In a remarkable example of grassroots innovation, a 17-year-old student from Mumbai has developed a low-cost artificial intelligence system that could significantly improve efficiency in India’s manufacturing sector.
Kcavyan Agarwal, a student of Dhirubhai Ambani International School, has created an AI-powered tool wear prediction system designed specifically for small and medium-sized machine shops (MSMEs). The device, which costs under ₹500, enables real-time monitoring of tool wear in lathe machines—eliminating the need for costly and time-consuming manual inspections.
Bridging a Critical Gap in MSMEs
In most Indian workshops, operators rely on experience or periodic microscopic inspection to assess tool wear—methods that often result in unnecessary downtime or premature replacement of tools. Agarwal’s innovation addresses this gap by providing continuous, real-time insights without interrupting production.
The system uses a physics-informed neural network (PINN) integrated with an affordable sensing setup. It collects vibration and acoustic data through sensors and processes it instantly to estimate tool wear during machining operations.
Built on Real-World Data
Unlike many student projects, Agarwal’s model is grounded in extensive fieldwork. During his summer break in 2025, he spent over 30 hours in machine shops across Mumbai and Navi Mumbai’s MIDC industrial belt, gathering more than 20 hours of machining data for training and additional data for testing.
Working under the mentorship of a PhD researcher from IIT Bombay, he identified key physical variables influencing tool wear and incorporated them into the model design. By September 2025, he had developed a fully functional prototype using an Arduino platform.
Strong Results in Trials
The system has shown promising results in validation tests:
• 22 out of 25 readings were within 10% of actual tool wear
• Operators were able to extend tool life by 3–4 hours
• Improved monitoring led to better planning and reduced downtime
The device uses an MPU6050 sensor for vibration data and an LM393 sensor for acoustic signals, which are analyzed by the AI model in real time.
Towards Industrial Deployment
Agarwal is now working towards deploying the system in live industrial environments and is collaborating with engineers at Hindustan Forgings. Future plans include expanding datasets, refining the model, and seeking validation from established manufacturing organisations.
A Step Towards Industry 4.0
With its combination of affordability, accuracy, and practical relevance, the innovation represents a significant step toward bringing Industry 4.0 technologies within reach of India’s MSME sector.
If successfully scaled, this low-cost AI system could help thousands of small manufacturers reduce costs, improve efficiency, and compete more effectively in a rapidly evolving industrial landscape.