摘要
研究产品性能退化规律并评估其剩余寿命和可靠性特征是维修决策的重要前提和依据。Gamma过程退化模型在性能退化建模和维修最优化方面得到了广泛研究和重视。但是很少考虑测量误差对模型估计的影响。由于观测数据常常受到测量不确定性的影响,所以如果在Gamma退化过程的构造中没有考虑测量误差的影响,就会导致退化动力学的有偏估计。为此,利用EM算法提出了1种统计方法,利用了可以获取的量测信息来克服以上不足。仿真研究说明了考虑测量不确定性的重要性和提出方法的有效性。另外,提出的统计方法也可以被应用于许多类型的随机过程退化模型。
The product performance degradation evolution and related reliability characters is the basis of maintenance decision. The Gamma process degradation model is widely applied to performance degradation modeling and maintenance optimization. However, the effect of measurement error on model estimation is considered by little researchers measurement uncertainty, this uncertainty can lead to a Since the observation is often affected by biased assessment of the degradation process, if it is not properly taken into account in the construction of the stochastic degradation process. Therefore, a statistical method based on EM algorithm is presented to overcome this difficulty by using the available knowledge on measurement uncertainty. The significance of the proposed method is illustrated through a simulated study. Additional, the proposed statistical method can be applied to several other types of stochastic processes degradation model.
出处
《系统仿真技术》
2010年第1期1-5,共5页
System Simulation Technology
基金
国家自然科学基金重点课题资助项目(60736026)
"教育部新世纪优秀人才支持计划"资助项目(NCET-07-0144)
