Multiscale local features learning based on BP neural network for rolling bearing intelligent fault diagnosis

Li, Jimeng; Yao, Xifeng; Wang, Xiangdong; Yu, Qingwen; Zhang, Yungang

doi:10.1016/j.measurement.2019.107419

Cited by 153 publications

(66 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the holdout procedure, 128 datasets were selected randomly as the training samples, and the remaining datasets were used for evaluating the accuracy of SVM classification. When C was set as 2 11 and g was set as 2 23 in the SVM model, a high SVM classification accuracy was obtained. The classification results obtained with the aforementioned parameter values for different sample sizes are listed in Table 1, which indicates that similar classification results were obtained for 256, 512, and 1024 samples.…”

Section: Weight-based Learning Algorithmmentioning

confidence: 99%

Dynamic weight-based learning method for data detection in manufacturing

Lin

2020

Advances in Mechanical Engineering

View full text Add to dashboard Cite

In this study, a dynamic weight-based method combined with principal component analysis (PCA) was developed for the first time for detecting measurement data in manufacturing. This weight-based learning technique can learn and train the measurement data sequence to isolate incorrect data sources for achieving high accuracy when detecting various types of data. Research has revealed that unsuitable image or data features might cause poor performance in industrial inspections. In contrast to the previous inspection methods, the weight-based learning method proposed in this study employs a dynamic learning algorithm for effectively and adaptively selecting optimal principle components to the support vector machine (SVM) algorithm and then establishes indicators. Finally, these PCA-based indicators act as substitutes for massive amounts of data in data processing and can be applied to timely detect data when the data contain redundant and incorrect inputs in a sequence. The experimental results indicate that the proposed method, which combines dynamic weight-based feature extraction with PCA, can provide useful indicators for detecting various types of manufacturing data and exhibited satisfactory performance in the data detection.

show abstract

Section: Weight-based Learning Algorithmmentioning

confidence: 99%

Dynamic weight-based learning method for data detection in manufacturing

Lin

2020

Advances in Mechanical Engineering

View full text Add to dashboard Cite

show abstract

“…Then these statistical features are input into a machine learning classifier for fault diagnosis, such as K-nearest neighbor (KNN), support vector machine (SVM), random forest (RF), gradient boosting decision tree (GBDT) and artificial neural network (ANN), etc. [ 6 , 7 , 8 , 9 , 10 ]. Bhakta et al used cepstrum analysis to pre-process the rolling bearing datasets provided by the Case Western Reserve University (CWRU) laboratory, then used the gradient boosting (GB) learning algorithm for fault diagnosis, and achieved good results [ 11 ].…”

Section: Introductionmentioning

confidence: 99%

Fault Diagnosis of Rolling Bearings Based on a Residual Dilated Pyramid Network and Full Convolutional Denoising Autoencoder

Shi

Chen

et al. 2020

Sensors

View full text Add to dashboard Cite

Intelligent fault diagnosis algorithm for rolling bearings has received increasing attention. However, in actual industrial environments, most rolling bearings work under severe working conditions of variable speed and strong noise, which makes the performance of many intelligent fault diagnosis methods deteriorate sharply. In this regard, this paper proposes a new intelligent diagnosis algorithm for rolling bearing faults based on a residual dilated pyramid network and full convolutional denoising autoencoder (RDPN-FCDAE). First, a continuous wavelet transform (CWT) is used to convert original vibration signals into time-frequency images. Secondly, a deep two-stage RDPN-FCDAE model is constructed, which is divided into three parts: encoding network, decoding network and classification network. In order to obtain efficient expression of data denoising feature of encoding network, time-frequency images are first input into the encoding-decoding network for unsupervised pre-training. Then pre-trained coding network and classification network are combined into residual dilated pyramid full convolutional network (RDPFCN) for parameter fine-tuning and testing. The proposed method is applied to bearing vibration datasets of test rig with different speeds and noise modes. Compared with representative machine learning and deep learning method, the results show that the algorithm proposed is superior to other methods in diagnostic accuracy, noise robustness and feature segmentation ability.

show abstract

“…Finally, they achieved a satisfactory fault diagnosis result using the HMM technology. Li et al [15] developed a new feature learning method for REB fault diagnosis. Firstly, they used wavelet multiscale transform to decompose the vibration signals.…”

Section: Introductionmentioning

confidence: 99%

“…Finally, they selected SVM as the fault classifier to detect the REB faults. Nevertheless, one limitation exists with conventional ML methods; they need to manually select fault features [15]. That is to say, the same fault classifier with different feature extraction methods can yield different diagnosis results.…”

Section: Introductionmentioning

confidence: 99%

Input Feature Mappings-Based Deep Residual Networks for Fault Diagnosis of Rolling Element Bearing With Complicated Dataset

et al. 2020

View full text Add to dashboard Cite

Most rolling element bearing (REB) fault diagnosis algorithms are evaluated on the Case Western Reserve University (CWRU) bearing dataset for its popularity and simplicity. However, the diagnosis accuracy on CWRU bearing dataset is overly saturated; it is nearly up to 100%. In this study, an input feature mappings (IFMs)-based deep residual network (ResNet) is proposed to conduct detailed and comprehensive fault diagnosis on REB with complicated bearing dataset. Firstly, a new data preprocessing method named as a signal-to-IFMs method is proposed to automatically extract features from raw signals without predefined parameters. Then, a deep ResNet is used as the fault classifier to learn the discriminative features from IFMs and identify the faults of REB. Finally, the proposed model is evaluated on the artificial, real, and mixed damages of the Paderborn university bearing dataset. The proposed method yields the average testing accuracies of 99.7%, 99.7%, and 99.81% in artificial, real, and mixed bearing damages, which outperforms other methods. INDEX TERMS Rolling element bearing, fault diagnosis, signal-to-input feature mappings, deep residual networks.

show abstract

Multiscale local features learning based on BP neural network for rolling bearing intelligent fault diagnosis

Cited by 153 publications

References 32 publications

Dynamic weight-based learning method for data detection in manufacturing

Dynamic weight-based learning method for data detection in manufacturing

Fault Diagnosis of Rolling Bearings Based on a Residual Dilated Pyramid Network and Full Convolutional Denoising Autoencoder

Input Feature Mappings-Based Deep Residual Networks for Fault Diagnosis of Rolling Element Bearing With Complicated Dataset

Contact Info

Product

Resources

About