研究目的
To develop a real-time spatter detection system in laser welding using machine vision and to investigate the influence of process parameters on spatter formation.
研究成果
The monitoring system developed allows for real-time spatter detection and shows significant differences in spatter number and size for varying process parameters. Beam oscillation is suitable for reducing spatter formation, and parameter combinations resulting in fewer spatters are more robust and repeatable.
研究不足
The standard deviation of the spatter size was larger than the corresponding mean value for all examined parameter combinations, indicating that the repeatability of the spatter size is not given.
1：Experimental Design and Method Selection:
The study uses a machine vision algorithm executed on a graphics processing unit for real-time spatter detection. The methodology includes a consumer-producer design pattern for data processing and an image processing algorithm for spatter segmentation.
2：Sample Selection and Data Sources:
The experiments were conducted on oxygen-free copper CW008A specimens, partially penetrated by a laser beam.
3：List of Experimental Equipment and Materials:
Welding optics elephant50 by ARGES, process camera (EoSens 3CXP) with a monochrome CMOS sensor, NVIDIA P5000 GPU, AMD Ryzen 7 1700x CPU, and an ytterbium-doped YAG single mode fiber laser by IPG Photonics.
4：Experimental Procedures and Operational Workflow:
The system detects spatters at a rate of 1 kHz with a resolution of 900 x 900 pixels. The algorithm splits into data processing and image processing parts running in parallel.
5：Data Analysis Methods:
The spatter number and size are evaluated using a quality vector. The reproducibility of spatter formation is investigated by calculating mean values and standard deviations over ten repetitions.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
