Hostname: page-component-8448b6f56d-tj2md Total loading time: 0 Render date: 2024-04-23T19:15:02.531Z Has data issue: false hasContentIssue false
  • English
  • Français

Encoder–decoder neural network for solving the nonlinear Fokker–Planck–Landau collision operator in XGC

Published online by Cambridge University Press:  24 March 2021

M. A. Miller Open the ORCID record for M. A. Miller [Opens in a new window] ,
R. M. Churchill Open the ORCID record for R. M. Churchill [Opens in a new window] ,
A. Dener Open the ORCID record for A. Dener [Opens in a new window] ,
C. S. Chang Open the ORCID record for C. S. Chang [Opens in a new window] ,
T. Munson Open the ORCID record for T. Munson [Opens in a new window]  and
R. Hager Open the ORCID record for R. Hager [Opens in a new window]
M. A. Miller*
Affiliation:
Plasma Physics Laboratory, 100 Stellarator Road, Princeton, NJ08540, USA Applied Physics and Applied Mathematics Department, Columbia University, New York, NY10027, USA Department of Nuclear Science & Engineering, Massachusetts Institute of Technology, Cambridge, MA02139, USA
R. M. Churchill
Affiliation:
Plasma Physics Laboratory, 100 Stellarator Road, Princeton, NJ08540, USA
A. Dener
Affiliation:
Argonne National Laboratory, 9700 S Cass Avenue, Lemont, IL60439, USA
C. S. Chang
Affiliation:
Plasma Physics Laboratory, 100 Stellarator Road, Princeton, NJ08540, USA
T. Munson
Affiliation:
Argonne National Laboratory, 9700 S Cass Avenue, Lemont, IL60439, USA
R. Hager
Affiliation:
Plasma Physics Laboratory, 100 Stellarator Road, Princeton, NJ08540, USA
*
†Email address for correspondence: mam2510@columbia.edu
Get access Rights & Permissions [Opens in a new window]

Abstract

An encoder–decoder neural network has been used to examine the possibility for acceleration of a partial integro-differential equation, the Fokker–Planck–Landau collision operator. This is part of the governing equation in the massively parallel particle-in-cell code XGC, which is used to study turbulence in fusion energy devices. The neural network emphasizes physics-inspired learning, where it is taught to respect physical conservation constraints of the collision operator by including them in the training loss, along with the $\ell _2$ loss. In particular, network architectures used for the computer vision task of semantic segmentation have been used for training. A penalization method is used to enforce the ‘soft’ constraints of the system and integrate error in the conservation properties into the loss function. During training, quantities representing the particle density, momentum and energy for all species of the system are calculated at each configuration vertex, mirroring the procedure in XGC. This simple training has produced a median relative loss, across configuration space, of the order of $10^{-4}$, which is low enough if the error is of random nature, but not if it is of drift nature in time steps. The run time for the current Picard iterative solver of the operator is $O(n^2)$, where $n$ is the number of plasma species. As the XGC1 code begins to attack problems including a larger number of species, the collision operator will become expensive computationally, making the neural network solver even more important, especially since its training only scales as $O(n)$. A wide enough range of collisionality has been considered in the training data to ensure the full domain of collision physics is captured. An advanced technique to decrease the losses further will be subject of a subsequent report. Eventual work will include expansion of the network to include multiple plasma species.

Keywords

fusion plasmaplasma simulation
Type
Research Article
Information
Journal of Plasma Physics , Volume 87 , Issue 2 , April 2021 , 905870211
Check for updates
Copyright
Copyright © The Author(s), 2021. Published by Cambridge University Press

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

Alom, Z., Taha, T. M, Yakopcic, C., Westberg, S., Nasrin, S. & Asari, V. K. 2018 The history began from alexnet: A comprehensive survey on deep learning approaches. arXiv:1803.01164.Google Scholar
Berg, J. & Nyström, K. 2018 A unified deep artificial neural network approach to partial differential equations in complex geometries. Neurocomputing 317, 28–41.CrossRefGoogle Scholar
Beucler, T., Pritchard, M., Rasp, S., Ott, J., Baldi, P. & Gentine, P. 2019 Enforcing analytic constraints in neural-networks emulating physical systems. arXiv:1909.00912.Google Scholar
Dener, A., Miller, M. A., Churchill, R. M., Munson, T. & Chang, C.-S. 2020 Training neural networks under physical constraints using a stochastic augmented Lagrangian approach. arXiv:2009.07330.Google Scholar
Dominski, J., Chang, C. S., Hager, R., Helander, P., Ku, S. & Yoon, E. S. 2019 Study of up–down poloidal density asymmetry of high- impurities with the new impurity version of XGCa. J. Plasma Phys. 85 (5).CrossRefGoogle Scholar
Hager, R., Yoon, E. S., Ku, S., D'Azevedo, E. F., Worley, P. H. & Chang, C. S. 2016 A fully non-linear multi-species Fokker–Planck–Landau collision operator for simulation of fusion plasma. J. Comput. Phys. 315, 644–660.CrossRefGoogle Scholar
He, S., Li, Y.n, Feng, Y., Ho, S., Ravanbakhsh, S., Chen, W. & Póczos, B. 2019 Learning to predict the cosmological structure formation. Proc. Natl Acad. Sci. USA 116 (28), 13825–13832.CrossRefGoogle ScholarPubMed
Helander, P. & Sigman, D. J. 2002 Collisional Transport in Magnetized Plasmas. Cambridge University Press.Google Scholar
Jiang, C. M., Marcus, P., Kashinath, K. & Marcus, P. 2020 Enforcing hard physical constraints in CNNs through differentiable PDE layer. In Eighth International Conference on Learning Representations, ICLR 2020 – Workshop on Integration of Deep Neural Models and Differential Equations.Google Scholar
Kao, J. C. 2017 Convolutional neural networks. Lecture Notes at UCLA.Google Scholar
Ku, S., Chang, C. S., Hager, R., Churchill, R. M., Tynan, G. R., Cziegler, I., Greenwald, M., Hughes, J., Parker, S. E., Adams, M. F., et al. 2018 A fast low-to-high confinement mode bifurcation dynamics in the boundary-plasma gyrokinetic code XGC1. Phys. Plasmas 25 (5), 056107.CrossRefGoogle Scholar
Landau, L. D. 1936 The transport equation in the case of coulomb interactions. Collected papers of LD Landau. pp. 163–170.Google Scholar
Liu, L., Jiang, H., He, P., Chen, W., Liu, X., Gao, J. & Han, J. 2019 On the Variance of the Adaptive Learning Rate and Beyond. arXiv:1908.03265.Google Scholar
Mohan, A. T., Lubbers, N., Livescu, D. & Chertkov, M. 2017 Embedding hard physical constraints in convolutional neural networks for 3D turbulence. arXiv:2002.00021.Google Scholar
Qin, H. 2019 Machine learning and serving of discrete field theories – when artificial intelligence meets the discrete universe. arXiv:1910.10147.CrossRefGoogle Scholar
Ronneberger, O., Fischer, P. & Brox, T. 2015 U-net: Convolutional networks for biomedical image segmentation. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9351, pp. 234–241.Google Scholar
Sanchez-Gonzalez, A., Bapst, V., Cranmer, K. & Battaglia, P. 2019 Hamiltonian Graph Networks with ODE Integrators. arXiv:1909.12790.Google Scholar
Simonyan, K. & Zisserman, A. 2015 Very deep convolutional networks for large-scale image recognition. In 3rd International Conference on Learning Representations, ICLR 2015 – Conference Track Proceedings. pp. 1–14, arXiv:arXiv:1409.1556v6.Google Scholar
Sirignano, J. & Spiliopoulos, K. 2018 DGM: a deep learning algorithm for solving partial differential equations. J. Comput. Phys. 375 (Dms 1550918), 1339–1364.CrossRefGoogle Scholar
Visin, F., Romero, A., Cho, K., Matteucci, M., Ciccone, M., Kastner, K., Bengio, Y. & Courville, A. 2016 ReSeg: A Recurrent Neural Network-Based Model for Semantic Segmentation. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, pp. 426–433. arXiv:1511.07053.Google Scholar
Yoon, E. S. & Chang, C. S. 2014 A Fokker–Planck–Landau collision equation solver on two-dimensional velocity grid and its application to particle-in-cell simulation. Phys. Plasmas 21 (3).Google Scholar