Multi-scale turbulence simulation suggesting improvement of electron heated plasma confinement

Abstract

Turbulent transport is a key physics process for confining magnetic fusion plasma. Recent theoretical and experimental studies of existing fusion experimental devices revealed the existence of cross-scale interactions between small (electron)-scale and large (ion)-scale turbulence. Since conventional turbulent transport modelling lacks cross-scale interactions, it should be clarified whether cross-scale interactions are needed to be considered in future experiments on burning plasma, whose high electron temperature is sustained with fusion-born alpha particle heating. Here, we present supercomputer simulations showing that electron-scale turbulence in high electron temperature plasma can affect the turbulent transport of not only electrons but also fuels and ash. Electron-scale turbulence disturbs the trajectories of resonant electrons responsible for ion-scale micro-instability and suppresses large-scale turbulent fluctuations. Simultaneously, ion-scale turbulent eddies also suppress electron-scale turbulence. These results indicate a mutually exclusive nature of turbulence with disparate scales. We demonstrate the possibility of reduced heat flux via cross-scale interactions.

Development of compact tokamak fusion reactor use cases to inform future transport studies

“…However, even in UCR-SS which has significantly more power flowing through the electrons, the short wavelength modes only contribute 25 % or less of the total thermal flux. Note that these results do not necessarily mean the shorter-wavelength instabilities are a priori negligible, as significant nonlinear cross-scale interactions have been observed in nonlinear multiscale gyrokinetic simulations (Howard et al 2014;Maeyama et al 2015;Howard et al 2016;Holland, Howard & Grierson 2017;Maeyama et al 2017;Bonanomi et al 2018;Howard et al 2021a;Maeyama et al 2022), and a key element of the TGLF SAT1 and SAT2 models is to capture these effects. Therefore, one goal for future studies should be to examine the strength and impact of such couplings at these (or similar) reactor-like parameters.…”

Multimodal convolutional neural networks for predicting evolution of gyrokinetic simulations

“…GKV has been extensively employed to investigate turbulent transport physics and quantify turbulent fluxes: finite‐ β dependence of electromagnetic turbulent transport, [ 7 ] the ion‐temperature‐gradient (ITG) driven turbulent tungsten transport, [ 8 ] and cross‐scale interaction between small‐ to large‐scale turbulence. [ 9–11 ]…”

“…Furthermore, in the studies described above, a multiscale gyrokinetic simulation simultaneously covering turbulence from low to high wavenumber spectra takes from tens of hours to weeks to complete, even when using tens of thousands of CPUs in a modern, powerful, large-scale supercomputer like Fugaku. Specifically, a GKV multiscale simulation shown in Reference [11] used 2560 nodes in Fugaku and took 250 hr, that is, 640 thousand nodehours, wherein each node in Fugaku had 48 computation cores. The vast amount of required computational resources hinders the systematic surveying of input parameter sets and from performing multiple cases for model validation activities.…”

“…GKV has been extensively employed to investigate turbulent transport physics and quantify turbulent fluxes: finite-𝛽 dependence of electromagnetic turbulent transport, [7] the ion-temperature-gradient (ITG) driven turbulent tungsten transport, [8] and cross-scale interaction between small-to large-scale turbulence. [9][10][11] Similar to other codes, GKV tracks the nonlinear evolution of a perturbed distribution function f at a single spatial location as an initial value problem by solving [6] 𝜕 f 𝜕t…”

Gyrokinetic Turbulent Transport Simulations on Steady Burning Condition in D-T-He Plasmas

“…There are several earlier works regarding the ITG driven turbulent transport of fuel hydrogen-isotope ions [1][2][3][4][5] and the light and heavy impurity transport [6,7], but the fuel and impurity ion transport has often been examined separately. The impacts of multi-ion species on the ion-and electron-scale microinstabilities have also been studied [8], and the reduction of turbulent heat flux through multi-scale interactions in the multi-ion species plasma has been revealed [9]. In addition, several recent experiments in the mixture [10][11][12] and/or the D-T burning plasmas [13] with dedicated controls of the ion density have revealed the significant isotopeauthor's e-mail: nakata.motoki@nifs.ac.jp Fig.…”