Nuclear fusion is the process that powers the stars. It is being investigated in multiple laboratories across the world. If it can be generated reliably on Earth, it could be a future power source that could provide clean and renewable energy free of the radioactive waste associated with current commercial nuclear fission plants.
The process of nuclear fusion forces nuclei of elements together to create heavier nuclei. In stars, the process begins with the compression of isotopes of hydrogen into helium, then the helium fuses, and so on. The nuclei are in an ultra-hot gas called a “plasma”. Great amounts of energy are released as nuclei fuse.
In some laboratories on Earth, the hot plasma is trapped by powerful magnetic fields as fusion occurs. Other laboratories use an array of powerful lasers to ignite tiny pellets of fuel. In any case, the energy generated is either used to heat water for steam turbines or, in some fusion reactors, it is converted directly to electricity.
One of the most important things that scientists must find in order to create reliable fusion reactors for power on Earth is the best mixture of isotopes of light elements to use as fuel. For reactors that run on hydrogen nuclei, there are three choices. Standard hydrogen has a single proton in the nucleus orbited by a single electron. Deuterium is a non-radioactive hydrogen isotope with one neutron in the nucleus. Tritium is a radioactive isotope of hydrogen with two neutrons in the nucleus. In fusion reactors called tokamaks, the mix of hydrogen isotopes undergoing fusion is monitored by spectroscopy. However, this process is complex and time-consuming.
A new paper was recently published in The European Physical Journal D by Mohammed Koubiti who is an association professor at the Aix-Marseille Universite in France. In the new paper, Koubiti assesses the use of machine learning in connection with plasma spectroscopy to determine how the ratios of hydrogen isotopes affect nuclear fusion plasma performance.
Koubiti said, “In terms of performances, fusion power plants will be operated with a mixture of deuterium and tritium since their nuclear fusion is optimal, but the content of tritium must be controlled and strictly managed to respect the limits imposed by the regulation authorities. Also, it may be necessary to know on a real-time scale the content of tritium to optimize the performances of the nuclear power plant.” Koubiti goes on to explain that to develop a way of doing this he decided to use a combination of machine learning and spectroscopy.
Koubiti adds that, "The ultimate aim is to avoid using spectroscopy, whose analysis is time-consuming, and replace it — or at least combine it — with deep learning to predict tritium contents in fusion plasmas," Koubiti explains. "This study is only a step towards that. I am still using spectroscopy as a means of allowing me to find other features that can be used by deep learning algorithms to predict as a function of time the tritium content in fusion plasmas."
Koubiti says that the next step is to identify the non-spectroscopic features that must be provided to any deep-learning algorithm. Then he will test the findings on several magnetic fusion devices that rely on external magnets to confine the plasma.