Environmental Radioactivity Gamma Spectroscopy Analysis Technology Based On Simulated Annealing Algorithm

With the nuclear power industry development and wide application of nuclear technology, the radioactive waste in emission to air, water and soil is increasing which has posed a serious threat to humans and the environment ,this has caused our government and public attention. Therefore ,it is of important significance of fast and accurate measurements of radionuclide activity.γspectrum analysis is one of important environmental radioactivity detection technology means ,including HPGeγspectrometer and NaIγspectrometer. At present, theγspectrum analysis software is on the basis for quantitative analysis including all-around peak or peaks including all-around spectrum HPGeγspectrum analysis software calculate the main round of the peak area by curve fitting. Usually the user is required to set the exact peak shape function and the related initial parameters, local optimization algorithm is generally used , if the parameter is not set properly, it is easy to fall into local minimum, which received only local optimal solution, especially for weak peaks and overlapping peaks, it is difficult to get all-round satisfactory results. The NaIγspectrum analysis software mainly makes use of spectral stripping method and inverse matrix method to do the numerical calculation in the peak area of the spectrum. Because NaI detector energy resolution is low and peaks overlap severely, spectral analysis result is in the severe impact of drift and statistical fluctuations. In addition, there is no analysis software which apply to HPGeγand NaIγspectrum and spectrum at the same time.Based on the linear superposition ofγspectrometer system, we proposed quantitative analysis method of full spectrum based on global optimization algorithm - simulated annealing algorithmγspectrum .Based on quantitative analysis of the whole spectrum, this method improves the utilization of spectral data, and applies to both HPGeγspectroscopy and NaIγspectrum data analysis. Mainly for the following work.The peak drift ofγspectrum is corrected, and the simulated annealing algorithm is applied to theγspectrum analysis, the whole spectrum of theγspectrum analysis increases the utilization of spectrum data and reduces errors.Some spectrum can not be gotten by HPGe detector and NaI detector, so we used Monte Carlo simulation ofγspectrum in this experiment and verified it is feasible to simulate annealing algorithm using Monte Carlo simulation ofγspectrum. The correct results of all methods are given, and of high precision, relative error of less than 1.2%. Forγspectrum measured by the NaI detector, the method has also been the same or even higher accuracyWe practice performance testing in the four aspects of lower limit of energy analysis, energy resolution capability, statistical errors, drift correction. Compared with the conventional peak area analysis method of all-around peak, this method should have a lower limit; For theγ-ray energy with varying degrees of overlapping, accurate results are given without significant change in the analysis of accuracy. The method has little effect by overlapping peaks; This method can measure theγspectrum that we want to analyze in very short time, which improves work efficiency; with the peak shift increasing, There was no significant changes among analysis precision of the methods. They have given accurate results of the analysis, errors are 3.46% or less.The actual environmental samples were analyzed and standard samples were provided by National Institute of Metrology ²²⁶Ra,²³²Th and ⁴⁰K, respectively, contain only ²²⁶Ra, ²³²Th and ⁴⁰K radionuclides. RTK378 mixed standard source, that is standard environmental samples, is taken as the work of analysis, which contains different activity radionuclides of ²²⁶Ra, ²³²Th and ⁴⁰K. The results show that the simulated annealing algorithm forγspectrum measured by the results of the analysis is more accurate, analysis of error is less than 10%, indicating that the simulated annealing algorithm withγspectrum with full spectral analysis is feasible and effective...