Application of AI technologies in nuclear power plants
summary
An internal LLM chatbot is built with Python backend, JavaScript front end, Azure AI Search vector indexing and SQL lookups, using a multi‑agent pipeline and periodic human evaluation; about 20% of 4,000 employees use it monthly. A predictive blowdown model runs every 12 hours in Databricks (inputs: MetoBlue weather, river forecasts, operator measurements) and claims ~€50,000/year saved. A pump blade‑angle optimizer (data in Databricks, Power BI alerts) reduces load by ~0.5 MW in certain conditions, estimated at ~€400,000/year.
AI-written summary · about
chapters (6)
- 0:00AI in nuclear plants
- 1:54Large language model chatbot for internal documentation
- 6:53AI for water treatment
- 11:09Cooling pump optimization
- 16:29Future AI in nuclear plants
- 17:13Q&A
related talks
24mAI for nuclear safety: Predicting component remaining useful life
50mAzure AI, Power new possibilities for every organization
31mAI Use Cases: Adding and Maintaining Value
23mBuilding AI-Driven Real-Time Lakehouse APIs
16mKey Points and Lessons Learned in AI Projects in Mechanical Engineering
17mAI Assisted Load Sharing for Improved Energy Efficiency
38mAI in action: How to use practically in iGaming beyond the hype?
50mBuilding AI ready applications