Search Results

A simulation framework with a validated system model capable of estimating fuel consumption is a valuable tool in analysis and design of the hybrid vehicles. In particular, the framework can be used for (1) benchmarking the fuel economy achievable from alternate hybrid powertrain technologies, (2) investigating sensitivity of fuel savings with respect to design parameters (for example, component sizing), and (3) evaluating the performance of various supervisory control algorithms for energy management. Presenter Chinmaya Patil, Eaton Corporation

A simulation framework with a validated system model capable of estimating fuel consumption is a valuable tool in analysis and design of the hybrid vehicles. In particular, the framework can be used for (1) benchmarking the fuel economy achievable from alternate hybrid powertrain technologies, (2) investigating sensitivity of fuel savings with respect to design parameters (for example, component sizing), and (3) evaluating the performance of various supervisory control algorithms for energy management. This paper describes such a simulation framework that can be used to predict fuel economy of series hydraulic hybrid vehicle for any specified driver demand schedule (drive cycle), developed in MATLAB/Simulink. The key components of the series hydraulic hybrid vehicle are modeled using a combination of first principles and empirical data. A simplified driver model is included to follow the specified drive cycle.

The approach towards building hybrid vehicles has evolved with time and requirements. What used to be direct prototype building activity has moved towards building mathematical models before the actual prototypes are built. These models are utilized in optimizing component sizes, design and calibrate controllers and to estimate fuel economy improvements. If model results show promise, the actual prototype building activity is started. But modeling of vehicles still has a long way to go before aligning with businesses and aiding them as decision making tools on R&D investments. The reason being - the model building activity itself is prolonged and expensive. In addition to this, a lot of proprietary information such as component efficiency maps is required in order to build the model. In absence of these component level data, extensive testing becomes necessary where again vast amounts of resources have to be allocated.