Deep Learning for Plasma Tomography and Disruption Prediction From Bolometer Data

Ferreira, Diogo R.; Carvalho, P.; Fernandes, H.

doi:10.1109/tps.2019.2947304

Cited by 40 publications

(29 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the present paper, we have developed a general deep learning capability to train multiple models with distinct architectures within a single software suite that serves to promote adaptability to different temporal and spatial learning tasks and also to enable ensemble schemes for highly accurate prediction. Various machine learning based algorithms targeting different learning and prediction tasks have been independently validated for modern tokamaks such as DIII-D [4,14] and JET [23][24][25][26][27][28][29][30][31][32][33][34][35]. The capability of utilizing effective ensemble models in realtime plasma control systems is an important task for successful operation of future machines.…”

Section: Discussionmentioning

confidence: 99%

Fully Convolutional Spatio-Temporal Models for Representation Learning in Plasma Science

Dong

Felker

Svyatkovskiy

et al. 2021

J Mach Learn Model Comput

View full text Add to dashboard Cite

We have trained a fully convolutional spatio-temporal model for fast and accurate representation learning in the challenging exemplar application area of fusion energy plasma science. The onset of major disruptions is a critically important fusion energy science issue that must be resolved for advanced tokamak plasmas such as the $25B burning plasma international thermonuclear experimental reactor (ITER) experiment. While a variety of statistical methods have been used to address the problem of tokamak disruption prediction and control, recent approaches based on deep learning have proven particularly compelling. In the present paper, we introduce further improvements to the fusion recurrent neural network (FRNN) software suite, which delivered cross-machine disruption predictions with unprecedented accuracy using a large database of experimental signals from two major tokamaks. Up to now, FRNN was based on the long short-term memory (LSTM) variant of recurrent neural networks to leverage the temporal information in the data. Here, we implement and apply the "temporal convolutional neural network (TCN)" architecture to the time-dependent input signals. This allows highly optimized convolution operations to carry the majority of the computational load of training, thus enabling a reduction in training time, and the effective use of highperformance computing resources for hyperparameter tuning. At the same time, the TCN-based architecture achieves better predictive performance when compared with the LSTM architecture for various tasks for a representative fusion database.

show abstract

Section: Discussionmentioning

confidence: 99%

Fully Convolutional Spatio-Temporal Models for Representation Learning in Plasma Science

Dong

Felker

Svyatkovskiy

et al. 2021

J Mach Learn Model Comput

View full text Add to dashboard Cite

show abstract

“…domain [5][6][7][8][9][10][11][12][13][14][15] . These techniques have been complemented with tools in the frequency domain 16 , based on Fourier transforms.…”

Section: And Jet Contributors*mentioning

confidence: 99%

Disruption prediction with artificial intelligence techniques in tokamak plasmas

Mailloux¹,

Abid

Abraham³

et al. 2022

Nat. Phys.

View full text Add to dashboard Cite

“…It is possible to observe that, a few hundred milliseconds before the disruption, the probability of disruption gets close do 1.0 and the time to disruption drops down to 0.0, both indicating that a disruption is imminent. In fact, both indicators should be taken into account for a more accurate prediction but, even then, the success rate (recall) on a more extensive test set is only about 70% [29], while other predictors currently being used at JET have a recall of about 85% [30]. It should be noted, however, that these predictors are based on a wide range of global plasma parameters (e.g.…”

Section: Deep Learning For Disruption Predictionmentioning

confidence: 99%

Using HPC infrastructures for deep learning applications in fusion research

Ferreira

2021

Preprint

Self Cite

View full text Add to dashboard Cite

In the fusion community, the use of high performance computing (HPC) has been mostly dominated by heavy-duty plasma simulations, such as those based on particle-in-cell and gyrokinetic codes. However, there has been a growing interest in applying machine learning for knowledge discovery on top of large amounts of experimental data collected from fusion devices. In particular, deep learning models are especially hungry for accelerated hardware, such as graphics processing units (GPUs), and it is becoming more common to find those models competing for the same resources that are used by simulation codes, which can be either CPU-or GPU-bound. In this paper, we give examples of deep learning models -such as convolutional neural networks, recurrent neural networks, and variational autoencoders -that can be used for a variety of tasks, including image processing, disruption prediction, and anomaly detection on diagnostics data. In this context, we discuss how deep learning can go from using a single GPU on a single node to using multiple GPUs across multiple nodes in a large-scale HPC infrastructure.

show abstract

Deep Learning for Plasma Tomography and Disruption Prediction From Bolometer Data

Cited by 40 publications

References 61 publications

Fully Convolutional Spatio-Temporal Models for Representation Learning in Plasma Science

Fully Convolutional Spatio-Temporal Models for Representation Learning in Plasma Science

Disruption prediction with artificial intelligence techniques in tokamak plasmas

Using HPC infrastructures for deep learning applications in fusion research

Contact Info

Product

Resources

About