Identification of a double decay length (λqt) heat flux deposition shape with embedded thermal measurement and neural network

Abstract

The estimation of the heat flux density distribution profiles in tokamak devices is a very important research topic for edge plasma physics purposes and also to ensure the safety of the machine. In the radial direction, the heat flux exhibits an exponential decay that could be captured by thermal sensors distributed in the plasma facing components. Radially distributed thermal sensors based on Fiber Bragg grating technology have been embedded in the WEST lower divertor to study the heat flux deposition profiles during plasma operation. The comparison between embedded measurements and a 3D finite element model shows a small decay length (5 – 10 mm) on top of a wider heat flux with a decay length around 30 to 50 mm. A tool using neural network has been developed in order to predict the values of the different parameters describing the deposited heat flux from embedded temperature measurements in steady state. A large span of deposited heat fluxes with maximum heat flux ranging from 1 to 9 MW/m² and decay length from 5 to 50 mm were characterized using this tool over a database of more than 250 experimental L-mode pulses performed in WEST in attached divertor configuration. The comparison of the predicted heat flux parameters values with macroscopic plasma parameters have revealed the appearance of the narrow component with the increase of the divertor power load (P_div) with a threshold dependant of the plasma current (I_P).