正交异性钢桥面板裂纹扩展的均质化方法

A Homogenization Method on Crack Growth on Orthotropic Steel Decks

正交异性钢桥面板往往在焊缝及其附近位置出现过早的疲劳损伤问题。考虑到这一结构形式所含的大量焊接接头,对其中的随机初始裂纹及气孔、夹杂等焊接缺陷对疲劳性能的影响展开研究。首先,利用扩展有限元法在求解几何不连续问题方面的高效性,生成了大量含随机缺陷的代表体积元;而后基于等效裂纹扩展长度准则,对随机缺陷进行了均质化处理,并通过数值算例证明了其在早期裂纹扩展中的适用性。作为重要的多尺度方法之一,该均质化方法可用于计算正交异性钢桥面板焊接接头的裂纹扩展。根据某正交异性钢桥面板的典型车辆荷载的现场实测数据,结合欧洲规范建议的年车流量及轮载横向分布,采用均质化的模型对U肋-顶板焊趾的宏观裂纹形成寿命(初始裂纹扩展至可检测尺度的寿命)进行了计算。结果表明:该方法可有效呈现宏观裂纹形成寿命的概率分布,但在现有假设下,初始裂纹尺度的影响明显高于气孔、夹杂等缺陷;若初始裂纹尺度呈均匀分布,将导致宏观裂纹形成寿命的分布较为离散;所得结果的统计特性依赖于所选择的Paris公式形式及其中的材料常数取值。

The orthotropic steel decks（OSDs）always suffered from premature fatigue damage at welded joints and the vicinities.Given the large number of welded joints on this type of structure,the effects of random initial cracks and weld defects on fatigue performance such as pore and inclusion were investigated in this paper.Taking advantage of the high efficiency on solving discontinuity problems of the extended finite element method（XFEM）,a large number of representative volume elements（RVEs）with random defects were generated in the first step.Based on the criterion of equivalent crack growth length,these defects were homogenized,which were proven to be feasible in the early stage of crack growth by numerical examples.As one of the most applied multi-scale methods,the homogenization method was quite useful on the calculation of crack growth on welded joints of OSDs.According to the on-site test data on a certain OSD,combined with the annual traffic volume and the transverse distribution of wheel loading suggested by Eurocode,the macro-crack initiation life（MCIL）（the life till crack growsto a detectable size）on rib-to-deck weld toe was calculated using homogenized models.The results show that in the form of possibility distribution,the method is able to demonstrate the variation of MCIL.Under the current hypotheses,however,the effect of initial crack size is apparently more important than that of pores and inclusions.Hence,the distribution of MCIL is obviously more scattered when the initial crack size is uniformly distributed.Besides,the statistic characteristics of the results are also dependent on the basic form of Paris law,as well as the material constants in it.