An improved deep convolution neural network for predicting the remaining useful life of rolling bearings

Guo, Yiming; Zhang, Hui; Xia, Zhijie; Dong, Chang; Zhang, Zhisheng; Zhou, Yifan; Sun, Han

doi:10.3233/jifs-201965

Cited by 10 publications

(2 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The vibration signals collected by sensors are typically highdimensional and contain a large amount of degradation-related information. The CNN has powerful feature extraction capabilities that allow them to automatically extract degradation information from high-dimensional signals [27], avoiding the trouble of extracting features manually. However, it is worth noting that the CNN cannot directly process time series data.…”

Section: Convolution Self-attention Lstm Networkmentioning

confidence: 99%

A domain adaptation network with feature scale preservation for remaining useful life prediction of rolling bearings under variable operating conditions

She,

Wang,

Zhang

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Transfer learning (TL) and domain adaptation (DA) methods have been utilized in bearing prognostic and health management (PHM), but most of the current domain adaptation methods do not take into account the feature scale change of degraded features when aligning the feature distribution, and these methods are more suitable for the classification problem, which is more robust to the feature scale change. However, they perform poorly in regression problems. In addition, most of the remaining useful life (RUL) prediction methods require preprocessing such as statistical feature extraction on the signal, which makes the prediction process complicated. To solve the above problems, a domain adaptation method based on the representation subspace distance (RSD) is proposed for predicting the bearing RUL under different operating conditions. First, the proposed CNN Self-Attention LSTM network (CSALN) model is utilized to extract the deep features from the original signal, which overcomes the limitations of the CNN in extracting time series. Then, the representation subspace distance in the Riemannian geometry of the Grassmann manifold is proposed as a domain transfer loss to learn domain invariant features. The modified method can align the feature distribution of the source domain and the target domain without changing the feature scale. At the same time, the bases mismatch penalization (BMP) is introduced to avoid destroying the semantic information of the features in the process of domain alignment. Finally, the effectiveness of the proposed method is verified by experiments on four types of transfer tasks, and its superiority is also demonstrated by comparison with other advanced methods.

show abstract

Section: Convolution Self-attention Lstm Networkmentioning

confidence: 99%

A domain adaptation network with feature scale preservation for remaining useful life prediction of rolling bearings under variable operating conditions

She,

Wang,

Zhang

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Bearings are one of the most important parts of mechanical equipment, and their working performance directly affects the health of the entire equipment [3][4][5][6][7][8]. At present, data-driven modeling has become the main method for bearing remaining life prediction.…”

Section: Introductionmentioning

confidence: 99%

Remaining useful life prediction for rolling bearings using correlation coefficient and Kullback–Leibler divergence feature selection

Liang¹,

Song²,

Wang³

et al. 2021

Meas. Sci. Technol.

View full text Add to dashboard Cite

The prediction accuracy of bearing remaining useful life (RUL) is not high due to the unreasonable stage division of bearing performance degradation and the blindness of feature selection. In order to solve this problem, RUL prediction for rolling bearings using Pearson product-moment correlation coefficient (PPMCC) and Kullback–Leibler divergence (KLIC) feature selection is proposed in this paper. First, in order to divide the bearing performance degradation status more accurately, a novel performance degradation state partition method is provided based on t-distributed stochastic neighbor embedding and a k-means clustering algorithm. Second, multi-scale time status feature selection is implemented using PPMCC and KLIC. Finally, the above-proposed feature selection method is validated on the performance degradation data of rolling bearings provided by the University of Cincinnati platform. The results show good and stable effectiveness of the proposed method, and its superiority is demonstrated by comparison with other approaches.

show abstract

Remaining useful life prediction of rolling bearing under limited data based on adaptive time-series feature window and multi-step ahead strategy

Kong

Li²

2022

Applied Soft Computing

View full text Add to dashboard Cite

An improved deep convolution neural network for predicting the remaining useful life of rolling bearings

Cited by 10 publications

References 26 publications

A domain adaptation network with feature scale preservation for remaining useful life prediction of rolling bearings under variable operating conditions

A domain adaptation network with feature scale preservation for remaining useful life prediction of rolling bearings under variable operating conditions

Remaining useful life prediction for rolling bearings using correlation coefficient and Kullback–Leibler divergence feature selection

Remaining useful life prediction of rolling bearing under limited data based on adaptive time-series feature window and multi-step ahead strategy

Contact Info

Product

Resources

About