Ein Ansatz zu Qualitätsvorhersage mittels intelligenter SMT-Lötstelleninspektion durch den Einsatz von Maschinellem Lernen

Schmidt K, Bönig J, Beitinger G, Thielen N, Franke J (2020)

Publication Language: German

Publication Type: Conference contribution, Conference Contribution

Publication year: 2020

Publisher: VDE Verlag

Conference Proceedings Title: GMM-Fb. 94: EBL 2020 – Elektronische Baugruppen und Leiterplatten

Event location: Fellbach DE

ISBN: 978-3-8007-5186-0

URI: https://www.vde-verlag.de/buecher/455185/gmm-fb-94-ebl-2020-elektronische-baugruppen-und-leiterplatten.html

Abstract

The success factors of modern, global competition are changing the framework conditions of industrial production in a sustainable way. The optimization of the manufacturing process of electronic assemblies must ensure that they meet the requirements for adapted strategies and new design concepts, in particular in the field of digitization. Due to the already low error rates, which are achieved by highly optimized manufacturing and testing processes, future process improve-ments must be realized by reducing non-value-adding processes without endangering error rates. This can be reached through the targeted evaluation of recorded operating and machine data and the use of machine learning procedures. Due to the high standardization of processes and advanced plant networking, electronics manufacturing offers high application potential for data-driven process optimization. The state of the art often provides for cost- and time-intensive X-ray in-spection of hidden solder joints. According to the literature, the majority of solder joint defects are already caused by solder paste printing. This article therefore describes an approach for using the inspection data of the solder paste inspec-tion for quality prediction of the X-ray process. For the training of the model, the optical measurements are used as characteristics, together with the corresponding quality labels of the X-ray inspection, to establish a relationship between the solder paste printing and the quality result of the X-ray inspection for a selected product. The results were evaluated on the basis of various metrics and analysed for transferability into practice.

Authors with CRIS profile

Konstantin Schmidt Lehrstuhl für Fertigungsautomatisierung und Produktionssystematik (FAPS) Nils Thielen Lehrstuhl für Fertigungsautomatisierung und Produktionssystematik (FAPS) Jörg Franke Lehrstuhl für Fertigungsautomatisierung und Produktionssystematik (FAPS)

Involved external institutions

Siemens AG DE Germany (DE)

How to cite

APA:

Schmidt, K., Bönig, J., Beitinger, G., Thielen, N., & Franke, J. (2020). Ein Ansatz zu Qualitätsvorhersage mittels intelligenter SMT-Lötstelleninspektion durch den Einsatz von Maschinellem Lernen. In VDE/VDI-Gesellschaft Mikroelektronik, Mikrosystem- und Feinwerktechnik (GMM) (Hrg.), GMM-Fb. 94: EBL 2020 – Elektronische Baugruppen und Leiterplatten. Fellbach, DE: VDE Verlag.

MLA:

Schmidt, Konstantin, et al. "Ein Ansatz zu Qualitätsvorhersage mittels intelligenter SMT-Lötstelleninspektion durch den Einsatz von Maschinellem Lernen." Tagungsband EBL Fellbach, Fellbach Hrg. VDE/VDI-Gesellschaft Mikroelektronik, Mikrosystem- und Feinwerktechnik (GMM), VDE Verlag, 2020.

BibTeX: Download