Scholarlink

<p>This study addresses the challenges of monitoring the melt pool during the wire arc additive manufacturing (WAAM) of magnesium alloys, which arise from high thermal conductivity and rapid cooling rates. A solution based on multi-band visual imaging is proposed. Firstly, spectral analysis was conducted to determine the radiation characteristics of the welding arc and the melt pool, revealing that the arc radiation is primarily concentrated in the visible-light spectrum. Based on this, the visual sensing system was optimized: A band-stop filter with a central wavelength of 532&#xa0;nm and a bandwidth of 490–550&#xa0;nm effectively suppresses arc interference. When the camera parameters are set to a gain of 2, exposure of 1500&#xa0;μs, and aperture of f/1.4, images of the melt pool with optimal edge contrast can be obtained. To address the instability issue during the arc burning phase of the cold metal transfer (CMT) WAAM process (which accounts for two-thirds of the process cycle), an innovative laser-triggered illumination system was developed: a 940-nm band-pass filter combined with a synchronized laser background light source successfully overcame arc interference, enabling uninterrupted melt pool imaging throughout the entire process cycle. Based on the optimized melt pool image data, a deep residual network model was constructed to achieve high-precision monitoring of weld beads formation quality, with an identification accuracy reach to 92.7%. The main contributions of this study are as follows: 1) establishing optimization criteria for multi-band imaging of the melt pool in magnesium alloy WAAM; 2) proposing a dynamic monitoring scheme that combines laser synchronized illumination with narrow-band filtering; and 3) achieving, for the first time, intelligent prediction of the entire process quality in magnesium alloy CMT-WAAM based on deep learning.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Study on Molten Pool Imaging and Weld Bead Formation Monitoring in the CMT-WAAM Process of Magnesium Alloy

  • 期刊论文

  • 发表日期：

    2026年1月9日

查看全文
  • Tianyang Zhang,
  • Jingjing Cheng,
  • Cheng Xu,
  • Lyuyuan Wang,
  • Kehong Wang

摘要

This study addresses the challenges of monitoring the melt pool during the wire arc additive manufacturing (WAAM) of magnesium alloys, which arise from high thermal conductivity and rapid cooling rates. A solution based on multi-band visual imaging is proposed. Firstly, spectral analysis was conducted to determine the radiation characteristics of the welding arc and the melt pool, revealing that the arc radiation is primarily concentrated in the visible-light spectrum. Based on this, the visual sensing system was optimized: A band-stop filter with a central wavelength of 532 nm and a bandwidth of 490–550 nm effectively suppresses arc interference. When the camera parameters are set to a gain of 2, exposure of 1500 μs, and aperture of f/1.4, images of the melt pool with optimal edge contrast can be obtained. To address the instability issue during the arc burning phase of the cold metal transfer (CMT) WAAM process (which accounts for two-thirds of the process cycle), an innovative laser-triggered illumination system was developed: a 940-nm band-pass filter combined with a synchronized laser background light source successfully overcame arc interference, enabling uninterrupted melt pool imaging throughout the entire process cycle. Based on the optimized melt pool image data, a deep residual network model was constructed to achieve high-precision monitoring of weld beads formation quality, with an identification accuracy reach to 92.7%. The main contributions of this study are as follows: 1) establishing optimization criteria for multi-band imaging of the melt pool in magnesium alloy WAAM; 2) proposing a dynamic monitoring scheme that combines laser synchronized illumination with narrow-band filtering; and 3) achieving, for the first time, intelligent prediction of the entire process quality in magnesium alloy CMT-WAAM based on deep learning.