Description
This study presents particle transport modelling results for D/T ratio control experiments conducted during the JET DTE3 campaign [1]. Interpretative TRANSP and predictive JETTO simulations were performed to analyse the evolution of the D/T mixture [2] in these experiments [3]. Using simplified Bohm–gyroBohm-based transport models, the simulations successfully reproduced electron density, steady-state D/T ratio, and neutron rate evolution. However, during the transient phases the predicted D/T ratio responded to control requests faster than observed experimentally. Additional scenarios, including cases which could not be realised experimentally where gas injection and pellet sources were swapped, were also investigated, and the corresponding simulation results are discussed.
The findings in this study demonstrate that simple models and assumptions regarding D and T transport and sources can be successfully adopted to predictively model the behaviour of Real Time (RT) controllers for D/T ratio. Detailed modelling of the SOL and core across all channels is not required for such tasks. It has been demonstrated that a simplified approach can be effective, using (i) measured Te, Ti, (ii) not modelling the SOL physics, and (iii) adopting simplifying particle transport assumptions. This conclusion is crucial for developing numerical tools to test future controllers.
In general, the proposed approach for developing and testing real-time D/T controllers should be directly applicable to future conventional tokamaks, such as ITER, SPARC, and DEMO, where simplified assumptions about energy and particle transport are valid. For devices where the transport models require validation — such as STEP — applicability largely depends on the extent to which transport can be reasonably simplified. Regarding source terms, the proposed model appears broadly applicable. The simulations presented here indicate that particle sources from pellet injection and NBI can be reliably captured by the models employed in this work. However, for gas injection sources, the simplified treatment used in this study code may not be adequate for devices or scenarios characterized by significantly different SOL physics.
[1] C. Maggi et al 2024 Nucl. Fusion 64 112012
[2] K. K. Kirov et al 2025 Nucl. Fusion 65 (2025) 106016
[3] M. Lennholm et al 2025 PRX Energy 4 023007