OPTIMUM DESIGN OF PRESTRESSED CONCRETE GIRDERS USING GENETIC ALGORITHMS

Abstract

Use of Prestressed concrete girders is gaining popularity in Bridge Engineering fraternity because of its low initial cost, minimal maintenance cost, durability, aesthetic appearance, structural efficiency and fire resistance. Prestressed concrete bridges have become increasingly popular, and presently comprise about two-thirds of all bridges with spans between 18 and 36m. The simply supported girders are quite suitable for both short and medium span bridges. In order to assist the designers optimization design procedure has been developed by coupling the design programming in FORTRAN language with the Standard GA tool. The design variables considered herein are the width of the top flange (bfl), depth of the top flange (dfl), width of the bottom flange (bf2), depth of bottom flange (df2), thickness of the web (bw) and overall depth of the girder (d). Minimization of mass of the structure is the main objective here after satisfying all the constraints like stresses at transfer stage and working stage, minimum sectional modulus, shear strength and loss ratio. In order to study the behavior of the optimized section, spans of 15m, 20m, 25m, 30m and 35m have been considered. And also the various type of loading like Class AA, Class 70R, and Class A -- 2 lane have been considered. These girders were analyzed, using the genetic algorithm tool, to evaluate the optimum weight of the girder.