Abstract Residual stresses can have a significant impact on the stability of structural members. In the case of I-section beam elements, such stresses can impact lateral–torsional buckling (LTB) capacity, particularly in the inelastic range. I-sections are typically fabricated by either rolling as a single shape or welding three plates together; residual stress distributions can differ considerably between the two section types. It is therefore possible for a built-up welded girder to have a different LTB capacity than that of a rolled one of identical cross-section. Concerns have been raised that such a difference may render North American steel design standards unconservative for welded girders. Because of the lack of recent physical LTB tests on welded girders, finite element modelling was used as a tool in making this assessment, and the assumed residual stress distributions were based on data from 1970 to 1980 that may not account for advancements in welding technologies and processes since the publication of these data. The paucity of recent residual stress data, however, prevents a direct assessment of these assumed distributions. In this study, residual stress measurements were carried out on a series of four welded steel test girders using the sectioning method. A comparison of the measured distributions with four existing models finds these models to be inaccurate in important ways; an updated model is therefore required for use in numerical simulations of LTB behaviour of welded girders used in industry today.

中文翻译：

---

焊接梁残余应力的测量和预测

摘要 残余应力会对结构构件的稳定性产生重大影响。在 I 型截面梁单元的情况下，这种应力会影响横向扭转屈曲 (LTB) 能力，特别是在非弹性范围内。I 型材通常通过轧制成单个形状或将三块板焊接在一起来制造；两种截面类型之间的残余应力分布可能会有很大差异。因此，组合焊接梁的 LTB 容量可能与具有相同横截面的轧制梁的 LTB 容量不同。有人担心这种差异可能会使北美钢结构设计标准对焊接梁变得不保守。由于最近缺乏对焊接梁的物理 LTB 测试，有限元建模被用作进行此评估的工具，假设的残余应力分布基于 1970 年至 1980 年的数据，这些数据可能没有考虑自这些数据发布以来焊接技术和工艺的进步。然而，近期残余应力数据的缺乏阻碍了对这些假设分布的直接评估。在这项研究中，残余应力测量是使用分段方法对一系列四个焊接钢试验梁进行的。将测得的分布与四个现有模型进行比较，发现这些模型在重要方面不准确；因此，需要一个更新的模型来对当今工业中使用的焊接梁的 LTB 行为进行数值模拟。