Particle Transport During the L-H transition Using Machine Learning*

Density rises following the low to high confinement mode transition (L-H transition) differ based on isotopically dependent transport properties. In hydrogen plasmas on DIII-D compared to deuterium, the linear increase in electron density is almost 50% faster for hydrogen during the L-H transition. The confinement mode change of low confinement (L-mode) to high confinement (H-mode) is helpful for fusion in tokamak plasmas as it decreases turbulent transport near the edge resulting in increased electron densities and sharp density and temperature gradients. We present transport differences between hydrogen and deuterium plasmas by comparing transport coefficients obtained through an optimization that matches a convective-diffusive fluid model along with an exponential fueling model to experiment in both time and space. A Machine Learning algorithm trained using the optimization algorithm replicates and predicts transport coefficients. Using the neural network, we analyze how differences in species affect electron transport during the L-H transition.