| The existence and the growing of cracks in welded steel bridge structures will undoubtedly affect the structural safety in use. The investigations of fatigue crack growth behavior and fatigue life prediction are very significative, and it provides fundamental knowledge for fatigue resistant design to ensure the safety of engineering structures. In this dissertation, the fatigue crack growth behavior of surface crack in 14MnNbq welded plates and the fatigue life prediction are investigated by using experimental method and finite element method. The achievements of the present study are as follows:(l)The FCG behaviors of through-thickness cracks in 14MnNbq steel and its weldments are investigated experimentally. The FCG rates and threshold under different stress ratios R=0.05,0.25 and 0.50 in base metal, heat-affected zone and welded metal are obtained, respectively. By fitting the test results, a simple FCG rate equation is proposed to account for the influences of stress ratio R,. and the advantage of this equation is that FCG rates with different stress ratios R can be determined directly from the Paris'material constants corresponding to R=0. By comparing the FCG behaviors among base metal, HAZ and weld metal, it can be concluded that HAZ is the control domain in welded steel structure for fatigue crack growth.(2)An efficient approach to determine fatigue crack closure factor by using thresholds and FCG rates is proposed. It is found that the FCG rate data with considernation of the proposed closure factor tend to merge regardless of R.(3)A dimensionless crack closure correction model is proposed to correct the differences in FCG rate between surface cracks and through-thickness cracks. This correction model is determined by the test results of surface crack under tensile loading with constant ratio R=0.05. By using the proposed correction model, the surface crack growth rate can be correlated with reasonable accuracy under tensile and bending loadings with various stress ratios-ranging from 0 to 0.5. Furthermore, predictions of fatigue life and crack aspect ratio for surface crack are discussed. The predicted results are compared with those obtained from other approaches. Comparison between a wide range of experimental data show that better predictions of fatigue life are obtained by using the proposed crack closure correction model.(4)A modified zigzag approach is proposed to approximate moving crack front with arbitrary shape based on the virtual crack closure-integral technique. An approximated zigzag line are used to represent the complex crack front which has the same general shape for the given crack front. Based on this zigzag approximate crack front, the approaches to determine the required virtually closed area, displacement openings are proposed. The differences of approaches between present study, and extraction approaches are discussed indetails. The comparison results indicate that the present approach provided a more detailed consideration to determine the virtually closed area, there is no blank space between the virtually closed areas for the consecutive crack tip nodes. Also, the present approach can be used to calculate G when the finite elements across the crack front have different sizes.(5)Following the proposed modified zigzag approach, two kinds of mesh patterns, exact mesh pattern and zigzag mesh pattern, can be used. With exact mesh pattern, the nodes lie precisely along the crack front, and the FEM results agree well with the analytical solutions. With zigzag mesh pattern, the nodes cross the curved crack front in a zigzag pattern. The advantages of this mesh pattern are that (?) a simple stationary finite element mesh can be used for crack front with arbitrary shape and (?) adaptive re-meshing technique is avoided in studying crack growth. The disadvantage of this mesh pattern is that, the G results oscillate around the analytical solutions. With zigzag mesh pattern, the G results evaluated by proposed approach are compared with these obtained from other approaches. The comparisons results indicated that the present approach provide much more accurate strain energy release rate G results than the extraction approaches.(6) The crack growth simulations using the proposed modified zigzag approach are presented for cracks with complex initial shapes. With exact mesh pattern, Paris law is used to determine the crack growth increament and fatigue life. In order to assess the validity of this approach, a case of surface crack growth is presented. Good agreement between FEM and experimental results is found, in which Paris material constants C and m are appropriately used. With zigzag mesh pattern, a mixed-mode fracture criterion is used to predict crack growth. In order to demonstrate the validity and accuracy of present approach, through-the-thickness crack, a corner crack and multiple cracks were presented. Reasonable agreement between present study and analytical solutions was obtained. Furthermore, the present approach is shown to be insensitive to the mesh density. The FEM results converge quickly after the mesh reaches a reasonable density. The shape changes during crack propagation can be tracked with ease by using the present simulation techniques. |
PDF Full Text Request