Scholarlink

<p>Predicting thermal cycles in multi-wire submerged arc welding is challenging due to complex arc interactions and flux coverage. Finite element models capture heat transfer and fluid flow, but their computational cost limits process optimization. We present a semi-analytical thermal model for a two-wire tandem SAW that superposes ellipsoidal sources within a Green’s function framework and includes temperature-dependent properties, latent heat via effective specific heat, and convective–radiative cooling. The solver predicts weld-pool geometry and full thermal histories; a CUDA implementation on an NVIDIA T4 runs full-plate cases in about five minutes. Validation against ASTM&#xa0;A36 bead-on-groove trials shows errors within <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(12\%\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>12</mn> <mo>%</mo> </mrow> </math></EquationSource> </InlineEquation> and accurate peak-temperature and <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(t_{8/5}\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>t</mi> <mrow> <mn>8</mn> <mo stretchy="false">/</mo> <mn>5</mn> </mrow> </msub> </math></EquationSource> </InlineEquation> cooling-time estimates, offering a fast alternative to full FEM for the tested SAW–T conditions.</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

Semi-analytical thermal modeling of two-wire tandem submerged arc welding incorporating latent heat and convective–radiative boundary effects

  • 期刊论文

  • 发表日期：

    2026年7月8日

查看全文
  • Y. Priyanka,
  • Adapa Mahanth Kumar,
  • Degala Venkata Kiran,
  • P. Mariappan

摘要

Predicting thermal cycles in multi-wire submerged arc welding is challenging due to complex arc interactions and flux coverage. Finite element models capture heat transfer and fluid flow, but their computational cost limits process optimization. We present a semi-analytical thermal model for a two-wire tandem SAW that superposes ellipsoidal sources within a Green’s function framework and includes temperature-dependent properties, latent heat via effective specific heat, and convective–radiative cooling. The solver predicts weld-pool geometry and full thermal histories; a CUDA implementation on an NVIDIA T4 runs full-plate cases in about five minutes. Validation against ASTM A36 bead-on-groove trials shows errors within \(12\%\) 12 % and accurate peak-temperature and \(t_{8/5}\) t 8 / 5 cooling-time estimates, offering a fast alternative to full FEM for the tested SAW–T conditions.