MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202611093242 A) filed by Indian Institute Of Technology Jodhpur on July 31, 2026, for System And Method For Generating Adaptive Trip Decision Through Virtual Replication Of Programmable Scheme Logic.
Inventors include Neha Kumari; Mahek Dharmesh Kumar Vanjani; Ravi Yadav; and Sandeep Yadav.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: The present invention discloses a machine learning-based (ML-based) system and an ML-based method for generating an adaptive trip decision through updatable virtual replication of programmable scheme logic (PSL) of a protection relay. A digital data bus (DDB) signal intelligence subsystem (206) obtains DDB signals and PSL configuration data, an embedding vector generation subsystem (208) generates embedding vectors and hidden features, an attention weight generation subsystem (210) generates trip-specific attention weights, a feature aggregation subsystem (212) generates weighted feature representations, a trip evaluation subsystem (214) generates trip decisions, a decision processing subsystem (216) generates a final trip decision, and a model update analysis subsystem (218) detects PSL mapping changes and incrementally updates machine learning models without complete retraining. The ML-based system (102) further comprises a label generation subsystem (220), a training subsystem (222), and a temporal validation subsystem (224) to maintain adaptive virtual replication of the protection relay. FIG. 2
Disclaimer: Curated by HT Syndication.
