Research On The α Spectra Analysis Technology Based On Portable Spectrometer With The SSB

With the launching of the reactor operation, and decommissioning activities, aswell as nuclear accidents, all of them produce a large amount of radioactivewastes.Because of the strong dangers of these wastes to the human environment, wemust dispose them properly to ensure the security of human life and production. Priorto recycling, treatment or disposal of these radioactive wastes, we must measure theradioactive level of the wastes for the classification and determining the treatmentoptions. Alpha measurement is one of the key measurements in nuclear radiation,which occupies an important position in the field of nuclear radiation monitoring.Specifically at the job site, it needs an accurate and fast measurement ofradioactivity, so the urgent need is to establish a set of alpha detection technologyapplied to the scene, to achieve accurate, effective alpha measurements of pollution atthe contaminated site.In order to achieve the accurate and rapid measurement at the site of pollution,the group has independently developed a portable alpha spectrometer, which bases onthe Au-Si surface barrier (SSB) detector. The group also developed a vacuum systemfor the spectrometer, so that the detector works under vacuum conditions to avoid thecomplex interactions between α-particle and air, which result in the energy loss of-particle, in order to achieve the accurate measurement of the total activity. But onthe total α activity is far from enough, it is necessary for the analysis of αspectroscopy for accurate analysis of the type and activity of α-radioactivity in thesamples.Since1987, the foreign to the alpha energy spectrum of numerical analysis hasmade many significant progresses. However, the traditional alpha spectroscopy analysis methods base on empirical formulas of energy spectrum fitting. Alphaspectroscopy analysis is a very complex non-linear problem, the first step of theanalysis is setting the initial value for fitting parameters. It needs many ways to solvethis problem, first some of the parameters of the linear fit to take the a priori value,while keeping all other parameters being constants. This process is so complicatedthat takes very large energy and consumes much time but may not be able to obtaingood results.In recent years, researchers have proposed the use of neural networks to analyzethe energy spectrum, artificial neural network (ANN) does not use the experimentaldatas to fit mathematical functions, but by using the actual α-spectroscopy peaks as amodel, then with the ANN for its forecast training. However, this method is still in itsinfancy, the research about them is very rare. Domestic research about alphaspectroscopy analysis technology is very little, and the artificial neural networkresearch is just in its infancy, so it restricts the extensive promotion and application ofalpha spectroscopy analysis.This thesis focuses on the key issues of the alpha spectroscopy analysistechnology, on the basis of the research status of research, relying on nuclearinstrument of geophysical exploration technology development and applied research-the National Natural Science Foundation of the National Outstanding Youth ScienceFoundation (41025015) and the ministry of science and technology innovationmethods work special (2008IM021500)"the emergency nuclide quick analysissystem development" project, carried out the analysis techniques basing on theα-energy spectrum of the Au-Si surface barrier semiconductor detector system.The main research contents and results are as follows:(1) The energy calibration and efficiency calibration of the alpha spectrometersystemUsing the preliminary experimental data, we do the energy calibration andefficiency calibration of the spectrometer system under different measurementconditions, and analysis the spectrometer in vacuum conditions on the detectionefficiency with the same energy in the detection of radioactive sources of differentintensity.The results showed that the detection efficiency of the same energyα-particle source of strength and the degree of vacuum meet the one-time index, thehigher the degree of vacuum, the better the detection efficiency of the instrument.(2) Study compared the rate of energy loss of alpha particles in the air By processing the experimental data, obtained the energy loss of alpha particlesin different vacuum conditions, and compared with the calculated results show thatthe experimental and theoretical values correspond very well, prove the results ofalpha spectrometer is true and reliable.(3) Use the experimental data of the BP network for the prediction of the energyspectrum ofTo build the BP neural network on the MATLAB platform based on LMalgorithm, through the network training, achieve the prediction of the interestingregion with-spectra.The comparative study between predicted spectrum and theexperimental spectrum showed that the BP network has a good pre-dicing results.(4)The study of spectroscopy solution spectrum technique based on theempirical formulaIn determining the solution spectrum model based on the empirical formula ofspectroscopy, try to set the initial values of model parameters, and ultimately to theoptimization of each parameter. Solution spectrum model is used in the experimentalspectra to obtain a better imitative affect, the correlation coefficients are all above0.99. Finally making comparative studies based on neural networks,, the resultsshow that the neural network has a simple, accurate advantages,which can be betterapplied to the alpha spectroscopy solution spectrum technology.This subject is to carry out basic research work at the same time there is a degreeof innovation in methods and technology, the final outcome of the purpose of accurateprediction for spectrometry, to establish a more accurate alpha spectroscopysolution spectrum technique.There are three main innovation of this paper:(1) Through the experimental study, we summed up the characteristics of thealpha spectrometer key parameters (vacuum conditions, different-particle energy)spectrometer output, and the establishment of a scale model;(2) Combined traditional test methods and advanced mathematical methods foralpha spectroscopy analysis technology to form a complete set of complex alphaspectrum solution spectrum theory and system;(3) Comparison studied the traditional method of fitting based on the empiricalformula of spectroscopy with neural network to predict the spectrum fittingmethod, to obtain better analytical methods. This paper starting from the basic theory of alpha-ray spectrometry, to theprevious independent research and development of portable alpha spectrometer,to thedifferent surface activity of α-point source as a measurement object, based on themeasurement conditions of experimental data for different vacuum conditions onportable α.The spectrometer energy calibration and detection efficiency calibration indifferent vacuum conditions; and then study the energy loss of alpha particles indifferent vacuum conditions, and compared with the calculated value. Finally,preliminary research results based on the use of neural network based on the empiricalformula obtained under different vacuum conditions to obtain the α spectrum analysisfor the decommissioning of our nuclear facilities, the accumulation of activitymonitoring in areas such as environmental monitoring more the data.