Search

• Images of vapor plume in high-speed laser welding are captured through high-speed imaging. • Influences of laser power and welding speed on the behavior of vapor plume are analyzed. • Humping occurrence are predicted with the random forest and a sliding time window. The increase of welding speed can improve the productivity of laser welding. The behavior of vapor plume is an important indicator reflecting the state of the laser welding process. However, the behavior of vapor plume in high-speed laser welding (whose welding speed is higher than 10 m/min) has not been studied adequately. In this research, images of the vapor plume in high-speed laser welding of SUS304 austenitic stainless steel are captured through high-speed imaging for the monitoring of the welding process. Characteristics of the vapor plume are extracted with image processing, and the influences of laser power and welding speed on these characteristics are analyzed. The relationships between the behavior of vapor plume and melt pool are discussed, and three ejection regimes of the vapor plume are proposed. In order to predict the occurrence of humping, which is a typical defect in high-speed laser welding, samples are built by applying a time sliding window and then a classification model is built with the random forest. In addition, the importance of features in the classification is analyzed. The test results show that the proposed method can predict the humping defect in high-speed laser welding accurately. [ABSTRACT FROM AUTHOR]

In order to ensure the production quality of high-speed laser welding, it is necessary to simultaneously monitor multiple state properties. Monitoring methods combining vision sensing and deep learning models are popular but most models used can only make predictions on single welding state property. In this contribution, we propose a multi-output model based on a lightweight convolutional neural network (CNN) architecture and introduce the particle swarm optimization (PSO) technique to optimize the loss function of the model, to simultaneously monitor multiple state properties of high-speed laser welding of AISI 304 austenitic stainless steel. High-speed imaging is performed to capture images of the melt pool and the dataset is built. Test results of different models show that the proposed model can achieve monitoring of multiple welding state properties accurately and efficiently. In addition, we make an interpretation and discussion on the prediction of the model through a visualization method, which can help to deepen our understanding of the relationship between the melt pool appearance and welding state. The proposed method can not only be applied to the monitoring of high-speed laser welding but also has the potential to be used in other procedures of welding state monitoring. [ABSTRACT FROM AUTHOR]