Parsimonious statistical techniques for the detection of drifts toward dangerous operational conditions in tokamaks

Abstract

The approaching initial operation of major new tokomaks is rendering more pressing the need for effective disruption prediction techniques. The required tools should be not only accurate but also capable of operating with a minimum number of signals, because in the first campaigns of new devices typically only a very limited number of diagnostics is available. The present work reports the results obtained with two advanced statistical techniques, chaos onset and concept drift. These methods are capable of detecting the plasma drifting towards dangerous regions of the operational space using only global signals such as the plasma current or the locked mode amplitude. The obtained performances meet the requirements of mitigation in devices such as JT60-SA and ITER at the beginning of their operation. The computational times of the chaos onset approach are already of the order of 1 ms on laptop computers, whereas the concept drift technique requires implementation on FPGAs or GPUs to be compatible with real time requirements