Description
Recent integrated modelling of L-mode discharges has progressively increased its physics fidelity,achieving consistent modelling of current diffusion, heat, particle, and impurity transport across the whole radius. This integrated modelling approach provides a physics-based prediction of the energy confinement time τE, reaching an accuracy comparable to empirical scaling laws [1].
Eight WEST L-modes discharges are selected for integrated modelling, following the successful validation of physics-based models on one of them [2]. They have similar power and density, are heated by Lower Hybrid heating and have plasma current from 0.3 to 0.5 MA to assess whether the integrated modelling captures this confinement dependence. These eight cases are simulated with the IMAS-coupled JINTRAC integrated modelling workflow, named the High Fidelity Pulse Simulator (HFPS). Within the HFPS framework, a stepwise approach is adopted, starting from current diffusion coupled to equilibrium reconstructions and a reduced Lower Hybrid model. Heat and particle transport are then predicted by the quasilinear physics model TGLFsat2 [3]. Finally, impurities (N and W) are also transported self-consistently, with radiation predicted accordingly. The dominant turbulent drive predicted by TGLFsat2 has been verified against linear simulations of the higher fidelity gyrokinetic code GKW. In these electron heated L-modes, Trapped Electron Modes dominate at 𝜌tor,N = 0.3-0.7,
leading to density peaking.
For the 8 pulses, the predicted temperature and density profiles of the simulations are validated against ECE, interferometry, bolometry, and neutron rate measurements, with mean relative errors ranging from
8 to 24% for the most consistent simulations including impurities. Notably, coupling heat and particle transport reduces the relative error on the electron temperature from 28% to 14%.The predicted energy
content lies within 11% of the measured one. Across the 8 pulses, the predicted τE has a similar relative error than the ITER-L96 scaling law prediction (resp. 12% and 10%).
The successful extended validation of the HFPS-TGLFsat2 settings are then used to explore the impact of Ip on the energy content. These investigations show that the energy content scales more strongly
with higher Ip without feedback on the line average density.
References
[1]: Angioni, C., et al. Nuclear Fusion 62.6 (2022): 066015.
[2]: Fonghetti, T., et al. Nuclear Fusion 65.5 (2025): 056018.
[3]: Staebler, G. M., et al. Physics of Plasmas 14.5 (2007): 055909.