Community Research and Development Information Service - CORDIS


In many plasma physics experiments the desirable line-of-sight access has to be balanced against technological constraints. In many instances the data are sparse and corrupted by noise. The use of linear inversion methods can exaggerate defects in the data. This scheme for analysing difficult inverse problems is based on the idea of describing the quantities to be reconstructed in terms of positive, additive distributions. Four different physical measurements, each involving an inverse transform in the analysis of diagnostic data from research in controlled thermonuclear fusion, are described and analysed within the framework of the maximum entropy (MAXENT) formalism. The generality of MAXENT is stressed; it is also applicable to many similar data analysis problems which are encountered outside the specific field of plasma diagnosis. The four cases dealt with here are: i) inversion of Abel-transformed data, ii) deconvolution of electromagnetic line spectra, iii) Fourier-transform spectroscopy and iv) two-dimensional tomographic reconstruction with sparse data. To illustrate some of the particular problems encountered in extracting information from laboratory plasmas, it is first necessary to give a brief overview of the essential features of experimental work in controlled magnetic fusion.

Additional information

Authors: COTTRELL G A, JET Joint Undertaking, Abingdon, Oxon. (GB)
Bibliographic Reference: Report: JET-P(90)04 EN (1990) 38 pp.
Availability: Available from the Publications Officer, JET Joint Undertaking, Abingdon, Oxon. OX14 3EA (GB)
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top