Remaining useful life prognosis of bearing based on Gauss process regression

Remaining useful life (RUL) prognosis of bearing is an enabling step for efficient implementation of condition based maintenance. Through intelligent tracking of features, status of bearing can be monitored and a rough RUL can be calculated. This paper presents an application of an important Bayesian machine learning method named Gaussian Process Regression (GPR) for bearing features tracking. The Gaussian process model can provide variance around its mean prediction to describe associated uncertainty in the evaluation and prediction. In this case, the GPR models with three different kinds of covariance functions are discussed for feature tracking and RUL evaluation. The dynamic model is introduced to realize a better accuracy prognosis of the bearing RUL by analyzing two important features. The experimental results show that using GPR for prognosis can achieve a high accuracy. In addition, the comparisons of prediction with different type of training covariance functions are discussed. Finally, this method provides a new way for prognosis of fluctuated signal of bearing features.