Swiss physicists and the company DeepMind have developed a fusion process controlled by artificial intelligence. Perhaps it is this combination of technologies that will give impetus to more active exploration of the Solar System.
Artificial intelligence will help physicists
École Polytechnique Fédérale de Lausanne (EFPL) together with DeepMind announced the start of development thermonuclear reactor, fully controlled by artificial intelligence. Swiss scientists hope that their experience can be combined with the innovative approach of the project founded by Google.
Thermonuclear fusion is a technology for obtaining energy by fusing the nuclei of light elements. Devices called “tokamak” are mainly used for this. EFPL has its own tokamak called TCV and has been operating for 20 years. Its feature is that it has many settings and it is possible to hold the plasma with a magnetic field in several modes.
The results of 20 years of research were used by DeepMind specialists in order to teach their artificial intelligence to manage thermonuclear fusion. It’s like how a neural network learns to play chess at grandmasters’ games and then beats them.
At first, artificial intelligence demonstrated its skills in a simulator. It was then attached to the TCV and operated the reactor in several different modes, including keeping two types of plasma in the same chamber with a magnetic field. EFPL specialists are very satisfied with the process and hope that the created system will allow them to finally get a controlled fusion reaction. DeepMind is also excited about the opportunity to involve the neural network in solving problems that humans cannot handle.
Thermonuclear reactors for space
The creation of effective thermonuclear reactors is extremely important for space exploration. Tests of ion and plasma rocket engines are in full swing. They are able to accelerate the spacecraft to speeds of tens and hundreds of kilometers per second, but require a large amount of energy to accelerate the working body, which creates thrust. As long as we are talking about automatic probes weighing hundreds of kilograms, this is not so important. But the main purpose of creating such engines is manned flight to outer planets. And there, only the residential module of the ship can weigh tens of tons. In this case, the power of the engine must be measured in tens of megawatts, and the traditional means of obtaining energy for modern cosmonautics cannot provide such a quantity. For example, the area of solar panels for this must reach square kilometers, but this will not matter, because the farther from the Sun, the less effective they will be.
A full-fledged reactor should be installed on the spacecraft, and it is better that it be thermonuclear. The energy sources we are used to, which work on nuclear fission, are less powerful and produce a large amount of waste, which simply has nowhere to go in space so as not to pollute it.
The neural network will perform a controlled reaction
The main problem of a controlled thermonuclear reaction is that engineers have not been able to implement it for 70 years. Now we are very close to its implementation, but it is done mainly by “brute force”. Thermonuclear reactors, on which it was closely approached, are large and powerful. They are not suitable for a spaceship.
EFPL and DeepMind’s approach seems to be better for this. If the magnetic fields and plasma in the reactor can be more finely controlled, it can be made smaller. Of course, if it matters at all. After all, we still don’t know why we can’t get energy from plasma. Perhaps artificial intelligence will figure this out.
Will artificial intelligence control the spaceship?
One of the most interesting questions that arises in connection with the experiments of Swiss physicists — what else can we entrust to artificial intelligence? After all, if we trust him to manage a thermonuclear reactor, which we are unable to manage ourselves, then why not also trust the engine that is powered by this reactor?
Artificial intelligence driving a spacecraft can provide huge benefits, such as collision avoidance. The problem of space debris is currently being solved by the fact that the trajectory of each piece is calculated in advance by observers on Earth. But its number is constantly growing.
It is not so difficult to evade a threat that has suddenly arisen. For this, you just need to make the right decision very quickly. But it may not be possible for a person. And from this, it is only one step to artificial intelligence, which will calculate the optimal trajectory for the ship. And the only question is why the right person will be on board.
