2025 – Enhanced RMP Hysteresis Achievement via omegaE based adaptive RMP control via real time turbulence analysis

Enhanced RMP Hysteresis Achievement via omegaE based adaptive RMP control via real time turbulence analysis

2025 Research Campaign, Transient Control

Purpose of Experiment

The main goal of this experiment is to achieve stable suppression of edge localized modes (ELMs)—harmful bursts of energy at the edge of fusion plasmas—while minimizing the negative impact on plasma performance. Specifically, this research investigates an advanced method known as Feedback Adaptive Resonant Magnetic Perturbation (RMP) Control, enhanced by real-time turbulence analysis through machine learning (ML). The core hypothesis is that controlling the rotation of the plasma at the edge of the device will significantly improve the effectiveness and robustness of ELM suppression, allowing the maintenance of high plasma confinement even at low RMP strengths. The experiment utilizes real-time data from the beam emission spectroscopy (BES) system, coupled with advanced ML algorithms, to continuously monitor and optimize plasma rotation and turbulence conditions. By detecting early indicators of reduced plasma rotation—which facilitates better penetration of magnetic perturbations—the adaptive controller can proactively maintain optimal conditions, avoiding the loss of ELM suppression and the associated confinement degradation. This work addresses a crucial challenge for future large-scale fusion reactors like ITER, where losing ELM suppression even briefly could result in significant damage to reactor components. By demonstrating a robust, predictive control strategy, this experiment offers practical solutions for improving fusion plasma performance and reliability, paving the way for more efficient and stable operation in next-generation fusion energy systems.

Experimental Approach

Step 1: Establish Standard RMP ELM Suppression (3-5 shots) We begin by loading a reference scenario based on shot #190738, employing a standard n=3 RMP configuration. This includes restoring the reference shot with an initial ELMy phase, followed by ramping up and maintaining the RMP current until ELM suppression is observed. This will involve identifying optimal plasma conditions, specifically density, safety factor (q95), and RMP current needed to sustain ELM suppression. Concurrently, the initial velocity spatial profile at the pedestal top is analyzed using the v–Net ML method via beam emission spectroscopy (BES). This analysis focuses on establishing baseline velocity evolution during the RMP current flat top and identifying the relevant BES columns corresponding to the pedestal top. Once suppression is achieved, we will use feedforward RMP control to determine an optimal velocity threshold through analyzing its evolution during two RMP ramp-down phases, providing critical information on response time and RMP current values associated with entry and loss of suppression. Step 2: Implement v-Based Adaptive RMP Control Scheme (4-6 shots) Utilizing the optimal velocity threshold identified in Step 1, we will implement the velocity-based adaptive RMP control scheme with target discharges featuring two ramp-down phases. In this adaptive scheme, the controller triggers a RMP current ramp-up whenever velocity exceeds the set threshold, updating the threshold based on the observed velocity at suppression loss. Simultaneously, the ELM–Net ML system will operate in the background to determine the optimal probability threshold for predicting suppression loss. If time permits, we will activate ELM–Net as the primary controller based on these predictions. Step 3: Optimization of Plasma Conditions for Enhanced RMP Hysteresis (2-4 shots) Finally, additional shots will be dedicated to further optimizing plasma conditions to enhance the RMP hysteresis effect. This optimization includes performing a safety factor (q95) scan to refine the alignment of rational surfaces with the pedestal top and conducting torque scans in increments of 0.5 Nm. These adjustments aim to further improve plasma confinement and robustness of the ELM suppression achieved through adaptive RMP control.

Interested in a behind-the-scenes look at DIII-D? Join us for a virtual OR in-person tour during Fusion Energy Week (May 5-9)! Sign up for a tour here.

Got it!
X