基于状态空间模型的可靠性评估方法  

Reliability prediction method based on a state space model

作　　者：李宏坤[1,2] 何德鲁 张志新[3] 任远杰 丛明[1,2] 

机构地区：[1]大连新宇理工科技开发中心有限公司,大连116024 [2]大连理工大学机械工程学院,大连116024 [3]大连大学机械学院,大连116622

出　　处：《振动与冲击》2016年第1期118-124,共7页Journal of Vibration and Shock

基　　金：国家数控机床重大专项(2013ZX04012071)

摘　　要：基于设备性能退化特征的可靠性分析是可靠性技术研究重要方向之一,当前许多研究是基于多样本进行分析。针对单个设备的可靠性预测问题非常有限,提出基于状态空间模型的可靠性方法进行小样本预测。首先通过在线监测技术获得反映设备状态的信号,运用小波分析方法提取监测信号的小波包能量,选取趋势明显符合设备状态变化的相关频带能量作为设备退化指标。然后对这些特征指标进行滑动平均滤波处理,提高退化特征的信噪比,将其作为状态空间模型的输入对模型参数进行估计,从而建立退化指标的状态空间预测模型,最后预测退化指标的概率分布并计算可靠度。结合滚动轴承试验数据和铣刀磨损数据验证方法的准确性和有效性,为小样本事件的可靠性预测提供一个有效方法。Reliability analysis based on equipment＇s performance degradation characteristics is one of important study directions of reliability technology. At the present, many studies are based on multi-sample analysis, there are few studies for single equipment reliability prediction. Here, the method of reliability prediction based on a state space model was proposed for small sample prediction. Firstly, signals about machine working conditions were determined with the on- line monitoring technology for an equipment. Secondly, the wavelet packet energy was used for characteristic extraction of the monitored signals. The frequency band energy to reflect the trend of the equipment＇s state changes was taken as the equipment＇s degradation index. Then, the signal to noise ratio for degradation characteristic was improved with the moving average filtering processing. Finally, the state space model was established to predict the degradation index＇s probability distribution, and the reliability level was calculated. Two real testing examples of bearing and milling cutter were used to demonstrate the correctness and effectiveness of this method. The results provided a useful method for reliability prediction of small-sample events.

关 键 词：可靠性预测 状态空间模型 特征提取 小波分析 滑动平均 

分 类 号：TG156[金属学及工艺—热处理]