用蒙特卡罗数值模拟评估可靠性试验中子样容量确定方法

THE MONTE CARLO EVALUATION OF THE METHOD USING CURRENT TEST DATA TO DETERMINE SAMPLE SIZE

针对长寿命高可靠性结构设计的特点，着重强调子样统计特性的无偏性，对可靠性试验（或抽样试验）中现有文献采用的由试验结果确定子样容量的方法进行评估，阐明了此方法在估计某些母体参数时的有偏性及其对可靠性分析的不安全影响，并且通过蒙特卡罗数值模拟的方法证实了这一观点。最后提出新的建议。

In the long life and high reliability design and analysis of structures, the unbiasedness of statistical property of sample is more important. In this paper, by means of Monte Carlo method, we evaluate the present method that uses current test data to determine sample size. The Monte Carlo simulations show the biasedness of this method and the possible unsafe effects on reliability analysis and design. Finally, we give a new proposal.