The method of using neural networks (NNs) for turbulent transport prediction in a simplified model of tokamak plasmas is explored. The NNs are trained on a database obtained via test-particle simulations of a transport model in the slab-geometrical approximation. It consists of a five-dimensional input of transport model parameters, and the radial diffusion coefficient as output. The NNs display fast and efficient convergence, a validation error below 2, and predictions in excellent agreement with the real data, obtained orders of magnitude faster than test-particle simulations. In comparison to a spline interpolation, the NN outperforms, exhibiting better predicting and extrapolating capabilities. We demonstrate the preciseness and efficiency of this method as a proof-of-concept, establishing a promising approach for future, more comprehensive research on the use of NNs for transport predictions in tokamak plasmas.
Design of HL-2A plasma position predictive model based on deep learning