The development of a quality prediction system for aluminum laser welding to measure plasma intensity using photodiodes

Lightweight metals have been used to manufacture the body panels of cars to reduce the weight of car bodies. Typically, aluminum sheets are welded together, with a focus on weld quality assurance. A weld quality prediction system for the laser welding of aluminum was developed in this research to ma...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of mechanical science and technology Jg. 30; H. 10; S. 4697 - 4704
Hauptverfasser: Yu, Jiyoung, Sohn, Yongho, Park, Young Whan, Kwak, Jae-Seob
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Seoul Korean Society of Mechanical Engineers 01.10.2016
Springer Nature B.V
대한기계학회
Schlagworte:
Algorithms
Aluminum
Automobile industry
Automotive bodies
Control
Dynamical Systems
Engineering
Feature recognition
Feed rate
Industrial and Production Engineering
Laser beam welding
Luminous intensity
Mathematical models
Mechanical Engineering
Neural networks
Pattern recognition
Photodiodes
Quality
Quality assurance
Vibration
Weight reduction
Welded joints
Welding
기계공학
Monitoring system
Plasma intensity
Aluminum laser welding
Neural network model
Fuzzy pattern recognition algorithm
ISSN:1738-494X, 1976-3824
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Lightweight metals have been used to manufacture the body panels of cars to reduce the weight of car bodies. Typically, aluminum sheets are welded together, with a focus on weld quality assurance. A weld quality prediction system for the laser welding of aluminum was developed in this research to maximize welding production. The behavior of the plasma was also analyzed, dependent on various welding conditions. The light intensity of the plasma was altered with heat input and wire feed rate conditions, and the strength of the weld and sensor signals correlated closely for this heat input condition. Using these characteristics, a new algorithm and program were developed to evaluate the weld quality. The design involves a combinatory algorithm using a neural network model for the prediction of tensile strength from measured signals and a fuzzy multi-feature pattern recognition algorithm for the weld quality classification to improve predictability of the system.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
G704-000058.2016.30.10.019
ISSN:1738-494X
1976-3824
DOI:10.1007/s12206-016-0940-9