The Internet of Things is booming in the aviation industry, thus proper collection and analysis methods for fault data could support the realization of intelligent maintenance support and management of components. Failure of aeronautical equipment is typically described by a Weibull distribution. Here we carry out a sensitivity analysis using an objective Bayesian method to investigate the effect of descriptive statistics on reliability evaluations of random right censored aeronautical equipment failure data. We design a multiple Markov chain algorithm with censoring rate- and sample size-dependent variables. The algorithm estimates the scale and shape parameters of a Weibull distribution, with prior information of a large variance gamma distribution, for different censoring rates and sample sizes. Estimation deviations can be evaluated by mean time between failures, and by the mean and the variation factor of the distribution parameter. The numerical results show that, in a Weibull distribution, the estimation accuracy of an objective Bayesian method is acceptable when the sample size is more than 10 or the censoring rate less than 0.5. Otherwise, a better reliability evaluation method for small samples and high censoring rates should be explored.

物联网在航空业中正在蓬勃发展，因此对故障数据的正确收集和分析方法可以支持实现智能维护支持和组件管理。航空设备的故障通常由Weibull分布描述。在这里，我们使用客观贝叶斯方法进行敏感性分析，以研究描述性统计数据对随机右删失航空设备故障数据的可靠性评估的影响。我们设计了一个多重马尔可夫链算法，其审查率和样本大小相关的变量。对于不同的删失率和样本量，该算法使用大方差伽马分布的先验信息来估计威布尔分布的比例和形状参数。估计偏差可以通过平均故障间隔时间以及分布参数的均值和变化因子来评估。数值结果表明，在Weibull分布中，当样本量大于10或检查率小于0.5时，客观贝叶斯方法的估计精度是可以接受的。否则，应该探索一种针对小样本和高审查率的更好的可靠性评估方法。