A method for rapidly evaluating reliability and predicting remaining useful life using two-dimensional convolutional neural network with signal conversion
- 96 Downloads
Real-time monitoring and rapid evaluation of bearing operating conditions, especially for the reliability evaluation and remaining useful life (RUL) prediction, are major challenges in the rotating machinery field. A two-dimensional (2-D) deep convolution neural network (CNN) is proposed for rapidly evaluating reliability and predicting RUL, in which a signal conversion method is proposed for converting the one-dimensional signal into the 2-D image to satisfy the input requirements of the 2-D CNN. Different activation functions are employed to implement the conversion of the input data to the output data for each layer of the network, and dropout is only adopted in the hidden layer to change the network structure to prevent overfitting. The maximum correlation entropy with regular terms is employed as the loss function of the model to obtain better training performance compared with the mean square error (MSE). Then, the rolling bearing degradation vibration data is applied to the proposed model to verify the accuracy and rapidity. The results show that the proposed method has good accuracy and fast calculation ability in bearing reliability evaluation and RUL prediction, especially, its time consumption is shorter than that by other deep learning networks.
KeywordsRolling bearing Remaining useful life Convolution neural network Signal conversion Correlation entropy
Unable to display preview. Download preview PDF.
This work was supported by the National Natural Science Foundation of China (Grant Nos. 51605361, 51875432); and the Natural Science Basic Research Plan in Shaanxi Province of China (Grant No. 2018JQ5034).
- B. Wang et al., An improved fusion prognostics method for remaining useful life prediction of bearings, IEEE International Conference on Prognostics and Health Management (2017) 18–24.Google Scholar
- X. Chen et al., Remaining life prognostics of rolling bearing based on relative features and multivariable support vector machine, J. of Mechanical Engineering, 227 (12) (2013) 2849–2860.Google Scholar
- F. Wang et al., An enhancement deep feature extraction method for bearing fault diagnosis based on kernel function and autoencoder, Shock and Vibration (2018) 1–12.Google Scholar
- J. Tao, Y. Liu and D. Yang, Bearing fault diagnosis based on deep belief network and multisensor information fusion, Shock and Vibration, 2016 (7) (2016) 1–9.Google Scholar
- L. Guo et al., Deep convolution feature learning for health indicator construction of bearings, 2017 Prognostics and System Health Management Conference (2017) 1–6.Google Scholar
- A. Krizhevsky, I. Sutskever and G. E. Hinton, ImageNet classification with deep convolutional neural networks, International Conference on Neural Information Processing Systems (2012) 1097–1105.Google Scholar
- K. Simonyan and A. Zisserman, Very deep convolutional networks for large-scale image recognition, Computer Science (2014).Google Scholar
- C. Szegedy et al., Going deeper with convolutions, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2014) 1–9.Google Scholar
- P. Nectoux et al., PRONOS-TIA: An experimental platform for bearings accelerated degradation tests, IEEE International Conference on Prognostics and Health Management (2012) 1–8.Google Scholar