考虑不确定测量的惯性平台剩余寿命自适应预测

Adaptive remaining useful life prediction for inertial platform with uncertain measurements

针对现有剩余寿命预测方法未考虑测量误差引起的不确定性,且存在预测不确定性大的问题,提出了一种基于Wiener过程的退化过程建模方法,并将测量的不确定性考虑到剩余寿命预测中,推导出了剩余寿命的概率分布。为了实时更新模型参数,利用Kalman滤波算法实时估计Wiener过程中的漂移系数,并利用期望最大化算法实时估计其它相关参数。以某型号惯性平台的退化测量数据为例,进行了实验验证,结果表明,相比其它算法,文中算法能够降低剩余寿命预测的不确定性,提高预测精度。

Current methods for the estimation of remaining useful life（RUL）ignore the random measurement errors,and thus lead to the difficulty in reducing the prediction uncertainty.Here we present a Wiener-based degradation process modeling to derive the RUL distribution, while considering the measurement errors.In order to update the model parameters in real-time,Kalman filter is used to predict the drift coefficient in the Wiener process and the Expectation Maximization（EM）algorithm to the other related parameters.The method is applied into a practical inertial platform and the results show that it can improve the estimation accuracy of the RUL comparing to other methods.