Study On Power Management Trategies For A Series Hydraulic Ybrid Vehicle

The challenge of improving fuel economy and reducing emission provides strongimpetus for pursuing ultra-efficient vehicle concept. In recent years, fuel consumed bytrucks grows faster, this is a consequence of an increase in the relative number of trucks,as well as higher demand for transportation of goods. As a result, truck systems callfor significantly improved fuel efficiency and hybridization of trucks has become nec-essary. The hydraulic pumps and motors are characterized by high power density andhigh efficiency, in addition, hydraulic accumulators have the ability to accept both highfrequencies and high rate of charging and discharging. These virtues make hydraulic hy-bridization very attractive in large mass associated trucks, thus hydraulic hybrid vehiclesappear to be one of the most viable technologies with significant potential to reduce fuelconsumption. However, the relatively low power density of the hydraulic accumulatorbrings about a unique control challenge, which makes hydraulic hybridization attractedgreat attention in studies.Hydraulic hybridization of vehicle opens up many possibilities related to systemarchitecture, which can be classified into two broad categories, series or parallel. Becausethere is no mechanical connection between the engine and the wheels in series system, fullflexibility in engine operation can be allowed. Furthermore, with the potential of furtherimproving vehicle fuel economy, hydraulic hybrid vehicle with the series architecture(i.e. Series Hydraulic Hybrid Vehicle, SHHV) is the object of this research. In this paper,research was conducted with respect to some deficiencies which currently exit within thecontext of SHHV studies.(1) High fidelity engine model. One of the most common methodologies used withinthe context of hybrid powertrain system models considers a purely static approach to es-timate the engine fueling. A steady state fueling map is used to estimate the engine fuelconsumption at different engine operating conditions. Even though this approach givesa satisfactory estimate of the fueling in steady state conditions, large discrepancies areobserved when the engine operates in transient conditions. The engine model developedin this paper focus on a simplified description of the air handling system dynamics inorder to estimate the real air mass flow rate in transient conditions. Similarly, the engine in-cylinder processes which lead to the torque production are described in a simplifiedway and only the low frequency bandwidth dynamics due to torque request variations aredescribed. The final estimation of the fuel mass flow rate is the result of the combinationof a map-based fuel mass flow rate to which a set of corrections accounting for specificphenomena are fully taken into consideration. The approach mimics the actual operationof an ECU (Electric Control Unite) algorithm. Comparison between engine model simu-lation results and experimental data demonstrated the high fidelity property of this enginemodel.(2) Design and application of hierarchical control architecture in SHHV system. Thehierarchical control architecture developed in this work includes two levels and threesub-module controllers. There is one top level controller, which is the vehicle powermanagement policy. There are two low level sub-module controller, one manages theengine and its associated hydraulic pump/motor, the other one operate the vehicle drivingmotors and the mechanical brakes. This architecture allows all the control variables andspecific phenomena in this SHHV system can properly managed.(3) Application of model predictive control based power management strategy andits comparison with other power management strategies. There is a rich literature on thepower management strategy study, which can be classified into three broad categories,i.e. rule based strategies, strategies with off-line optimization (e.g. deterministic dynamicprogramming, stochastic dynamic programming, etc.) and strategies with on-line opti-mization (i.e. model predictive control based strategies). In this research, three powermanagement strategies that are typical in their respective categories are developed andthoroughly studied. The three different strategies are thermostatic SoC (of rule based cat-egory, denoted with RB), stochastic dynamic programming based (of the category withoff-line optimization, denoted with SDP) and model predictive control based (of the cat-egory with on-line optimization, denoted with MPC). In particular, currently no studyemploying MPC in SHHV exits, and this work is the first instance. Simulation resultson the federal urban driving schedule demonstrate that MPC>SDP>RB (here">" meas"better than on fuel economy improvement and power performance"), and more precisely,MPC can significantly expand the operational envelope of SHHV.(4) Real time optimization. One main disadvantage of MPC for application is thatMPC requires on-line real time optimization, which is computational expensive. To over- come this problem, a Lyapunov based dynamic optimization approach is proposed inthis study. Compared with conventional real time optimization methods (typically thequadratic programming approach), this Lyapunov based approach converges faster andneeds fewer computation time. More importantly, the computation time of the Lyapunovbased approach increases linearly with the number of optimization variables. Further-more, its computing stability can be as good as that of quadratic programming.