Font Size: a A A

Research On The Nuclide Identification Algorithm And Digital Spectra Acquisition System

Posted on:2010-07-14Degree:DoctorType:Dissertation
Country:ChinaCandidate:L ChenFull Text:PDF
GTID:1102360308957497Subject:Nuclear Science and Technology
Abstract/Summary:PDF Full Text Request
Spectra measurement and nuclide identification is key technology widely used in nuclear monitoring. Nuclides identification algorithm and digital spectra acquisition system are core technology of the spectroscopy system. According to the test requirement for portable RID (Radioisotope Identification Device) drawn up by IAEA, we research several probloms on nuclide identification algorithm and digital spectra acquisition system; it includes the following four parts:(1) NaI(Tl), CZT, and LaBr3(Ce) detectors are modeled using Geant4 Monte Carlo simulation program package. Theγresponses of the three types of detectors are calculated according to the IAEA's requirement. It is shown that the calculated data are reliable by comparing to the experiment data,and it provides reliable basis for the detector choice and nuclide identification algorithms research.(2) Design and implement a series of nuclide identification algorithms suitable for portable RID. Smooth and peak searching for the spectra data,"weighted count calibration"digital stabilizing method and nuclide identification method based on characteristic peak matching are discussed. The nuclide library special for the algorithms are established. The algorithms are applied in self-designed devices, and tested by authority department. The test results show better performance than similar devices.(3) Propose a novel nuclide identification algorithm based on feature extraction using K-L transform and artificial neural network. The algorithm introduces the pattern recognition idea into the nuclide identification. The whole shape of the spectra is used as the judgement of identification. It makes full use of the the spectra data imforamtion. We test the performace of the algorithm by experimented data, and discusse the results in single nuclide, mixed nuclides and unknown nuclide situations. The new algorithm has features of low level computation, wide tolerance, good applicability and insensitive to noise. It provides a new idea for spectra analysis and nuclide identification. (4) A FPGA based digital spectra acquisition system is designed. All modules of multi-channel pulse height analysis, including digital pulse shaping, threshold denoising, pileup rejection, baseline estimation and hight extraction are integrated in a FPGA chip. We obtained spectra data using the acquision system and a self-designed monitoring softeware on PC. The system has much better throughput under same accuracy.The research work of this article is applied in first digitial portable spectroscopy in China, and good results are obtained. It also lays the foundation for the research and development of high performance digitial spectroscopy.
Keywords/Search Tags:nuclide identification, Monte Carlo simulation, feature extraction, artificial neural network, spectra acquisition
PDF Full Text Request
Related items