Vehicle simulation models are essential throughout the development process in the automotive industry. The benefit starts when benchmarking, continues when target setting and component selection and permits model-based development of controllers and strategies to ease the calibration of the vehicle.
This paper studies the suitability of different vehicle performance and consumption simulation methodologies based on longitudinal dynamics for the variety of applications on vehicle development. These methodologies can be applied to architectures ranging from quadricycles to trucks and from combustion to hybrid. The main difference between methodologies is the solver, which influences the results and the area of application. The two main trends, namely forward and backward simulation, have features that make them not suitable for all the applications. Consequently, simulation methods that combine the virtues of both for a specific objective or to a wider field of application appear in literature.
This paper demonstrates that the combined methodology developed by IDIADA for the software vemSim (Vehicle Energy Management SIMulator) in Matlab/Simulink environment performs properly in all the applications during the development process. The base of the method is a solver that combines the advantages of backward and forward methods to cover the different application cases (correlation, calibration, target setting, controllers development…) with a single method. The advantage over forward is that the solver itself does not require a driver model that affects results, but it can also work together with driver models. The paper compares the different methods based on the ability to describe a target velocity profile and simulate dynamic phenomena in a reference model.