/2026.
Abstract: ABSTRACT [0018] The present invention discloses an explainable artificial intelligence framework designed for predictive maintenance and anomaly detection inside large-scale industrial cyber-physical systems. The framework continuously ingests multi-modal sensor streams, control signals, and operational logs from distributed physical assets and their digital counterparts. It employs a hybrid architecture that first extracts temporal and spatial features through specialized neural modules, then performs anomaly scoring with an ensemble of unsupervised detectors, and finally generates human-readable causal explanations by combining local feature attributions with physics-informed residual analysis and rule extraction. A dedicated verification layer cross-checks each anomaly decision against both statistical consistency metrics and domain knowledge constraints before issuing maintenance recommendations. The resulting system reduces unplanned downtime, improves operator trust through transparent reasoning, and scales to plants containing thousands of interconnected devices while preserving real-time performance. Experimental validation on representative industrial data streams demonstrates early detection of degradation patterns, low false-alarm rates, and explanations that align with known physical failure modes, thereby enabling proactive intervention and optimized resource allocation across manufacturing, energy, and process industries.
Disclaimer: Curated by HT Syndication.
