Z箍缩中子编码诊断系统模拟平台搭建

Development of Monte Carlo code for Z-pinch driven fusion neutron imaging diagnosis system simulation

开发了基于Geant4的Z箍缩中子编码成像系统模拟平台,实现聚变中子编码成像诊断系统各关键部件的完整模拟。获得了低中子产额(约1010量级)下,中子经编码孔编码后在闪烁体阵列中形成的发光分布图像。利用维纳滤波、Richardson-Lucy(RL)及遗传算法(GA)对低中子产额下获得的极低信噪比图像进行重建,并对信噪比、中子产额及重建效果进行了对比研究,结果表明:遗传算法对低信噪比中子编码图像的重建具有很强的鲁棒性;中子编码图像的信噪比与遗传算法重建结果的准确性呈正比。

The model of Z-pinch driven fusion imaging diagnosis system was set up by a Monte Carlo code based on the Geant4 simulation toolkit. All physical processes that the reality involves are taken into consideration in simulation. The light im- age of low neutron yield （about 101~） pill was obtained. Three types of image reconstruction algorithm, i.e. Richardson-Lucy, Wiener filtering and genetic algorithm were employed to reconstruct the neutron image with a low signal to noise ratio （SNR） and yield. The effects of neutron yields and the SNR on reconstruction performance were discussed. The results show that genetic al- gorithm is very robust for reconstructing neutron images with a low SNR. And the index of reconstruction performance and the image correlation coefficient using genetic algorithm, are proportional to the SNR of the neutron coded image.