This paper investigates the full-range behaviour of T-stubs through both experimental and numerical approaches. Nineteen T-stub specimens with six different yield line patterns were loaded to the complete failure. Most of the specimens failed due to rupture of the bolts after significant yielding of the flanges. A new method is proposed to determine the points of initial yielding and ultimate state on a nonlinear full-range force-deformation curve, and by using this method the mechanical properties of the specimens were obtained. Benchmarked against the experimental results, the validities of three T-stub analytical models are examined. None of the models is proven to be able to accurately capture the full-range behaviour of the T-stubs, with the Jaspart model showing a comparatively better accuracy. Finite-element models capable of reproducing the experimental phenomena are developed and are extended to cover all possible failure modes. It is observed that the ratio between the resistances of Failure modes 1 and 3, β, can be used to predict the behaviour and failure mode of the T-stub. The actual β-value dividing Failure modes 1 and 2 is equal to the theoretical value, but that dividing Failure modes 2 and 3 is much smaller than the theoretical value of 2.0. Taking the resistance, stiffness and ductility into comprehensive consideration, it is suggested that a T-stub be designed with a β-value around 0.67 to achieve a balanced design with a good overall performance.

本文通过实验和数值方法研究了 T 型短截线的全范围行为。具有六种不同屈服线模式的 19 个 T 形短柱试样被加载到完全破坏。大多数试样由于在法兰显着屈服后螺栓断裂而失效。提出了一种确定非线性全范围力-变形曲线上的初始屈服点和极限状态点的新方法，并通过该方法获得了试件的力学性能。以实验结果为基准，检验了三个 T 型存根分析模型的有效性。没有一个模型被证明能够准确地捕捉 T 型短截线的全范围行为，而 Jaspart 模型显示出相对更好的准确性。开发了能够重现实验现象的有限元模型，并将其扩展到涵盖所有可能的故障模式。据观察，失效模式 1 和 3 的电阻之间的比率，β可用于预测 T 型桩的行为和故障模式。实际β值除以失效模式1和2等于理论值，但除以失效模式2和3比理论值2.0小得多。综合考虑阻力、刚度和延展性，建议T型桩设计β值在0.67左右，以达到综合性能良好的平衡设计。