Integrating spatiotemperporal features into fault prediction using a multi-dimensional method
This study proposes a method to validate multidimensional fault prediction models. It integrates vibration and current data, analyzes spatiotemporal characteristics, and uses support vector machines and random forest algorithms to analyze fault characteristics. The short-time Fourier transform is us...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-09-01
|
| Series: | Results in Engineering |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2590123025019279 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This study proposes a method to validate multidimensional fault prediction models. It integrates vibration and current data, analyzes spatiotemporal characteristics, and uses support vector machines and random forest algorithms to analyze fault characteristics. The short-time Fourier transform is used to convert spatiotemporal data into the frequency domain for classification, and high-order features are extracted through convolutional networks. The model considers three spatial dimensions and three vibration measurement sources to form a nine-dimensional data structure, and a fault prediction algorithm based on these dimensions is established. The study evaluates the model using metrics such as accuracy, precision, recall, F1 score, and receiver operating characteristic area under curve (ROC-AUC), and visualizes performance through confusion matrices, ROC curves, and precision-recall curves. In order to further verify the significance of the model, the experiment is conducted on mercury induction motors of industrial water pumps, and an McNemar’s Chi-Square Test was used to statistically test the fault discrimination ability of the model under various conditions. Results show that this model achieves the stability and accuracy of the fault prediction under multi-dimensional data, effectively identifies abnormal conditions, and the technical support for smart monitoring and maintenance can be provided. |
|---|---|
| ISSN: | 2590-1230 |