Optimization of a Steel Floor Framing System Using Evolver

摘要

This paper examines the application of Evolver, a genetic algorithm (GA) solving program, in a three parameters optimization of a steel truss floor system with a concrete slab decking. The floor system is comprised of truss girders supporting beams running in a direction perpendicular to the truss girders with a composite floor deck along the top. Using Evolver three parameters are optimized for two truss girder topologies in order to find the least cost floor system. The weight of the structural members is correlated to the expense of material, labor, and equipment required for the construction of the floor by Means Building Construction Cost Data and interviews with some steel fabricators. This procedure can be modified to optimize the cost of any floor area that may use different truss girder topologies, beam sections, and/or connections. Parametric optimization is defined as the combination of configuration, size, and topology optimization of a truss girder, the size optimization of beams, and the spacing of both beams and truss girders in the system. A discrete set of values has been selected for each variable that make up the search space from which all solutions to the problem exist. A GA is incorporated to quickly explore a wide range of answers and focus on better areas of the search space to find the global optimum.