(A) study on gamma-ray spectra unfolding based on monte carlo simulation for water monitoring system

Abstract

Direct matrix inversion and Maximum Likelihood fitting by the Expectation Maximization (ML-EM) deconvolution methods are presented to analyze gamma-ray spectra recorded with an NaI(Tl) detector for a water monitoring system (WMS). The ML-EM iteration method based on a Poisson distribution is known as the best approach to estimate the parameters and errors of the original gamma-ray fluence energy distribution, especially when the number of counts is small. For applications in which it is not only necessary to obtain an accurate estimate of the number of counts in a particular full-energy peak, but it is also imperative to preserve the accuracy of the peak locations, the maximum likelihood method has been shown to be superior among the deconvolution methods found in the literature. The applicability of the two methods were tested by deconvolving measured spectra taken with an industry-standard 3`` × 3`` cylindrical NaI(Tl) detector in a model water tank using several calibration sources. The results show significant removal of the Compton continuum counts and efficient transfer of the counts into the corresponding photo-peaks. The peak-to-total count ratio and the number of counts in the photo-peaks in the deconvolved spectra increased approximately 4.67 and 5.29 times, respectively, compared to those of measured spectra taken with an NaI(Tl) scintillation detector for the case of $^{137}Cs$. In this work, only the feasibility study has been performed with a model water tank containing several calibration sources. The feasibility study confirmed that the developed ML-EM iteration algorithm can be practically applied to the WMS in YoungKwang NPP to analyze the radioisotopes in the exhausted cooling water under normal operation or any incident or accidental cases every couple minutes.