Scholarlink

<p>The cross-sectional morphology of a laser deep melting welding seam is a key factor influencing the mechanical properties of the joint. Traditional destructive testing methods cannot meet the demands of high-quality production, making real-time prediction of seam morphology based on process signals a research hotspot. This paper systematically reviews recent progress in weld morphology prediction using plasma visual monitoring. The dynamic behavior of plasma is analyzed through multimodal signals—including acoustic, optical, electrical, and imaging data—and the correlation mechanisms between plasma dynamics and morphological parameters such as penetration depth and width are discussed. Significant correlations are reported, for instance, between light signal intensity and electron temperature, as well as between plasma height and penetration depth. To address the challenges of image noise and complex information in high-speed photography, innovative algorithms such as filtering, histogram equalization, and gravitational law edge detection are applied to enhance image quality. A data characterization framework is then established by combining geometric features with time–frequency characteristics. In terms of prediction methods, empirical formulas are suitable for simple working conditions, numerical simulations dynamically reproduce the welding process using heat source models, and machine learning demonstrates high accuracy and stability in handling complex nonlinear relationships. Overall, this study provides theoretical guidance and technical references for the intelligent quality control of laser deep melting welding and supports its engineering applications in advanced manufacturing fields such as aerospace and rail transit.</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

Advances in the prediction of cross-section morphology of laser welded seams based on plasma vision monitoring

  • 期刊论文

  • 发表日期：

    2026年1月19日

查看全文
  • Zifan Bai,
  • Yi Zhang,
  • Cong Chen

摘要

The cross-sectional morphology of a laser deep melting welding seam is a key factor influencing the mechanical properties of the joint. Traditional destructive testing methods cannot meet the demands of high-quality production, making real-time prediction of seam morphology based on process signals a research hotspot. This paper systematically reviews recent progress in weld morphology prediction using plasma visual monitoring. The dynamic behavior of plasma is analyzed through multimodal signals—including acoustic, optical, electrical, and imaging data—and the correlation mechanisms between plasma dynamics and morphological parameters such as penetration depth and width are discussed. Significant correlations are reported, for instance, between light signal intensity and electron temperature, as well as between plasma height and penetration depth. To address the challenges of image noise and complex information in high-speed photography, innovative algorithms such as filtering, histogram equalization, and gravitational law edge detection are applied to enhance image quality. A data characterization framework is then established by combining geometric features with time–frequency characteristics. In terms of prediction methods, empirical formulas are suitable for simple working conditions, numerical simulations dynamically reproduce the welding process using heat source models, and machine learning demonstrates high accuracy and stability in handling complex nonlinear relationships. Overall, this study provides theoretical guidance and technical references for the intelligent quality control of laser deep melting welding and supports its engineering applications in advanced manufacturing fields such as aerospace and rail transit.