Engineering topics

Xu, Xue-Qiao

Publications and source records attributed to Xu, Xue-Qiao.

3 records

DOE OSTI2022

A BOUT++ extension for full annular tokamak edge MHD and turbulence simulations

For tokamak edge plasma simulation, a plasma simulation framework BOUT++ employs a dual coordinate system to simulate moderate-n and high-n plasma instability with reasonable computational cost, where n is the toroidal mode number. This coordinate system however limits the computational domain to the toroidal wedge (full torus divided into N parts in the toroidal direction) for computational efficiency and the use of flute-ordering approximation in the field solver calculating the flow potential from the vorticity which may not be valid for low-n modes. Improving numerical treatment of low-n modes is however indispensable to address simulations of low-n current-driven edge localized mode (ELM), ELM control by resonant magnetic perturbations (RMPs), edge turbulence with RMPs and so on. In this work, BOUT++ is extended to simulate the interplay between $n=0$, low-n and high-n plasma components in a full annular tokamak edge domain through hybrid modeling of the flow potential and the vorticity. Low-n modes of flow potential are calculated in an orthogonal flux surface coordinate and high-n modes in the dual coordinate system separately in Fourier space. Finally, the proposed scheme can capture an interplay between $n=1$ global modes and high-n turbulence during pedestal collapse in a full annular torus domain with a circular cross section.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2020

Machine learning surrogate models for Landau fluid closure

The first result of applying the machine/deep learning technique to the fluid closure problem is presented in this letter. As a start, three different types of neural networks (multilayer perceptron (MLP), convolutional neural network (CNN) and two-layer discrete Fourier transform (DFT) network) were constructed and trained to learn the well-known Hammett-Perkins Landau fluid closure in configuration space. We found that in order to train a well-preformed network, a minimum size of training data set is needed; MLP also requires a minimum number of neurons in the hidden layers equals to the degrees of freedom in Fourier space despite training data is fed in configuration space. Out of three models DFT performs the best for the clean data most likely due to the existence of nice Fourier expression for Hammett-Perkins closure but it is least robust with respect to input noise. Overall, with appropriate tuning and optimization, all three neural networks are able to accurately predict Hammett-Perkins closure and reproduce the inherit nonlocal feature, suggesting a promising path to calculate more sophisticated closures with the machine/deep learning technique.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2020

Achieving a robust grassy-ELM operation regime in CFETR

We have identified a robust grassy-edge localized mode (ELM) operation regime for future tokamak reactors. The regime exists within a pedestal top electron collisionality (ν*) window at high global poloidal beta (β p ). The existence of an upper ν* limit for grassy-ELMs is consistent with results previously reported in experiments (Oyama et al 2010 Nucl. Fusion 50 064014), while the existence of a lower ν* limit has not been reported previously. Using EPED and BOUT++, a theoretical model that quantitatively explains the physics of the grassy-ELMs within the window, which distinguishes them from the small mixed-ELMs at lower ν*, is presented for the first time. A peeling-ballooning stability boundary is obtained by scanning the operating density space. The change in density corresponds to a change in ν* that affects the pedestal bootstrap current. High β p leads to a strong Shafranov shift, which affects the flux surface averaged pressure drive. The two effects combine to create a peeling-dominated window in intermediate ν* buffered by ballooning-dominated regimes. Only the peeling-dominated regime shows a cyclic behavior in the perturbed pressure during the nonlinear simulation of an ELM crash, reminiscent of grassy-ELM dynamics. Similarly, the energy released across the separatrix is demonstrated to be significantly smaller. The quick recovery of the ELM crash is explainable by the rapid rise of a low n kink-peeling instability when the pedestal current I ped exceeds a threshold at high β p . It minimizes the excursion beyond marginal stability and is absent in the ballooning-dominated regime. Comparison with recent experiments over a range of β p and ν* strongly supports the physical picture proposed by the modeling.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗