OPTIMUM DESIGN OF COMPOSITE FLOOR SYSTEMS

Abstract

Steel-concrete composite floors are a structurally efficient combination of constituents as they exploit the tensile resistance ofthe steel and compressive resistance of the concrete in an effective manner. The current design procedures and horizontal shear strength of the composite slabs are assessed by the expensive full scale experiments. Hence, the design automation ofthe composite floor is complicated. The main objective of this research programme is to develop an optimum design procedure for composite floor systems to tackle a large number ofdesign variables and constraints effectively at various levels by using the Genetic Algorithm (GA) and Artificial Neural Network (ANN). The optimization procedure based onGA for the composite slab was developed in the two stages. In the construction stage and composite stage, the minimum mass design of the profiled deck and the cost of the composite slabs were considered as the objective function respectively. The design studies were carried out by varying the span, width and design loads. Finally, the design ofthe steel beam was incorporated for the system level optimization. The ANNs were incorporated in the genetic algorithm to assess the horizontal shear strength of the composite slab. Two ANN had been developed by using the available m-k values for the existing profiles collected from the manufacturers and slip block test results. The trained neural network predicted the m-k values or ru values of the composite slab. Further, in the experimental work, 96 specimens were prepared for conducting the small scale slip block test. These specimens were tested for assessing the horizontal shear strength of the composite slab. These values were used to develop a simplified approach for the design of composite slabs, with a simple calculation model to obtain the moment of resistance based on the partial interaction method. These data were used to train ANN. In this research, the optimization tool was developed for the design of composite floor system and the problems related to horizontal shear strength was over ridded by developing the ANN and slip block test. Thus, the dependency on the many expensive and time consuming full-scale experiments would be minimized. 11