LAUSR

Abstract We consider a sandwich plate with face sheets made of unidirectional fiber-reinforced composite with fibers being either glass, or carbon or aramid and the core made of balsa wood loaded by a blast pressure, and find optimal geometries and materials for maximizing the first failure load. While analyzing the problem, we assume that the areal density is fixed and use the Nest-Site Selection optimization algorithm, a third-order shear and normal deformable plate theory, a one-step stress recovery scheme, and the Tsai-Wu failure criterion. We also delineate the effect on the first failure load of inertia forces and uncertainties in values of various parameters. For a sandwich plate optimally designed for the first failure load, we find the ultimate load by progressively degrading elasticities of failed elements. We find that the optimal single-core sandwich designs are symmetric about the mid-surface with thick face sheets and the optimal two-core sandwich designs have a thin middle face sheet, and thick top and bottom face sheets. The first failure load of the optimal clamped single-core (two-core) design is approximately 20% (30%) more than that of the corresponding simply-supported plate. For simply-supported (clamped) sandwich structures, the failure initiates at the centroid (center of the clamped edge) of either the top or the bottom surfaces. It is also found that the first failure occurs in a face sheet (core) due to the in-plane transverse axial stress (transverse shear stress) exceeding its critical value. The collapse load of a clamped (simply-supported) sandwich structure is approximately 15%–30% (0%–17%) higher than the first failure load. The maximum deflection of the collapsed structure may be in a direction opposite to that of the applied load. A novelty of the work is in considering inertia effects in ascertaining the first and the ultimate failure loads and quantifying the benefits, if any, of two-core over one-core sandwich structures, and determining effects of uncertainties in values of parameters.