Enhancing Crimp Curve Monitoring in Wiring Harness Production: A Machine Learning Approach with Emphasis on Diverse Data

Hofmann B, Scheck A, Nguyen HG, Meiners M, Franke J (2024)

Publication Type: Conference contribution

Publication year: 2024

Journal

Lecture Notes in Mechanical Engineering Springer

Publisher: Springer

City/Town: Singapur

Pages Range: 3-13

Conference Proceedings Title: Proceedings of the 11th International Conference on Industrial Engineering and Applications

Event location: Hiroshima JP

ISBN: 9789819764914

DOI: 10.1007/978-981-97-6492-1_1

Abstract

With emerging trends toward electromobility and autonomous driving, rigorous standards on the quality, safety, and reliability of underlying signal and power distribution networks are becoming increasingly important. These developments necessitate significant adjustments in manufacturing methodologies, particularly in the production process of wiring harnesses, which are characterized by a high variety of components. Crimping, one of the most complex processes within the wiring harness value chain, is commonly used for joining wire and connecting elements. To ensure elevated quality standards throughout the manufacturing process and to address the complexity associated with the high variety of crimp connections, a deeper investigation of data-driven monitoring approaches during production is necessary. Machine learning-based techniques, which diverge from conventional approaches by not relying on fixed programming rules but being trained on sample data, are gaining prominence. This paper further investigates the capabilities of machine learning for monitoring crimp curves, placing a special emphasis on a diverse dataset. The dataset, encompassing process curves derived from various machines and wire cross-sections, highlights the diverse conditions encountered in manufacturing environments. It is analyzed by evaluating several supervised machine learning techniques, with results compared to the conventional monitoring system. Achieving accuracies up to 95%, the results underline the suitability of this approach for handling diverse process curves in the domain of crimping.

Authors with CRIS profile

Bernd Hofmann Lehrstuhl für Fertigungsautomatisierung und Produktionssystematik (FAPS) Albert Scheck Lehrstuhl für Fertigungsautomatisierung und Produktionssystematik (FAPS) Huong Giang Nguyen Lehrstuhl für Fertigungsautomatisierung und Produktionssystematik (FAPS) Jörg Franke Lehrstuhl für Fertigungsautomatisierung und Produktionssystematik (FAPS)

Involved external institutions

Schäfer Werkzeug- und Sondermaschinenbau GmbH DE Germany (DE)

How to cite

APA:

Hofmann, B., Scheck, A., Nguyen, H.G., Meiners, M., & Franke, J. (2024). Enhancing Crimp Curve Monitoring in Wiring Harness Production: A Machine Learning Approach with Emphasis on Diverse Data. In Proceedings of the 11th International Conference on Industrial Engineering and Applications (pp. 3-13). Hiroshima, JP: Singapur: Springer.

MLA:

Hofmann, Bernd, et al. "Enhancing Crimp Curve Monitoring in Wiring Harness Production: A Machine Learning Approach with Emphasis on Diverse Data." Proceedings of the 11th International Conference on Industrial Engineering and Applications, Hiroshima Singapur: Springer, 2024. 3-13.

BibTeX: Download