Product Code: ICAL07_1503

Estimation of Penetration Depth by using Coaxial Image Monitoring during Laser Lap Welding
Authors:
Cheolhee Kim, KITECH; Yeonsu-gu Incheon South Korea
Junghak Kim, KITECH; Incheon South Korea
Jeonghan Kim, KITECH; Incheon South Korea
Sehun Rhee, HanYang University; Seoul South Korea
Presented at ICALEO 2007

Laser welding are well-known welding processes applied in the automotive. Because the process monitoring is beneficial for maintaining quality and improving productivity, various monitoring techniques have been proposed for laser welding. In this study, we coaxially monitored the welding phenomena during laser lap welding. Some welding characteristics pool width and length, keyhole shape, etc were extracted from coaxial images and the relationship between these characteristics and the penetration depth were investigated. For example, as the welding speed increases with a constant laser power, the penetration mode changes from the full penetration mode to the partial penetration one and the length of weld pool suddenly changes at the transition point. The neural network model was established to estimate the penetration depth from the pool images and showed a good performance.