Research On Fast Nuclide Recognition Method Based On Background Comparison Model

Radioactive material inspection in public places is an important part of nuclear safety,as well as a key part of security for holding large-scale events,and fast and efficient radioactive material detection methods are an important technical guarantee for nuclear safety.In the contemporary era of rapid development of nuclear technology,there are more and more cases of smuggling,proliferation,and illegal and malicious use of radioactive materials.Strengthening the management of radioactive materials and strengthening the tracking and monitoring of nuclear materials have become important aspects in the field of nuclear safety.The spectrum-based nuclide identification method is based on the measurement and comparison of the gamma spectrum.It needs to measure for a long enough time and detect enough photons to obtain better statistics,and then to effectively judge whether a certain nuclide exists.In the security inspection equipment of railway stations,airports,and customs,the passage time of items is only a dozen seconds.In such a short detection time,the number of photons that can be collected is very limited,and it is difficult to identify nuclides with gamma spectroscopy.In the process of radioactivity monitoring with nuclear safety,nuclear signal detection is the foundation,and rapid nuclide identification is the goal.In order to achieve the requirement of fast nuclide identification,this thesis proposes a fast nuclide identification algorithm based on detector particle event and background comparison model.The radionuclide identification method established in this paper integrates radionuclide half-life,characteristic?-ray energy and branching ratio information,proposes EMS(Event Mode Sequence)sequence,establishes a new background comparison model of the improved sequential Bayesian detection method,and then through the event screening algorithm,parameter estimation based on conjugate prior distribution and particle filtering,and real-time update of the decision function The radionuclide identification is achieved,and fast identification of radionuclides in natural radiation fields is realized.The correctness of discrimination and response speed of low activity radionuclides in natural radiation background are improved.Based on the theoretical study,a simulation environment was built using Matlab software.In the simulation experiments,EMS were formed using Monte Carlo method sampling,and then the event screening algorithm,parameter estimation and decision function update were used to identify whether the randomly sampled EMS contained nuclides or not.The equivalent activity is used to characterize the activity of nuclides and the relative detector distance during the simulation experiments,and the detection time and detection rate of the background comparison model-based nuclide identification method in different environments are obtained by varying the equivalent activity or detector resolution.Simulation experiments were conducted to verify the feasibility of the method and evaluate the theoretical performance of the method,and the results demonstrate that the detection time performance of the background comparison model-based nuclide identification method is better than that of the general nuclide identification method.To realize the application of rapid identification method of radionuclides in the field,radiation particle event acquisition hardware and radionuclide identification software are developed.In order to accurately capture the particle event information,the hardware design uses an FPGA to control the analog constant ratio timing circuit to capture the particle event arrival time,and the analog peak hold circuit to control the energy of particle events in conjunction with a low-power and medium speed ADC.This design reduces the power consumption of the system design and lays the foundation for localized ADC replacement.This thesis is based on CeBr₃detector with nuclide identification hardware and software platform to carry out experimental validation of nuclide identification method based on background comparison model,including tests on different radionuclides,different detection distances,different radioactivity and motion radioactive sources.The experiments show that the average detection time reaches6.14s when using a 9200Bq ⁶⁰Co target source at a relative distance of 55cm from the detector,6.51s when using a 18100Bq ¹³⁷Cs target source at a relative distance of75cm from the detector,and 9.10s when using a mixed ¹³⁷Cs and ⁶⁰Co target source at a distance of 65cm from the detector.The average detection time reaches 9.10s;when a 8700Bq ¹³⁷Cs target source is 20cm away from the detector,the radioactive source can be detected in 3s when it passes quickly.