Recently a new design of the CS for the next-generation experimental fusion device has been proposed. The CS system comprises six stacked subcoils, each split into two modules. The inner modules located in the high-field region use a YBCO high-temperature superconductor (HTS), while the outer low-field submodules utilize Nb₃Sn. Each HTS module is composed of 5 hex‑pancakes wound with the identical conductor based on the HTS Conductor-on-Round Core (CORC) strand concept. We performed simulations of normal operation of four selected HTS conductors during the whole CS current scenario, using the THEA code by CryoSoft. We took into account heat loads due to magnetization, coupling and eddy current AC losses. These were estimated based on the evolution of the magnetic field distribution within the CS modules which was obtained from an electromagnetic analysis using the 2D axisymmetric ANSYS Parametric Design Language (APDL) model. The coupling losses and eddy current losses were evaluated using the nτ values dependent on the magnetic flux density, which were computed for the considered conductor geometry with a dedicated electrostatic FEM model. The goal of the study was the estimation of the minimum temperature margin (ΔTmarg) in the considered conductors to verify the potential possibility of safe operation of the HTS modules. For all considered conductors the global minimum of ΔTmarg occurred in the first second of the current scenario (when the current and magnetic field were close to the maximum) and it was sufficiently large (above the 1.5 K criterion).
The accurate conversion of CAD models into MCNP geometry is fundamental to the design and development of nuclear fusion facilities such as ITER, DEMO, JT-60SA, and IFMIF-DONES, where radiation transport simulations play a critical role. As system geometries grow increasingly complex, advanced computational tools are essential for generating reliable MCNP models. This study evaluates the performance of the GEOUNED code and RadModeling tools for CAD-to-MCNP conversion. The CAD model of the Vacuum Ultra Violet (VUV) spectrometer, developed to monitor the divertor region of JT-60SA, diagnostic system for JT-60SA has been simplified and processed using these tools. The validity of the resulting MCNP model has been confirmed through stochastic volume calculations and benchmarked against results obtained using the SuperMC code. Preliminary nuclear response simulations, including neutron and proton fluxes, total dose, total heating, and neutron flux difference (%), have been performed for the integrated JT-60SA model with VUV diagnostic, demonstrating the effectiveness of the proposed conversion workflow.
To improve the computational efficiency and statistical reliability of deep-penetration neutron transport calculations in complex fusion geometries, a global variance reduction (VR) method based on an unstructured mesh (UM) importance map was developed in RMC. A mesh-wise importance map was generated from the neutron flux distribution and employed to perform particle splitting and Russian roulette during transport. The proposed method was verified using a simplified half-tokamak UM model. Compared with an analog calculation using 5 × 109 particle histories, the VR calculation using only 1 × 107 particle histories significantly improved the statistical quality in deep-penetration regions while preserving statistical unbiasedness, achieving an overall computational efficiency improvement of approximately 77.8 times. Furthermore, an interval-based importance merging strategy further increased the computational efficiency to approximately 99.1 times. To reduce the dependence on high-particle-number reference calculations, an iterative importance map generation strategy based on density reduction was also proposed, enabling effective importance maps to be gradually constructed and optimized using only low-particle-number calculations. The proposed method extends the applicability of RMC by enabling global importance-based VR on UM, providing a promising approach for deep-penetration fusion neutronics calculations and other complex radiation shielding analyses.
DTT (Divertor Tokamak Test facility) is a superconducting tokamak currently under construction at the ENEA Research Center in Frascati. As a DEMO-relevant device, DTT is aimed at addressing the problem of heat exhaust, which is critical for future fusion reactors.Vacuum pumping is essential to ensure appropriate vacuum levels across subsystems, providing a clean environment for plasma formation and control plasma density during the pulse phase.This paper outlines the preliminary design of the vacuum pumping system, considering the operational and integration requirements with interfacing systems.DTT pumping infrastructure comprises three main subsystems - divertor, vacuum vessel and cryostat - combining cryopumps, turbomolecular pumps, roots and screw pumps. The present work focuses on the vacuum vessel and cryostat pumping systems with the aim of providing the required vacuum levels for both environments under various operating conditions and with specific arrangements aimed at maximizing operational uptime. The regeneration of the divertor cryopumps is also studied considering inventory limitations and hydrogen safety constraints.Key aspects addressed include target pumping speeds for cleaning procedures, equipment compatibility with pumped gases, and magnetic-field-related constraints.Preliminary performance is presented in terms of pump-down times and achievable ultimate pressures for the vessel and cryostat.
Next-generation tokamak fusion devices span broad orders of magnitude in parameter spaces from current devices, requiring reduced models to accelerate the engineering design and scenario optimization of the plasma. Additionally, the strict performance targets of advanced tokamaks necessitate precise and efficient control over the plasma current profile using methods such as off-axis electron cyclotron current drive (ECCD) to prevent the formation of dangerous instabilities, such as Neoclassical Tearing Modes (NTMs). NTMs pose a substantial threat in future devices, due to their ability to significantly degrade core confinement and lead to disruptions which have the potential to cause catastrophic damage to plasma-facing components. Ray-tracing codes, such as TORAY, offer considerable advantages in runtime over high-fidelity RF solvers for obtaining electron cyclotron (EC) heating and current drive (H/CD) profiles. Yet, large-scale parameter scans for H/CD optimization are still time-consuming due to the high-dimensional input space and are still far too slow for real-time control applications. This work expands upon previous efforts (Irvin et al., 2026) to develop a surrogate model by leveraging machine learning via Gaussian process regression (GPR) to reliably predict key EC H/CD profile characteristics with uncertainty quantification (UQ) for cases within a broad parameter space including future reactor-scale devices and many present experimental devices, in orders of magnitude less time than TORAY.
Lasers and compact filter polychromators were developed for real-time Thomson scattering (TS) measurements in tokamak plasmas under intense background light. The multi-wavelength laser system is comprises a 3 ns / 2 J / 50 Hz 1064 nm Nd:YAG laser, a 1047 nm Nd:YLF laser and a 946 nm Nd:YAG. The long-term stability of 1064 nm and 1047 nm lasers was experimentally validated over 108 shorts. Development of 946 nm laser is still in progress, and its design implementation is currently under discussion. The laser amplifiers employ stimulated Brillouin scattering phase-conjugate (SBS-PC) mirrors to compensate for thermally induced wavefront distortions and to reduce the risk of unintended beam focusing onto the vacuum window. A compact nine-channel filter polychromator integrates 5 GHz waveform digitizers and embedded signal processing. Recording of the short laser pulse shape allows the useful TS signal to be separated from stray-light artifacts with a delay of only 10 ns in between. Testing on the Globus-M2 tokamak demonstrated the temporal separation of TS signals from delayed stray light, electron-temperature measurements down to 1 eV, and an average real-time processing latency of 0.72 ms. These results demonstrate the applicability of the developed equipment for plasma diagnostics and its utility in a feedback system for plasma control.
In a tokamak magnetic confinement nuclear fusion device, real-time precise feedback control of the plasma current density is critical for maintaining steady-state operation. The Motional Stark Effect (MSE) diagnostic, which measures the polarization angle variations of spectral lines excited by Neutral Beam Injection (NBI), directly reflects the evolution of the plasma current density and provides essential data for plasma equilibrium and control. To meet the signal processing requirements of MSE diagnostic in the Experimental Advanced Superconducting Tokamak (EAST), a real-time multi-harmonic signal processing system based on Field Programmable Gate Array and Advanced Reduced Instruction Set Computer Machine (FPGA+ARM) architecture is developed. This system employs dual phase detection technology and the Coordinate Rotation Digital Computer (CORDIC) algorithm to synchronously demodulate the amplitudes of two target harmonic signals and directly compute the polarization angle from the relationship between the amplitude and polarization angle. The system offers high integration, rapid processing speed, and the direct output of polarization angle without additional signal processing. Field tests in the EAST demonstrated a polarization angle error of less than 1.06%, with a total signal processing cycle of approximately 205 μs, satisfying the diagnostic requirements for accuracy and real-time processing.
In this study, we developed a finite element method (FEM)-based equilibrium reconstruction code for the KSTAR tokamak to analyze plasma behavior during highly transient phases, including early startup, pre-disruption, and vertical displacement events (VDEs), where conventional equilibrium fitting (EFIT) reconstructions often fail to converge. The code minimizes a weighted least-squares cost function fitting magnetic diagnostic signals, with the currents induced in the vacuum vessel and passive structures fitted simultaneously with the plasma current rather than prescribed through a separate eddy current circuit model. Validation against EFIT for stable limited and diverted discharges yielded plasma position agreement to within 20–40 mm, a level consistent with the intrinsic uncertainty of EFIT-based reconstruction during the current ramp-up and plasma shaping phase. Applied to the early startup phase, cross-validation with fast-frame camera imaging confirmed the code's ability to capture initial plasma breakdown well before EFIT becomes available. The reconstruction was further extended through the VDE and current quench phases, with the results cross-validated against halo current monitors, deuterium-alpha emission, and diamagnetic flux measurements, confirming the physical consistency of the reconstructed plasma evolution throughout the event. These results demonstrate the capability of the FEM-based reconstruction code to reliably track plasma behavior across a wide range of transient operational phases, and the framework can serve as a valuable tool for analyzing transient plasma phases in future tokamak devices.
EU-DEMO (European Demonstration Fusion Power Plant) is considered the successor to the experimental fusion reactor ITER (”The Way”). EU-DEMO will have to operate under significantly more demanding conditions, reaching the energy production of an industrial-scale plant, producing energy continuously. This case study examines the EU-DEMO divertor, a critical component of nuclear fusion reactors and one of the most significant engineering challenges to be addressed.The divertor has to handle extremely high heat fluxes, due to the discharge of exhausted ion particles, and it is subjected to intense neutron fluxes, while providing shielding to the vacuum vessel and the vacuum pumping system in order to reduce plasma contamination. The development of such a complex component requires the integration of multi-physics analysis, such as plasma or computational fluid dynamics (CFD) simulations, which are more computationally expensive, with EM/FEM analysis, which are generally less time-consuming. A BPMN-based (Business Process Model and Notation) model was developed to comprehensively map the multi-disciplinary phases of the divertor design process. This work highlights the differences in terms of time consumption among various simulation approaches, suggesting the exploration of alternative methodologies.This paper investigates the feasibility of integrating a pre-existing high-fidelity database of plasma physics simulations into the modeFRONTIER software, as a case study, to evaluate the capability of predicting the heat flux on divertor targets through mathematical models, thereby accelerating the assessment of different scenarios and reducing computational time and resources.
In nuclear fusion systems, the gyrotron is the source of electromagnetic waves, exploiting the electron maser instability to initialize and stabilize the fusion reaction. A simulator that predicts the system’s behaviour based on the fundamental physics is crucial for its control. This requires modelling each component by integrating the relevant physical phenomena for that component, and linking the resulting component models together. However, current models of fusion system components like the breeding blanket or the gyrotron are too computationally intensive for real-time simulations if high accuracy is sought and therefore high-fidelity and high-dimensional models are needed, as lumped approaches may not offer high enough accuracy for safety-critical applications such as monitoring and control. Reduced Order Models (ROM) can overcome this problem while maintaining accuracy. This paper focuses on the Magnetron Injection Gun (MIG), responsible for generating the electron beam by thermionic effect. It is essential to calculate the heat transfer within the MIG to predict the electron density on the emitter surface. This methodological paper studies the suitability of Dynamic Mode Decomposition (DMD) in capturing the non-linear dynamics and time-scales of the beam current emission considering a representative pulse provided by the FALCON facility. Results show the computational efficiency and capability of such technique in capturing temporal dynamics, performing spectral analysis, denoising data and forecasting features beyond the training dataset, positioning this paper as a foundation from which more complex models and ROM techniques can be considered.
