Bayesian modelling of fusion diagnostics
Integrated data analysis of fusion diagnostics is the combination of different, heterogeneous diagnostics in order to improve physics knowledge and reduce the uncertainties of results. One example is the validation of profiles of plasma quantities. Integration of different diagnostics requires systematic and formalized error analysis for all uncertainties involved. The Bayesian probability theory (BPT) allows a systematic combination of all information entering the measurement descriptive model that considers all uncertainties of the measured data, calibration measurements, physical model parameters and measurement nuisance parameters. A sensitivity analysis of model parameters allows crucial uncertainties to be found, which has an impact on both diagnostic improvement and design. The systematic statistical modelling within the BPT is used for reconstructing electron density and electron temperature profiles from Thomson scattering data from the Wendelstein 7-AS stellarator. The inclusion of different diagnostics and first-principle information is discussed in terms of improvements.
Bibliographic Reference: An article published in: Plasma Physics and Controlled Fusion 45 (July 2003), pp.1095-1111
Availability: This article can be accessed online by subscribers, and can be ordered online by non-subscribers, at: http://stacks.iop.org/0741-3335/45/1095
Record Number: 200316587 / Last updated on: 2003-08-21
Original language: en
Available languages: en