Accelerated Workflow for Advanced Kinetic Equilibria

Bechtel, Torrin; Nelson, A.; Lao, L. L.; Xing, Zhe; Smith, S. P.; Nazikian, R.; Flanagan, S.; Schissel, D.P.; Stephey, L.; Thomas, R.; Williams, Samuel; Antepara, O.; Dart, Eli; Koleman, E.; Tang, Wanhu

doi:10.1109/ciw-ius56691.2022.00008

Cited by 1 publication

(1 citation statement)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the procedure is traditionally time consuming and generally manual, requiring substantial human intervention to obtain high quality fits. Recently, there have been efforts underway to merge high-performance computing centers with the existing code infrastructure at DIII-D to perform automatic kinetic EFITs for between-shot analysis [18]. In addition, there have been recent efforts to create an automated kinetic EFIT procedure (CAKE [19]) to generate large-scale kinetic EFIT databases and have been used for real-time neural network (NN) surrogate deployment to aid in plasma operations on DIII-D [20].…”

Section: Introductionmentioning

confidence: 99%

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

Sun,

Akçay,

Bechtel Amara

et al. 2024

Nucl. Fusion

View full text Add to dashboard Cite

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.

show abstract

Section: Introductionmentioning

confidence: 99%