The extremely low cycle fatigue failure of steel tubular joints under strong seismic action may lead to progressive collapse of the structures. The prediction models of extremely low cycle fatigue fracture focused on the tensile fracture models and often ignored the influence of shear effect. Within the scope of fracture model based on microscopic mechanism, a Lode parameter enhanced cyclic void growth model (LCVGM) was proposed by improving the cyclic void growth model (CVGM) in the study. The LCVGM can comprehensively consider stress triaxiality sensitivity and Lode parameter dependence in extremely low cycle fatigue fracture prediction. The monotonic tensile and extremely low cycle fatigue tests were conducted on the different types of base metal and weld metal specimens in circular steel tubular joints. The micro fracture morphology of the specimens was observed by scanning electron microscope, and the fracture properties of both materials were compared. Moreover, LCVGM parameters of base metal and weld metal were calibrated, and the model was validated against extremely low cycle fatigue tests results of plate specimens. The fracture failure tests of circular steel tubular X-joints under monotonic loading and extremely low cyclic loading were completed. The joints exhibit chord punching failure mode. Finally, the numerical simulations of fracture tests of the joints were performed by means of the LCVGM and CVGM. The analysis results demonstrate that the fatigue life predicted by LCVGM is consistent with the test results, while the predicted life obtained from CVGM is significantly longer than the test result due to neglecting the influence of shear effect.

钢管节点在强地震作用下的极低周疲劳破坏可能导致结构逐步倒塌。极低周疲劳断裂预测模型侧重于拉伸断裂模型，往往忽略剪切效应的影响。在基于微观机理的裂缝模型范围内，通过改进研究中的循环孔隙生长模型（CVGM），提出了Lode参数增强循环孔隙生长模型（LCVGM）。LCVGM在极低周疲劳断裂预测中可以综合考虑应力三轴敏感性和Lode参数依赖性。对圆形钢管接头中不同类型的母材和焊缝金属试样进行了单调拉伸和极低周疲劳试验。用扫描电镜观察试样的显微断口形貌，比较两种材料的断裂性能。此外，校准母材和焊缝金属的 LCVGM 参数，并根据板试样的极低周疲劳试验结果验证模型。完成了单调加载和极低循环加载下圆形钢管X型接头的断裂破坏试验。接头表现出弦冲压失效模式。最后，利用LCVGM和CVGM对接头的断裂试验进行数值模拟。分析结果表明，LCVGM预测的疲劳寿命与试验结果一致，