Proceedings of the
World Congress on Micro and Nano Manufacturing (WCMNM 2022 )
19–22 September 2022, Lueven, Belgium
doi:10.3850/978-981-18-5180-3_RP39-0041
Study of AE and Sound Signals in Micro Laser Welding
Department of Mechanical Engineering, National Chung Hsing University
ABSTRACT
Several studies regarding the application of acoustic emission (AE) to the quality monitoring in micro-welding have been reported in past decades. However, these studies focus only on the development of a data-driven model to predict the welding quality. The detailed information regarding the relationship between the weld pool behavior and corresponding AE signal features has not been carefully investigated. In this study, single-layer spot welding tests were conducted with different laser power delivery to study the relationship between the weld pool motion and the features of the AE signal corresponding to each stage of weld pool generation. The potential mechanism of AE generation in each heating stage was studied based on the investigation of the AE signal, audible signal, and the photo taken by a high-speed camera, along with the fundamental theory of AE generation. The results show that two different AE generation mechanisms could be observed at the early stage when the weld pool is generated by heating. The AE signals respective to these two stages are found to concentrate on different frequency range bands. Moreover, the signals respective to the second stage will increase as the laser power increase. As the heat is continuously delivered to the material, the burst signals could be observed randomly during the whole heating process. However, as the conduction mode welding is conducted with laser power lower than 275 W, these randomly generated AE signals could not be observed. Because of the weld pool behavior, the high pressure of the gas inside the keyhole will be accumulated and generate a bubble randomly to cover the weld pool before it collapsed. Referring to the bubble generation and collapse, a burst AE signal with a high energy level could be observed. These results provide valuable information for the feature extraction in the development of a monitoring system to identify the welding quality, especially for spot welding. It demonstrates that the observed signals in the early stage of heating could be used to identify the failure of the spot welding if the penetration of the weld pool fails to reach the bottom layer of the material.
Keywords: Laser Welding, AE Signal, Sound Signal.
Department of Mechanical Engineering, National Chung Hsing University
ABSTRACT
Several studies regarding the application of acoustic emission (AE) to the quality monitoring in micro-welding have been reported in past decades. However, these studies focus only on the development of a data-driven model to predict the welding quality. The detailed information regarding the relationship between the weld pool behavior and corresponding AE signal features has not been carefully investigated. In this study, single-layer spot welding tests were conducted with different laser power delivery to study the relationship between the weld pool motion and the features of the AE signal corresponding to each stage of weld pool generation. The potential mechanism of AE generation in each heating stage was studied based on the investigation of the AE signal, audible signal, and the photo taken by a high-speed camera, along with the fundamental theory of AE generation. The results show that two different AE generation mechanisms could be observed at the early stage when the weld pool is generated by heating. The AE signals respective to these two stages are found to concentrate on different frequency range bands. Moreover, the signals respective to the second stage will increase as the laser power increase. As the heat is continuously delivered to the material, the burst signals could be observed randomly during the whole heating process. However, as the conduction mode welding is conducted with laser power lower than 275 W, these randomly generated AE signals could not be observed. Because of the weld pool behavior, the high pressure of the gas inside the keyhole will be accumulated and generate a bubble randomly to cover the weld pool before it collapsed. Referring to the bubble generation and collapse, a burst AE signal with a high energy level could be observed. These results provide valuable information for the feature extraction in the development of a monitoring system to identify the welding quality, especially for spot welding. It demonstrates that the observed signals in the early stage of heating could be used to identify the failure of the spot welding if the penetration of the weld pool fails to reach the bottom layer of the material.
Keywords: Laser Welding, AE Signal, Sound Signal.
