基于改进比例失效率模型的设备剩余寿命预测

产品的剩余寿命预测能够为产品的维修和更换提供重要的决策依据.传统的比例失效模型方法在剩余寿命预测中得到了广泛的应用,然而此方法没有充分利用产品的历史寿命信息,对产品的退化过程也没有很好地描述.针对此问题,提出了一种融合退化过程与失效率建模的产品剩余寿命预测方法.该方法首先利用线性过程对产品的退化过程建模,然后利用比例失效模型融合退化过程对失效率的影响,达到充分利用产品历史信息的目的.此外,与传统比例失效模型方法不同,模型中的比例参数分为两部分,分别将产品退化的初始信息和产品的退化增量联系起来,进一步利用产品的当前退化信息对产品的参数进行Bayes更新,基于此进行剩余寿命预测,从而实现产品历史数据和当前退化数据的有效融合.通过激光发生器的退化数据验证了所提方法的有效性.

Remaining Useful Life Prediction Based on Extended Cox Proportional Hazard Model

Remaining useful life prediction of the item can provide important decision-making basis for its replacement and maintenance. Traditional proportional hazards model has been widely used in remaining useful life prediction. However % the previous life information is not fully taken into account in his method % nor is the underlying degradation process of the item described in an appropriate way. To solve this problem % a method for remaining useful life prediction was presented % involving the degradation process and hazard rate modeling of the item. Specifically % the degradation process was first modeled as a linear process and then the proportional hazards model was adopted to model the effect of the degradation process on hazard rate % which aimed at making full use of the previous information of the items. In addition % this model was different from the traditional proportional hazards model because the proportion parameter fell into two parts linking the initial value of the degradation signal and its increase with the hazard rate, respectively. Furthermore % the parameters of the degradation model were updated with the current degradation information by Bayesian method % and based on that the remaining useful life could be predicted accordingly % and the previous data and the current data were fused by the proposed method. Finally % the effectiveness of this method was demonstrated through the degradation data of laser.