Reducing the squeak and rattle risk by improving the dynamic response and geometric variation in an assembly using topometry optimisation

Karl Lindkvist and Minh Tang (Lund University and Volvo Cars)

Squeak & Rattle are two types of undesired noise that occur in cars when components come in contact with each other, either impacting or sliding. It reduces the sense of quality and comfort for users, which are important aspects for premium cars. Two of the major contributors for generating squeak and rattle are geometric variation and insufficient stiffness, causing induced vibrations. This thesis aims to increase that stiffness by adding thickness to the inner door panel of a car. This is done by using a multi-disciplinary topometry optimisation approach which takes the geometric variation and the dynamic response into account. The optimisation begins with an initial sensitivity analysis in order to apply the thickness to the most influential areas of the model. Topometry optimisation is then performed in multiple stages with constraints set to create a pattern of optimal thickened areas. Each stage of the optimisation uses a multi-objective genetic algorithm with an initial design of experiments population. Due to the time it takes to optimise a detailed model of the inner door panel, the development of the methodology is done by applying the optimisation on a simplified model. The adding of thickness to a component changes the modal behaviour, resulting in a very sensitive optimisation process. However, the approach proved to be successful for both the simplified and complete model with a reduction of both of the optimisation objectives. The resulting thickness differed between the models, although for both models it had better effect on the objectives to increase the global stiffness rather than increasing the stiffness locally, near the fasteners