Increased understanding of hybrid vehicle design through modeling, simulation, and optimization

Vehicle design is constantly changing and improving due to the technologically driven nature of the automotive industry, particularly in the hybridization and electrification of vehicle drivetrains. Through enhanced design vehicle level design constraints can result in the fulfillment of system level design objectives. These constraints may include improved vehicle fuel economy, all electric range, and component costs which can affect system objectives of increased national energy independence, reduced vehicle and societal emissions, and reduced life-cycle costs. In parallel, as computational power increases the ability to accurately represent systems through analytical models improves. This allows for systems engineering which is commonly quicker and less resource consuming than physical testing. As a systems engineering technique, optimization has shown to obtain superior solutions systematically, in opposition to trial-and-error designs of the past. Through the combination of vehicle models, computer numerical simulation, and optimization, overall vehicle systems design can greatly improve. This study defines a connection between the system level objectives for advanced vehicle design and the component-and vehicle-level design process using a multi-level design and simulation modeling environment. The methods and information pathways for vehicle system models are presented and applied to dynamic simulation. Differing vehicle architecture simulations are subjected to a selection of proven optimization algorithms and design objectives iii such that the performance of the algorithms and vehicle-specific design information and sensitivity is obtained. The necessity of global search optimization and aggregate objective functions are confirmed through exploration of the complex hybrid vehicle design space. Whether the chosen design space is limited to available components or expanded to academic areas, studies can be performed for numerous design objectives and constraints. The combination of design criteria into quantifiable objective functions allows for direct optimization comparison based on any number of design goals. Integrating well-defined objective functions into high performing global optimization search methods provides increased probability of obtaining solutions which represent the most germane designs. Additionally, key interactions between different components in the vehicular system can easily be identified such that ideal directions for gain relative to minimal cost can be achieved. Often times vehicular design processes require lower order representations or consist of time and resource consuming iterations. Through the formulation presented in this study, more details, objectives, and methods become available for comparing advanced vehicles across architectures. The main techniques used for setting up the models, simulations and optimizations are discussed along with results of test runs based on chosen …