A Deep Learning-Based Remaining Useful Life Prediction Approach for Bearings

In industrial applications, nearly half the failures of motors are caused by\nthe degradation of rolling element bearings (REBs). Therefore, accurately\nestimating the remaining useful life (RUL) for REBs are of crucial importance\nto ensure the reliability and safety of mechanical systems. To tackle this\nchallenge, model-based approaches are often limited by the complexity of\nmathematical modeling. Conventional data-driven approaches, on the other hand,\nrequire massive efforts to extract the degradation features and construct\nhealth index. In this paper, a novel online data-driven framework is proposed\nto exploit the adoption of deep convolutional neural networks (CNN) in\npredicting the RUL of bearings. More concretely, the raw vibrations of training\nbearings are first processed using the Hilbert-Huang transform (HHT) and a\nnovel nonlinear degradation indicator is constructed as the label for learning.\nThe CNN is then employed to identify the hidden pattern between the extracted\ndegradation indicator and the vibration of training bearings, which makes it\npossible to estimate the degradation of the test bearings automatically.\nFinally, testing bearings' RULs are predicted by using a $\\epsilon$-support\nvector regression model. The superior performance of the proposed RUL\nestimation framework, compared with the state-of-the-art approaches, is\ndemonstrated through the experimental results. The generality of the proposed\nCNN model is also validated by transferring to bearings undergoing different\noperating conditions.\n