针对先验信息为加速退化数据的情况,提出了利用非共轭先验分布进行Bayesian统计推断的剩余寿命预测方法。不预先假定Wiener过程参数值的分布类型,利用加速系数将加速应力下的参数值折算到工作应力水平下,进而使用Anderson-Darling方法...针对先验信息为加速退化数据的情况,提出了利用非共轭先验分布进行Bayesian统计推断的剩余寿命预测方法。不预先假定Wiener过程参数值的分布类型,利用加速系数将加速应力下的参数值折算到工作应力水平下,进而使用Anderson-Darling方法确定参数值的最优拟合分布类型。在对参数值进行折算时,根据周源泉提出的理论对Wiener过程参数与加速应力之间的关系进行了推导。参数估计时,通过极大似然法得到超参数的估计值,利用WinBUGS软件实现Markov Chain Monte Carlo仿真得到参数的后验均值。通过某型军用电连接器寿命预测实例验证了所提方法的实用价值和研究意义,结果表明本方法可有效解决先验信息为加速退化数据时进行剩余寿命预测的难题。展开更多
An aviation hydraulic axial piston pump＇s degradation fiom comprehensive wear is a typical gradual failure model. Accurate wear prediction is difficult as random and uncertain char- acteristics must be factored into ...An aviation hydraulic axial piston pump＇s degradation fiom comprehensive wear is a typical gradual failure model. Accurate wear prediction is difficult as random and uncertain char- acteristics must be factored into the estimation. The internal wear status of the axial piston pump is characterized by the return oil flow based on fault mechanism analysis of the main frictional pairs in the pump. The performance degradation model is described by the Wiener process to predict the remaining useful life （RUL） of the pump. Maximum likelihood estimation （MLE） is performed by utilizing the expectation maximization （EM） algorithm to estimate the initial parameters of the Wiener process while recursive estimation is conducted utilizing the Kalman filter method to estimate the drift coefficient of the Wiener process. The RUL of the pump is then calculated accord- ing to the performance degradation model based on the Wiener process. Experimental results indi- cate that the return oil flow is a suitable characteristic for reflecting the internal wear status of the axial piston pump, and thus the Wiener process-based method may effectively predicate the RUL of the pump.展开更多
电弧侵蚀引起交流接触器触头质量损耗,使其电接触性能不断退化,根据其性能退化过程可以对交流接触器剩余寿命进行预测。该文分析可线性化的非线性 Wiener 过程,采用统计回归的建模方法,利用触头质量损耗历史数据建立三阶函数退化轨迹模...电弧侵蚀引起交流接触器触头质量损耗,使其电接触性能不断退化,根据其性能退化过程可以对交流接触器剩余寿命进行预测。该文分析可线性化的非线性 Wiener 过程,采用统计回归的建模方法,利用触头质量损耗历史数据建立三阶函数退化轨迹模型,进而建立时间尺度变换模型,对分段累积操作次数进行时间尺度变换;以累积触头质量损耗首次达到失效阈值作为失效判据,提出基于非线性 Wiener 过程的可靠性评估方法和剩余寿命预测模型;在交流接触器电寿命试验基础上,进行试验数据变化规律分析,证明可以采用本文建立的退化模型和预测方法进行交流接触器剩余寿命预测;最后,通过试验数据进行交流接触器剩余寿命的极大似然估计,并比较分析不同数据处理方法的预测精度,验证了本文剩余寿命预测方法的准确性。展开更多
文摘针对先验信息为加速退化数据的情况,提出了利用非共轭先验分布进行Bayesian统计推断的剩余寿命预测方法。不预先假定Wiener过程参数值的分布类型,利用加速系数将加速应力下的参数值折算到工作应力水平下,进而使用Anderson-Darling方法确定参数值的最优拟合分布类型。在对参数值进行折算时,根据周源泉提出的理论对Wiener过程参数与加速应力之间的关系进行了推导。参数估计时,通过极大似然法得到超参数的估计值,利用WinBUGS软件实现Markov Chain Monte Carlo仿真得到参数的后验均值。通过某型军用电连接器寿命预测实例验证了所提方法的实用价值和研究意义,结果表明本方法可有效解决先验信息为加速退化数据时进行剩余寿命预测的难题。
基金supported by the National Natural Science Foundation of China(No.51305011)the National Basic Research Program of China(No.2014CB046402)the 111 Project of China
文摘An aviation hydraulic axial piston pump＇s degradation fiom comprehensive wear is a typical gradual failure model. Accurate wear prediction is difficult as random and uncertain char- acteristics must be factored into the estimation. The internal wear status of the axial piston pump is characterized by the return oil flow based on fault mechanism analysis of the main frictional pairs in the pump. The performance degradation model is described by the Wiener process to predict the remaining useful life （RUL） of the pump. Maximum likelihood estimation （MLE） is performed by utilizing the expectation maximization （EM） algorithm to estimate the initial parameters of the Wiener process while recursive estimation is conducted utilizing the Kalman filter method to estimate the drift coefficient of the Wiener process. The RUL of the pump is then calculated accord- ing to the performance degradation model based on the Wiener process. Experimental results indi- cate that the return oil flow is a suitable characteristic for reflecting the internal wear status of the axial piston pump, and thus the Wiener process-based method may effectively predicate the RUL of the pump.
文摘电弧侵蚀引起交流接触器触头质量损耗,使其电接触性能不断退化,根据其性能退化过程可以对交流接触器剩余寿命进行预测。该文分析可线性化的非线性 Wiener 过程,采用统计回归的建模方法,利用触头质量损耗历史数据建立三阶函数退化轨迹模型,进而建立时间尺度变换模型,对分段累积操作次数进行时间尺度变换;以累积触头质量损耗首次达到失效阈值作为失效判据,提出基于非线性 Wiener 过程的可靠性评估方法和剩余寿命预测模型;在交流接触器电寿命试验基础上,进行试验数据变化规律分析,证明可以采用本文建立的退化模型和预测方法进行交流接触器剩余寿命预测;最后,通过试验数据进行交流接触器剩余寿命的极大似然估计,并比较分析不同数据处理方法的预测精度,验证了本文剩余寿命预测方法的准确性。
