抄録
The competition in today vehicle industry demands a fast design process with the best solution for the resources available. The best designs can be achieved by using optimization techniques developed in the last decades, in conjunction with multibody tools. The use of multibody tools, with flexible bodies in particular, can predict the performance of vehicle dynamic systems in an efficient and accurate procedure. Multibody design optimization has a great variety of design parameters, ranging from the bodies geometric dimensions, position of joints or structural characteristics of the body itself or of the connection components. The different types of cost functions required, involve various forms of mathematical functions, not easy to define, ranging from max values, peaks of dynamic responses, or accumulated time responses, represented by time integrals. The methodology proposed here for flexible multibody models allows for the representation of complex shaped bodies using general finite element descriptions, which deform during the dynamic loading of the system, while the gross rigid body motion of these bodies is still captured using body fixed coordinate frames. The vehicle optimum design is achieved through the use of an optimization algorithm with finite differences sensitivities, with vehicle components as design variables and constraints are imposed. The ride optimization is achieved by defining a metric that accounts for the acceleration in several key points in the vehicle and attributing weights for each one. Simulations with different road profiles are performed for different speeds, to account for diverse ride situations. The results are presented and discussed in view of the different methodologies with emphasis on models and algorithms used.
