焊接
正交异性材料
材料科学
甲板
结构工程
残余应力
有限元法
应力集中
压力(语言学)
复合材料
极限抗拉强度
耐久性
疲劳极限
开裂
疲劳开裂
振动疲劳
瞬态(计算机编程)
灵敏度(控制系统)
热影响区
应力场
巴黎法
作者
Saman Karimi,Olivia Mirza,Fidelis Mashiri,Christophe Camille
标识
DOI:10.1016/j.ijfatigue.2026.109725
摘要
Orthotropic steel decks (OSDs) provide a high strength-to-weight ratio and structural efficiency, yet fatigue cracking at welded joints remains a major concern. While traffic loading is the dominant external action, welding residual stress (WRS) induces a fundamental shift in mean stress that can significantly accelerate fatigue damage accumulation and lead to non-conservative life estimates when omitted from design assessments. This paper presents a numerical investigation into the fatigue life of an OSD by integrating transient welding simulations with an advanced Normal Strain-Morrow durability assessment. A three-dimensional finite element model (FEM) was validated against experimental strain data and reproduced the gas metal arc welding (GMAW) process using a Goldak double-ellipsoid heat source. The welded U-rib-to-deck plate connection is explicitly modelled to capture its effect on local stress distribution. The results demonstrate that a peak residual tensile field of 688 MPa acts as a substantial tensile preload, shifting the cyclic traffic response from a constant stress range of 8.63 MPa to a sustained high-damage regime. Under an Australian standard traffic spectrum, the model incorporating WRS identified fatigue initiation at 102,000 cycles, whereas the WRS-excluded model showed no damage. Sensitivity analysis further reveals an imbalance in the joint’s durability: a marginal 15% increase in tensile WRS results in an estimated 43% reduction in initiation life. This identifies a critical stress gap in current design protocols, which may underestimate failure risks in high-restraint weldments. This framework provides a rigorous methodology for refining fatigue life predictions and supporting performance-based maintenance strategies for steel bridge infrastructure worldwide.
科研通智能强力驱动
Strongly Powered by AbleSci AI