Search

In the incremental launching method employed for steel bridge construction, the girder is subjected to patch loading which occurs at the piers’ position. This loading significantly affects the girder resistance in the construction stage. Therefore, prediction of the girder resistance under this loading is important. This paper proposes a new approach for predicting the patch load resistance of stiffened plate girders using an extreme gradient boosting algorithm (XGBoost). A total of 170 experimental data on stiffened plate girders under patch loading collected from the literature serves as the training and testing data to build the predictive model. To demonstrate the efficiency of the proposed model, its predictions were compared with those obtained from other Machine Learning (ML) methods such as support vector machines (SVM), decision tree (DT), random forest (RF), adaptive boost (Adaboost), and deep learning (DL). The accuracy of the proposed model was validated against the existing equations taken from the design standards (EN-1993-1-5 and BS 5400) as well as existing formulae in the literature. The comparative results reveal that the proposed model provides better and more accurate predictions than the existing formulae.

This paper aims to derive new simple and consistently reliable residual strength equations for predicting the residual ultimate strength of steel dented stiffened cylinders under combined axial compression and radial pressure including hydrostatic pressure. For this, two types of the stiffened cylinders, namely, stringer- and/or ring-stiffened cylinders, which are generally main components of offshore marine structures used in semi-submersibles, submarines, tension legs of platforms (TLPs), and other various types of floating offshore structures, are investigated. Firstly, the various collision damage scenarios which illustrated the collision accidents of offshore stiffened cylinders with supply vessels or floating subjects, were analysed. Then, a finite element modelling of the stiffened cylinders was developed using a commercial software package ABAQUS/Explicit after being validated against available test results. Next, rigorous parametric studies on full-scaled design examples of stiffened cylinders were carried out by changing the design variables. Based on the rigorous case of the study results, new simple design equations, which are used to predict the ultimate residual strength of the damaged stiffened cylinders under combined loads are proposed through a regression study. Finally, the accuracy and reliability of the proposed equations have been compared with available test data and a nonlinear finite element analysis (NFEA). These equations can be useful for the initial design stage of stiffened cylinders under the risk conditions of offshore structures in the case of collision incidents.

This paper aims at investigating the optimum positions and the required flexural rigidity (γrq) of a single and two longitudinal web stiffeners of unsymmetric plate girders subjected to pure bending by using gradient-based optimization algorithms. The optimization procedure is performed into two steps: The first step is to maximize the bend-buckling coefficient kb generated from eigenvalue buckling analyses, and then the second step aims to minimize the flexural rigidity of the stiffeners (γ). This procedure is implemented using the Abaqus2Matlab toolbox which allows for the transfer of data between Matlab and Abaqus and vice versa. Based on this research, the optimum locations of the stiffeners are recommended for the unsymmetric plate girders under pure bending. Additionally, equations are suggested to determine γrq of the longitudinal web stiffeners of the unsymmetric plate girders. These proposed equations are more realistic in the practical designs than those recommended by AASHTO LRFD specifications and several works in the literature in which γrq of the longitudinal web stiffeners is determined based on the simply supported plates but not the unsymmetric plate girders under bending.