Recent Publications

An explicit algorithm in the quasi-static approximate PIC program QPAD

Rong Tang, Weiming An, Hainan Wang, Weiyu Meng, Zhihao Xu, Fei Li, Thamine Dalichaouch, Viktor K Decyk, Warren B Mori

Plasma wake field acceleration uses a drive particle beam or an intense laser to excite a wake field in the plasma and uses the wake field to accelerate another bunch of particles. Due to the acceleration gradients as high as 10 GeV/m or even higher, it is expected to become a key acceleration technology for future high energy colliders and X-ray free electron lasers. In response to the need for simulating the plasma wake field acceleration, large-scale parallel computing programs such as QuickPIC and QPAD have been developed. These programs use a quasi-static approximation PIC algorithm and QPAD additionally applies the azimuthal Fourier decomposition. Currently, both QuickPIC and QPAD use a predictor-corrector method to solve the electromagnetic field equations under the quasi-static approximation. In certain cases, multiple iterations are required to achieve convergence. In this work, we propose a new explicit algorithm in QPAD that is based on the azimuthal Fourier decomposition. By starting from the motion equation of each particle, we analytically extract the transverse magnetic field contributions out of the longitudinal derivative of the transverse current, thereby deriving an explicit equation for the transverse magnetic field. The resulting formulae introduce evolving coefficients that couple different azimuthal modes, which leads to a set of global linear equations incorporating all retained modes. The comparison of the simulation results show excellent agreement between the explicit algorithm and the predictor-corrector method.

Forecasting the first edge localized mode (ELM) after LH-transition with a neural network trained on Doppler Backscattering data from DIII-D

Nathan Qi Xuan Teo, Kshitish Kumar Barada, Valerian Hongjie Hall-Chen, Lin Gu, Terry L Rhodes

In H-mode tokamak and stellarator plasmas, edge localized modes (ELMs) lead to the expulsion of heat and particles beyond the edge transport barrier. ELMs cause a loss of energy and have the potential to damage the divertor and other plasma facing components, which motivates efforts to forecast such events to work alongside mitigation systems. In this paper, we use the Doppler backscattering (DBS) diagnostic data as input to train a neural network model, adapted from DeepHit [Lee et al., Deephit, AAAI 2018], to forecast the first ELM crash of H-mode discharges in DIII-D. The model takes 50 ms of DBS spectrogram data and predicts the probability of an ELM crash occurring within set time windows. Training and testing on shots found in the DIII-D database, we find the initial results promising, with the model reliably forecasting the first ELM 100 ms before it occurs. This successful proof-of-concept lays a strong foundation for a predictive tool that can deploy ELM-mitigation techniques before an ELM crash occurs. Future work will expand the training set with carefully selected shots and refine the neural network architecture to improve model robustness to noise and data variation.

Investigation of the magnetic flux pumping effect in MAST Upgrade

Sam Blackmore, Christopher J Ham, Daniele Brunetti, Clive A Michael, Bhavin S Patel, Koki Imada, Lucy Kogan, cameron Olde, Fulvio Auriemma, Chiara Piron, et al.

Stationary safety factor q profiles with qmin ≈ 1 and without sawtooth activity have been measured for the first time in MAST Upgrade (MAST-U) plasmas, indicative of the magnetic flux pumping mechanism. This anomalous current re-distribution phenomenon is observed in MAST-U plasmas with 2/1 tearing MHD instabilities. Simulations using the ideal MHD code MISHKA indicate that q profiles, measured using the motional Stark effect (MSE) diagnostic, with zero magnetic shear are also ideally unstable to n = 1, m = 1 MHD modes. Database analysis of MAST-U pulses demonstrate overlap in βN ≈ 2 in pulses which exhibit magnetic flux pumping and sawtooth activity. The operational space for robust access to the flux pumping regime on MAST-U is characterised by operating at q 95 > 6 and β p ≈ 0.6.

Exploration of an intrinsic low-collisionality, high performance, grassy ELM regime in the DIII-D tokamak

Zeyu Li, H Q Wang, Xi Chen, Xueqiao Xu, Rongjie Hong, Filipp Khabanov, V. S. Chan, Patrick H Diamond, Brian S Victor, Nami Li, et al.

Recent DIII-D tokamak experiments have demonstrated the integration of intrinsic low-collisionality grassy edge-localized mode (ELM) regime with high-performance hybrid core scenario, possibly offering a core-edge compatible solution for ITER and future fusion reactors. This regime features grassy ELMs (with ELM energy loss over pedestal stored energy <2%) at ITER-relevant pedestal top collisionality (ν_e^*~0.1), ITER similar shape, while maintaining high core confinement (H_98y2~1.5) in non-inductive hybrid scenarios. A small ELM-focused database is constructed to investigate the parametric dependence of the small/grassy ELM regimes. Access and sustainment of this regime appear to be favored by high poloidal beta (β_p>1.5), higher ratio of separatrix-to-pedestal density (n_(e,sep)⁄n_(e,ped) >0.4) and low pedestal top collisionality (υ_(e,ped)^*=0.1-0.4). In this scenario, the pedestal width exceeds the prediction from EPED–KBM scaling, consistent with the expectation of a turbulence-limited pedestal. Linear modeling using ELITE indicates that this grassy ELM regime is along the peeling boundary. Relative to a large-ELM phase, the grassy-ELM phase exhibits a broader inner-target heat-flux width and a substantially reduced ELM-induced transient heat-flux increment. These results motivate further evaluation of low-collisionality grassy ELMs as a potentially reactor-relevant operating regime, while full-duration sustainment and compatibility with divertor detachment remain unresolved.

Stochastic Cluster Dynamics Model for Simulating the Irradiation Response of Austenitic Stainless Steels in the Athena Fusion Pilot Plant

Sicong He, Francisco Ogando, Silvia De Dea, Susana Reyes, Conner Galloway, Jaime Marian

Neutronic Analyses for the Design of the Chamber of Xcimer’s Athena Fusion Pilot Plant

Francisco Ogando, Robert Earley, Wayne Meier, Rene Raffray, Patrick Sauvan, Jaime Marian, Susana Reyes, Conner Galloway

Multi-field turbulence and transport barrier measurements and validation of predictive codes for high-performance, negative triangularity ELM-free DIII-D plasmas

Guiding Wang, Terry L Rhodes, Julius Damba, Rongjie Hong, William A Peebles, Quinn Pratt, L Zeng, Max E Austin, Kathreen E Thome

A recent DIII-D campaign demonstrated high confinement, ELM-free plasmas in strongly shaped, diverted negative triangularity (NT) configurations. This paper presents new multi-field turbulence and flow measurements near the edge to understand the physics of edge transport of these plasmas. Correlation electron cyclotron emission and Doppler backscattering measurements provide electron-temperature fluctuation levels, density-fluctuation levels, radial correlation lengths, and poloidal turbulence velocity profiles. The measurements show a modest edge electron temperature/pressure pedestal, a localized poloidal-velocity well near the pedestal-top region, and the peaking of both temperature and density fluctuation levels near the region of minimum poloidal velocity shear, consistent with edge-transport-barrier formation. Linear TGLF calculations indicate a transition from predominantly ITG-like modes in the inner region to TEM-like modes near the pedestal top, approximately coincident with the observed increase in fluctuation levels. These results provide new quantitative constraints for testing predictive models of turbulence and transport in high-performance NT tokamak plasmas.

Pre-conceptual design of compact-fusion prototypical neutron source target for optimized fusion materials testing

Jiankai Yu, Nesrin Cetiner, Pei Biorn-Hansen, Brian D Wirth, Jaime Marian, Ethan Peterson, Daniel Winklehner, Lance Snead

This work describes a design study of a cyclotron-driven Fusion Prototypic Neutron Source (FPNS) target optimized to meet US specifications for a fusion materials irradiation target. The analysis supports the use of multiple compact deuterium cyclotrons beams at a total current under 40 mA in the range of 35 MeV to 60 MeV converging on flowing annulus of lithium surrounding the sample target. The sample target region is modeled as miniaturized tensile and Charpy bend-bar specimens located within the central region of the annulus. A specific focus of this work is to provide a highly uniform flux trap of fusion-relevant neutrons maximizing the average neutron damage per unit of beam current. This design achieves the desired sample volume (>50 cm3) and damage (>10 dpa/yr) metrics with damage gradient (<40%) within the sample target region. The proposed design is modeled using the neutronic and transmutation analysis codes, PHITS [1] and FISPACT [2], with the assistance of a customized coupling computation platform. Sensitivity analyses are presented investigating the impact of the deuteron beam (height and energy), dimension of the various system elements , and the chosen nuclear data library (deuteron library, neutron library, Damage library) on the ultimate figure of merits including damage dose rate (DPA), gas production. Additionally, a complete transmutation calculation has been carried out on relevant fusion materials: Eurofer97, Silicon Carbide, Alloy 316, V-4Cr-4Ti, and pure tungsten, to ensure relevance of this C-FPNS elemental transmutation spectrum to that of a fusion reactor, especially for high deuterium energies. Implications to materials response of using deuterium energies in the 35-60 MeV range as compared to self-same materials exposed to a true fusion reactor neutron spectrum are presented.

GPU-MPI parallelisation for QuickPIC, a three-dimensional quasi-static particle-in-cell algorithm

Yueran Tian, Yueluo Wang, Thamine Dalichaouch, Viktor Decyk, Frank S. Tsung, Warren B. Mori, Weiming An

QuickPIC is a quasi-static PIC program for simulating plasma wakefield acceleration. It supports a hybrid parallel simulation on multiple central processing units via OpenMP and message passing interface (MPI). We now present a version of QuickPIC that runs on graphics processing units (GPU), with MPI for communication between different GPUs, based on the UCLA Parallel Partice-In-Cell framework. By rewriting relevant Fortran77 functions with CUDA C, we port the two-dimensional particle and field computations onto GPU platforms. QuickPIC-GPU supports simulation with a single GPU for up to a resolution of 1024 cubed 1024 3 $1024^3$ and 25 particles per cell to avoid time spent on cross-processor communications, and simulation with multiple GPUs for larger problem sizes.

Reduced fast-ion transport calculations of m = n = 1 fishbone-like instabilities in MAST-U

H. H. Wong, P. J. Bonofiglo, M. Podesta, C. A. Michael, S. Thomas, D. Dunai, A. R. Field, K. G. McClements, M. Cecconello, N. A. Crocker, et al.

Fast-ion transport associated with an m=n=1 fishbone-like burst in MAST-U discharge 47128 is investigated using a reduced guiding-center-based transport model (ORBIT-Kick) constrained by multi-diagnostic measurements. The two-dimensional beam-emission spectroscopy system provides measurements of the core poloidal mode structure and fluctuation amplitude, while EFIT++ reconstructions constrained by the motional Stark effect diagnostic indicate a flat q-profile with q0>1, indicating the absence of a resonant q=1 surface and supporting a pressure-driven infernal-mode interpretation. Analytic m=n=1 displacement profiles consistent with the measured core mode structure and equilibrium constraints are used as the mode structure inputs to ORBIT-Kick. The calculations show that the dominant resonances occur between the mode and co-passing fast ions, producing redistribution localized near the magnetic axis. Synthetic neutron camera signals from TRANSP-Kick recover up to 90% of the experimentally observed neutron deficit at the time of peak mode amplitude, indicating that the measured m=n=1 mode is a dominant contributor to core fast-ion transport. However, the synthetic neutron signals recover rapidly, whereas the measured neutron emission continues to decrease after the peak amplitude. The remaining discrepancy may arise from contributions not included in the present single-harmonic model, including higher-m and higher-n harmonics, multi-harmonic interactions, and additional transport mechanisms, motivating future diagnostic development and modeling efforts to resolve and incorporate these additional contributions.

An explicit algorithm for the three-dimensional quasi-static particle-in-cell code: QuickPIC

Hainan Wang, Weiming An, Rong Tang, Weiyu Meng, Zhihao Xu, Fei Li, Thamine N. Dalichaouch, Viktor K. Decyk, Warren B. Mori

The three-dimensional quasi-static particle-in-cell code QuickPIC is widely used for simulating plasma-based acceleration but relies on predictor-corrector methods for solving electromagnetic fields. This paper presents an explicit algorithm for QuickPIC. We derive field equations under the quasi-static approximation that incorporate plasma ion contributions and provide an expression for the ξ derivative of the transverse plasma current. To support this explicit formulation, we introduce a finite difference field solver to calculate the transverse magnetic field and adopt a non-staggered particle pusher scheme, thereby eliminating the reliance on predictor–corrector methods. Systematic benchmarks across linear, nonlinear regimes, and positron-driven regimes as well as hosing instability scenarios demonstrate that the explicit algorithm achieves excellent agreement with the standard iterative QuickPIC. These results validate the physical fidelity and numerical stability of the proposed method in simulating complex beam-plasma interactions.

Suppression of ion temperature gradient modes by Alfvén activity above a drive threshold in DIII-D

X. D. Du, W. W. Heidbrink, Z. Yan, P. H. Diamond, G. R. McKee, L. Schmitz, R. Hong, M. A. Van Zeeland, K. J. Callahan, H. Q. Wang, et al.

A recent study demonstrates that the suppression of ion temperature gradient (ITG) modes can occur during the nonlinear evolution of toroidicity-induced Alfvén eigenmodes (TAEs) in the DIII-D tokamak [Du et al., Phys. Rev. Lett. 135, 265101 (2025)]. In that work, ITG was suppressed by the formation of a narrow, TAE-induced shear flow layer, whose shearing rate exceeded the ITG decorrelation rate. The shear flow arises from an imbalance between Reynolds and Maxwell stress forces, as the TAE departs from the conventional shear Alfvén wave polarization. This follow-up paper systematically identifies the plasma conditions required for robust ITG suppression through a series of comparative experiments. The results show that TAEs routinely suppress ITG turbulence in plasmas with higher local safety factor (q), elevated fast-ion beta, and larger populations of fast ions on passing orbits, i.e., conditions consistent with the substantial TAE drive. Database analysis further reveals the existence of a threshold in TAE drive for ITG suppression. That is, once the drive exceeds a critical value within a favorable q window, the system undergoes the nonlinear bifurcation process, characterized by a sharp increase in TAE saturated amplitudes at a nearly fixed fast ion drive, suppression of ITG turbulence, and the formation of an internal thermal transport barrier at the localized radii.

A Doppler backscattering diagnostic for the EXL-50U spherical tokamak: Plasma considerations and preliminary quasioptical design

Ying Hao Matthew Liang, Valerian Hall-Chen, Terry Rhodes, Yumin Wang, Yihang Zhao