Learning Plasma Dynamics and Robust Rampdown Trajectories with Predict-First Experiments at TCV
This work addresses the problem of plasma instability during rampdown for fusion energy research, representing an incremental advance by applying existing SciML methods to a specific domain.
The paper tackled the challenge of simulating and controlling the rampdown phase in tokamak plasmas, which often leads to instabilities, by developing a neural state-space model (NSSM) that predicts dynamics and uses reinforcement learning to design robust trajectories, resulting in statistically significant improvements in metrics and a successful 20% increase in plasma current in experiments.
The rampdown phase of a tokamak pulse is difficult to simulate and often exacerbates multiple plasma instabilities. To reduce the risk of disrupting operations, we leverage advances in Scientific Machine Learning (SciML) to combine physics with data-driven models, developing a neural state-space model (NSSM) that predicts plasma dynamics during Tokamak à Configuration Variable (TCV) rampdowns. The NSSM efficiently learns dynamics from a modest dataset of 311 pulses with only five pulses in a reactor-relevant high-performance regime. The NSSM is parallelized across uncertainties, and reinforcement learning (RL) is applied to design trajectories that avoid instability limits. High-performance experiments at TCV show statistically significant improvements in relevant metrics. A predict-first experiment, increasing plasma current by 20% from baseline, demonstrates the NSSM's ability to make small extrapolations. The developed approach paves the way for designing tokamak controls with robustness to considerable uncertainty and demonstrates the relevance of SciML for fusion experiments.