Direct measurement of JET local deuteron densities by neural network modeling of Balmer alpha beam emission spectra

The analysis of Balmer alpha beam emission spectra represents a challenging task, both in terms of the wealth of information hidden in them, and in terms of complex spectral features. The shape of a spectrum depends on many local plasma parameters, such as magnetic field strength and direction, beam density, effective charge and deuteron density. This paper is concerned with the deduction of local plasma deuteron densities from Balmer alpha emission from plasma atoms following charge exchange with the beam atoms. The model we will use is statistical, and is based on multi-layer perceptron neural networks (Hertz er al 1991, Bishop 1995). The use of neural networks makes the deconvolution task fully automatic and fast enough for real-time calculation of complete deuteron density profiles. It is shown that the spectra themselves and local electron densities are the only data necessary for accurate inference of Local deuteron densities. This result is partly inferred from a sensitivity analysis of dependences on different plasma parameters. Proper error bars for the model predictions will be derived using Bayesian probability theory.