基于Kalman滤波和Wiener过程的系统性能退化数据建模方法

摘要

At present, modeling for lifetime prediction seldom takes the effect of the drift coefficient change into consideration, which is neither reasonable nor truthful. To solve the problem, a degradation model is proposed based on Kalman filter and Wiener process. Based on which, Kalman filter and EM algorithm are utilized for parameter estimation and updating. Finally, an example of remaining useful life prediction of a gyroscope verifies the effectiveness of the proposed method.%目前寿命预测的建模过程中很少考虑漂移系数变化产生的影响，这并不合理且不符合实际情况，为解决这个问题，建立了基于Kalman滤波和Wiener过程的系统性能退化模型，在此基础上采用Kalman滤波和EM算法实现了模参数的估计和更新，最后通过某陀螺仪的寿命预测实例验证了方法的有效性。