The Fifth Elephant 2025 Annual Conference (18 & 19 July)

Hydrogen explosions are an inherent risk in nuclear reactors due to the unavoidable generation of hydrogen during its operation. From Chernobyl to Fukushima, several major nuclear accidents have been significantly worsened by hydrogen explosion. In sodium-cooled fast breeder reactors, due to the use of sodium as a coolant, hydrogen production is prevalent. The disposal of used radioactive sodium involves spray injection, which can unintentionally lead to hydrogen build up and explosions if not properly controlled.

This talk presents how Visual AI techniques are used to monitor and mitigate such risks. Through real-time video analysis, deep learning models detect early signs of hydrogen accumulation and flame anomalies. These insights are then used to trigger automated control systems based on reinforcement learning techniques to optimize the disposal operation, enabling intelligent decision-making in dynamic environments. The session will cover data generation, model development, and how AI solutions are deployed in safety critical nuclear applications.

Outline of Talk

• Introduction to Fast Breeder Reactors (3 mins)
  • Overview of three stage nuclear program
  • Role of fast breeder reactors and its challenges
• Need for Sodium Disposal (2 mins)
  • Why disposal of used radioactive sodium is essential
  • Spray injection method and its risks
• Sodium Fire Detection (8 mins)
  • Data generation: Controlled experiments simulating sodium fires using experiments
  • Data annotation: Building labelled datasets from raw footage
  • Modeling: Training Deep learning models for real time fire detection
• Video captioning of hydrogen deflagration event (6 mins)
  • Real time video analysis
  • RNN based models to caption the events
• Reinforcement Learning for autonomous control (5 mins)
  • Reinforcement learning to optimize the spray injection process
  • Intelligent decision-making to prevent hydrogen accumulation
• Future Work (1 min)
  • Deployment on edge device

Key Takeaways
• Role of AI in Nuclear Safety
• Experimental data generation in industrial environment
• End to End pipeline of Vision based automation system

Speaker Info
Muthukumar Ganesan is currently working as a Scientist at the Atomic Research Centre, Government of India. He brings over 11 years of professional experience in AI application development across the automotive and nuclear industries. His work focuses on enhancing safety and automation in sodium-cooled fast breeder reactors using AI-driven solutions, including hydrogen explosion detection, sodium leak identification, predictive maintenance, and autonomous fire mitigation systems. He has published 3 research papers in the field of AI and is currently working on 2 more. An experienced technical speaker, he has delivered talks on AI applications in industrial systems and actively contributes to the GitHub and MathWorks communities, where he shares innovative tools and solutions for real-world engineering challenges.
LinkedIn
Git