MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641072806 A) filed by Bert Labs Private Limited on June 11, 2026, for Self- Updating Digital Twin Of Kiln In Real Time Using Reinforcement Learning Feedback”.
Inventors include Rohit Kochar; and Riya Kumari.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: ABSTRACT SELF- UPDATING DIGITAL TWIN OF KILN IN REAL TIME USING REINFORCEMENT LEARNING FEEDBACK The invention relates to a system and method for real-time updating of a plant-level 10 digital twin of a rotary kiln and associated equipment in a cement plant. Sensor/DCS inputs, including temperatures, airflow, oxygen, fuel and feed rates, are preprocessed and supplied to a kiln environment model forming the digital twin. A reinforcement learning (RL) agent trained within the digital twin generates optimized operational parameters. An error/deviation detection module computes 15 performance differences between plant measurements and twin predictions to localize faulty equipment sub-models. A model-update module selectively retrains only the affected sub-models using recent plant data, restoring prediction fidelity without full retraining. The optimized parameters are deployed via ICS/SCADA/DCS for real-time or near-real-time closed-loop control. The system 20 improves fuel efficiency, process stability, safety, emissions compliance and product quality by continuously aligning the digital twin with real-world operations. [
Disclaimer: Curated by HT Syndication.
