284

Papers

9/9

Roadblocks Active

3

Connections

Injecting boron powder into the edge of the DIII-D tokamak at rates up to 12.7 mg/s progressively suppresses the violent plasma instabilities called ELMs, achieving ~300 ms ELM-free phases by decoupling two competing instability boundaries in a way that standard H-mode operation cannot reach. The boron cools the plasma edge and triggers low-frequency turbulence that limits how steep the pressure gradient gets before it would trigger an ELM, essentially keeping the plasma in a stable zone that previously required exotic configurations. Because ELMs are the dominant cause of intense heat pulses to the divertor — the component most at risk in ITER — demonstrating a powder-injection route to suppressing them is a practically scalable finding.

██████████ 1.0 elm-control Preprint

ReadSaveConnections

Nonlinear 3D MHD simulations using the NIMROD code model how a resistive wall mode (RWM) — a long-wavelength plasma kink driven by the finite conductivity of the surrounding metal wall — triggers full disruptions in a CFETR-class reactor (13.3 MA, 6.5 T) operating above its no-wall stability limit. The RWM rapidly stochastizes magnetic flux surfaces, causing a thermal quench followed by a current quench, and the growth rate is found to depend strongly on wall conductivity but weakly on plasma resistivity at high Lundquist number. Since CFETR is designed to operate in this high-beta advanced scenario, the result quantifies the disruption risk and the importance of active RWM feedback control for any reactor operating near maximum stored energy.

██████████ 0.9 plasma-disruption Preprint

ReadSaveConnections

At a critical plasma beta (ratio of plasma pressure to magnetic pressure), electromagnetic turbulence driven by ion temperature gradients (ITG) undergoes an abrupt transition from a well-behaved state stabilized by zonal flows to a strongly turbulent streamer-dominated state with much higher heat loss. The transition happens because magnetic and pressure-gradient stresses overwhelm the sheared flows that normally suppress turbulence — a mechanism identified by decomposing the stress tensor in nonlinear fluid simulations. This matters for reactor design because it implies there may be a hard confinement ceiling near a specific beta value, and crossing it could catastrophically degrade energy confinement before any disruption trigger is reached.

██████████ 0.9 turbulence-modeling Preprint

ReadSaveConnections

This paper replaces the most computationally expensive step in simulating how hydrogen isotopes move through tungsten — calculating energy barriers for each atomic hop using quantum-mechanical methods — with a three-stage deep learning pipeline, reducing calculation cost from thousands of CPU hours per barrier to effectively real-time prediction. The framework reproduces the known physics that hydrogen preferentially accumulates at grain boundaries in polycrystalline tungsten, which is where tritium retention and embrittlement nucleate. For fusion, this unlocks the ability to simulate tritium inventory and permeation through plasma-facing materials at reactor-relevant timescales, directly informing how long tungsten components remain safe before replacement.

██████████ 0.9 plasma-wall Preprint

ReadSaveConnections

Plasma turbulence generates local hot spots with non-Maxwellian particle velocity distributions — meaning more particles at very high energies than a thermal equilibrium plasma would have — which can boost fusion reaction rates above standard predictions. This study compares two collision models (BGK vs. Fokker-Planck) and finds the simpler BGK model overestimates the enhancement; the more physically accurate Fokker-Planck operator predicts a ~50% smaller boost. For inertial confinement fusion (ICF) targets, this means turbulence-driven reactivity enhancement is real but more modest than previously estimated, requiring more precise kinetic modeling in yield predictions.

██████████ 0.8 turbulence-modeling Preprint

ReadSaveConnections

This paper implements physically realistic wall boundary conditions into the PIXIE3D extended MHD simulation code — specifically accounting for the fact that real tokamak walls have finite electrical resistance that allows magnetic perturbations to slowly penetrate them. The Boundary Integral Method handles the mathematically singular integrals that arise in toroidal geometry, and the implementation is verified against analytical solutions including an ITER-geometry vertical displacement event (VDE). Accurate resistive wall physics in MHD codes is essential for predicting how fast disruptions evolve in ITER and future reactors, where the wall's electromagnetic response governs whether active control systems have enough time to intervene.

██████████ 0.8 plasma-disruption Preprint

ReadSaveConnections

Different coordinate systems used to describe tokamak magnetic geometry produce different Fourier coefficients for the same physical magnetic perturbation, which has caused inconsistencies in how researchers calculate resonant coupling between externally applied fields and internal plasma modes. This paper proves analytically — and verifies numerically using the GPEC code on a DIII-D equilibrium — that area-weighting the Fourier integrals with the square root of the flux-surface Jacobian makes the coupling matrix invariant to coordinate choice. The fix is important for ELM suppression and disruption avoidance calculations, where quantitatively wrong coupling coefficients can lead to miscalibrated coil current prescriptions for 3D field applications.

██████████ 0.8 elm-control Preprint

ReadSaveConnections

Stellarators avoid plasma disruptions by using twisted external coils instead of induced plasma current, but designing compact versions that still confine particles well (quasiaxisymmetric, or QA) is mathematically complex. This paper develops a perturbative analytical theory treating compact QA stellarators as small deviations from perfect axisymmetry, revealing that ridge-like features on flux surfaces concentrate field lines in a way that mimics divertor X-point geometry without requiring a rational magnetic surface at the edge. This has practical implications for how exhaust heat can be managed in compact stellarators, which are being actively pursued as disruption-free alternatives to tokamaks.

██████████ 0.7 divertor-thermal Preprint

ReadSaveConnections

Fast ions produced by neutral beam injection or fusion reactions can drive Alfvénic instabilities (like toroidal Alfvén eigenmodes) that expel them before they deposit their energy usefully — a key concern for achieving Q>1. This paper uses kinetic simulations to show that these instabilities can grow concurrently with the buildup of the fast-ion distribution that drives them, rather than only after the distribution is fully formed, and that near-marginal saturation is achievable even in this dynamic regime. The finding refines how reactor-relevant fast-ion transport should be modeled, particularly in transient phases of a discharge when neutral beams or alpha heating first turn on.

██████████ 0.7 q-engineering Preprint

ReadSaveConnections

Gyrotrons are the high-power microwave sources used to heat tokamak plasmas via electron cyclotron resonance — ITER requires 24 gyrotrons delivering 170 GHz at 1 MW each. This paper identifies the diocotron instability (a rotating charge-density wave in the electron cloud) as the cause of cyclical electron cloud collapse and reformation in the magnetron injection gun region of gyrotrons, with 3D FENNECS simulations matching experimental current measurements. Understanding and controlling this instability is relevant to improving gyrotron efficiency and beam quality, directly affecting how much heating power reaches the fusion plasma.

██████████ 0.6 q-engineering Preprint

ReadSaveConnections