Machine-Learning-Based System for the Detection of Entanglement in Dyeing and Finishing Processes

Chien-Chih Wang and Yu-Hsun Li
Additional contact information
Chien-Chih Wang: Department of Industrial Engineering and Management, Ming Chi University of Technology, New Taipei City 24301, Taiwan
Yu-Hsun Li: Department of Industrial Engineering and Management, Ming Chi University of Technology, New Taipei City 24301, Taiwan

Sustainability, 2022, vol. 14, issue 14, 1-12

Abstract: Many dyeing and finishing factories generally use old-fashioned dyeing machines. A key issue when using these machines is that the dyeing tank cannot detect entanglement problems, which may result in a lower dyeing quality. In this paper, imbalanced data with ensemble machine learning, such as Extreme Gradient Boosting (XGBoost) and random forest (RF), are integrated to predict the possible states of a dyeing machine, including normal operation, entanglement warning, and entanglement occurrence. To verify the results obtained using the proposed method, we worked with industry−academia collaborators. We collected 1,750,977 pieces of data from 1848 batches. The results obtained from the analysis show that after employing the Borderline synthetic minority oversampling technique and the Tomek link to deal with the data imbalance, combined with the model established by XGBoost, the prediction accuracy of the normal operation states, entanglement warning, and entanglement occurrence were 100%, 94%, and 96%, respectively. Finally, the proposed entanglement detection system was connected with the factory’s central control system using a web application programming interface and machine real-time operational parameter data. Thus, a real-time tangle anomaly warning and monitoring system was developed for the actual operating conditions.

Keywords: dyeing process; anomaly detection; ensemble learning techniques; Web API; empirical research (search for similar items in EconPapers)
JEL-codes: O13 Q Q0 Q2 Q3 Q5 Q56 (search for similar items in EconPapers)
Date: 2022
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/2071-1050/14/14/8575/pdf (application/pdf)
https://www.mdpi.com/2071-1050/14/14/8575/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jsusta:v:14:y:2022:i:14:p:8575-:d:861901

Access Statistics for this article

Sustainability is currently edited by Ms. Alexandra Wu

More articles in Sustainability from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().