摘要
步进应力加速退化试验已发展成高可靠、长寿命产品可靠性评估与剩余寿命估计的主要试验方法。为研究测量误差在基于步进应力加速退化试验方法剩余寿命估计中的影响,建立了基于Wiener过程的考虑个体差异和测量误差的退化模型。将Wiener过程的漂移系数随机化描述个体差异,在首达时间意义下得到了寿命分布的概率密度函数。基于极大似然估计法对模型中引入的未知参数进行估计。采用蒙特卡洛方法对激光器的性能退化进行了仿真研究。研究结果表明,考虑测量误差的退化模型的模型拟合性和剩余寿命估计精度都优于不考虑测量误差的方法,可以提高可靠性估计的精度与剩余寿命预测的准确性。
Accelerated degradation test （ADT） has been developed into a main test method, which can be used to estimate the reliability and remaining useful life （RUL） of products with high reliability and long life. In order to study the effect of measurement error on the estimation of RUL in the step stress ac- celerated degradation test, a degradation process modeling based on Wiener process model considering the measurement error and the individual variation is proposed. The drift coefficient of Wiener process is randomized to describe the individual variation in different equipment, and the probability density func- tion of life distribution is obtained at first hitting time. The maximum likelihood estimation method is used to estimate the unknown parameters introduced in the model. Monte Carlo method is used to simulate the performance degradation of laser. The results show that the fitting of model and the accuracy of RUL esti- mation in the degradation model considering the measurement error are better than those in the model without considering the measurement error, which can enhance the estimation accuracy of reliability and the prediction accuracy of RUL.
出处
《兵工学报》
EI
CAS
CSCD
北大核心
2017年第8期1586-1592,共7页
Acta Armamentarii
关键词
概率论
随机过程
剩余寿命估计
步进应力
测量误差
极大似然估计
probability theory
stochastic process
remaining useful life estimation
step stress
meas- urement error
maximum likelihood estimation
