1.4.1 Objective
The goal of this study is to develop a systematic hybrid electric powertrain design methodology that includes the automatic generation of design candidates, their com- puterized steady-state and dynamic equation derivations, automated feasibility and powertrain type determination algorithms, and design candidate performance and fuel economy evaluation algorithms suitable to component sizing. The methodology will be applied to exploring the viable hybrid electric powertrain concepts with two planetary gearsets, two electric machines, one battery pack, one internal combustion engine, and at least two operational modes for light-duty truck applications.
1.4.2 Approach
The design methodology aims to explore hybrid electric powertrain concepts that meet or exceed light-duty truck performance requirements while having superior fuel economy benefits compared to an equivalent conventional powertrain. Performance requirements of light-duty trucks are demanding as they need to be able to tow loads greater than the vehicle curb weight. An HEV powertrain shall meet these requirements with electric machines, whose maximum power is limited due to weight, cost, efficiency, and packaging constraints.
Before deriving the design methodology, the first question to be answered is which components need to be included in the design process. Assuming the target appli- cation is either a front-wheel or rear-wheel drive vehicle, one component must be a vehicle output shaft. Since the focus of this study is a hybrid electric powertrain design, one engine should also be included. If one electric machine is used to provide electric propulsion functionality, the series powertrain type and the eCVT operation of the power-split powertrain types, which require two electric machines cannot be realized. To cover a wide range of design candidates, two electric machines are used in this study. Planetary gearsets with their simplicity and functionality in generat- ing design candidates are also included in the component list. In determining the number of PGs, the first option is to use a single PG, which has three nodes. Since the number of components is four (engine, output shaft, two electric machines), and only one of the electric machines can share a node with other components, 54 modes can be generated. Moreover, a one PG design excludes the possibility of designing a compound-split mode, which has a competitive advantage in providing eCVT opera- tion with low power requirements for the electric machines at medium to high vehicle speed levels. In contrast to single PG designs, two-PG designs not only facilitate the inclusion of compound-split modes but also considerably increase the number of fea- sible modes in the design space. Thus, the design process begins with two PGs in this
dissertation. If two PGs do not provide enough feasible designs to meet all the vehicle requirements at the end of the process, then three-PG designs can be explored. The maximum number of brakes that will be used in the generation of design candidates is determined to be three because more than three brakes in a two-PG mode would lock every PG node.
After determining the number of PGs in the design, the next research question is how many modes to include in a design. A fixed-gear mode not mated with a trans- mission cannot meet all the performance requirements at a wide vehicle speed range. EV and series HEV powertrain modes are similar, as electric machines serve as the primary propulsion elements. Since electric machines used in the design have limited power, they cannot meet towing requirements alone. Moreover, since the energy sup- ply in EVs is a battery with limited energy capacity, they cannot be used in long haul towing. Series HEVs do not have this problem, as there is a generator. A problem that both EVs and series HEVs share is the inflexibility of controlling the operating points of the propulsion electric machine as they are completely dependent on the driver’s torque demand and current vehicle speed. Hence, operating the powertrain at its most efficient point is impossible, considering the various towing conditions rang- ing from an empty to a fully loaded vehicle at any vehicle speed. Similar arguments are valid for parallel HEV modes as well. Since the speed of the electric machine is dependent on the vehicle speed in a parallel HEV mode, the degrees of freedom in control authority are reduced. As a result, a parallel HEV can only be efficient under all towing conditions if it is mated to a transmission with multiple gears.
Power-split HEV modes are the right candidates for meeting the performance re- quirements of a light-duty truck since they allow the flexibility of controlling operating points of the components to desired states for changing operating conditions. As will be shown, an input-split type mode might meet performance requirements at low to medium vehicle speeds very well. As vehicle speed increases beyond a certain level,
however, electric machine power requirements exceed the allowed power limits. On the other hand, the output-split and compound split types are weak in meeting per- formance requirements at low vehicle speed but their performance increases as vehicle speed rises. Even if a power-split mode meets all performance requirements, a second mode will be needed to provide backward speed capability. Thus, the hypothesis in this study is that all performance requirements can be met across a wide range of vehicle speeds by using at least two modes.
The proposed design approach in this dissertation first creates all possible modes that can be generated with one vehicle output shaft, one engine, two electric ma- chines, two PGs and at most three brakes. PG gear ratios are also included as design variables. Selecting powertrain modes as the design candidates creates a relatively smaller design space compared to other methods in the literature. The process then performs an exhaustive search through the strategic ordering of evaluation steps, where as many modes as possible are eliminated in the early design stages. An ex- haustive search eliminates the risk of missing any competent design candidate in the process. Computationally heavy performance evaluations are accelerated through the introduction of new algorithms. Moreover, extremely time-consuming fuel economy simulations are conducted at the last stage of the process so as to handle just a small group of designs.