摘要
提出了基于混合高斯输出贝叶斯信念网络模型的设备退化状态识别与剩余使用寿命预测新方法,将变量消元和期望最大化算法相结合对模型进行推理,应用聚类评价指标对状态数进行优化,通过计算待识别特征向量的概率值来确定设备当前的退化状态,在退化状态识别的基础上,提出了剩余使用寿命预测方法。最后,分别应用50组轴承全寿命仿真数据和3组轴承全寿命实验数据对模型进行验证。结果表明,该模型可有效地识别设备的退化状态并对剩余使用寿命进行预测。
A new approach for equipment health state identification and residual life prediction based on mixture of Gaussian Bayesian belief network （MoG-BBN）was presented.The inference algorithm was established based on variable elimination （VE ）and expectation maximization （EM）.State number was optimized based on cluster validity indexes. The equipment degradation state was determined through calculating the probability of eigenvectors to be identified.Then, the residual life prediction method was presented based on identifying the degradation state.Finally,50 bearings whole life simulation data and 3 bearings whole life test data were used to demonstrate the proposed methods.The results showed that the proposed method can be used to identify degradation states of an equipment and predict its residual life effectively.
出处
《振动与冲击》
EI
CSCD
北大核心
2014年第8期171-179,共9页
Journal of Vibration and Shock
基金
国家自然科学基金(51035008)
军队"十二五"武器装备预先研究项目(51327020101
51327020304)
