Research on demagnetization fault diagnosis method of permanent magnet wind turbine based on ResNet-50 network transfer learning model
ID：92
Submission ID：154 View Protection：ATTENDEE
Updated Time：2024-10-23 10:33:51
Hits：53
Oral Presentation
Start Time：2024-11-02 09:10 （Asia/Shanghai）
Duration：20min
Session：[P4] Parallel Session 4 » [P4-2] Parallel Session 4（November 2 AM）
No files
Abstract
Abstract—Residual networks (ResNet) in convolutional neural networks (CNN) were used for demagnetization fault image recognition and classification. Initially, we collected a dataset of demagnetization fault characteristics in permanent magnet wind generators. Using this dataset, a ResNet50 fault diagnosis model was developed for a 25kW wind turbine. This model was then transferred to a 2MW permanent magnet wind generator using deep transfer learning. Compared to models without transfer learning, this approach significantly improved fault diagnosis accuracy and efficiency.
Keywords
Permanent Magnet Wind Turbines,Fault Diagnosis,Transfer Learning,Convolutional Neural Networks Introduction
Speaker
DuSiyu
Student Xi'an Jiaotong UniversitySubmission Author
All comments
15th August 2024 31st August 2024- Manuscript Submission
15th September 2024 - Acceptance Notification
1st October 2024 - Camera Ready Submission
1st October 2024 – Early Bird Registration
Contact Us
Website:
https://icsmd2024.aconf.org/
Comment submit