Determination of experimental tokamak plasma profiles using maximum‐entropy analysis

摘要

The diagnostician commonly faces the problem of extracting the maximum information on plasma spatial profiles from only a very limited and noisy data set. Conventionally, functional forms (e.g., a parabola raised to a power) are fitted to the measured data using nonlinear least‐squares regression. This immediately biases the interpretation and constrains results to a narrow family of curves for which there may be no real evidence in the data. In this paper, we examine the consequences of abandoning this assumption and, instead, seek the constrained maximum entropy (ME) profile making no such assumptions. Results are compared with conventional analysis. The ME method has the inherent advantages of (i) yielding only positive (and therefore physically meaningful) profiles, (ii) suppression of noise on the data, and (iii) producing a profile which contains the least amount of spurious detail. The method is illustrated with reference to measured chord‐average CO2 laser interferometer data taken on the Doublet III tokamak.