Machine Learning Based Anomaly Clustering Study for Bolt Tightening System

In the context of Industry 4.0, the tightening quality of bolts is still critical to product performance and safety. In order to improve the abnormal data identification and classification ability of bolt tightening system, this paper adopts principal component analysis to reduce the dimension and a...

Full description

Saved in:
Bibliographic Details
Published in2024 5th International Conference on Electronic Communication and Artificial Intelligence (ICECAI) pp. 199 - 204
Main Authors Liu, Tiehui, Hu, Shan, Bing, Zhigang, Zhao, Di
Format Conference Proceeding
LanguageEnglish
Published IEEE 31.05.2024
Subjects
automatic state identification
bolt tightening
Clustering algorithms
Dimensionality reduction
Fasteners
Feature extraction
improved k-means clustering method
Machine learning algorithms
Production
profile coefficients
Real-time systems
Online AccessGet full text

Cover

More Information
Summary:In the context of Industry 4.0, the tightening quality of bolts is still critical to product performance and safety. In order to improve the abnormal data identification and classification ability of bolt tightening system, this paper adopts principal component analysis to reduce the dimension and adds residual network structure on the basis of deep self-encoder to improve the feature expression. The k-means++ algorithm is used to optimize the initialization of the clustering center, which in turn improves the k-means clustering algorithm. The experimental results show that the improved algorithm achieves significant improvement in contour coefficient and David Bunting index, which effectively improves the automatic identification accuracy of bolt tightening state.
DOI:10.1109/ICECAI62591.2024.10674921