Determining the configuration of snowflake divertor plasmas in real-time with OFIT

Abstract

One of the outstanding problems in fusion research is the heat load on the divertor wall. The current method to handle the heat load is based on a magnetic configuration in which the plasma exhaust is distributed over two divertor legs. This method may not provide sufficient heat load handling for large future devices. Alternative divertor concept, such as the snowflake divertor configuration, are therefore investigated in present day machines. The snowflake divertor has a second order null point in the poloidal magnetic field. This leads to two additional divertor legs and flux expansion which reduces the heat flux per surface area on the wall. The pure snowflake configuration is topologically unstable. However, by locating the two x-points close to each other, more stable configurations can be achieved. These configurations are called the snowflake (+) and snowflake (-). To identify the snowflake configuration, visible light images are considered.This thesis has two main questions: 1) Can the snowflake divertor configuration be determined from visible light images? 2) Can this analysis be executed in real time?For the first question, images of the snowflake configuration are acquired with a visible light camera. The plasma edge will be reconstructed with the plasma edge reconstruction method OFIT. OFIT is developed to interpret toroidaly symmetric objects and is therefore suited for this type of analysis. The intensity distribution of the snowflake plasma edge was examined, and the relation between the intensity, contrast and location of the features and the corresponding configuration is studied.The visibility study shows that a single null divertor plasma evolves to a full snowflake divertor plasma at ?=0.5~0.6 at ?=75°~100°. To extract the snowflake divertor configuration parameters, two more advanced methods are proposed for future investigation.For the second question, the OFIT algorithm needs to run in real time with a frequency of 1kHz. To achieve this, part of the OFIT code must be ported to an FPGA. A proof of concept of the real-time OFIT code is implemented. The results from this proof of concept are used for a feasibility study of the planned hardware for the real-time implementation on the tokamak. It was found that the algorithm can run on the test FPGA at a frequency of 250Hz. Optimisations are proposed to reach 1KHz with the hardware used at the final realisation.

Date of Award	30 Sept 2013
Original language	English
Supervisor	G. Hommen (Supervisor 1) & M.F.M. de Bock (Supervisor 2)