Tokamak edge localized mode onset prediction with deep neural network and pedestal turbulence

Abstract

A neural network, BES-ELMnet, predicting a quasi-periodic disruptive eruption of the plasma energy and particles known as edge localized mode (ELM) onset is developed with observed pedestal turbulence from the beam emission spectroscopy system in D\rom{3}-D. BES-ELMnet has convolutional and fully-connected layers, taking two-dimensional plasma fluctuations with a temporal window of size 128\:$\mu$s and generating a scalar output which can be interpreted as a probability of the upcoming ELM onset. As approximately labelled inter-ELM broadband (15\:kHz\:$\leq f\leq$\:150\:kHz) fluctuations are given to the network, BES-ELMnet learns by itself ELM-related precursors arising before the onsets through supervised learning scheme. BES-ELMnet achieves the gradually increasing ELM onset probabilities between two consecutive ELMs during the inter-ELM phases and can forecast the first ELM onsets which occur after the high confinement mode transition. We further investigate the network generality in terms of the selected frequency band to ensure the use of BES-ELMnet for various operation regimes without changing the trained architecture. Therefore, our novel prediction method will enhance a proactive high confinement mode control of fusion-grade plasmas.

Real-time confinement regime detection in fusion plasmas with convolutional neural networks and high-bandwidth edge fluctuation measurements

“…This potential to extract meaningful physics from the data may be necessary as we look toward more nuanced plasma states and sub-regimes such as WP QH-mode. ( 2) Given (1), it is reasonable to expect 2D BES to hold more predictive power than the signals used in previous work, and we thereby anticipate the ability to utilize the BES data for many real-time downstream tasks in addition to confinement regime detection, such as confinement transition prediction, ELM prediction [51], and radial electric field calculation to name a few. We view this work as a first step toward such advanced physics-informed control operations.…”