Survey of power management control technologies for HEVs and PHEVs suggests future need to consider vehicle as part of larger system
A comprehensive survey of major power management control algorithms for hybrid-electric (HEVs) and plug-in hybrid electric vehicles (PHEVs) proposes that future work will need to consider the vehicle as part of a larger system which can be optimized at an even larger scale.
This type of large-scale optimization will require the acquisition and processing of additional information from the driver and conditions outside the vehicle itself, suggests Dr. Andreas Malikopoulos, Deputy Director of the Urban Dynamics Institute and an Alvin M. Weinberg Fellow in the Energy and Transportation Science Division with Oak Ridge National Laboratory (ORNL).
The research reported in the literature to date has aimed at enhancing our understanding of power management control optimization in HEVs and PHEVs. While much progress has been made, some improvements have been incremental, and there has been considerable repetition of a limited number of basic concepts. It appears that the current state of the art is now at a point where new and significantly different approaches are needed.
… There is a solid body of research now available that has aimed at enhancing our understanding of power control optimization in HEVs and PHEVs. … The biggest remaining uncertainties are related to external factors, including the driver’s driving style, the surrounding traffic environment, and the driving terrain. It appears that future research studies need to be devoted to considering the vehicle as part of a larger system, which can be optimized at an even larger scale. Such large-scale optimization will require the acquisition and processing of additional information from the driver and conditions outside the vehicle itself.
This is likely to require addition of new sensors and/or better utilization of information generated by existing sensors. However, the processing of such multiscale information will require significantly new approaches in order to overcome the curse of dimensionality. One particular area where new sensors will be needed is in vehicle-to-vehicle communication.
… we can assume that these technologies will be available in a few years. The question is whether we could take advantage of these technologies and optimize the power management control in HEVs and PHEVs. What if we would consider the problem of optimizing fuel economy and emissions for a fleet of vehicles rather than a single vehicle, thus eliminating the uncertainty related to traffic? What would be the appropriate conceptual approaches for modeling and optimization?—Malikopoulos (2014)
Malikopoulos starts his paper by reviewing the power management control problem and presenting control algorithms that can be used to derive the optimal control policy—e.g., dynamic programming (DP); equivalent fuel consumption minimization strategy (ECMS); and model predictive control (MPC).
Dynamic programming has been widely used as the principal method for analysis of sequential decision-making problems such as deterministic and stochastic optimization and control problems, Markov decision problems, minimax problems, and sequential games, Malikopoulos notes. DP relies on the principle of optimality—i.e., regardless of the initial state of the system and initial decision, the remaining decisions must constitute an optimal policy with regard to the state resulting from the first decision.
Although DP can yield a global optimal solution in closed form, the associated computational requirements are often overwhelming, and for many problems, a complete solution by DP is impossible, Malikopoulos says.
Model predictive control uses prediction models to obtain a control action by solving an online optimization problem; it is often used in constrained regulatory related control problems of large-scale multivariable systems, where the objective is to operate the system in a certain desired way.
The instantaneous equivalent fuel consumption minimization strategy (ECMS) allows the battery SOC to be taken into account.
Malikopoulos then enters into his survey of the applications of major power management control algorithms reported in the literature to date within four discrete categories: parallel HEVs; series HEVs; power-split HEVs; and PHEVs, with subcategories of parallel, series and power-split architectures.
In the parallel HEV, both the engine and the motor are connected to the transmission, and can power the vehicle either separately or in combination. Parallel HEVs can use a smaller battery pack as they rely more on regenerative braking and the engine can also act as a generator for supplemental recharging; thus, they are more efficient for highway driving than in urban stop-and-go conditions or city driving, he notes.
Parallel HEVs offer two architectures: pre-transmission and post-transmission. In a pre-transmission architecture, the electric machine can start the engine; thus, a starter/alternator is not necessary, and there are some savings associated with the reduced weight. In the post-transmission architecture, the regenerative braking efficiency is maximized due to the physical location of the motor. There are also fewer to no spinning losses through the transmission.
In a series HEV, the electric motor is the only means of providing the power to the wheels; the motor draws electric power in combination from the battery and from a generator run by the engine. The engine is typically smaller in series HEVs as it only has to meet on average the driver’s power demand, and the battery pack is generally more powerful than the one in parallel HEVs to provide remaining peak driving power needs.
The larger battery and motor, which are required by series HEVs, along with the generator, add to the cost, making series HEVs more expensive than parallel HEVs. While the engine in a conventional vehicle may inefficiently operate to satisfy the driver’s power demand, e.g., stop-and-go driving, in a series HEV, the engine operates only at its most efficient speeds and loads as it is not coupled to the wheels.
Since series HEVs are superior in stop-and-go driving, they are primarily being considered for buses and other utility vehicles.
The power split HEV combines the advantages of both series and parallel configurations; series HEVs are more efficient at lower vehicle speeds, whereas parallel HEVs are more efficient at high speeds. The power split HEV costs more than a parallel HEV as it needs two electric machines acting as both a motor and a generator and a larger battery pack.
The “power split” name comes from the power split device (PSD), which is a planetary gear set that replaces the traditional gearbox and acts as a continuously variable transmission with a fixed gear ratio. The PSD allows the smaller of the two electric machines to act as a starter for the engine, thereby eliminating another component of a traditional gasoline engine.
The engine can both power the vehicle directly, as in the parallel drivetrain, and be effectively disconnected from the wheels so that only one of the electric machines acting as a motor propels the vehicle, resembling a series HEV.
PHEVs are hybrid vehicles with rechargeable batteries that can be restored to full charge by connecting a plug to an external electric wall socket. A PHEV shares the characteristics of both an HEV, i.e., having a battery, an electric motor, and an engine, and an all-electric vehicle, i.e., having a plug to connect to the electrical grid.
|Hybrid architectures. Malikopoulos (2014) Click to enlarge.|
The numerous control algorithms described cover the period from 1998 to the present, and they are distinguished by the HEV or the PHEV architecture in which they were implemented and their approximate chronological order.
Investigating a new optimization framework that considers a fleet of vehicles could aim to compute the most efficient vehicle speed in centralized locations and communicate this with driver information systems to the driver to avoid congestion, thus improving overall efficiency and reducing emissions in conventional vehicles.
In HEVs and PHEVs, the power management controller would have to account for limited uncertainty about surrounding traffic and commute and be able to optimize fuel economy, pollutant emissions, as well as battery lifetime and range. The detailed investigation of these issues could provide policymakers with unique new tools to assess the implications in promoting the development of technologies and infrastructure in new directions.—Malikopoulos (2014)
Malikopoulos, A.A. (2014) “Supervisory Power Management Control for Hybrid Electric Vehicles: A Survey,” IEEE Transactions on Intelligent Transportation Systems doi: 10.1109/TITS.2014.2309674