Unfolding neutron spectra obtained from BS–TLD system using genetic algorithm

Due to the variability of neutron spectrum within the same environment, it is essential that the spectral distribution as a function of energy should be characterized. The precise information allows radiological quantities establishment related to that spectrum, but it is necessary that a spectrometric system covers a large interval of energy and an unfolding process is appropriate. This paper proposes use of a technique of Artificial Intelligence (AI) called genetic algorithm (GA), which uses bio-inspired mathematical models with the implementation of a specific matrix to unfolding data obtained from a combination of TLDs embedded in a BS system to characterize the neutron spectrum as a function of energy. The results obtained with this method were in accordance with reference spectra, thus enabling this technique to unfold neutron spectra with the BS–TLD system.

► The unfolding code used the artificial intelligence technique called genetic algorithms.
► A response matrix specific to the unfolding data obtained with the BS–TLD system is used by the AGLN.
► The observed results demonstrate the potential use of genetic algorithms in solving complex nuclear problems.