论文部分内容阅读
为解决探测器校正因子MonteCarlo模拟中,两个随机变量比值计算效率过低和探测器尺寸过小,粒子难以到达探测器并发生碰撞的问题,在MCNP-4C平台上结合已有的Dxtran球和强迫碰撞两种减方差技巧,以及新实现的相关抽样,共同形成了一种高效的校正因子计算方法。选用了4种不同的技巧分别与相关抽样结合,对一个简化的探测器校正因子计算模型进行了计算,与该方法进行了比较。计算结果表明,相关抽样无论与哪种减方差技巧结合,都提高了校正因子计算的FOM(figureofmerit)值。而该文方法则比其中效果最好的方法的FOM值还要高两个数量级。
In order to solve the problem of Monte Carlo simulation of detector calibration factor, the calculation of the ratio of two random variables is too inefficient and the detector size is too small, the particles hardly reach the detector and collide with each other. On the MCNP-4C platform, the existing Dxtran ball and Forced collision two subtraction techniques, as well as the newly implemented correlation sampling, together form an efficient method of calculating the correction factor. Four different techniques were selected, respectively, in combination with the correlation sampling. A simplified calculation model of detector calibration factor was calculated and compared with this method. The calculation results show that the correlation sampling increases the figure of merit (FOM) calculated by the correction factor, no matter which reduction subtraction technique is used. This method is two orders of magnitude higher than the FOM value of the best performing method.