Operating cycle representations for road vehicles

Abstract: This thesis discusses different ways to represent road transport operations mathematically. The intention is to make more realistic predictions of longitudinal performance measures for road vehicles, such as the CO2 emissions. It is argued that a driver and vehicle independent description of relevant transport operations increase the chance that a predicted measure later coincides with the actual measure from the vehicle in its real-world application. This allows for fair comparisons between vehicle designs and, by extension, effective product development. Three different levels of representation are introduced, each with its own purpose and application. The first representation, called the bird's eye view, is a broad, high-level description with few details. It can be used to give a rough picture of the collection of all transport operations that a vehicle executes during its lifetime. It is primarily useful as a classification system to compare different applications and assess their similarity. The second representation, called the stochastic operating cycle (sOC) format, is a statistical, mid-level description with a moderate amount of detail. It can be used to give a comprehensive statistical picture of transport operations, either individually or as a collection. It is primarily useful to measure and reproduce variation in operating conditions, as it describes the physical properties of the road as stochastic processes subject to a hierarchical structure. The third representation, called the deterministic operating cycle (dOC) format, is a physical, low-level description with a great amount of detail. It describes individual operations and contains information about the road, the weather, the traffic and the mission. It is primarily useful as input to dynamic simulations of longitudinal vehicle dynamics. Furthermore, it is discussed how to build a modular, dynamic simulation model that can use data from the dOC format to predict energy usage. At the top level, the complete model has individual modules for the operating cycle, the driver and the vehicle. These share information only through the same interfaces as in reality but have no components in common otherwise and can therefore be modelled separately. Implementations are briefly presented for each module, after which the complete model is showcased in a numerical example. The thesis ends with a discussion, some conclusions, and an outlook on possible ways to continue.