In magnetically confined plasmas, low-frequency (typ. 1-10 kHz) large-scale magnetohydrodynamic (MHD) instabilities represent a risk for performance and plasma stability. During long pulses in the WEST tokamak, deleterious MHD modes appear frequently inducing a drop of central temperature and a higher plasma resistivity that result in lower performances and shorter discharge duration. The real-time detection of such instabilities and the application of mitigation strategies is therefore of great importance for plasma control in WEST but also for future devices like ITER.
These MHD instabilities induce coherent temperature/density perturbations. Instruments like Electron Cyclotron Emission (ECE) radiometer or reflectometrer provide localized, high time resolution of temperature or density fluctuations. However, MHD analysis is currently performed offline, after the discharge. Real-time capability is crucial for control applications. The modes must first be identified before applying a mitigation strategy based on the knowledge of the MHD stability criteria. MHD stability is strongly affected by local heating and current drive, for which Electron Cyclotron Resonance Heating and Current Drive systems (ECRH/ECCD) are especially well suited.
The objective of this PhD is to develop a control strategy for WEST long pulse operation. The first step is the real-time detection of low frequency MHD instabilities using first ECE radiometer, then adding instruments like ECE-imaging or reflectometry to enhance reliability and accuracy. Integrated plasma modelling will then be performed to explore MHD mitigation strategies. ECCD is an obvious actuator, but other tools such as a temporary change of the plasma parameters (current, density or temperature) will also be evaluated. The mitigation strategy will be integrated in WEST Plasma Control System. Initial strategy will rely on simple control loop, then Neural Network or deep-leaning algorithms will be tested.