Optimum design of welded connections of steel structures using metaheuristic methods

In this thesis, welded connections of steel structures were optimized by using metaheuristic artificial intelligence algorithms. Three different numerical models were created from the fillet welded connection chosen to be used in the study, according to the loading conditions. These models were coded separately for Load and Resistance Factor Design (LRFD) and Allowable Strength Design (ASD) principles in ÇYTHYE-2018. The coded numerical models were optimized for three different population sizes with Jaya, TLBO and Sine-Cosine algorithms written in MATLAB language. The minimization of the total welding area were chosen as the objective function in optimizations. The results of 54 analyzes obtained for different scenarios such as loading status, design principle, population size and solution algorithm were compared for the relevant parameters. Furthermore, the "Fair Penalty Method", which is a unique penalty method, was developed to be used in the constrained optimization method. As a result, it is understood that all of the algorithms used in the study were suitable for the optimization of welded connections of steel structures. The advantages and disadvantages of algorithms were understood by comparing them with each other. The fair penalty method proposed in this study was found to be usable for constrained optimization problems.