Generating Proper Integrated Dynamic Models for Vehicle Mobility Using a Bond Graph Formulation

This work presents the development, validation, use and reduction of an integrated vehicle model composed of the engine, drivetrain and vehicle dynamics. Integrated models that include all vehicle subsystems are essential for many driving scenarios that induce rich dynamic responses. The model is developed with the assumption that it will be used for a wide range of excitations, and therefore, all possible complexity is included in the model, e.g., drivetrain flexibility and large rigid body motions. The bond graph formulation is used for model development because it facilitates the integration of component/subsystem models, provides the user with physical insight, and allows easy manipulation of models. The engine model is a steady state torque map generated from an engine thermodynamic model. The drivetrain consists of the torque converter, transmission and driveline. A nonlinear planar model of the vehicle is used to predict the dynamics in the longitudinal, heave and pitch degrees of freedom. For illustration, the model is configured for a Class VI, International 4700 series delivery truck, and implemented in the 20SIM modeling and simulation environment. The integrated vehicle simulation is validated against transient data measured on the proving ground. An energy-based model reduction methodology is applied in order to produce proper vehicle models that provide more design insight and can be more computationally efficient. This provides a systematic approach to address the modeling assumptions and generate reduced models that are valid under specific scenarios. A reduced system model is generated, which produces results very similar to the full (baseline) model. In addition to its predictive quality, the utility of the reduced model to study trade-offs involved in redesigning components and control strategies for improved performance of the vehicle system is demonstrated.