Towards Fast and Accurate Predictions of Radio Frequency Power Deposition and Current Profile via Data-driven Modeling
Name
22ja004_full.pdf
Size
6.29 MB
Format
Adobe PDF
Checksum (MD5)
3f7681d99601af3b6a874540d71153ef
Author(s)
Wallace, Greg M.
Bai, Z.
Sadre, R.
Perciano, T.
Bertelli, N.
Shiraiwa, S.
Bethel, E.W.
Wright, John C.
Date Issued
April 2022
Journal
Journal of Plasma Physics
Publisher
Cambridge University Press
Abstract
Three machine learning techniques (multilayer perceptron, random forest, and Gaussian process) provide fast surrogate models for lower hybrid current drive (LHCD) simulations. A single GENRAY/CQL3D simulation without radial diffusion of fast elec- trons requires several minutes of wall-clock time to complete, which is acceptable for many purposes, but too slow for integrated modeling and real-time control applications. The machine learning models use a database of 16,000+ GENRAY/CQL3D simulations for training, validation, and testing. Latin hypercube sampling methods ensure that the database covers the range of 9 input parameters (ne0, Te0, Ip, Bt, R0, n||, Zeff , Vloop, PLHCD) with sufficient density in all regions of parameter space. The surrogate models reduce the inference time from minutes to ∼ms with high accuracy across the input parameter space.
Description
Submitted for publication in Journal of Plasma Physics
MIT Department
Massachusetts Institute of Technology. Plasma Science and Fusion Center
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1017/s0022377822000708
