Applications of neural networks for free unfolding of experimental data from fusion neutron spectrometers. (25th January 2010)

• Record Type: Journal Article
• Authors: Ronchi, Emanuele; Conroy, S.; Sunden, E. Andersson; Ericsson, G.; Johnson, M.; Hellesen, C.; Sjostrand, H.; Weiszflog, M.; contributors, JET-EFDA
• Is Part Of: International journal of nuclear knowledge management. Volume 4:Number 1(2010)
• Page Start: 25
• Page End: 31
• Publication Date: 2010-01-25
• Subjects: neural networks -- free unfolding -- unfolding -- NE213 -- magnetic proton recoil upgrade -- MPRu -- fusion neutron spectrometers -- deconvolution -- real-time; Nuclear energy -- Periodicals; Nuclear energy -- Management -- Periodicals; Nuclear industry -- Periodicals; Nuclear industry -- Management -- Periodicals; 333.792405
• Languages: English
• ISSNs: 1479-540X

Abstract

Free unfolding in neutron spectroscopy means reconstructing energy spectra from experimental data without a priori assumptions regarding their shape. Due to the ill-conditioned nature of the problem, this cannot be done analytically. Neural Networks (NNs) were applied to this task and synthetic data was used for training and testing. Results showed very consistent performance especially in the region of low and medium counts, where they fall near the Poisson statistical boundary. Comparison with other unfolding methods validated these results. Application time on the order of ms makes NNs suitable for real-time analysis. This approach can be applied to any instrument of which the response function is known.