Abstract
The present study proposes a multi-objective optimization method to maximize the fundamental frequencies and minimizing production cost for the fibrous composite plate with curvilinear fibers. The production cost is represented by the average curvature of the curvilinear fibers. Curvilinear fiber shapes are defined by the projections of contour lines for a cubic polynomial surface, and the coefficients of the polynomial are employed as design variables. A finite element code with isoparametric 8-node rectangular element based on the first order shear deformation theory is employed to exploit the frequencies of plates. An improved non-dominated sorting genetic algorithm (NSGA-II) based on the non-dominated sorting and crowded distance metric sorting is used as an optimizer with a mutation whose probability is controlled to maintain the variety of individuals and to keep searching ability strong. The numerical results showed that the present optimization method resulted in the widely distributed Pareto-optimum front ranging from parallel fibers to largely curved fibers.
