Thermal modelling for ITER First Wall Heat Load Control

The International Thermonuclear Experimental Reactor (ITER) plays a pivotal role in energy fusion research based on tokamak technology and will magnetically confine a hot toroidal plasma, whose heat exhausted will impact the internal surface of the tokamak chamber closest to the plasma itself, namely the First Wall (FW), potentially leading to accidental conditions. The FW will be protected in real-time (RT) by the First Wall Heat Load Controller (FWHLC), which evaluates its surface temperature through a combination of measurements from infrared (IR) cameras and reconstruction of the impacting heat load thanks to the coupling with a thermal model, which is the focus of this thesis. This work has been at the center of a research and development process as ITER is presently undergoing a re-baseline involving a series of proposals, including the installation of inertially cooled First Wall Panels (FWPs) in an early operation phase called Augmented First Plasma (AFP). Thermal models present in the literature lack the necessary flexibility to cope with the FW design evolution imposed by the re-baseline. Moreover, the literature lacks a systematic approach to identify a trade-off between accuracy and running time requirements for RT applications. Finally, no FW temperature map data are presently available in the ITER database, which are needed for the planned development of IR cameras synthetic diagnostics. In light of these research gaps, this work aims first at finding the most suitable thermal model for the FWHLC in the AFP phase in terms of trade-off between accuracy and running time requirements for RT response, secondly at analysing the thermal response of the FW in the AFP phase, and thirdly at finding a simple yet reliable thermal model for IR cameras synthetic diagnostic. The first two aims have been fulfilled by developing a control-oriented model kept flexible enough to adapt to evolving FW geometries and either active or inertial cooling. Numerical heat transfer simulations have been carried out in MATLAB with a focus on individual fingers composing FWPs. Simulations included fingers’ 1D and 2D models, considering both constant and temperature-dependent properties. A comparative assessment between these simulations allowed to find models at the same time simple and accurate. Additionally, 2D models have been validated against higher fidelity simulations to test their accuracy. The comparative assessment has shown that a 1D model is suitable for actively cooled FWPs while, in the case of inertially cooled panels, a 2D model is recommended. Also, the assumption of constant thermal properties for the temperature range of interest has been verified. Furthermore, a convergence analysis has been carried out to find time steps and computational grid sizes suitable for RT applications. Finally, realistic heat load data coming from simulations including plasma dynamics have been employed to investigate FW thermal response time scales essential for the controller design, which resulted in the order of seconds. The third aim has been fulfilled by developing a simple steady-state 1D model for actively cooled FWPs. The problem allows an analytical solution, which also considers the volumetric heat generated by neutrons produced by fusion reactions. Results show that for most of the FWPs the expected surface temperature is lower than IR cameras lower detection threshold and that volumetric heat generation is negligible.