| American based on time correlation coincidence detection technology developed nuclear material identification system, using ext ernal neutron source to determine uranium components mass abundance information, which can effectively detect different levels of uranium components, but the parameters has not yet been made public. For the cost, many institutions are using the simulation mode to collect data. Monte Carlo method is with the atomic energy industry and the development of a random sampling method, realistic simulation of atomic physics process consistent with the actual experimental results are obtained. GEANT4 as a mainstream of the Monte Carlo open source software package, by adding thermal neutron reaction cross sections and process to simulated uranium components quality abundance detection data acquisition process.This paper firstly evaluated neutron international section are obtained the corresponding data, and then rewrite the neutrons and uranium components reaction process in Geant4; according to the set of input parameters and abstract model construction of uranium components and detectors and other entities and define the material and the detector sensitive range, and sampling value definition of neutron source energy; the control variable method, each time only change the single parameters, multiple data collection and comparison, determine the parameters can effectively distinguish the detection result; through the use of buffer pool and timing sampling method with Open GL and Win Form show the process; the results of data are saved to the corresponding document; at the same time introduced the evaluation index, and us e a variety of optimization scheme.In this paper, through the experimental comparison, the parameter setting can distinguish the experimental results obtained. In the use of the group setting parameters, by fixing abundance was 90%, collecting the quality for 8kg to 25 kg of uranium components detection data; 17 kg by fixed quality, acquisition of the abundance of 20% to 90% of the testing data. |