What question did this study set out to answer?

To improve gearbox fault diagnosis accuracy under variable speed conditions by establishing a deep transfer learning framework based on non-contact acoustic sensing.

February 11, 2023ProcessesOpen Access

Fault Diagnosis Algorithm of Gearboxes Based on GWO-SCE Adaptive Multi-Threshold Segmentation and Subdomain Adaptation

View Full Paper

Ask AI

Bookmark

Authors

YLYangshuo LiuQingdao University JKJianshe KangPLA Army Engineering University LWLiang WenChinese PLA General Hospital

Discussion

Member takes

Implication

Experimental study demonstrates improved fault diagnosis accuracy in variable-speed gearboxes using deep transfer learning, highlighting robust condition monitoring without contact sensors.

Key Points

To improve gearbox fault diagnosis accuracy under variable speed conditions by establishing a deep transfer learning framework based on non-contact acoustic sensing.
Acquired acoustic data using non-contact sound pressure sensors and converted raw vibration signals into time–frequency images using continuous wavelet transform (CWT).
Segmented target fault characteristics adaptively using the gray wolf optimization (GWO) algorithm combined with symmetric cross entropy (SCE).
Extracted shared domain features with a convolutional neural network (CNN) and introduced local maximum mean discrepancy (LMMD) into deep fully connected layers to align subclass feature distributions.
Fault preset experiments on gearboxes operating under variable speed conditions demonstrated superior diagnostic accuracy compared with alternative diagnostic methods.
The integration of LMMD and adaptive multi-threshold segmentation successfully reduced distribution differences among subclass fault features across varying working conditions.

Cite This Study

Liu et al. (2023) studied this question.

synapsesocial.com/papers/6a1bbfbf0a1f7575939cd14b https://doi.org/10.3390/pr11020556

View Full Paper

Ask AI

Bookmark

Implication

Experimental study demonstrates improved fault diagnosis accuracy in variable-speed gearboxes using deep transfer learning, highlighting robust condition monitoring without contact sensors.

Fault Diagnosis Algorithm of Gearboxes Based on GWO-SCE Adaptive Multi-Threshold Segmentation and Subdomain Adaptation

Discussion

Member takes

Implication

Key Points

Cite This Study

Fault Diagnosis Algorithm of Gearboxes Based on GWO-SCE Adaptive Multi-Threshold Segmentation and Subdomain Adaptation

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider