数论方法在统计公差分析中的应用

Monte-Carlo仿真方法通常用于复杂装配组件的公差分析,其主要缺点是计算量大、效率低.对此,提出以基于数论方法产生的伪随机数代替由Monte-Carlo方法所产生的随机数进行仿真.与随机数相比,伪随机数的分布更为均匀.算例结果表明基于伪随机数的公差分析具有Monte-Carlo方法简单实用的优点,同时计算量小,在计算精度相同的情况下计算效率可以提高10～20倍.

APPLICATION OF NUMBER THEORETIC METHODS IN STATISTICAL TOLERANCE ANALYSIS

Monte-Carlo simulation is used to determine the sets of tolerancce in the process of statistical tolerance analysis and synthesis. A large amount of samples are used to increase calculation accuracy because of the inhomogeneity of the random number based on the Monte-Carlo method. Instead of random numbers, the pseudo random numbers are used for statistical tolerance analysis based on number theoretic method. The distribution of pseudo-random numbers is more homogeneous than that of random numbers based on Monte Carlo method. Consequently, it is indicated that the method presented is simple and practical as the Monte-Carlo Method, and the calcuslation amount amount compared with that of random numbers may be decreased 10-20 times with the same accuracy.