Experimental and computational magnetic surface mapping of the HIDRA stellarator

Rizkallah, Rabel

Permalink

https://hdl.handle.net/2142/105087 Copy

Description

Title

Experimental and computational magnetic surface mapping of the HIDRA stellarator

Author(s)

Rizkallah, Rabel

Issue Date

2019-04-25

Director of Research (if dissertation) or Advisor (if thesis)

Andruczyk, Daniel

Committee Member(s)

Curreli, Davide

Department of Study

Nuclear, Plasma, & Rad Engr

Discipline

Nuclear, Plasma, Radiolgc Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• Fusion
• stellarator
• magnetic confinement
• magnetic mapping
• magnetic flux surface
• magnetic islands
• plasma facing components
• plasma-material interaction.

Abstract

HIDRA, the Hybrid Illinois Device for Research and Applications, is a newly acquired fusion device by the University of Illinois at Urbana-Champaign. It is an l = 2 machine with a five-fold toroidal symmetry, that was previously run by the Max-Planck-Institut für Plasmaphysic in Greifswald under the name WEGA. Even with its much lower plasma temperature and density when compared to bigger machines such as EAST and W7-X, HIDRA can be used to conduct plasma-material interaction experiments. Its particle and heat fluxes should be high enough to test and develop novel plasma facing components, with a main focus on liquid lithium designs such as the FLiLi (Flowing Liquid Lithium) and the LiMIT (Liquid-Metal Infused Trenches) concepts. With the planning of new experiments to test both FLiLi and LiMIT plates on HIDRA, characterizing the magnetic configuration of the device is necessary. While previous campaigns were conducted on WEGA for the same purpose, with the disassembly and re assembly of HIDRA, the physical disposition of the coils and/or vacuum vessel may have been offset by even the slightest of margins. This can lead to additional perturbations of the magnetic field, changing the magnetic topology of the device. The electron gun and fluorescent detector technique was employed to experimentally measure the magnetic surfaces for rotational transform values of 1/3, 1/4 and 1/5. The electron gun used consists of a tungsten filament and fires free electrons at adjustable radial positions in the direction of the field. A moving rod covered with a zinc oxide (ZnO) powder was chosen as the fluorescent detector. The rod was used to sweep across fixed poloidal cross sections, at toroidal angles of φ = 144○ and φ = 216○ . Because of the five-fold toroidal symmetry of HIDRA, both positions display identical magnetic surfaces. A long-time integrating CCD-camera was used to capture the electron traces from the luminescent traces on the rod. The camera images were then used to build up Poincaré plots of the electron trajectories. This was carried out for the various rotational transforms, with and without the addition of a vertical field. The resulting images didn’t have the same resolution as the WEGA images, but still exhibited similar features. In particular, the n = 1 non-natural islands were visible on the HIDRA magnetic flux surfaces in almost the same position as they were seen on the WEGA ones. While the electron trajectories are subject to drifts from the curvature and gradient of the magnetic field, the induced displacement from the field lines is relatively small on the order of a few millimeters. Hence, the experimentally obtained surfaces derived from the electron traces are close enough to the magnetic flux surfaces. Computational magnetic flux surface images were also generated. For this, the FIELDLINES code of the STELLOPT suite of codes was used. The expected n = 1 error field was added to the code source files to reproduce the physical system at hand. The code follows field lines from an initial set of generators for 1000 toroidal transits. A poloidal section is then taken at a computational toroidal angle corresponding to the physical φ = 216 . Processing the coordinates of the magnetic surfaces in the Poincaré section allowed to estimate the location of the magnetic axis as well as determine the last closed flux surface (LCFS). To obtain smooth and closed magnetic surfaces, the ones inside the LCFS were filtered with a Savitzky-Golay filter after applying a change of coordinates to the local polar coordinate system centered at the magnetic axis. The filtered magnetic surfaces are then traced back into FIELDLINES. From the new output, the magnetic axis is found, and Poincaré sections are generated. This routine is repeated for every magnetic configuration tested during the experimental runs. Because of the limitations of the experimental apparatus and little time available to run the experiments, the obtained experimental images had a much lower quality than those generated computationally through FIELDLINES. The experimental images suffered in particular from heavy outgassing of the walls, which led to a background noise that was very difficult to filter out. Therefore, the comparison between the experimental and computational images was mainly qualitative. The two showed similar features from the location, size and shape of the magnetic islands, to the location of the magnetic axis and overall size of the magnetic flux surfaces. The results are also very close to those reported from the WEGA campaigns, suggesting that the experienced error field is similar and probably coming from the same origin. The effect of the vertical field was also investigated and shown to effectively suppress the non-natural islands and shift the position of the magnetic axis. However, because the camera didn’t have a direct perpendicular view to the fluorescing rod, the experimentally measured magnetic axis radial shift was smaller than the predicted ∆R ≈ 2 cm. Similarly, the measured magnetic axis radial position from the experimental images was higher than the expected R ≈ 70.5 cm at R = 71.2 cm. This misalignment of the camera perspective was hinted at by the higher density of electron traces observed near the low-field side on the experimental images. Taking this into account explains the difference between the experimental and computational images. The additional control given by the vertical coils is very useful to carry out the upcoming tests of plasma facing components inside HIDRA. Apart from removing the low-order rational resonances, shifting the magnetic axis towards the high-field side by ≈ 2 cm creates additional space to maneuver and install various systems along with their support structure. Manipulating the rotational transform in combination with adding a vertical field can displace the plasma and change its angle and point of contact with the surface of the tested component. Knowing the magnetic topology of HIDRA and the control that can be applied on its magnetic structure, a LiMIT and FLiLi setup to be installed inside the vacuum vessel is being designed. A smaller collector and LiMIT plate than the ones planned for the EAST runs have already been machined, with plans to further reduce the plate size being considered. The mapping of the HIDRA magnetic flux surfaces will also be used to create a magnetic grid for the machine that would allow computational calculations of the particle and heat fluxes the plates will be subject to. Experiments to verify these fluxes are being planned.

Graduation Semester

2019-05

Type of Resource

text

Permalink

http://hdl.handle.net/2142/105087

Copyright and License Information

Copyright 2019 Rabel Rizkallah

Owning Collections