The coupon specimen with a transverse butt weld in the middle could be used for determining local constitutive properties of the heat-affected zone (HAZ) and the weld metal (WM) based on the digital image correlation (DIC). However, limited research is reported to demonstrate how to identify the boundary of each region in DIC results. Accordingly, it is difficult to determine the adequate gauge length for measuring the strain of each region and establishing a stress–strain curve for Finite Element Analysis (FEA) of structural problems. In this paper, a method for identifying the region's boundary is proposed based on coupon tests using three steel grades, S355, S500, and S700, corresponding to three weld matching types, match, overmatch, and undermatch, respectively. First, the hardness and the microstructure investigation were conducted to determine the region's boundary. Then, the boundary was identified based on the DIC result using the proposed method. Finally, the identified HAZ regions were verified against hardness results. Using the proposed method, the gauge length for measuring the strain of a single region, such as HAZ and WM, are determined based on the DIC results. The primary purpose is to establish the stress–strain relationship for FEA of welded joints.

Failure of bolts exposed to tension is generally avoided in the design of bolted connections due to the smaller deformation capacity of bolts than the connected plates. This is one of the reasons why few studies focus on the tensile failure behaviour of bolts. However, failure behaviour of bolts is essential for the advanced finite element analysis especially relevant to the deformation capacity and failure mode of bolted connections. This paper presents a numerical study on the fracture of partially threaded bolts under tension incorporating damage models, with which the failure mechanism of bolts can be better understood. The post-necking stress–strain relation is firstly calibrated to describe the behaviour of bolt threaded parts at large deformation. Then, direct tension tests on partially threaded bolts with different threaded lengths within the grip are modelled using ABAQUS with the explicit solver. Two criteria for fracture are investigated: the void growth model (VGM) and a model proposed by Bao and Wierzbicki (BW). The former is adopted to simulate the tensile fracture of bolts and the latter is used to predict the thread stripping failure. Results indicate that bolt failure modes of tensile fracture and thread stripping can be well predicted by combining the calibrated post-necking stress–strain relation and a suitable fracture criterion in the analysis. It is revealed that the large plastic strain in the threads introduced by thread rolling process is a major reason for the thread stripping failure.

Numerical analyses incorporating appropriate damage models provide an opportunity to predict the strength and deformation capacity of steel structures. This paper presents a practical calibration for the ductile damage model of S355 and high-strength steel S690Q, S700MC, S960Q based on tensile coupon test results. A combined linear and power expression is adopted to calibrate the post-necking damaged stress–strain relations of the investigated steels, upon which the undamaged stress–strain relations are estimated further. Damage initiation criterion is based on the Rice-Tracey model and damage evolution law is related to the calibrated damaged stress and the estimated undamaged stress. Fracture of the tensile coupons is modelled using a critical damage variable. Tensile coupon tests on the investigated steels are modelled in ABAQUS with the explicit solver. Results show that combining the proposed post-necking stress–strain relations and ductile damage model generates very good predictions for strain localization and final fracture of the tensile coupons. Numerical engineering stress–strain curves agree well with the experimental results. It also indicates that high-strength steels are more susceptible to damage than S355. The damage variable of S960Q is about 2 times as large as that of S355 from the onset of necking to the final fracture.

To predict the flexural strengths of steel girder-concrete abutment connections in fully integral abutment bridges with perfobond connectors, this paper analytically and numerically investigates the ultimate flexural behavior of these connections. A nonlinear finite element model was first established and validated against experimental results. Subsequently, another 89 nonlinear finite element models with different connector quantities, connector arrangements, abutment widths, and shear-span ratios were studied. An analytical calculation method considering the moment contribution of perfobond connectors and the shear force distribution on perfobond connectors at ultimate was then proposed, verified against experimental and numerical results, and was compared with current calculation methods. Results show that, compared with current methods, the proposed method predicts the flexural strengths of the connections with better accuracy, with the average predicted-actual ratio being 0.96 and the coefficient of variation being 0.09. The flexural strengths of girder-abutment connections are significantly affected by girder embedded length, abutment width, and the quantity and arrangement of perfobond connectors. The flexural strengths are higher when connectors are set close to the girder end. When increasing the quantity of perfobond connectors, the flexural strengths of the connections could increase by 86% at most. With the same connector quantity, the flexural strengths could vary 20% with different connector arrangements.