B4_MichaelStapelbroek_Paper_English

Title: “Hybrid Operation Strategy of a new Axle-Split Hybrid” Authors: Dr.-Ing. Michael Stapelbroek (FEV GmbH, Aachen, Germany) Dipl.-Ing. Hans-Peter Lahey (FEV GmbH, Aachen, Germany) Dipl.-Ing. Jürgen Ogrzewalla (FEV GmbH, Aachen, Germany) Contact: Michael.Stapelbroek@fev.com Herta.Dumitru@fev.com Subject Group: Propulsion/ Powertrain Key Words: Operation strategy, Optimization, Hybrid Electric Vehicle, Axle-Split-Hybrid, HEV, Discrete Dynamic Programming, DDP Abstract Hybrid electric vehicles offer the potential to minimize fuel consumption and emissions compared to conventional vehicles. Moreover Plug-In Hybrid electric vehicles offer the advantage of local emission free driving while overcoming the disadvantage of a short pure electric operating range and the need of a time-consuming battery charge procedure. All hybrid propulsion concepts use the additional degree of freedom – a second energy converter – to improve the powertrains efficiency in all operation points. Consequently, the operational strategy of those vehicles must be adapted to the corresponding concept to meet conflicting demands of fuel or energy consumption, engine emissions and acoustics as well as driving comfort in the most optimal way. In this area FEV uses a continuous development process to determine the concept layout of a hybrid powertrain and further for the design of hybrid operating strategies from concept to series production vehicle. With specific powertrain simulations and numerical optimization methods (De-sign of Experiments, DoE and discrete dynamic programming, DDP) the optimal hybrid power-train as well as the globally optimal operational strategy can be determined. This enables FEV to find an optimal solution within the range of available components, the customer vehicle targets and fuel consumption benefit. The results will be used as reference for the design and implementation of a rule-based operation strategy meeting fuel consumption benefits as well as soft objectives like NVH and drivers’ comfort requirements. This paper describes the development process for an exemplary axle-split hybrid concept. Beginning with an optimization of all hybrid powertrain components parameters the best powertrain layout was found. Several boundaries like the limited brake energy recuperation of the rear axle electric motor, engine warm up and catalyst heating, increased transmission friction due to low transmission oil temperature, the required 12V board net power and efficiencies of battery, electric motors, engine and transmission were considered. Further the control strategy model of the HEV includes hybrid functions like engine start/stop control, e-motor boosting, load point shift, range extender operation, pure electric driving, brake energy recovery, etc. to ensure a realistic simulation result. Due to the optimization method a globally optimal operation strategy was determined offline. Taking these results as a reference a rule- based operation strategy was derived, able to be used online in the vehicle. During vehicle tests the operation strategy could be validated successfully achieving the customer’s fuel consumption benefit targets. The presented approach is demonstrated on basis of a typical hybrid application. Paper Introduction In the last ten years China has become the largest car market in the world with an incredible average growth of 26 % per annum [1]. But the rapid growth of individual mobility has also tightened the CO2 and smog problems in china’s mega cities. In China’s capital Beijing the continuous air pollution is even made responsible for the drop of tourism1 in the first half of 2013 [2]. Furthermore, it has put increasing pressure on the national goals of oil independence and leadership in e-mobility. Realizing these challenges, China introduced two phases of vehicle fuel consumption regulations for light-duty vehicles in 2005, and is planning to implement the third phase by the end of 2015. [1]. Figure 1 - Major Chinese manufacturer corporate-average fuel consumption (CAFC) [1] In Figure 1 the current situation of the major Chinese automotive manufacturers is shown. For each manufacturer the average fleet fuel consumption over average vehicle weight is plotted. The figure can be differentiated between independent and joint venture automakers by its color and sold vehicles by the size of the circles. While Chinese independent automakers sell vehicles with a lower curb weight compared to joint venture automakers their average fuel consumption is not significantly lower. The 2010’s fleet average fuel consumption is 7.7 L/100 km even though China has already decreased its CO2 emissions by 15 % between 2002 and 2010 [3]. Nevertheless, most of the manufacturers have to further reduce their average fleet fuel consumption to reach the preliminary legislative target of 6.9 L/100km (Phase 3 target) in 2015. Moreover, the Chinese Ministry of Industry and Information Technology (MIIT) even aims to lower the target to 5.0 L/100km by oder until 2020. 1 The number of tourists visiting China’s capital fell by more than 14 per cent in the first half of this year compared to last year. This scenario is very similar to the European situation even though the average fleet fuel consumption is lower. The market share of vehicles in the same segment is comparable apart from some differences2 and the industry is being driven by the same incentive: Climat situation, legislative pressure and customer demands. A well-known solution is to substitute conventional vehicles by hybrid and pure electric vehicles to significantly reduce the fleet average fuel consumption. Although electric vehicles are already available on the market and offer the possibility for local emission-free mobility, they still suffer from well-known disadvantages like limited battery capacity, high production costs, long charging times and restricted performance at low and high temperatures. Despite of these drawbacks, automakers worldwide consider hybrid and plug-in hybrid electric vehicles as a promising approach for fuel consumption reduction and bridging technology for future pure electric vehicles. Axle-Split Hybrid Concepts When speaking of hybrid vehicles actually different types of powertrain solutions are meant. Besides the main classification into parallel, serial and mixed-hybrid topologies which are shown in Figure 2, further subclasses and combinations are possible. Figure 2 – Overview of main hybrid powertrains topologies Thus parallel hybrid concepts are divided into speed, torque and traction force addition concepts. Torque addition concepts are widely popular as they provide an easy entry in hybrid technology since they allow reuse of most of the conventional powertrain parts in the hybrid application. Serial hybrids are not further classified although one may distinguish between Range-Extender and Serial hybrid 2 In China almost 20 % of vehicles are so-called Mini Vans that are rarely present on European streets. All other vehicle classes have the same market share in China as well as in Europe. concepts. Mixed-hybrid concepts may be subdivided into combined hybrids as a combination of parallel and serial hybrid and well-known power distribution hybrids like the Toyota Prius Hybrid. Taking into account all hybrid combinations, a huge number of hybrid topologies have to be considered. Furthermore, a classification can be made according to the logical alignment of energy converters3 in the powertrain as shown in Figure 3. This helps to understand the differences between the various concepts more precisely. The main challenge however, to find the optimal powertrain solution in terms of fuel economy, costs and customer benefit, remains. Figure 3- Classification of hybrid topologies Starting with conventional powertrains as a basis, the position of the 2nd energy converter can be divided into four sections. Concepts are identified as P0 hybrids when the electric energy converter is directly attached to the engine. Attaching the energy converter to the crankshaft of the engine, but without the possibility to decouple it from the engine, is known as a P1 hybrid. Concepts with a decoupling mechanism like a clutch but still mounted to the engines crankshaft are identified as P2 hybrids. As the engine can be decoupled they already offer the possibility to drive solely electrically. If the energy converter is linked to the output shaft of the transmission it is identified as P3a hybrid. Finally, when linking the energy converter to the 2nd axle it is identified as P3b hybrid. This paper is focused on axle-split hybrids as a variant of traction force addition concepts. These concepts are distinguished by having linked energy converters on both axles. Therefore their wheel torque or better said traction force is added to drive the vehicle. Related to the classification in Figure 3 these concepts principally are identified as a combination of sections. Therefore P03, P13 or P23 hybrid concepts are known depending on the logical alignment of their first electric energy converter. Advantages and disadvantages of different axle-split hybrid concepts will be discussed later. In Figure 4 the concept differences by their energy converters location is shown. All concepts are already in series production. Volvo for example uses a 7.5 kW belt-driven starter generator (BSG) as primary 3 Nowadays, electric motors are almost exclusively being used as energy converters. electric motor in its V60 plug-in hybrid [4]. Flywheel starter generators in contrast to BSG solutions are mostly used only as 12 V fallback solutions in hybrid vehicles and not as a full starter generator solution. PSA for example uses a redundant starter in their Hybrid4 vehicles [5]. On the right side of Figure 4 an integrated starter generator (ISG) solution by GAC is selected as an example. Integrated starter generator solutions offer the advantage of high power and high torque with the disadvantage of higher integration effort. If an ISG offers high low end torque by taking a permanent magnet synchronous motor (PMSM), for instance a 12 V, fallback solution might be obsolete. In conclusion the selection of a primary electric motor solution has to be done carefully and in relation to the OEM’s vehicle targets. Figure 4 - Different starter generator solutions At the other powertrains end the secondary electric motor can be linked in different ways to the differential as shown in Figure 5. Figure 5 - Different rear-axle concepts In the Lexus RX450h the electric motor is directly linked to the differential [6]. This means the complete vehicles speed range has to be covered by the rear-axle electric drive (ERAD). Normally, this comes along with small low end torque capabilities due to the characteristics of electric motors. In contrast BMW i8 provides a 2-speed transmission at the front axle to cover both, high low end torque and maximum speed range with one electric motor. The disadvantage of torque interruption during electric driving may be compensated by a second electric motor on the rear axle with a decoupled engine. But compared to the Lexus solution a 2-speed transmission adds more complexity to the powertrain. PSA is taking a different approach by adding a dog clutch between the differential and the electric motor [5]. This allows high low end torque and prevents e-motor drag losses but it cannot cover the complete speed range solely electrical. Alternatively, one may use different electric motor technologies (i.e. asynchronous motors) with overspeed capability and without drag losses of a PMSM. Simulation approaches may help to understand the differences between the main and subclasses in greater detail. In this area FEV uses a continuous development process to determine the concept layout of a hybrid powertrain. With detailed powertrain simulations and the use of design of experiments (DoE) approaches the optimal hybrid powertrain can be determined fitting the specific customer requirements. Several boundaries like the limited brake energy recuperation of the rear axle electric motor, engine warm up and catalyst heating, increased transmission friction due to low transmission oil temperature, the required 12V board net power and efficiencies of battery, electric motors, engine and transmission were considered. This enables FEV to find an optimal solution within the range of available components, the customer vehicle targets and fuel consumption benefit. In this paper the development of a hybrid operation strategy of an axle-split hybrid is highlighted. This axle-split hybrids optimal system topology is already given and shown in Figure 6. Figure 6 – Focused axle-split hybrid powertrain The focused axle-split hybrid powertrain consists of a four-cylinder, 1.8L displacement gasoline engine with 94 kW of power and 158 Nm of maximum torque. Compared to the conventional vehicle the 2.0L engine was replaced by the 1.8L engine. The missing power and torque were substituted by the additionally installed electrical power. The integrated starter generator (ISG) with maximum 14 kW of power and 90 Nm of torque is mounted directly to the engine crankshaft. Its main task is to smoothly restart the engine, to recharge the HV battery and to support fast engine acceleration or deceleration during shifting. To ensure safe engine starts even during cold conditions the conventional 12 V starter remains in the powertrain. The conventional 5-speed automated manual transmission (AMT) with its dry launch clutch remains unchanged in the powertrain even if the transmission control unit (TCU) was adapted to the hybrid application`s needs. To overcome the disadvantages of the AMT’s torque interruption during shifting an electric rear-axle motor (ERAD) with 28 kW of power and 150 Nm of torque was mounted to the rear-axle. It is able to cover the complete vehicles speed range with one speed reduction gear. The electrical energy storage is a 310 V and 1.24 kWh Lithium-Ion battery with a maximum of 41 kW discharge power. Hybrid Operation Strategy Development With a given powertrain architecture a hybrid operation strategy has to fulfill further targets besides only fuel consumption reduction. Generally, hybrid operation strategies are focusing on the following main targets:  Reduction of fuel consumption  Compliance to emission regulations  Driving comfort and added customer value  Performance and dynamic vehicle handling While reduction of fuel consumption and compliance to emission regulations are tied to each other, driving comfort and added customer value as well as performance and dynamic vehicle handling are more subjective targets. Electric driving, increased performance by electric boosting or electric All- wheel-drive (AWD) are examples to added customer values. Additional comfort can be realized for instance by smoother gear shifting or engine starting. Looking at fuel consumption reduction hybrid electric vehicles offer additional degrees of freedom compared to conventional vehicles and even more there are further boundaries by the emission regulations. To achieve all targets the following parameters have to be varied for each step:  Decision on the hybrid operation mode like E-Drive, serial hybrid or parallel hybrid driving  Selection of the right transmission gear ratio,  Torque split of all energy converters like Engine, ISG and ERAD and  Boundaries like the HV battery charge sustaining. All these variables have to be taken into account to achieve the best fuel economy and emission results. It becomes apparent that the control strategy has a major influence on the targets. The key question is: how to find the optimal hybrid operation strategy within a large spectrum of variable parameters and boundaries? Each solution is characterized by a specific development of battery SOC. It is evident that only one solution shows the full potential of the analyzed powertrain. To be able to objectively compare different powertrains in a specific use case, FEV performs an optimization of the hybrid operation strategy shown in Figure 1 [7], [8]. Figure 7 – FEV hybrid operation strategy development process First of all, a simulation model for all relevant powertrain modules has to be implemented. The simulation model will further be validated with real vehicle measurement data, for example from NEDC driving. Simulation results accuracy increases with the availability of measurement data. One could think of adding temperature dependent engine, catalyst and transmission models. These will be added as soon as measurement data becomes available. In a next step a numerical optimization method, the discrete dynamic programming method (DDP), will be configured to the focused axle-split hybrid powertrain. The DDP is based on the “Bellman’s Principle of Optimality” explained in [8] and used by FEV to determine global fuel economy optima. It offers the possibility to reduce the complexity of an optimization problem. Given a degree of freedom of N different control modes and M time steps, a brute force approach leads to a complexity of N^M. This is not applicable for current computer systems. By use of “Bellman’s Principle of Optimality”, the complexity can be reduced to N*M. Therefore the different operation modes E-Drive, serial and parallel hybrid driving require separate calculation tasks. Furthermore, the maximum power, torque and speed limitations of the powertrain are taken into account for each mode. Finally, the DDP parameters battery start SOC, end SOC, minimum SOC, maximum SOC and delta SOC during the cycle as well as the time step size and the torque split delta torque are adjusted. These parameters influence the optimized results accuracy. After configuration the optimization routine can be executed for a predetermined driving cycle. Here the NEDC was chosen as a common and simple driving cycle allowing simple explanation of the results. The results will represent the optimal operation of the powertrain. Hereby an objective comparison of different powertrain concepts without the influence of an operation strategy can be performed. Furthermore, the results might be taken as a benchmark for the later online operation strategy. Thus the online operation strategy’s efficiency becomes measureable and can be tuned as optimal compromise for all driving cycles. This is an advantage compared to just optimizing an online strategy without the knowledge of the global optimum. The online hybrid operation strategy will be derived from the optimized operation and implemented in the vehicle. With adaptation of the online strategy to further hybrid operation targets like driving comfort, performance and dynamic vehicle handling the optimum might not be reached. Nevertheless, the optimized results of DDP can be used as benchmark. This allows carefully tuning the online strategy to all targets but still with best fuel economy. In the next chapter the axle-split hybrid is analyzed according to the described process with the DDP method. Results The simulation model for the conventional powertrain was implemented and validated with measurement data from the customer. To validate the influence of the shift strategy the conventional vehicle has been optimized. Here the optimization param