研究目的
Investigating the challenges of in situ quality control in laser keyhole welding, focusing on pore dynamics and the correlation between optical and acoustic signals with momentary events during the welding process.
研究成果
The combination of AE and optical sensing techniques, along with high-speed X-ray imaging, provides a comprehensive approach to studying laser welding processes. However, the dynamic and complex nature of these processes necessitates advanced signal processing and machine learning techniques for effective in situ and real-time quality control.
研究不足
The complexity and short time spans of events such as keyhole fluctuation, pore formation, and spattering make human visual inspection of signals inefficient and unreliable. Advanced signal processing and statistical techniques are required for accurate correlation.
1：Experimental Design and Method Selection:
The study employed high-speed X-ray imaging at the European Synchrotron Radiation Facility (ESRF) to visualize the welding process, alongside optical and acoustic emission (AE) sensors to record signals during welding.
2：Sample Selection and Data Sources:
Aluminum samples were chosen for their high X-ray transmissivity, with dimensions of 2x20x50 mm
3：List of Experimental Equipment and Materials:
A single-mode fiber laser source StarFiber 150P, optical sensors (Si, Ge, AlGaAs), an acoustic piezoelectric sensor Pico, and a high-speed X-ray imaging setup at ESRF.
4：Experimental Procedures and Operational Workflow:
Laser welding experiments were conducted with varying laser powers and pulse durations, with simultaneous recording of optical and AE signals and X-ray imaging.
5：Data Analysis Methods:
Wavelet decomposition and machine learning techniques were used for signal processing and event classification.
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
