Optimization of CFRP prepreg composite rim by using MOGAII Genetic algorithm

Abstract

In the automotive industry, lightweight design is of great importance, as in other industries. Especially in performance vehicles, different material technologies have become widespread with the development of simulation and production techniques. Automotive rims are an example of these structures. Lighter rims can be obtained by using composite materials. This thesis study was carried out in order to determine the most suitable lamina sequence and number that will ensure the strength of the automobile rim under various loads, which is aimed to be developed within the scope of Carre Wheel Design Company's composite wheel development project. In the study, a 19-inch diameter and 8.5-inch wide ten-spoke rim were used as a reference. A plain weave and resin-impregnated fabric prepreg material were used in the composite wheel. The composite wheel was designed according to ETRTO standards using CATIA V5 software and the required reference surfaces for analysis were obtained. The rim is designed with a metallic plate and is inserted in the hub joint area. The finite element model of the rim was modeled with Hypermesh software. Radial, cornering, and torsion loads are applied to the composite wheel model. For the analysis to be carried out using the MSC Nastran solver, the prepreg material card was used as MAT8 in the model, and the definition of composite laminate was made with PCOMP cards. Rim optimization was performed with MOGAII (Multi Objective Genetic Algorithm II) provided by the modeFrontier software. In the optimization study conducted with elitism, ply orientation and ply number were used as input variables. In the multi-objective optimization study, minimization of the maximum deformation and minimization of the total composite mass were performed. The maximum Tsai-Wu failure index value was defined for three different load cases as constraints. A total of 60 generations of evolutionary optimization were carried out with the initial population produced with the Sobol design of experience algorithm. The Pareto front set was created from the designs obtained as a result of 60 generations. The created Pareto set was analyzed with linear multi-criteria decision making and genetic algorithm multi-criteria decision making methods. The analysis of the design with the highest grade was repeated and the results were shown.