A5050651874- Machine Fault Diagnosis Techniques
- Advanced Battery Technologies Research
- Fault Detection and Control Systems
- Advancements in Battery Materials
- Imbalanced Data Classification Techniques
- Evaluation and Optimization Models
- Engineering Diagnostics and Reliability
- Evaluation Methods in Various Fields
- Non-Destructive Testing Techniques
- Advanced Decision-Making Techniques
- Welding Techniques and Residual Stresses
- Advanced Battery Materials and Technologies
- Gear and Bearing Dynamics Analysis
- IoT-based Smart Home Systems
Kunming University of Science and Technology
2023-2025
Abstract A boundary enhancement and Gaussian mixture model (G) optimized synthetic minority oversampling technique (SMOTE) algorithm (BE-G-SMOTE) is proposed to improve diagnostic accuracy under imbalanced bearing fault data conditions. It designed solve the problem that diversity of samples generated by original SMOTE limited, as well deep learning limited size training processing speed. Firstly, a few are clustered G achieve cluster division. Secondly, according density distribution...
Synthetic minority oversampling (SMOTE) has been widely used in dealing with the imbalance classification mechanical fault diagnosis field. However, classical SMOTE model generates poor quality data, which leads to a low diagnostic accuracy of model. This article proposes an generation based on Gaussian mixture (GMM) and boundary joint optimization (BDOP-GMM-SMOTE). First, GMM is utilized cluster class bearing weights different classes should be distributed according density distribution...
The vibration signals of rolling bearings under the influence strong background noise and large fluctuation rotational speed often show characteristics in non-stationarity spectral dispersion. These will bring about problems such as failure traditional filtering methods, error low efficiency order analysis methods. A fault diagnosis method normalized time-frequency entropy spectrum (NTFES) combined with characteristic coefficient template (FCCT) for variable operating conditions is proposed...
Deep learning models can automatically capture the state in a device from complex data. However, it is typically used for end-to-end fault diagnosis, and visualization results of convolutional layer are abstract. A neural network model with maximization mutual information (MMI) between inputs outputs proposed. It basic idea to enhance feature extraction capability through improved loss deconstraints improve interpretability at level. First, bearing vibration data converted an envelope...
The imbalance of bearing fault samples can bring about the problems unstable learning process classification model and low accuracy. A Wasserstein generative adversarial network (V AE-WGAN-GP) that fuses a variational auto-encoder data with Gradient Penalty (GP) is proposed in this work. First, structure generator improved to extract hidden variables by feature coding through encoding-decoding latent information; Then, training adopts distance measure difference between generated real score,...