具有离群值的性能退化产品可靠性评估  

The reliability estimation of performance degradation product with outliers

作　　者：张一弛[1] 

机构地区：[1]西北工业大学机电学院,陕西西安710072

出　　处：《电子设计工程》2015年第12期95-97,共3页Electronic Design Engineering

摘　　要：运用性能退化数据对高可靠长寿命产品进行可靠性评估整体效果良好。但是,在进行产品性能检测过程中,由于随机误差的影响,导致性能退化数据中可能存在一些离群值,使得评估结果不够稳健。在这种情况下,利用模糊聚类的最小二乘估计方法对退化轨迹参数进行估计,从而削弱离群值的影响,提高了评估的精度。本文利用实例进行分析,验证了方法的正确性与有效性。Performance degradation data can provide with useful information about the reliability assessment. Especially for the high reliability and long life products, the overall effect is good using of performance degradation data. However, there are some outliers in the testing process of product performance because of the influence of random error, which make the assessment is not robust. In this case, we make use of the fuzzy clustering least squares method to evaluate the parameters, which impair the influence of outliers and improve the stability. This paper gives the actual example to show that the method is correct and effective.

关 键 词：离群值 性能退化 模糊聚类 最小二乘估计 退化轨迹 

分 类 号：TN406[电子电信—微电子学与固体电子学]