Experimental and analytical study of nonlinear bending response of sandwich beams

The development of optimization design procedures has emerged as an essential component primarily to overcome the challenge of high initial construction costs associated with the fiber reinforced polymer (FRP) bridge deck systems. This paper devises a genetic algorithm-based optimization procedure to minimize the weight of FRP web core sandwich bridge deck systems. A Ritz-based simplified analysis method is applied to couple with the proposed optimization procedure as the structural analysis engine. The design procedure involves the parallel optimal search of mixed continuous and discrete parameters––geometrical parameters, sandwich configuration and material architecture––for the whole sandwich structure. A novel data structure is proposed for the genetic representation in a one-gene-one-variable format. Consequently, specific genetic operators are tailored to accommodate the nonstandard data structure. A problem-independent scheme that does not require any penalty parameter is developed to handle the design constraints. The resulting design procedure provides an effective optimal design framework for the preliminary design stage before performing detailed finite element analysis.