Integrated Design Optimization Of Structure And Material Of Braided Composites

During the traditional design of composite structures, the mechanical properties of the composites were designed at first, and then the composites were made into complex components. However, for the braided composites, the mechanical properties and the shape of the components were manufactured at the same time. It is difficult to get perfect properties of the braided composite components using the traditional design approach. The microstructure and macrostructure both have large influence on the performance of the composite components. In order to realize the large potential of the braided composites completely, a new optimization design method considering scales in both structure and material should be developed. In this thesis, the mechanical properties prediction approach, microstructure optimization design approach and integrated optimization design approach based on structure and material were investigated.The microstructure model of the two-dimensional plain weave and braided composites were derived based on the microstructure analysis. The sinusoid curve was used to model the undulation of the yarns. The cross section shape was assumed to be cubic B-spline. For the 2.5-dimensional braided composites, the cross section of the warp yarn was assumed to be rectangle, and the undulation perform was modeled by the cubic B-spline and bitangent. The direction and cross section of weft yarn were modeled by beeline and B-spline respectively. The Stiffness Average Method (SAM), Micromechanical Finite Element Method (MFEM), High-fidelity Generalized Method of Cells (HFGMC) and Node Interpolation Cell Method (NICM) were used to predict the composites'mechanical property. The feasibility and validity of these four methods were also investigated.By combining the Level Set Method with HFGMC, transferring the discrete variables optimization into continuous variables optimization, and changing the microstructure topological optimization design into shape optimization design, a new micro-structural optimization design method of multiple phases composite was proposed. To perform the random transformation of topological microstructure, another new composite topological optimization design method was developed. In this method, the microstructure was described by Digital Array, and the cross operation of genetic algorithm was improved. A microstructure optimization method of braided composite was proposed by establishing the relationship between the macro mechanical properties and micro structure parameters using the microstructure models and micromechanical finite element method.The relationship between the microstructure and macrostructure mechanical properties was derived using HFGMC and structural analysis of composite components. An integrated optimization design method of multi-phase composites based upon structure and material was established. By combining the mechanical properties prediction with the components structural analysis, uniting the structure optimization with the material optimization, and taking the microstructure parameters and macrostructure parameters as the design variables, the integrated optimization design method of braided composites based upon structure and material was established.The integrated optimization design of turbine guide vane made of two-dimensional plain weave and 2.5-dimensional angle-interlock braided composites based upon structure and material was investigated. By incorporating the advantages of Complex method's local search capability with the Genetic Algorithm's global search capability, the hybrid Genetic Algorithm method was developed. The Complex method, Genetic Algorithm method and hybrid Genetic Algorithm method were all used to perform the integrated optimization design of the composite turbine guide vane. The effects of stress and displacement constraints on optimization design were also analyzed.The computation accuracy of the analysis results for the two-dimensional plain weave composite turbine vanes using polynomial response surface model and radical basis function (RBF) neural network model was analyzed. An integrated optimization design method of braided composites turbine guide vane based upon structure and material using RBF neural network model was proposed.