| The demands imposed by the optimal design process form a unique set of criteria for the development of a computational model for vehicle simulation. Due to the large number of simulations that must be performed to obtain an optimized design the model must be computationally efficient. A competing criterion is that the computational model must realistically model the vehicle.; Current trends in vehicle simulation codes have tackled the problem of realism by constructing elaborate full vehicle models containing dozens if not hundreds of distinct bodies. Each body in a model of this type is associated with six degrees of freedom. Numerous constraint equations are applied to the bodies to represent the physical connections. While the formulation of the equations is not particularly difficult, and in fact has been automated in several software packages, the resulting model requires a considerable amount of computational time to run. This makes the model unsuitable for the application of computational optimal design techniques.; Past research in the field of vehicle dynamics has produced numerous computational models which are small enough and fast enough to satisfy the speed demands of the optimal design process. These models typically use less than a dozen degrees of freedom to model the vehicle. They do a good job of predicting the general motion of the vehicle and they are useful as design tools but they lack the accuracy required for optimal design.; A model that bridges the gap between these two existing classes of models and is suitable for performing optimal design was developed. The model possesses twenty-eight degrees of freedom and consists of eight bodies which represent the sprung mass, the rear suspension, the left front spindle, the right front spindle, and the four wheels. A driver control algorithm was developed which is capable of driving the car near its handling limits. The NCSU Legends race car was modeled and an attempt was made to optimize the vehicle setup for the Kenley, NC race track. |
