Artificial neural networks in neutron dosimetry

Abstract

An artificial neural network (ANN) has been designed to obtain neutron doses using only the count rates of a Bonner spheres spectrometer (BSS). Ambient, personal and effective neutron doses were included. One hundred and eighty-one neutron spectra were utilised to calculate the Bonner count rates and the neutron doses. The spectra were transformed from lethargy to energy distribution and were re-binned to 31 energy groups using the MCNP 4C code. Re-binned spectra, UTA4 response matrix and fluence-to-dose coefficients were used to calculate the count rates in the BSS and the doses. Count rates were used as input and the respective doses were used as output during neural network training. Training and testing were carried out in the MATLAB environment. The impact of uncertainties in BSS count rates upon the dose quantities calculated with the ANN was investigated by modifying by +/-5% the BSS count rates used in the training set. The use of ANNs in neutron dosimetry is an alternative procedure that overcomes the drawbacks associated with this ill-conditioned problem.

Spectral pre and post processing for infrared and Raman spectroscopy of biological tissues and cells

“…Inspired by biological neural networks, their popularity stems from their general applicability to any problem [95], and they have seen applications in radiation science to dose-dependent models for flow cytometric analysis [96], for the prediction of depth dose in radiotherapy [97], and for neutron dosimetry [98]. Udelhoven [75,94], and careful training and rigorous evaluation of the network is required to prevent this.…”

Robust Design of Artificial Neural Networks Methodology in Neutron Spectrometry

“…To unfold the neutron spectrum, Φ, several methods are used. [43,44,47] ANN technology is a useful alternative to solve this problem; [25][26][27][28][29][30][31] however, several drawbacks must be solved in order to simplify the use of these procedures.…”

“…Recently, the use of ANN technology has been applied with relative success in the research area of nuclear sciences, [3] mainly in the neutron spectrometry and dosimetry domains. [25][26][27][28][29][30][31] …”

Evolutionary Artificial Neural Networks in Neutron Spectrometry

“…Because the number of unknowns overcome to the number of equations, this is an ill-conditioned system and has an infinite number of solutions. The procedure of selecting the solution that has meaning for the problem type, is part of the unfolding process (Vega-Carrillo et al, 2005;2006). The spectral information needs to be unfolded from the BSS system detector responses by using a suitable computational code, most of them are based in some of these methods: least square, iterative (Bedogni et al, 2007;Miller, 1993), bayesian and maximum entropy (Reginatto & Zimbal, 2008), and Monte-Carlo (Vega-Carrillo et al, 2007a).…”

“…This genetic search capability is much more effective than random searching, as the genetic process of recombining features vastly improves the speed of identifying highly fit networks. To train the evolved networks, a data set of 187 neutron spectra, compiled by the International Atomic Energy Agency (IAEA) (IAEA, 2001) was used (Iñiguez & Vega-Carrillo, 2002;Vega-Carrillo et al, 2005;2006). The ANN genetically evolved were designed by meas of the NeuroGenetic Optimizer (NGO) software (BiocompSystems, 2010).…”