In this paper, a numerical simulation method of mixed-mode fatigue crack propagation was explored using the extended finite element method (XFEM) and the Virtual Crack Closure Technique (VCCT). Both 2D and 3D numerical models were selected to simulate the fatigue crack propagation of steel specimens. Two coefficients were proposed to calculate the equivalent energy release rate (G_eq) for a better simulation of the mixed-mode fatigue crack propagation of S355 steel. The Walker equation and the calculation formula of G_eq were realized by a user-defined subroutine. A set of optimal correction coefficients of mode II energy release rate (G_II) and mode III energy release rate (G_III) were quantitatively comparing the simulation results and test data. The results will contribute to fatigue crack propagation prediction of steel structures in the civil engineering field.

In steel structures, a lot of attention is paid to lightweight structures, i.e. reduction of dead load without compromising structural safety, integrity and performance. Thanks to modern steel aluminium foam sandwich panel manufacturing technology a new possibility became available for lightweight structural design. Assessment and understanding of the behaviour of this sandwich panel under in-plane compression or flexure is crucial before its application in steel structures. Column buckling theory is considered and applied to the steel aluminium foam sandwich panel to evaluate its behaviour under in-plane compressive load. In this work, various assumptions are made to generalise Euler's buckling formula. The generalisation requires modification of the buckling stiffness expression to account for sandwich panel composite properties. The modified analytical expression is verified with finite element simulation employing various material models specific to steel face-plates and aluminium foam as well as various geometric imperfections. Based on this study, it can be concluded that Euler's buckling formula can be successfully modified and used in the prediction of the load-carrying capacity of a sandwich panel.

There is a significant number of old metallic bridges with high levels of structural degradation due to their long service period. Fatigue problems are especially important in these structures since the majority of them were not designed taking into account this phenomenon. Several investigations showed that riveted joints are critical details since several fatigue cracks were found in these joints. In this sense, strengthening methodologies need to be studied. The strategy that has been considered a good solution is the implementation of injection bolts to replace faulty rivets. The structural performance of injection bolts has been demonstrated essentially under quasi-static conditions presenting good results. This paper intends to contribute to the scientific knowledge regarding the fatigue behavior of connections with preloaded injection bolts in the context of a bridge strengthening scenario. An experimental investigation was conducted to compare the fatigue performance of connections with preloaded injection bolts and preloaded standard bolts. Single and double shear connections were tested. New S–N design curves were proposed based on a statistical analysis of the results and compared with the S–N curves proposed in EC3-1–9. The obtained results showed that the use of injection bolts lead to lower scatter and improvement of fatigue life. It was verified that the Eurocode 3 is not able to represent the fatigue strength of connections whose performance is influenced by old metallic materials. Additionally, the fatigue behavior of these connections was assessed by numerical analysis. The relevance of the fatigue crack initiation was evident.

Nominal stresses have been used for a long time for the assessment of fatigue resistance of welded joints, however, this approach has strong limitations since the definition of the nominal stress may be subjective for complex welded details and/or complex loading. On the other hand, the hot-spot stress approach has been proposed to overcome these limitations considering the structural geometrical discontinuities. However, the hot-spot stress methods also present certain limitations, and the present study aims at evaluating the available numerical and analytical hot-spot stress methods proposed by DNVGL (2016) and IIW (2014). The particular case of an offshore tubular KT joint has been considered herein and discretized in two planes. It has been studied numerically using the ABAQUS software coupled with the hot-spot stress extrapolation methods described in IIW (2014) and DNVGL (2016). The influence of the weld geometry has been considered and evaluated. In addition to the numerical method, the present study has also considered the analytical approach proposed in DNVGL (2016) derived from the combination of Efthymiou solutions for the stress concentration factor with the method of superposition of stresses. The numerical models according to IIW (2014) have been found to be more conservative when compared with the mesh-size methods proposed by DNVGL (2016), both in numerical modelling without the weld or with weld. For the numerical models with weld cord, the mean values of normalized difference index obtained for all braces together, as a result of comparing numerical results with analytical solutions, are lower, when compared with results obtained from the numerical models without weld cord.

In this work, three classes of fatigue models are reviewed according to the fatigue regimes commonly considered in the current components design. Particular attention is devoted to the so-called Class III fatigue models, covering the three fatigue regimes, namely, LCF, HCF and VHCF. The applicability and limitations of the proposed analytical sigmoidal solutions are discussed from the viewpoint of practical design. The compatible Weibull S-N model by Castillo and Canteli is revisited and improved by considering a new reference parameter GP = E·σ_M ·(dε/dσ)|_M as the driving force alternative to the conventional stress range. In this way, the requirement, σ_M ≤ σ_u, according to the real experimental conditions, is fulfilled and the parametric limit number of cycles, N₀, recovers its meaning. The probabilistic definition of the model on the HCF and VHCF regimes is maintained and extended to the LCF regime. The strain gradients may be calculated from the monotonic or cyclic stress–strain curve of the material although a direct derivation from the hysteresis loop is recommended. Some Class III fatigue models from the literature and another one improved by the authors are applied to the assessment of one experimental campaign under different stress ratios conditions and the results compared accordingly. Finally, the new probabilistic GP-N field is evaluated. The results confirm the practical confluence of the stress- and the strain-based approaches into a single and advantageous unified methodology.