In previous work, a serial hybrid powertrain concept with a phlegmatised ICE has been described. Drivability is to be ensured through an innovative predictive operating strategy. Battery State-of-Charge (SoC) is controlled using a backend-based prediction of energy consumption on a given route based on road map and traffic data. In this paper, a spotlight is thrown on the proposed control architecture. On the top level of the controller, a Dynamic Programming algorithm finds an optimal reference trajectory for the SoC over a known route with the goal of avoiding certain Worst-Case scenarios commonly associated with the serial hybrid powertrain topology. Close adherence to the reference trajectory is ensured on a lower level through Model Predictive Control, taking into account additional factors such as battery stress. These control layers closely represent the map data distributed on the on-board bus network of state-of-the-art road vehicles under the current ADASIS standard. The necessary input data for the proposed controller is therefore available at no extra cost or engineering effort to OEMs. A simulation framework based on Matlab/Simulink and AVL CruiseM enables testing of the operating strategy using high-quality, open-source map data. Thus, the viability of the proposed control architecture is demonstrated in a selection of challenging driving scenarios on real-road speed and gradient profiles. It is shown that this quite basic prediction algorithm outperforms classical, non-predictive serial hybrid operating strategies in terms of drivability. Thus, systematic optimisation of the ICE towards high efficiency and low emissions is enabled, reducing requirements for transient behavior and high power density. Potential for future development, especially further improvements of efficiency and emissions behavior of the ICE through predictive thermal management, is also elucidated.
Session:
HYBRID POWERTRAIN
|
| 17:00 - 17:30