Answers are grounded in this paper and cite their sources.
Divertor thermocouples can provide robust, in-vessel monitoring of heat loads during short attached or reattachment phases, but their number and accessible locations are limited. We present a Bayesian experimental design workflow to optimise thermocouple placement applied to MAST Upgrade (MAST-U) divertor tiles 3 and 4 for conventional and elongated magnetic configurations. A maximum-entropy design in heat flux decay length ( λ q ), Gaussian spreading width ( S ) and scrape-off-layer power ( P S O L ) generates a representative ensemble of attached plasma states spanning configuration-specific parameter ranges and strike-point locations ( R s p , Z s p ). For each state we compute synthetic thermocouple readings using the Heat Flux Engineering Analysis Toolkit and OpenFOAM. A Gaussian process surrogate, coupled to digiLab’s Bayesian experimental designer in the Uncertainty Engine, selects sensor subsets (1–10 sensors) that maximise expected information gain on quantities of interest θ = ( λ q , S , P S O L , R s p , Z s p ) . The resulting designs concentrate around the strike region while retaining off-peak coverage to resolve profile shape, and exhibit clear diminishing returns with increasing sensor count. This provides quantitative, configuration-specific layouts for MAST-U tiles 3 and 4 as a case study, with a transferable methodology for diagnostics design in future divertor-equipped devices and power plants.
No takes yet. Share an insight, caveat, or question.
Battye et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: