MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641058185 A) filed by Mr. M. Ramesh; Mr. C. Thirumoolar; Ms. A. Abarna; Mr. V. Sujith Kumar; and Mr. V. Sorna Mugash on May 07, 2026, for Ai Natural Language Web Platform For Automated 3d Physics Simulation.
Inventors include Mr. M. Ramesh; Mr. C. Thirumoolar; Mr. V. Sujith Kumar; and Mr. V. Sorna Mugash.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: Abstract The invention presents an AI-Assisted Simulation Platform that democratizes engineering analysis by allowing any user to perform CFD, FEA, and thermal simulations through natural language and 3D model upload. Conventional CAE tools require expert knowledge, expensive licenses, and high-end hardware, limiting access for students and startups. The proposed web-based system integrates a fine-tuned LLM that interprets plain-English requests (e.g., "simulate airflow at 60 m/s"), automatically configures meshes, boundaries, and solver settings, executes lightweight open-source engines (OpenFOAM/PyBullet) in Docker containers, and delivers interactive Three.jsNVebGL visualizations of velocity, pressure, temperature, or stress fields directly in the browser. An AI explanation module translates numerical results into simple educational insights. Developed with React frontend, Flask/Node.js backend, and MongoDB/Azure storage, the platform achieves sub-minute automated setup, browser accessibility, and acceptable accuracy for educational and conceptual design use. By eliminating manual configuration and providing real-time 3D exploration with AI summaries, the invention bridges the gap between powerful physics simulation and intuitive human interaction, offering a scalable, cost-free tool for learning, rapid prototyping, and early-stage innovation in aerospace, mechanical, and thermal engineering.
Disclaimer: Curated by HT Syndication.
