MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202641107792 A) filed by Pragati Engineering College on September 08, 2026, for Agentic Ai-Based Digital Twin Framework For Autonomous Industrial Asset Management.
Inventors include Dr. Avinash Gudimetla; Mr. Akella Yeswanth; Dr. Balla Satyanarayana; and Dr. K Saroja Rani.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: The present invention provides a system and method of autonomous industrial asset management using an AI digital twin system, or Agentic AI. It creates a dynamic virtual image of a physical industrial asset, machine, production unit or infrastructure component by collecting real-time data from sensors, controllers, maintenance systems, operational databases and enterprise platforms. The digital twin stores the current operating conditions, equipment health, detected anomalies, remaining useful life estimates, simulation results, maintenance history, manuals, rules, knowledge graphs and previous operating cases. Agentic AI core capabilities involve perception and diagnosis, planning and optimization, decision orchestration, and learning functions. The perception and diagnosis function is used to interpret the state of the digital twin, put the events in context, identify abnormal behavior, identify potential causes and root causes and estimate the operational risk. The planning and optimization function creates alternative maintenance, operational and control solutions and ranks them by their urgency, safety, production requirements, resource availability, workforce schedules, part availability, energy usage, etc., and asset performance. The selection of the safe action is done by the decision orchestrator, who coordinates the functions and evaluates proposed strategies. Enforces configurable operating limits, organizational policies, permissions, confidence thresholds, regulatory requirements and fail-safe conditions prior to authorizing action using safety guardrails. Human control is applied to the review of explanations, acceptance or rejection of recommendations, plan variations, and the overturning of autonomous decisions, especially where there is lack of confidence or where the level of operational risk is in excess of a defined limit. Authorized actions could include predictive maintenance plans, service schedules, work orders, allocation of technicians and tools, spare parts requests, process set points, energy optimization commands, operator alerts and explanations as to the actions. The results of execution are fed back to a learning function which modifies analytical models, decision rules, memory records, and future planning strategies. The system can run on edge, cloud, or hybrid computing infrastructures and can remain to provide critical monitoring and decision-making capabilities when network connectivity is limited. The invention includes real-time digital twin synchronisation, collaborative AI reasoning, safety constrained autonomy, explainable decisions and continuous feedback, which reduce unexpected failures and downtime, increase asset availability and resource utilisation, enable timely maintenance, increase operational safety, and allow for autonomous control and operation of a wide range of industrial assets in manufacture, energy, transportation, mining, utilities and process industries with authorised human oversight, control and accountability.
Disclaimer: Curated by HT Syndication.
