• Energy Research

Machine protection is a core task of real-time image diagnostics aiming for steady-state operation in nuclear fusion devices. The paper evaluates the applicability of the newest low-power NVIDIA Jetson Xavier NX platform for image plasma diagnostics. This embedded NVIDIA Tegra System-on-a-Chip (SoC) integrates a Graphics Processing Unit (GPU) and Central Processing Unit (CPU) on a single chip. The hardware differences and features compared to the previous NVIDIA Jetson TX2 are signified. Implemented algorithms detect thermal events in real-time, utilising the high parallelism provided by the embedded General-Purpose computing on Graphics Processing Units (GPGPU). The performance and accuracy are evaluated on the experimental data from the Wendelstein 7-X (W7-X) stellarator. Strike-line and reflection events are primarily investigated, yet benchmarks for overload hotspots, surface layers and visualisation algorithms are also included. Their detection might allow for automating real-time risk evaluation incorporated in the divertor protection system in W7-X. For the first time, the paper demonstrates the feasibility of complex real-time image processing in nuclear fusion applications on low-power embedded devices. Moreover, GPU-accelerated reference processing pipelines yielding higher accuracy compared to the literature results are proposed, and remarkable performance improvement resulting from the upgrade to the Xavier NX platform is attained.

Plasma–wall interactions play a crucial role for the performance of fusion devices and the lifetime of plasma-facing components. In this work the results of some plasma–wall interaction processes, namely the erosion/deposition pattern and the fuel content of divertor baffle tiles exposed in Wendelstein 7-X in the initial phase of island divertor, are presented. Net-deposition of carbon with co-deposited oxygen and hydrogen is determined on the graphite tiles after about 1 hour of plasma operation in hydrogen and helium in operation phase OP 1.2a. C is predominately a result of net-erosion of the graphite target plates and oxygen is the strongest intrinsic impurity in Wendelstein 7-X in OP 1.2a. The hydrogen content distribution on a set of tiles exposed in equivalent positions in three of the five modules of the stellarator was quantified for the first time. Ex-situ performed laser-induced breakdown spectroscopy measurements show the depth-resolved fuel content in deposited layers as well as implantation and diffusion in the base material. Complementary, gas analysis after laser-induced ablation offers quantitative hydrogen content determination in the deposited layers up to 1018 hydrogen atoms/cm2that is non-uniformly distributed over the 95mm×125 mm sized tiles. The results show a toroidal asymmetry with 60% more hydrogen in the stellarator module 2 with respect to module 1 as well as a slight top-down asymmetry for the baffle tiles in upper and lower half modules. A clear dependence of the hydrogen content on the surface temperature during plasma exposition in stellarator discharges in standard magnetic divertor configuration with edge transformation 5/5 was not observed.

Nuclear materials and energy 37, 101514 - (2023). doi:10.1016/j.nme.2023.101514 Published by Elsevier, Amsterdam [u.a.]

During the interaction of the scrape off layer (SOL) plasma with the first wall the evolution of both wall and plasma are tightly coupled: The erosion of the first wall leads to an impurity concentration in the plasma which affects the particle and power balance in the plasma. In turn the impurities, when leaving the plasma via transport, can form deposits and mixed materials, far away from their initial source location. These deposits can be eroded, allowing the impurities to stepwise migrate through the fusion device until they end up at a location where the plasma at the wall is cold enough and no further erosion occurs. To describe these processes an integrated model of surface evolution and plasma transport of impurities is needed. The WallDYN code achieves this required coupling of processes by parameterising the output of surface evolution- and plasma impurity-migration-codes by analytical models. For a given fixed background plasma it evolves the surface composition, derives impurity flux into and from the plasma and can from this derive the impurity densities in the plasma. This paper will show the importance of including this recycling of impurities at the wall in impurity migration modelling: The 13CH4seeding experiment performed in the Wendelstein 7-X Stellerator is modelled using the recent extension of the WallDYN code to 3D plasma and wall geometries. A comparison with post mortem analysis of the 13C deposition shows both qualitative and quantitative agreement with the WallDYN calculations.

During its second operational campaign (OP 1.2a) W7-X operated with an edge island divertor. Much higher densities (1020m−3) than in to the previous, limiter campaign were achieved. In order to asses this improved performance and the nature of transport with regard to the expected higher densities and other parameters such as the electron and ion temperature, the electric field and plasma flow, measurements at the plasma edge were conducted with reciprocating probe, mounted on a Manipulator. The probe contains a set of Langmuir pins, Mach probes, compensation coils and a set of 3D pick up coils for the measurement of the local magnetic field and fluctuations, and an ion sensitive probe. It has been observed, in addition to differences due to the magnetic configurations, that the magnetic topology is sensitive to the plasma conditions, especially to changes due to a evolving toroidal plasma current. Measurements from divertor cameras showed that an increasing toroidal current in the standard configuration broadened the heat flux and shifted the strike line considerably. These down-stream observation can be complemented with the measurements at the mid-plane with the manipulator mounted probes. During the experiments the edge island remnant was identified with the measured edge plasma profiles and compared with the predictions from the field line tracing code. Those electron densities and temperatures have been used as an input for understanding edge transport with the EMC3-EIRENE modeling. The results from EMC3-EIRENE also help improve the caculated density profiles, since the modelling provides a profile of the effective charge Zeff. Keywords: W7-X, Langmuir probes, Magnetic islands, Plasma edge transport

Wendelstein 7-X, the world largest superconducting advanced stellarator, aims to demonstrate high-performance steady-state experiments lasting up to 30 min. To this purpose, high heat flux (HHF) divertors capable of withstanding steady-state heat fluxes up to 10 MW/m2 are being installed on the machine, in preparation for the next experimental campaign (OP2.1). The real-time heat flux estimation is pivotal for controlling the divertor heat loads during the experiments. Currently, the THEODOR (Thermal Energy Onto DivertOR) code computes the heat flux offline by numerically solving the heat equation, but the computation time does not allow the application of this approach to the real-time operation of the device. In this work, a new approach based on Physics Informed Neural Networks (PINNs) is proposed for solving the heat equation and estimating the heat fluxes on the divertor tiles in real time. PINN models are Neural Networks that learn Partial Differential Equations (PDEs) by minimizing the PDE loss. The inputs of a PINN are the independent variables of the PDE, e.g., spatiotemporal coordinates, while the loss of the model is designed to make the neural network satisfy the PDE and related initial and boundary conditions. Hence, the model can be trained without any experimental data. Only the initial and boundary conditions of the PDE are necessary for constructing the model. First intermediate results are discussed considering a normalized tile and starting from a constant thermal diffusivity.

One of the main aims of Wendelstein 7-X (W7-X), an advanced stellarator, is to investigate the quasi-steady state operation of magnetic confinement devices for nuclear fusion, for which power exhaust is an important issue. A dominant fraction of the energy leaving from the confined plasma region will be removed by 10 so-called island divertor units, which are designed to sustain a maximum heat flux of up to 10 MWm−2. An essential prerequisite for the safe operation of a steady-state device is automatic detection of hot spots and other abnormal events. Simple temperature limits in infrared (IR) thermographic images will not be enough on their own, because of plasma-generated surface coatings and other effects summarized in the following. To protect divertor elements from overheating, and to monitor power deposition onto the divertor elements, near real-time hotspot detection algorithms for the analysis of carbon plasma facing components (PFCs) were implemented and tested in the GLADIS facility.One of the difficulties in hotspot detection in a carbon-based machine is the deposition of plasma impurities as layers with a reduced thermal connection to the underlying bulk material. We have developed and successfully tested a method to classify surface layers and benchmarked the performance of the method with the Tore Supra IR data operating with actively cooled carbon PFCs. The surface layers can be detected in a steady plasma discharge during the initial rise and decay in temperature when a strike line touches parts of the divertor or wall. It can also be detected by modulating electron cyclotron resonance heating (ECRH) input power. This feature allows detection of overheated areas while reducing false positives. For the recent operational campaign, inertially cooled test divertor units (TDU) were installed to prepare for steady-state operation with water-cooled divertor units. Automatic, near real-time detection of hot spots and identification of surface layers in the W7-X divertor are presented. Results are compared with a best fit estimate of the heat transmission coefficient α which is used to calculate heat flux onto the divertor in the presence of surface layers. Keywords: W7-X, Divertor, Plasma facing components, Heat load, Infrared, Surface layers, Tore supra

Hot spots and bright patterns on plasma-facing components (PFC) in visible light are observed in electron cyclotron resonance heated (ECRH) plasmas at low densities in the stellarator Wendelstein 7-X. The events are often located far outside of any convective plasma loads and led in some cases to damages of diagnostic components, where the interaction zones qualitatively agree with fast particle loss simulations. Reciprocating electric probes indicate that this phenomenon is related to a non-thermal electron population that can have a beam like character, being directed in one parallel direction. Further, the electrons can be trapped on rational flux surfaces and be used to map magnetic islands.