Optimum design of box girder bridges with metaheuristic algorithms

Composite systems offer economical and safe solutions by combining the advantages of concrete and steel materials. Composite box girder bridges are systems consisting of rectangular or trapezoidal box girder steel beams and reinforced concrete slabs. The advantages of composite box girder bridges such as being safe, economical and aesthetic are effective in choosing this bridge type for the thesis study. Choosing the appropriate optimization technique is a very important issue in solving optimization problems. Building systems are often large and complex problems, and also there are so many solution options, therefore these problems have too many local minimum points. For the most appropriate design, the global minimum point must be selected from among all minimum points. It is one of the advantages of metaheuristic algorithm methods that the metaheuristic algorithm methods reach the global minimum point without getting stuck in the local minimum points and that the problems can be solved quickly and easily with these methods. In this thesis, optimum design of 6 reinforced concrete-steel composite box-girder bridges with different spans was made. Before the optimum design process, the bridges were designed with the finite element model and the maximum loads affecting the bridges were obtained. Optimum design of these bridges was made with four different metaheuristic algorithm methods (Artificial Bee Colony Algorithm, Harmony Search Algorithm, Particle Swarm Optimization and Grey Wolf Optimization). In the optimum design problem, it is aimed to keep the amount of steel material used to minimum. While finding the optimum value in the design, the objective function must satisfy the stress, deflection and geometric constraints. In order to compare the classical design and optimum design methods, a three-span bridge was designed and the amount of savings obtained from the steel material was determined.