卫星动量轮的污染数据的Bayes参数估计

Bayesian Estimation of Parameter Based on Contaminated Data for Momentum Wheel of Satellites

由于试验条件等的限制,在寿命试验中,所获得的数据可能是污染数据。这时,采用传统方法分离污染项并对产品的寿命进行估计非常困难。针对指数型寿命数据受正态污染的情况,建立定数截尾情况下的数据模型,并采用Bayes方法对模型进行统计分析;在验前分布确定方面提出运用Bootstrap方法确定超参数。实例分析表明,Bayes方法对有污染情况寿命数据的统计推断是有效的。

Limited by experimental situations and some other factors, the data obtained in life test are possibly con-taminated, and it is extremely difficult to separate contaminated items through traditional means at this time. Upon thesituation that an exponential life model is often contaminatednormally, this thesis establishes a data model through TypeⅡ censoring, analyses the model by means of Bayes andpoints out adopting Bootstrap methods to figure out the super parameters, the practical experiment proves that Bayes means iseffective toward statistical inference of contaminated life data.