In this article, we demonstrate the inference of two-dimensional tungsten concentration profiles in tokamak plasmas, using Gaussian process tomography applied to bolometry and assuming a specific model for the tungsten cooling factor. In ITER, tungsten has been selected for divertor material due to its low tritium retention and ability to handle large heat and particle flux loads. On the other hand, this will cause tungsten impurities to enter the bulk plasma through various plasma-wall interaction processes. Therefore, a detailed understanding of tungsten impurity transport and active control of the impurity transport in tokamaks is crucial. The computational complexity of the method described in their article O(n(2)m) compares favorably to a simple least-squares approach O (n(3)), n represents the number of pixels and m the number of measurement channels, hence bringing real-time tungsten profile monitoring within reach (<10 ms repetition time). The feasibility study of this method has been demonstrated here to a discharge in HL-2A for observing the entire process of tungsten impurities entering the bulk plasma from the scrape-off layer area, tungsten pump-out by edge-localized modes, as well as the formation of a poloidally asymmetric tungsten distribution.
展开▼
