Posted: 5/5/2019
Nuclear fusion is still one of humanities best bets to provide limitless, renewable, reliable, and fairly environmentally friendly energy supply for the next millennia.
Just read Artificial intelligence speeds efforts to develop clean, virtually limitless fusion energy
It appears that nuclear fusion is an almost ideal domain to apply machine learning, because there are petabytes of high quality data available from many U.S. and European nuclear fusion experiments.
One of the reasons, why humanity has not succeeded yet are frequent disruptions (“the sudden loss of confinement of plasma particles and energy”) when running experiments in the Tokamak designed reactors. For decades, these disruptions have stymied success. Applying, machine learning, we may finally be much better able to predict and prevent the occurrence of such disruptions. If there are not other, perhaps more serious obstacles to successful nuclear fusion power generation, then we ought to be optimistic ...
However, we may also finally realize that the Tokamak design was wrong from the beginning and other designs may work better.
No comments:
Post a Comment