Minimum Weight Design of Transversely Stiffened Plate Girder Using Genetic Algorithm

Abstract

Fundamental advantage of using transversely stiffened plate girders (TSPG) rather than plane web plate girders is in achieving increased strength and stiffness of the web panel. As these girders are made of steel plates, they can be custom designed to suit consumer's requirements. Weight minimization of TSPG by satisfying strength and serviceability conditions, further enables in achieving higher strength to weight ratios. The utilisation of these girders in the building sector in India is growing, and optimising these girders can lead to more cost-effective and structurally sound designs. No significant research has been carried out in the field of optimization of plate girders based on specifications according to IS 800 (2007): General Construction in Steel - Code of Practice (Bureau of Indian Standards). Therefore, in the present study, the main goal is to develop a formulation for optimum weight design of TSPG using genetic algorithm (GA) with weight minimization as objective function. The breadth of flange (bf), thickness of flange (t_f), depth of web (d), thickness of web (t_w), breadth of stiffener (b_s), thickness of stiffener (t_s) and spacing between stiffeners(c) are adopted as the design variables. The optimum girder weight is compared with weight obtained through conventional design using IS 800 (2007). The results demonstrate that conventional design overestimates the girder weight by about 30% on an average and therefore the developed formulation helps in attaining economical girder weight. The viability of the GA formulation is verified by comparing the optimized results with existing experimental and numerical data from literature. Parametric studies are also carried out to understand the behaviour of design variables. Results from these studies show that deeper web reduces flange thickness but requires thicker web in order to ensure ample shear capacity and to prevent shear buckling.