PSPar Strand

Par Strand

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Par Strand has a total of 115 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    On the impact of isotope mass on the stored core thermal energy in JET H-modes

    by E Fransson, L-G Eriksson, D B King, H Nordman, P Strand, E Viezzer, D Yadykin, Jet Contributors

    Understanding how the stored thermal energy scales with the main ion isotope, or mixtures of isotopes, in a tokamak is a key question for predicting the performance of future deuterium–tritium operations in ITER and fusion power plants. Although this remains an active area of research, a complete understanding has yet to be achieved. In this study, three JET H-mode discharges with matched engineering parameters are analysed to provide further experimental and analytical input into the investigation of mass scaling of the plasma turbulent transport. The discharges comprise one almost pure hydrogen, one mixed hydrogen–deuterium and one almost pure deuterium plasma. The analysis employs both linear gyrokinetic simulations and an integrated modelling framework. Particular attention is given to three mechanisms through which the mass of the main ion species is expected to influence the core thermal stored energy: (i) the boundary conditions set by the H-mode pedestal, especially in relation to stiff behaviour of temperature and density profiles; (ii) E × B shearing arising from neutral beam injection driven rotation; and (iii) the contribution of electron temperature gradient-modes to plasma transport. The results presented indicate that the pedestal and E × B shearing was key in explaining the increased thermal stored energy in the core in the analysed discharges.

  • Open Access

    Interpretability guided transfer learning approaches for tritium pedestal predictions

    by Andreas Gillgren, Dimitriy Yadykin, Pär Strand

    We explore transfer learning approaches to extend data-driven pedestal models trained on deuterium (D) plasmas to tritium (T) and DT mixtures. Specifically, we use models pre-trained on JET D pulses, and JET T/DT data for the transfer learning. We use model interpretability to guide our choice of transfer learning strategy. Analysis of model behavior post transfer learning reveals that sparsity and multicollinearity in the T/DT data lead to severe overfitting when fine-tuning the weights of the pre-trained neural network-based models. Therefore, we instead use a more robust and simple output calibration approach to facilitate the impact of isotope composition. This yields models with R2 between 0.66-0.87, performing significantly better than uncalibrated models though not matching the original D-only performance. The scaling coefficients obtained qualitatively agree with previous research, namely that the pedestal density scales positively with increased isotope mass, while pedestal temperature exhibits a weak negative scaling with isotope mass. This work highlights the importance of understanding model behavior in transfer learning to ensure reasonable functional mappings, which is particularly relevant for fusion research where sparse, multicollinear data are encountered.

  • Open Access

    Investigating characteristics of the growth rates from QuaLiKiz using an interpretable surrogate model

    by A. Gillgren, E. Fransson, A. Ludvig-Osipov, W. Enström, L. Flyckt, M. Green, M. Kvartsén, Y. Liljegren, E. Olsson, A. Orthag, et al.

    We present an interpretable, machine learning-based surrogate model for the eigenvalue solver in QuaLiKiz, a model that simulates turbulent transport in fusion plasmas. The aim is to exploit prediction transparency to gain insight into the anticipated behavior of QuaLiKiz-based surrogates and the underlying eigenvalue solver, a task that is more challenging when using black-box surrogate models. Specifically, we focus on predicting the growth rate of turbulence driving ion temperature gradient instabilities computed by QuaLiKiz for the normalized poloidal wavenumber kθρs=0.325. We split the task into a classification task, to determine whether the growth rate is positive (unstable mode) or not, and a growth rate prediction task, knowing the mode is unstable. The dataset used is a QuaLiKiz dataset based on JET pulses. The method used is the NeuralBranch method, a neural network-based method that reveals how the inputs of the models, in this case plasma parameters, impact the output. Results show that NeuralBranch models outperform linear models and match dense neural networks (traditional black-box models) in accuracy while being interpretable. By analyzing the NeuralBranch models, we identify parameter dependencies that cannot be captured by linear models. For instance, the models indicate that the stabilizing effect of ExB shear on the growth rate is suppressed at low magnetic shear, which can be attributed to how ExB shear influences the eigenfunction width in QuaLiKiz. In summary, this work demonstrates how interpretable methods can shed light on the behavior of surrogates and their underlying counterpart, thus enhancing both model credibility and understanding.


  • Open Access

    Investigating pedestal dependencies at JET using an interpretable neural network architecture

    by A. Gillgren, A. Ludvig-Osipov, D. Yadykin, P. Strand, Jet Contributors

    We present NeuralBranch, an interpretable neural network framework. In this work, we use it specifically to predict the pedestal from key engineering parameters in tokamak fusion experiments. The main goal is to uncover intricate relationships that traditional power scalings, with their limited expressive capacity, fail to capture. A secondary objective is to provide a transparent alternative to current opaque, black-box machine learning models used to predict the pedestal in integrated modeling frameworks. By using the proposed method, we obtain a novel global overview of several intricate dependencies in the JET pedestal database. For instance, while both input power and plasma current are positively correlated with pedestal top pressure and temperature, NeuralBranch reveals an attenuating interaction. This means that increasing power weakens the impact that current has on pedestal pressure and temperature, and vice versa. Further investigation of this interaction may be important to avoid overestimating pedestal stored energy at future machines like ITER when using established power scalings. We also identify an amplifying interaction between plasma current and triangularity, where higher triangularity amplifies the effect of plasma current on pedestal density, and vice versa. In addition to these findings, NeuralBranch matches the accuracy of black-box neural networks, with R 2 values as high as 0.88. This demonstrates that interpretability, with its associated benefits, can be achieved without sacrificing accuracy, making NeuralBranch a promising alternative for pedestal predictions.


  • Open Access

    Access and sustainment of ELMy H-mode operation for ITER pre-fusion power operation plasmas using JINTRAC

    by E. Tholerus, L. Garzotti, V. Parail, Y. Baranov, X. Bonnin, G. Corrigan, F. Eriksson, D. Farina, L. Figini, D.M. Harting, et al.

    In the initial stages of ITER operation, ELM mitigation systems need to be commissioned. This requires controlled flat-top operation in type-I ELMy H-mode regimes. Hydrogen or helium plasma discharges are used exclusively in these stages to ensure negligible production of neutrons from fusion reactions. With the expected higher L–H power threshold of hydrogen and helium plasmas compared to corresponding D and D/T plasmas, it is uncertain whether available auxiliary power systems are sufficient to operate in stable type-I ELMy H-mode. This has been investigated using integrated core and edge/SOL/divertor modelling with JINTRAC. Assuming that the L–H power threshold is well captured by the Martin08 scaling law, the presented simulations have found that 30 MW of ECRH power is likely required for the investigated hydrogen plasma scenarios, rather than the originally planned 20 MW in the 2016 Staged Approach ITER Baseline. However, past experiments have shown that a small helium fraction (∼10%) can considerably reduce the hydrogen plasma L–H power threshold. Assuming that these results extrapolate to ITER operation regimes, the 7.5 MA/2.65 T hydrogen plasma scenario is likely to access stable type-I ELMy H-mode operation also at 20 MW of ECRH.


  • Open Access

    Simulations of the stationary Q = 10 and the exit phase from the flat-top of an ITER 15MA baseline scenario: predictive JINTRAC simulation with a consistent treatment of D and T in the whole plasma

    by F. Eriksson, E. Tholerus, G. Corrigan, Y. Baranov, X. Bonnin, D. Farina, L. Figini, L. Garzotti, S.H. Kim, F. Koechl, et al.

    Designing a robust termination scenario for a burning ITER plasma is a challenge that requires extensive core plasma and divertor modelling. The presented work consists of coupled core/edge/SOL/divertor simulations, performed with the JINTRAC code, to study the Q = 10 flat-top phase and exit phase of the ITER 15 MA/5.3 T DT scenario. The modelling utilizes the recently implemented option to treat deuterium and tritium separately in the SOL/divertor, enabling a consistent treatment of deuterium and tritium in the whole plasma volume, which is a unique capability of JINTRAC. In addition, these are the first JINTRAC simulations of this scenario that use a first-principles transport model to self-consistently model the ECRH power deposition and to include tungsten while keeping track of tungsten sputtering and accumulation. The flat-top simulations demonstrate the possibility of sustaining a steady state fusion Q of 10 using pure deuterium gas puffs together with DT mixed pellets, which is an option to make a more effective use of tritium. Simulations of the exit phase are set up sequentially, with each phase providing initial conditions for the next, starting with a density decay at full current and auxiliary power, and demonstrate the possibility of reducing the density robustly within a few seconds. Following the density decay, a subsequent auxiliary power ramp-down in H-mode is performed with a late H–L transition at low auxiliary power, which may provide an option for the optimization of the plasma termination. The final ramp-down phase consists of a current ramp-down in L-mode to 3.75 MA.


  • Open Access

    Data-driven models in fusion exhaust: AI methods and perspectives

    by S. Wiesen, S. Dasbach, A. Kit, A.E. Jaervinen, A. Gillgren, A. Ho, A. Panera, D. Reiser, M. Brenzke, Y. Poels, et al.

    A review is given on the highlights of a scatter-shot approach of developing machine-learning methods and artificial neural networks based fast predictors for the application to fusion exhaust. The aim is to enable and facilitate optimized and improved modeling allowing more flexible integration of physics models in the light of extrapolations towards future fusion devices. The project encompasses various research objectives: (a) developments of surrogate model predictors for power & particle exhaust in fusion power plants; (b) assessments of surrogate models for time-dependent phenomena in the plasma-edge; (c) feasibility studies of micro–macro model discovery for plasma-facing components surface morphology & durability; and (d) enhancements of pedestal models & databases through interpolators and generators exploiting uncertainty quantification. Presented results demonstrate useful applications for machine-learning and artificial intelligence in fusion exhaust modeling schemes, enabling an unprecedented combination of both fast and accurate simulation.


  • High temporal resolution of pedestal dynamics via machine learning on density diagnostics

    by Diogo R Ferreira, Andreas Gillgren, Andrei Ludvig-Osipov, Pär Strand, Jet Contributors

    At the Joint European Torus, the reference diagnostic to measure electron density is Thomson scattering. However, this diagnostic has a low sampling rate, which makes it impractical to study the temporal dynamics of fast processes, such as edge localized modes. In this work, we use machine learning to predict the density profile based on data from another diagnostic, namely reflectometry. By learning to transform reflectometry data into Thomson scattering profiles, the model is able to generate the density profile at a much higher sampling rate than Thomson scattering, and more accurately than reflectometry alone. This enables the study of pedestal dynamics, by analyzing the time evolution of the pedestal height, width, position and gradient. We also discuss the accuracy of the model when applied on experimental campaigns that are different from the one it was trained on.


  • Open Access

    A fast neural network surrogate model for the eigenvalues of QuaLiKiz

    by E. Fransson, A. Gillgren, A. Ho, J. Borsander, O. Lindberg, W. Rieck, M. Åqvist, P. Strand

    We introduce a neural network surrogate model that predicts the eigenvalues for the turbulent microinstabilities, based on the gyrokinetic eigenvalue solver in QuaLiKiz. The model quickly provides information about the dominant instability for specific plasma conditions, and in addition, the eigenvalues offer a pathway for extrapolating transport fluxes. The model is trained on a 5 × 106 data points large dataset based on experimental data from discharges at the joint European torus, where each data point represents a QuaLiKiz simulation. The most accurate model was obtained when the task was split into a classification task to decide if the imaginary part of eigenvalues were stable (≤0) or not, and a regression model to calculate the eigenvalues once the classifier predicted the unstable class.


  • Open Access

    Validation of D–T fusion power prediction capability against 2021 JET D–T experiments

    by Hyun-Tae Kim, Fulvio Auriemma, Jorge Ferreira, Stefano Gabriellini, Aaron Ho, Philippe Huynh, Krassimir Kirov, Rita Lorenzini, Michele Marin, Michal Poradzinski, et al.

    JET experiments using the fuel mixture envisaged for fusion power plants, deuterium and tritium (D–T), provide a unique opportunity to validate existing D–T fusion power prediction capabilities in support of future device design and operation preparation. The 2021 JET D–T experimental campaign has achieved D–T fusion powers sustained over 5 s in ITER-relevant conditions i.e. operation with the baseline or hybrid scenario in the full metallic wall. In preparation of the 2021 JET D–T experimental campaign, extensive D–T predictive modelling was carried out with several assumptions based on D discharges. To improve the validity of ITER D–T predictive modelling in the future, it is important to use the input data measured from 2021 JET D–T discharges in the present core predictive modelling, and to specify the accuracy of the D–T fusion power prediction in comparison with the experiments. This paper reports on the validation of the core integrated modelling with TRANSP, JINTRAC, and ETS coupled with a quasilinear turbulent transport model (Trapped Gyro Landau Fluid or QualLiKiz) against the measured data in 2021 JET D–T discharges. Detailed simulation settings and the heating and transport models used are described. The D–T fusion power calculated with the interpretive TRANSP runs for 38 D–T discharges (12 baseline and 26 hybrid discharges) reproduced the measured values within 20 % . This indicates the additional uncertainties, that could result from the measurement error bars in kinetic profiles, impurity contents and neutron rates, and also from the beam-thermal fusion reaction modelling, are less than 20% in total. The good statistical agreement confirms that we have the capability to accurately calculate the D–T fusion power if correct kinetic profiles are predicted, and indicates that any larger deviation of the D–T fusion power prediction from the measured fusion power could be attributed to the deviation of the predicted kinetic profiles from the measured kinetic profiles in these plasma scenarios. Without any posterior adjustment of the simulation settings, the ratio of predicted D–T fusion power to the measured fusion power was found as 65%–96% for the D–T baseline and 81%–97% for D–T hybrid discharge. Possible reasons for the lower D–T prediction are discussed and future works to improve the fusion power prediction capability are suggested. The D–T predictive modelling results have also been compared to the predictive modelling of the counterpart D discharges, where the key engineering parameters are similar. Features in the predicted kinetic profiles of D–T discharges such as underprediction ofneare also found in the prediction results of the counterpart D discharges, and it leads to similar levels of the normalized neutron rate prediction between the modelling results of D–T and the counterpart D discharges. This implies that the credibility of D–T fusion power prediction could bea prioriestimated by the prediction quality of the preparatory D discharges, which will be attempted before actual D–T experiments.