基于改进 Bayes-Bootstrap 方法的系统可靠性仿真评估

摘要

Aiming at the characteristics of system reliability assessment under small-sampling conditions and the shortcomings in the analysis of reliability data with traditional Bayes,the paper presentes an assessment method based on improved Bayes-Bootstrap.The field test data annihilation question existing in system reliability assessment by traditional Bayes method is analyzed firstly,and then an improved sampling arithmetic is raised on the foundation of an experience function.At last,the effectiveness of method is proved through an example.The research result has shown that the method is more accurate than traditional Bayes,effectively reducing the objectivity and complexity in the assessment of complex system.%针对小样本条件下系统可靠性评估的特点，以及传统 Bayes 方法在进行可靠性数据分析时存在的不足，提出了一种基于改进 Bayes-Bootstrap 的系统可靠性仿真评估方法。首先，对基于 Bayes 方法进行系统可靠性评估时易发生的现场试验数据“湮灭”问题进行了分析；其次，针对传统 Bayes-Bootstrap 方法的局限性，提出了一种基于经验函数修正的 Bayes-Bootstrap 数据抽样算法；最后，结合算例对方法的有效性进行了验证。结果表明：该仿真评估算法具有较高的评估精度，有效降低了传统 Bayes 方法进行可靠性评估的主观性和处理过程的复杂性。