Variable stiffness optimization algorithm for vibration optimization of variable-stiffness composite plates

Abstract

A novel variable stiffness optimization (VSO) algorithm is proposed for vibration optimization of variable-stiffness composite (VSC) plates uisng linear variation fiber path function. Firstly, an independent 2D-sampling optimization method (2DSOind) is employed to generate a good initial point. Then, optimization is conducted in multiple phases, within each phase, the problem is simplified to an approximately one-dimensional linear search problem. By dividing the fiber angles into two groups, VSO algorithm designs the laminate between two fiber angle groups sequentially and iteratively, making the stiffness of VSC plate stiffened part by part. Lastly, the VSC plate is redesigned accounting for the curvature constraint. The sequential permutation search (SPS) algorithm is coupled into VSO to optimize one group of fiber angles whilst the angles of another group are fixed during the iteration. Under a variety of boundary conditions, the free vibration frequencies of VSC plates are optimized using VSO. In comparison to conventional layerwise optimization (LO) and genetic algorithm (GA), the present results demonstrate much higher efficiency and similar robustness. Moreover, the maximum fundamental frequency of VSC plates can be improved by about 10% in comparison to straight fiber laminates.