Multiple objective optimization design of hybrid composite structures considering multiple-scale uncertainties

Abstract

A multi-objective uncertainty optimization design methodology of hybrid composite structures considering multiple-scale uncertainties is presented in this paper. The hybrid composite structures comprise of multiple materials with different types and volume ratios of matrix and fibers. The objective is to maximize the first-order natural frequency and frequency gap with the cost constraint, the micro-scale material parameters of matrix and fibers are considered as uncertain but bounded variables, and the optimization variables are stacking sequence and material patches (material selection in every patch of composite structures). Firstly, the uncertainty propagation analysis is implemented based on representative volume element method, and the macro material uncertainties are quantified based on neural network model. Furthermore, the multiple objective robust optimization design function is constructed based on the uncertainty analysis results of first-order natural frequency, frequency gap, and the total cost. An adaptive nondominated sorting genetic algorithm II (NSGA-II) method is proposed to optimize the stacking sequence and material patches simultaneously. The engineering examples of 9-layer laminated plate and conical cylindrical shell show that the proposed methodology can make full use of hybrid composite structures to improve the vibration characteristics while maintaining the low material cost.