Online power management with embedded optimization for a multi-source hybrid with real time applications
The focus of this thesis is to develop a suitable power management optimization strategy for a three-source hybrid vehicle powertrain. This strategy takes into account the integration of optimized parameters that limit the battery and fuel cell current by utilizing a third power source, namely supercapacitor. The goal is to develop a modular structure with decoupled online and offline parts such that implementation in case of real driving conditions is feasible. Based on the literature review it can be concluded that providing optimal solutions in terms of multiple objectives online is an issue. Adaption of optimized control strategy to real driving data is another concern. The developed strategy employs an online rule-based control with embedded offline-optimized parameters. The parameters are optimized with respect to multiple and conflicting objectives such as fuel consumption and state-of-charge deviation minimization. By a suitable selection of parameters, operation of all three sources within desired working ranges is possible, keeping in mind the load demand. By varying the weights between the objectives, one or more objectives can be given more priority than others. The application of this concept to fuel cell-battery-supercapacitor hybrid is discussed in this thesis. Detailed modeling of all components along with verification and plausibility assessment is done. For the purpose of experimental validation, the real powertrain components are replaced by controllable power sources and sinks that emulate the dynamics of real components. Finally, a brief concept is presented to integrate the developed power management optimization in real driving scenarios. For validation/verification purposes, a driving simulator environment is connected to the experimental hybrid electric vehicle set-up and with the help of an illustrative example, the desired predicted optimal values are calculated online and displayed to the human driver by a suitable interface. The absence of online tuning of controller parameters in this example is counteracted by developing a concept based on literature. With the help of this concept, the adaption of the power management control concept, developed in this thesis, can be realized.
Use and reproduction:This work may be used under a
Creative Commons Attribution - NonCommercial 4.0 License (CC BY-NC 4.0)