HOMOTOPY OPTIMIZATION FOR THE COMPRESSION OF A MULTIBODY MODEL OF A VEHICLE

Abstract

For tasks like design optimization, sensitivity analysis, parameter identification, and controller tuning, which can require hundreds or thousands of simulations, the reduction of complex multibody dynamic models remains an important topic of investigation, despite the increasing computational power of modern processors. Initially, we create a detailed Adams/Car model of a mass-produced SUV. Then, in Maples, we create front and rear suspension models called single-link equivalent kinematic quarter-cars (SLEKQ, pronounced "sleek"). All suspension connections are combined into one unsprung mass at each corner of the vehicle to reduce the computational complexity of including bush ings or kinematic loops. The high- fidelity Adams model is used to generate lookup tables or polynomial functions that are then used in the SLEKQ models to simulate the kinematic behaviour of a complete suspension model. Each SLEKQ model's capacity to forecast behaviour is contingent on the accuracy with which its nonlinear spring and damper parameters, such as the bushings' stiffness and damping contributions and the unsprung mass, are accounted for. Parameters that allow for the smallest gap between Adams and MapleSim model responses are found through homotropy optimization. Four-post heave and pitch tests are used to verify the dynamic performance of the SLEKQ models integrated into a reduced 10- degree-of-freedom model of the whole vehicle compared to the high-fidelity Adams model.