Bayesian analysis of electron cyclotron emission measurements at Wendelstein 7-X

Electron cyclotron emission spectroscopy (ECE) is a standard diagnostic technique on the optimized stellarator Wendelstein 7-X (W7-X) that can record data with a high resolution in time. The spatial assignment via the blackbody emission of a plasma layer depends on the optical depth and thus plasma parameters as well as the magnetic field along the line of sight. The measurements with a multichannel radiometer contain a large amount of information about the electron temperature profile, as well as being very sensitive to the magnetohydrodynamic equilibrium at W7-X. First, the diagnostic was comissioned and absolutely calibrated. At W7-X this is achieved by an optical system identical to the plasma measuring system, which alternately measures room temperature and liquid nitrogen temperature of a microwave blackbody radiator by means of a rotating mirror. The signal difference associated with the temperature change then permits the determination of the calibration factors, the accuracy of which represents the most important source of uncertainty for the diagnostic. In order to allow a systematic treatment of the uncertainties, a completely new, general Bayesian forward model of a calibration unit with rotating mirror was developed and tested within the Bayesian modeling framework Minerva. The calibrated data then allow to obtain a radiation temperature spectrum. The actual desiderata, i.e. the sought-after quantities, however, are the electron temperatures on the effective plasma radius. Traditionally, the emission region is approximated by the cold resonance location. However, this method is inaccurate if, for example, relevant plasma pressure is reached that leads to a modification of the magnetic field along the line of sight. For the more precise determination of the emission region and the underlying electron temperature profiles, forward modelling must therefore be carried out taking into account the radiation transport along the line of sight. Furthermore, the ray should be determined by raytracing, since ray deflection via the plasma parameter-dependent refractive index can have a serious influence on the model predictions, especially at higher densities. Both is achieved by incorporating the tracing visualized (TRAVIS) code into Minerva, in which the forward model of the electron cyclotron emission (ECE) for W7-X is written. The model includes a prediction of line-integrated electron density via interferometry. One of the advantages of this completely new model is that it is relatively general and should allow easy transferability to other machines, as well as compatibility with the »plug’n’play« neural network generator currently in development. As examples for applications, the model is used to obtain information about the absolute values of the electron density profile during low and high density plasma discharges, which is a good addition to the already existing possibilities of density measurements by, for example, the Thomson scattering diagnostic. Finally, the ECE data is used on a simple Bayesian heatwave analysis model in an attempt to obtain the electron heat diffusivity.

Author dc:creator

• Höfel, Udo

Advisor dc:contributor.advisor

• Hirsch, Matthias

Licence dc:rights.uri

• https://creativecommons.org/licenses/by/4.0/

Language dc:language.iso

en

Identifier URI

http://dx.doi.org/10.14279/depositonce-9621

OAI identifier oai:identifier

oai:depositonce.tu-berlin.de:11303/10724