The purpose of this project is to develop a motorcycle rear swing arm using finite element analysis and response optimization. This paper aims to discuss the specific features, benefits, and precautions when using design optimization to develop a specific project. Design Optimization has been an evolving process for many years. The latest versions of finite element software allow users to develop, analyze, and optimize structural designs within one program quickly and efficiently. A single shock absorber mounted close to the chassis and centrally located was the design selected to be analyzed. The design was selected for use in a variety of motorcycle types. This project consisted of a unique set of design attributes that were ideal to exemplify design optimization techniques. Static structural models were created to refine the design before using response optimization. These models finalized the material selection and initial sizes. A central composite design type was generated with selected boundary conditions for four parametric dimensions of the model. The ideal design of this component would include the resulting stress below a safe allowable value, minimal deflection, and the least amount of weight. It is evident that these three parameters will oppose one another as geometry is changed. Conceptually, an ideal candidate can be created that is a balance of the three parameters. Using the parameters of the selected candidate, a new model was generated for analysis. The final model was further refined by removing unnecessary material that was identified in the structural models.
The first step in a thorough optimization is generating an appropriate amount of design parameter values that are an acceptable representation of all the possible outcomes, or design of experiments (DOE). The DOE tool used to generate the parameters in this project was central composite design (CCD), since it is the most appropriate for second order response models . The second order relationship was confirmed using a trade-off plot. The two level, four input parameter DOE produces twenty five potential candidates that were refined using response surface. The response surface method used in this design process [2,3] to make judgement calls on the final design is examined during this development, and proves useful. Initial static structural models are created and used to set up the model for optimization. Material selection was also accomplished in this phase of development. This process aides in the overall design process by identifying areas of concern as well as the range of parameters that will be analyzed. Multiple acceptable candidates were selected through the use of the optimization tools and a final candidate was selected based on the output of the design attributes and the values of the corresponding parametric geometry. The final selection was also made with the consideration for cost, ease of fabrication, and standardization.