The team in charge of building an Iter reactor is using AI tools to significantly increase the speed and accuracy during the assembly.
Iter furnace is in the process of construction. Image: Innovation Newsnetwork
In the Southeast of France, a team of 2,000 physicists, scientists and workers are busy building the world’s largest lymphadenopic plants called International Test Nuclear Thermal reactor (Iter). Their goal is to put the facility into operation in 2033 and a series of artificial intelligence tools (AI) being implemented to complete the job that requires high accuracy on unprecedented scale, Interesting Engineering On March 27 reported.
The heat -conditioning reaction simulates the condition of the sun to unite the Hydro deuterium and tritium isotopes, forming a helium and releasing large energy. Most of the process takes place in the reactor cavity called Tokamak. In this bread -shaped cavity, lymph node fuel is baked to hot plasma state, where the temperature is 10 times higher than the sun core, causing the lymph node reaction to become feasible.
At Iter, researchers are working to build the largest tokamak furnace in the world to study in detail the process of thermonuclear reaction. However, completing this project is a huge task due to the large scale of construction and the number of large parts, requiring whose assistance.
The Tokamak furnace in Iter will be assembled with 9 parts, 5 parts made in Europe, 4 parts made in Korea, while Russia and India will provide some components. Made from a special stainless steel, these parts will be arranged and welded together. A series of complex ultrasound scanners will ensure that the weld is not faulty. The process of producing large amounts of data is needed, so the research team has built a AI model in the Visual Studio Code editor of Microsoft, where many encrypted languages can be used to explore data and save hours of quality checks and accuracy of the weld. The team can also use the model to determine the material to be used as a reactor chamber lining from the inside.
Collaboration efforts in Iter project gather more than 30 countries, requiring a standardized document management system for more than 20 years. More than 1.5 million documents were created during this time, and the search for information was like “groping the bottom of the tank”. In the first quarter of 2024, the ITER ITER project tool department built a chatbox AI) to do this job and gradually improved in the year. Earlier this year, AI tools summarize documents and classify information in a large database. The advantage of this is that the user can send a request to the database and collect the useful feedback instead of receiving the answers to the keyword.
Using tools from Openai, Chatbot supports questions in Chinese, Korean, Japan, Russia, Hindu and can be approached by 120 project partners. Because Iter document also contains many abbreviations, the team development team develops another chatbox to answer the question about that issue. They also work to make the document system easier to connect to other large language models, which allows easier information exchange.