Impact of various DIII-D diagnostics on the accuracy of neural network surrogates for kinetic EFIT reconstructions

Sun, Xuan; Akcay, Cihan; Bechtel, Torrin; Kruger, Scott; Lao, Lang L.; Liu, Yueqiang; Madireddy, Sandeep; McClenaghan, Joseph

doi:10.1088/1741-4326/ad5d7b

Title: Impact of various DIII-D diagnostics on the accuracy of neural network surrogates for kinetic EFIT reconstructions

Journal Article · 30 June 2024 · Nuclear Fusion

DOI: https://doi.org/10.1088/1741-4326/ad5d7b · OSTI ID:2395955

^[1]; Akcay, Cihan ^[2]; Bechtel, Torrin ^[2]; Kruger, Scott ^[3];

^[2];

^[4];

^[5]

General Atomics, San Diego, CA (United States); General Atomics
General Atomics, San Diego, CA (United States)
Tech-X Corporation, Boulder, CO (United States)
Argonne National Laboratory (ANL), Argonne, IL (United States)
Oak Ridge Associated Universities (ORAU), Oak Ridge, TN (United States)

Kinetic equilibrium reconstructions make use of profile information such as particle density and temperature measurements in addition to magnetics data to compute a self-consistent equilibrium. They are used in a multitude of physics-based modeling. This work develops a multi-layer perceptron (MLP) neural network (NN) model as a surrogate for kinetic Equilibrium Fitting (EFITs) and trains on the 2019 DIIID discharge campaign database of kinetic equilibrium reconstructions. We investigate the impact of including various diagnostic data and machine actuator controls as input into the NN. When giving various categories of data as input into NN models that have been trained using those same categories of data, the predictions on multiple equilibrium reconstruction solutions (poloidal magnetic flux, global scalars, pressure profile, current profile) are highly accurate. When comparing different models with different diagnostics as input, the magnetics-only model outputs accurate kinetic profiles and the inclusion of additional data does not significantly impact the accuracy. When the NN is tasked with inferring only a single target such as the EFIT pressure profile or EFIT current profile, we see a large increase in the accuracy of the prediction of the kinetic profiles as more data is included. These results indicate that certain MLP NN configurations can be reasonably robust to different burning-plasma-relevant diagnostics depending on the accuracy requirements for equilibrium reconstruction tasks.

View Accepted Manuscript (DOE)

Cite

Export

Save

Research Organization:: General Atomics, San Diego, CA (United States)

Sponsoring Organization:: USDOE Office of Science (SC), Fusion Energy Sciences (FES)

Grant/Contract Number:: FG02-95ER54309; SC0021203; SC0021380; FC02-04ER54698

OSTI ID:: 2395955

Alternate ID(s):: OSTI ID: 2382851; OSTI ID: 2403389; OSTI ID: 2484165

Journal Information:: Nuclear Fusion, Journal Name: Nuclear Fusion Journal Issue: 8 Vol. 64; ISSN 0029-5515

Publisher:: IOP ScienceCopyright Statement

Country of Publication:: United States

Language:: English

References (44)

Equilibrium Reconstruction in EAST Tokamak Jinping, Qian; Baonian, Wan; Lao, L. L. Plasma Science and Technology, Vol. 11, Issue 2 https://doi.org/10.1088/1009-0630/11/2/03	journal	April 2009
Fast equilibrium reconstruction by deep learning on EAST tokamak Lu, Jingjing; Hu, Youjun; Xiang, Nong AIP Advances, Vol. 13, Issue 7 https://doi.org/10.1063/5.0152318	journal	July 2023
Real time equilibrium reconstruction for tokamak discharge control Ferron, J. R.; Walker, M. L.; Lao, L. L. Nuclear Fusion, Vol. 38, Issue 7 https://doi.org/10.1088/0029-5515/38/7/308	journal	July 1998
Normalization methods for input and output vectors in backpropagation neural networks Kim, Daehyon International Journal of Computer Mathematics, Vol. 71, Issue 2 https://doi.org/10.1080/00207169908804800	journal	January 1999
Reconstruction of current profile parameters and plasma shapes in tokamaks Lao, L. L.; St. John, H.; Stambaugh, R. D. Nuclear Fusion, Vol. 25, Issue 11 https://doi.org/10.1088/0029-5515/25/11/007	journal	November 1985
Equilibrium analysis of current profiles in tokamaks Lao, L. L.; Ferron, J. R.; Groebner, R. J. Nuclear Fusion, Vol. 30, Issue 6 https://doi.org/10.1088/0029-5515/30/6/006	journal	June 1990
Polarimetry of motional Stark effect and determination of current profiles in DIII‐D (invited) Wróblewski, D.; Lao, L. L. Review of Scientific Instruments, Vol. 63, Issue 10 https://doi.org/10.1063/1.1143463	journal	October 1992
A Linux cluster for between-pulse magnetic equilibrium reconstructions and other processor bound analyses Peng, Q.; Groebner, R. J.; Lao, L. L. Review of Scientific Instruments, Vol. 72, Issue 8 https://doi.org/10.1063/1.1384451	journal	August 2001
An upgrade of the magnetic diagnostic system of the DIII-D tokamak for non-axisymmetric measurements King, J. D.; Strait, E. J.; Boivin, R. L. Review of Scientific Instruments, Vol. 85, Issue 8 https://doi.org/10.1063/1.4891817	journal	August 2014
Augmenting machine learning of Grad–Shafranov equilibrium reconstruction with Green's functions McClenaghan, J.; Akçay, C.; Amara, T. B. Physics of Plasmas, Vol. 31, Issue 8 https://doi.org/10.1063/5.0213625	journal	August 2024
Equilibrium analysis of iron core tokamaks using a full domain method O'Brien, D. P.; Lao, L. L.; Solano, E. R. Nuclear Fusion, Vol. 32, Issue 8 https://doi.org/10.1088/0029-5515/32/8/I05	journal	August 1992
CAKE: Consistent Automatic Kinetic Equilibrium reconstruction Xing, Z. A.; Eldon, D.; Nelson, A. O. Fusion Engineering and Design, Vol. 163 https://doi.org/10.1016/j.fusengdes.2020.112163	journal	February 2021
Overview of physics basis for ITER Mukhovatov, V.; Shimada, M.; Chudnovskiy, A. N. Plasma Physics and Controlled Fusion, Vol. 45, Issue 12A https://doi.org/10.1088/0741-3335/45/12A/016	journal	November 2003
Hydromagnetic equilibria and force-free fields Grad, Harold; Rubin, Hanan Journal of Nuclear Energy (1954), Vol. 7, Issue 3-4 https://doi.org/10.1016/0891-3919(58)90139-6	journal	September 1958
Real‐time estimation of the electron temperature profile in DIII‐D by leveraging neural‐network surrogate models Morosohk, Shira; Schuster, Eugenio Contributions to Plasma Physics, Vol. 63, Issue 5-6 https://doi.org/10.1002/ctpp.202200153	journal	January 2023
On the potential of physics-informed neural networks to solve inverse problems in tokamaks Rossi, Riccardo; Gelfusa, Michela; Murari, Andrea Nuclear Fusion, Vol. 63, Issue 12 https://doi.org/10.1088/1741-4326/ad067c	journal	November 2023
Overview of the Tritium Technologies for the EU DEMO Breeding Blanket Cristescu, Ion; Priester, F.; Rapisarda, D. Fusion Science and Technology, Vol. 76, Issue 4 https://doi.org/10.1080/15361055.2020.1716456	journal	May 2020
Initial results of the high resolution edge Thomson scattering upgrade at DIII-D Eldon, D.; Bray, B. D.; Deterly, T. M. Review of Scientific Instruments, Vol. 83, Issue 10 https://doi.org/10.1063/1.4738656	journal	October 2012
Deep neural network Grad–Shafranov solver constrained with measured magnetic signals Joung, Semin; Kim, Jaewook; Kwak, Sehyun Nuclear Fusion, Vol. 60, Issue 1 https://doi.org/10.1088/1741-4326/ab555f	journal	December 2019
Integrated modeling applications for tokamak experiments with OMFIT Meneghini, O.; Smith, S. P.; Lao, L. L. Nuclear Fusion, Vol. 55, Issue 8 https://doi.org/10.1088/0029-5515/55/8/083008	journal	July 2015
Magnetic diagnostic system of the DIII-D tokamak Strait, E. J. Review of Scientific Instruments, Vol. 77, Issue 2 https://doi.org/10.1063/1.2166493	journal	February 2006
Spatial and temporal analysis of DIII-D 3D magnetic diagnostic data Strait, E. J.; King, J. D.; Hanson, J. M. Review of Scientific Instruments, Vol. 87, Issue 11 https://doi.org/10.1063/1.4960419	journal	August 2016
Technological challenges of ITER diagnostics Costley, A. E.; Sugie, T.; Vayakis, G. Fusion Engineering and Design, Vol. 74, Issue 1-4 https://doi.org/10.1016/j.fusengdes.2005.08.026	journal	November 2005
Advanced control of neutral beam injected power in DIII-D Pawley, C. J.; Crowley, B. J.; Pace, D. C. Fusion Engineering and Design, Vol. 123 https://doi.org/10.1016/j.fusengdes.2017.02.106	journal	November 2017
Equilibrium properties of spherical torus plasmas in NSTX Sabbagh, S. A.; Kaye, S. M.; Menard, J. Nuclear Fusion, Vol. 41, Issue 11 https://doi.org/10.1088/0029-5515/41/11/309	journal	November 2001
GS-DeepNet: mastering tokamak plasma equilibria with deep neural networks and the Grad–Shafranov equation Joung, Semin; Ghim, Y. -C.; Kim, Jaewook Scientific Reports, Vol. 13, Issue 1 https://doi.org/10.1038/s41598-023-42991-5	journal	September 2023
Surrogate models for plasma displacement and current in 3D perturbed magnetohydrodynamic equilibria in tokamaks Liu, Yueqiang; Akcay, Cihan; Lao, Lang L. Nuclear Fusion, Vol. 62, Issue 12 https://doi.org/10.1088/1741-4326/ac9d4c	journal	November 2022
Boosting with early stopping: Convergence and consistency Zhang, Tong; Yu, Bin The Annals of Statistics, Vol. 33, Issue 4 https://doi.org/10.1214/009053605000000255	journal	August 2005
Neural net modeling of equilibria in NSTX-U Wai, J. T.; Boyer, M. D.; Kolemen, E. Nuclear Fusion, Vol. 62, Issue 8 https://doi.org/10.1088/1741-4326/ac77e6	journal	July 2022
Current Profile Measurement on the DIII-D Tokamak Jayakumar, R. J.; Allen, S. L.; Burrell, K. H. Fusion Science and Technology, Vol. 48, Issue 2 https://doi.org/10.13182/FST05-A1044	journal	October 2005
On Early Stopping in Gradient Descent Learning Yao, Yuan; Rosasco, Lorenzo; Caponnetto, Andrea Constructive Approximation, Vol. 26, Issue 2 https://doi.org/10.1007/s00365-006-0663-2	journal	April 2007
Thomson scattering diagnostic upgrade on DIII-D Ponce-Marquez, D. M.; Bray, B. D.; Deterly, T. M. Review of Scientific Instruments, Vol. 81, Issue 10 https://doi.org/10.1063/1.3495759	journal	October 2010
KSTAR equilibrium operating space and projected stabilization at high normalized beta Park, Y. S.; Sabbagh, S. A.; Berkery, J. W. Nuclear Fusion, Vol. 51, Issue 5 https://doi.org/10.1088/0029-5515/51/5/053001	journal	April 2011
Chapter 7: Diagnostics Donné, A. J. H.; Costley, A. E.; Barnsley, R. Nuclear Fusion, Vol. 47, Issue 6 https://doi.org/10.1088/0029-5515/47/6/S07	journal	June 2007
MHD Equilibrium Reconstruction in the DIII-D Tokamak Lao, L. L.; John, H. E. St.; Peng, Q. Fusion Science and Technology, Vol. 48, Issue 2 https://doi.org/10.13182/FST48-968	journal	October 2005
Machine learning-based real-time kinetic profile reconstruction in DIII-D Shousha, Ricardo; Seo, Jaemin; Erickson, Keith Nuclear Fusion, Vol. 64, Issue 2 https://doi.org/10.1088/1741-4326/ad142f	journal	December 2023
EAST discharge prediction without integrating simulation results Wan, Chenguang; Yu, Zhi; Pau, Alessandro Nuclear Fusion, Vol. 62, Issue 12 https://doi.org/10.1088/1741-4326/ac9c1a	journal	November 2022
Importance of input data normalization for the application of neural networks to complex industrial problems Sola, J.; Sevilla, J. IEEE Transactions on Nuclear Science, Vol. 44, Issue 3 https://doi.org/10.1109/23.589532	journal	June 1997
Separation of β̄ _p and ℓ _i in tokamaks of non-circular cross-section Lao, L. L.; John, H. St.; Stambaugh, R. D. Nuclear Fusion, Vol. 25, Issue 10 https://doi.org/10.1088/0029-5515/25/10/004	journal	October 1985
Overview of equilibrium reconstruction on DIII-D using new measurements from an expanded motional Stark effect diagnostic Holcomb, C. T.; Makowski, M. A.; Allen, S. L. Review of Scientific Instruments, Vol. 79, Issue 10 https://doi.org/10.1063/1.2955711	journal	October 2008
An overview of the ITER project Holtkamp, N. Fusion Engineering and Design, Vol. 82, Issue 5-14 https://doi.org/10.1016/j.fusengdes.2007.03.029	journal	October 2007
Multilayer feedforward networks are universal approximators Hornik, Kurt; Stinchcombe, Maxwell; White, Halbert Neural Networks, Vol. 2, Issue 5 https://doi.org/10.1016/0893-6080(89)90020-8	journal	January 1989
Measurements of the neutron source strength at DIII-D Heidbrink, W. W.; Taylor, P. L.; Phillips, J. A. Review of Scientific Instruments, Vol. 68, Issue 1 https://doi.org/10.1063/1.1147646	journal	January 1997
Application of machine learning and artificial intelligence to extend EFIT equilibrium reconstruction Lao, L. L.; Kruger, S.; Akcay, C. Plasma Physics and Controlled Fusion, Vol. 64, Issue 7 https://doi.org/10.1088/1361-6587/ac6fff	journal	June 2022

Similar Records

Application of machine learning and artificial intelligence to extend EFIT equilibrium reconstruction

Journal Article · 2022 · Plasma Physics and Controlled Fusion · OSTI ID:1873871

Lao, L. L.; Kruger, S.; Akcay, C.; +15 more

Neural net modeling of equilibria in NSTX-U

Journal Article · 2022 · Nuclear Fusion · OSTI ID:1880110

Wai, J. T.; Boyer, M. D.; Kolemen, E.

EFIT-Prime: Probabilistic and physics-constrained reduced-order neural network model for equilibrium reconstruction in DIII-D

Journal Article · 2024 · Physics of Plasmas · OSTI ID:2477694

Madireddy, S.; Akçay, C.; Kruger, S. E.; +8 more

Related Subjects

70 PLASMA PHYSICS AND FUSION TECHNOLOGY

Title: Impact of various DIII-D diagnostics on the accuracy of neural network surrogates for kinetic EFIT reconstructions

Citation Formats

References (44)

Similar Records

Related Subjects