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.
出处
《电子设计工程》
2015年第12期95-97,共3页
Electronic Design Engineering
关键词
离群值
性能退化
模糊聚类
最小二乘估计
退化轨迹
outliers
performance degradation
fuzzy clustering
least-squares estimation
degradation path