Real-time virtual circuits for plasma shape control via neural network surrogates: dynamic validation in closed-loop simulations

Kamran Pentland, Alasdair Ross, Nicola C. Amorisco, Pedro Cavestany, Timothy Nunn, Adriano Agnello, George K. Holt, Charles Vincent · 2026

Read Real-time virtual circuits for plasma shape control via neural network surrogates: dynamic validation in closed-loop simulations by Kamran Pentland, Alasdair Ross, Nicola C. Amorisco, Pedro Cavestany, Timothy Nunn, Adriano Agnello, George K. Holt, Charles Vincent (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-00781.txt.

Introduction

Pentland and colleagues test neural-network emulators of virtual circuits for real-time tokamak control. Each circuit is a coil-current pattern that adjusts one plasma shape parameter on its own. In closed-loop simulations with the FreeGSNKE Pulse Design Tool, the emulators steered MAST Upgrade scenarios well and stayed robust under measurement uncertainty and varied update rates. The results support moving these real-time virtual circuits toward deployment in the MAST-U control system.

Catalog record

CREATORKamran Pentland, Alasdair Ross, Nicola C. Amorisco, Pedro Cavestany, Timothy Nunn, Adriano Agnello, George K. Holt, Charles Vincent
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE22601
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.00781

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Real-time virtual circuits for plasma shape control via neural network surrogates: dynamic validation in closed-loop simulations Kamran Pentland, Alasdair Ross, Nicola C. Amorisco, Pedro Cavestany, Timothy Nunn, Adriano Agnello, George K. Holt, Charles Vincent Abstract Reliable confinement and stable performance of tokamak fusion plasmas require accurate real-time magnetic shape control. A promising route to reduced latency and increased flexibility in plasma control systems (PCS) is to emulate physics-based controllers using neural networks. In prior work, we have demonstrated that virtual circuits (VCs), which define the poloidal field coil current vectors able to modify each plasma shape parameter independently, can be accurately emulated with neural network models trained on a large library of simulated Grad-Shafranov equilibria. This enables magnetic controllers to accurately adapt to evolving plasma equilibria, in contrast to pre-set VC schedules whose performance degrades upon departure from their reference equilibria. Here, we investigate the performance and robustness of these emulators in closed-loop simulations using the FreeGSNKE Pulse Design Tool (FPDT): a framework that couples the FreeGSNKE evolutive equilibrium solver with a virtual PCS. The FPDT models the coupling between controllers, plasma current and shape response, and actuator constraints. Using the emulated VCs within the FPDT, we demonstrate effective in-silico control of MAST Upgrade (MAST-U) plasma scenarios and show that the emulators are robust in the presence of input measurement uncertainty and under different update frequencies. These results establish the viability of neural network emulated VCs for closed-loop plasma shape control, representing a key step toward real-time deployment in the MAST-U PCS. I Introduction Accurate real-time magnetic control of tokamak fusion plasma position and shape is essential for maintaining reliable confinement and enabling access to high-performance operating regimes and advanced divertor configurations [ 1 , 2 ] . To this end, the plasma control system (PCS) regulates deviations in geometrical shape parameters (e.g. midplane radii, X-point locations, strikepoints—typically under the assumption of toroidal symmetry) from prescribed reference waveforms by computing the required adjustments to poloidal field (PF) coil currents. These adjustments are enacted in real-time via PF coil voltage actuation, with sub-millisecond latency [ 3 , 4 ] . Here, our focus is on calculating the adjustments in PF coil currents using virtual circuits (VCs) [ 5 , 6 ] . A VC matrix decouples the shape control problem around a specific plasma equilibrium, i.e. it determines the PF coil current changes required to move each shape parameter independently, by one physical unit. Building a VC matrix is not currently feasible in real time as it requires a linearisation of the Grad-Shafranov (GS) equilibrium [ 7 , 8 ] to obtain the sensitivities of the shape parameters. In classical plasma scenario design, VCs for a small number of target GS equilibria are calculated offline, to be used throughout a plasma discharge in a pre-set, phase-segmented manner. As their validity is only local, the accuracy of a VC matrix rapidly degrades as the plasma equilibrium evolves away from the reference state. As a result, the pre-calculated VCs may lose their effectiveness in producing orthogonal plasma responses, causing coil current adjustments to interfere with one another and potentially drive unnecessary stress on the PF coil power supplies. Prior work has demonstrated VCs can be accurately emulated using feedforward neural networks (NNs), based on real time tokamak diagnostic measurements [ 9 , 10 ] . With inference times on the order of milliseconds, NN emulators enable state-aware decoupling of plasma shape control by providing VCs that are continuously updated with the evolving plasma state. This can improve shape controller responsiveness and accuracy, particularly for complex shape displacements that would otherwise require segmented control strategies with a large number of pre-set time-phases and pre-computed VCs at the corresponding equilibrium waypoints. This contribution complements an increasing body of work related to plasma control using NNs [ 11 , 12 , 13 , 14 ] and reinforcement learning [ 15 , 16 , 17 , 18 , 19 , 20 ] . In this paper, we validate the performance and test the robustness of emulated VCs for real-time shape control by deploying the NN models presented by [ 10 ] within closed-loop plasma equilibrium simulations using the open-source FreeGSNKE Pulse Design Tool (FPDT) [ 21 , 22 ] . Previously validated against MAST Upgrade (MAST-U) discharges, the coupled evolutive equilibrium solver and virtual PCS within the FPDT provide an ideal test bed for assessing the emulated VCs in silico, in preparation for deployment. This ‘dynamic validation’ complements previous work on purely ‘static validation’ (more details in Section II ), by accounting for the evolving currents in passive conducting structures and the interaction with other controllers in the PCS, both of which are neglected in static tests. It also enables monitoring of the plasma and coil responses to frequently changing VCs over time, providing a more realistic assessment of their impact on closed-loop plasma control. In Section II , we briefly summarise the process by which the emulated VCs are generated and how they were statically validated. We then explain, in Section III , how the emulated VCs are seamlessly integrated into the virtual PCS within the FPDT. In Section IV , we perform robustness tests and verify that emulated VCs can successfully control MAST-U plasma discharges in silico. Fig. 1: Poloidal cross-section of a MAST-U plasma (shot 53278 t=0.5 s) equilibrium generated using FreeGSNKE. Displayed are the contours of total poloidal flux (yellow/green), the separatrix (red), and the shape parameters (labelled black dots). Also shown are the PF coils (blue), vessel structures (grey), and first wall outline (black). II Emulated virtual circuits In the following, we adopt cylindrical coordinates (R,\phi,Z) under toroidal ( \phi ) axisymmetry, with R denoting the major radius and Z the vertical distance from the tokamak midplane. II-A Definitions and notation For any given GS equilibrium, we describe its core shape and divertor leg characteristics with a set of geometric shape parameters \bm{P}\in\mathbb{R}^{m} . Plasma equilibria are fully defined by the currents in any active and passively conducting metals \bm{I}_{m}=(\bm{I}_{\text{act}},\bm{I}_{\text{pas}}) ; the total plasma current I_{p} ; and the plasma current density profile parameters \bm{\theta} . Therefore, we have \bm{P}=\bm{P}(\bm{I}_{m},I_{p},\bm{\theta}) . Note, however, that the passive currents \bm{I}_{\text{pas}} are not actively controlled, and only the active PF coil currents \bm{I}_{\text{act}}\in\mathbb{R}^{n} ( m\leqslant n ) can in fact be modulated. For this work, \bm{P} contains the inner and outer midplane radii (R_{\text{in}},R_{\text{out}}) ; the lower X-point position (R_{\text{X}},Z_{\text{X}}) ; the “gap” between the plasma core and a specific point on the wall of the main chamber S_{\text{gap}} (i.e. squareness); the lower outboard strikepoint R_{\text{s}} ; and the radial intersection of lower divertor leg and a virtual line crossing the divertor entrance R_{\text{nose}} (refer to Fig. 1 and [ 10 ] for more details). The VC matrix \mathcal{V} is defined as the pseudoinverse of the sensitivity matrix \mathcal{S} , which itself is the Jacobian of shape parameters with respect to PF coil currents 1 1 1 Note that Ohmic coils are typically not included in VC calculations as their primary purpose is to modulate I_{p} via a different controller. for a specific GS equilibrium: \displaystyle\mathcal{V}=(\mathcal{S}^{\intercal}\mathcal{S})^{-1}\mathcal{S}^{\intercal},\quad\text{}\quad\mathcal{S}=\frac{\partial\bm{P}}{\partial\bm{I}_{\text{act}}}\Bigg|_{(\bm{I}_{m},I_{p},\bm{\theta})}. (1) The columns of \mathcal{V} represent the vectors of PF coil current changes required to modify a shape parameter

Read the complete text