Stacking sequence optimization of composite laminates for maximum fundamental frequency using Bayesian optimization computational framework

Abstract

A computational framework combined with the commercial finite element software Abaqus and Bayesian optimization algorithm is proposed. The proposed computational framework leverages the Gaussian process based-probabilistic capability in Bayesian optimization as a surrogate model for minimizing the computational cost of objective function evaluations in the finite element analysis. The optimization problem in this work is to enhance the maximum fundamental frequency of the composite laminates, which is one of the critical parameters in the design for composite structures. The optimization of stacking sequence selection is investigated and validated by the results obtained from literatures. The effectiveness and efficiency of the proposed Bayesian optimization computational framework in maximizing the fundamental frequency of composite laminates are demonstrated by comparing the optimized results from different optimization techniques through a series of cases including the 8-layer rectangular plates with 11 different boundary conditions and the 10 to 20-layer trapezoidal plates with the same boundary condition. The proposed framework developed in this work has highly potential as an engineering tool to address a broader range of structural optimization in vibration problems.