Performance assessment of the KSTAR MSE diagnostic

  • A.G.G. Lange

Student thesis: Master

Abstract

For relevance in energy production, fusion operation should be steady-state with high output power. This requires regimes of operation in which the amount of noninductively driven current is optimised. The KSTAR tokamak has been built with the aim to develop these advanced scenarios, which can be characterised with the current density profile, j. To measure and to control j, a Motional Stark Effect (MSE) diagnostic is in development at KSTAR. Such a diagnostic measures locally the magnetic pitch angle, giving the unknown poloidal magnetic field. This thesis investigates the ability of this diagnostic to distinguish the various operating scenarios. To establish a means to assess this ability, MSE measurements are simulated withthe use of two simulation codes BEAMCODE and MSESIM. Firstly, there was investigated whether these codes account for all the relevant effects. This has been done by comparing their output to spectral measurements from the 2012 and 2013 KSTAR campaign. This benchmark is divided into (a) the shape: here the Doppler shift, the Stark Shift, the power fractions and the spectral asymmetry are analysed. Then the residual difference between measurement and simulation was observed to be in the order of the measured noise, completing the benchmark to achievable accuracy. (b) the intensity (including the emission and throughput): This comparison has been done by integrating sphere calibration and Bremsstrahlung calculation. Both were consistent with the MSE measurement/simulation. However, for the plasma density, which is an input parameter in the emission intensity simulation, no trustworthy profile data was available. It was concluded that by adjusting the input plasma density profile, agreement between measurement and simulation could be reached. During the performance assessment, the simulation codes MSESIM and BEAMCODE were used to simulate MSE light resulting from artificial plasma equilibria. To synthesise future scenarios, a tool, EQUINU, has been developed that computes the poloidal magnetic field based on an analytical expression for the q-profile. With the MSE output a simulated voltage signal was generated that would be measured by the future MSE diagnostic. Several input equilibria, defined by their q-profile, have been analysed in this way and the output pitch angle profile has been compared to the input. These are well in agreement along the range of measurement (from 10 cm inside the magnetic axis to almost the plasma edge). Two types of deviations of result from input were found: (1) the systematic deviation that is introduced by the act of filtering and by the simulation of volume effects (view rays, beam emission volume); (2) the statistical deviation that is introduced by the noise on the signal. The first one is in the order of 0.2° and it can hardly be reduced, as the magnitude is also dependent on varying effects as the beam energy and the magnetic field. The statistical error can be reduced by time integration and has been found to reduce as ?? µ (?t)-1/2. Within a few milliseconds, the statistical error is below the systematical error. From this work, the MSE diagnostic seems well capable in performing its task.
Date of Award31 Aug 2014
Original languageEnglish
SupervisorRoger J.E. Jaspers (Supervisor 1) & Y.S. Bae (Supervisor 2)

Cite this

'