Test Beds
Highway Vehicles - Hydraulic Hybrid Passenger Vehicle (Test Bed 3)
NEW!
See video of the HHPV test bed and learn about the CCEFP's infrastructure for this project.
Leader
Prof. Perry Li (UMN)
Statement of Test Bed Goals
The overall goal of this project is to realize a hydraulic hybrid power-train with drastic improvement in fuel economy and good performance to be competitive with other technologies such as electric hybrid, for the passenger vehicle segment. As a test bed project, it also drives and integrates associated projects by identifying the technological barriers to achieving that goal. The design specifications for the vehicle include: fuel economy of 70 mpg under the federal drive cycles; an acceleration rate of 0-60 mph in 8 seconds; the ability to climb a continuous road elevation of 8%; emissions meeting California standards; and size, weight, noise, vibration and harshness comparable to similar passenger vehicles on the market.
Test Bed Role in Support of Strategic Plan
Test Bed 3 directly supports goal 2: improving the efficiency of transportation. Efficiency is obtained by utilizing fluid power to create novel hybrid powertrains for passenger vehicles. The powertrains integrate high efficiency components (goal 1), compact energy storage (goal 2) and methodologies for achieving quiet operation (goal 4) from related CCEFP projects.
Description and Explanation of Research Approach
The high power density of hydraulics make it an attractive technology for hybrid vehicles, since they should be able to provide both high mileage and high performance. A few hydraulic hybrid vehicles have been developed for heavy, frequent stop-and-go applications such as garbage or delivery trucks. However, hydraulic hybrids have not yet reached the much larger passenger vehicle market. In order to realize their potential for small vehicles, hydraulic hybrid drive trains must overcome limitations in component efficiency, energy storage density, and noise. These barriers represent worthwhile challenges that stretch the envelope of existing fluid power technologies.
Three possible architectures for hybrid drive trains are series, parallel and power split. A series drive transmits all power from the engine to the wheel with hydraulic pumps and motors. This architecture enables running the engine at its most efficient combination of torque and speed; however, it cannot take advantage of the high efficiency of purely mechanical power transmission through a shaft. A parallel architecture augments the engine with a pump/motor. It sends a high percentage of wheel power through the efficient mechanical shaft, but it has less ability to keep the engine at its best operating point. TB3 focuses on power split architectures which are the less studied hydraulic hybrid architectures. Power-splits combine the positive aspects of the series and parallel drive train. In addition, all architectures can be used to regenerate braking energy.
This test bed is currently developing two hydraulic hybrid passenger vehicles, each of which offers unique research benefits. The "Generation 1" vehicle was built in-house using the platform of an off-road all terrain vehicle (a Polaris "Ranger" which was donated to the CCEFP). An input-coupled power-split architecture is utilized in this vehicle. The vehicle has been outfitted with a modular power train. This enables experimenting with different pump, motor and energy storage technologies, including those developed in complementary CCEFP projects. However, the first vehicle cannot be driven at speeds higher than about 25 MPH due to concerns about vehicle stability.
Test Bed 3 - Generation 1: Polaris Ranger utility vehicle retrofitted with a hydraulic hybrid drive train.
The "Generation 2" vehicle is being developed. It is built on the platform of a 2010 production pickup truck, which has refined vehicle dynamics capable of highway speeds. Its power-train utilizes a custom-built continuously variable output-coupled power-split hydraulic transmission developed by Folsom Technologies which will be complemented with hydraulic accumulators for enable hybrid operation. The power-train is attractive in that it is built as a compact, highly integrated, self-contained package. It will be capable of rigorous testing and presents an opportunity for the study of an alternate power-split architecture. Nevertheless, the integrated package prevents changing out the hydraulic pump/motors. Also, since it is not originally designed for hybrid operation, the transmission not necessarily optimally sized and presents some control restrictions when operating in hybrid modes. Therefore, the "Generation 1" vehicle is being continued despite the pending availability of the roadworthy "Generation 2" vehicle.
Our ultimate goal will be a "Generation 3" vehicle with a true passenger vehicle chassis. We expect this development to begin in 2012.
Achievements
Achievements and Plans Applicable to Generation 1 & Generation 2 Vehicles
Three achievements apply to both vehicles: replacement of the controls firmware, a study of input and output coupled hybrid transmission architectures, and a comparison of hydraulic and electric hybrid architectures. These studies utilize the 3 level hierarchical control/analysis approach that was developed in previous years [5]. They are described in order below.
Controls firmware upgrades: The Generation 1 vehicle has previously used "xPC Target" firmware to interface the controller with the powertrain. We are now converting to firmware that is popular for automotive systems, "Micro-Autobox", to improve both the hardware and software robustness. In addition, the standard system will simplify migration of the controller to the Generation 2 vehicle.
Micro-Autobox utilizes a specialized software, "DSpace". The control algorithms developed in-house are written in "MATLAB". Therefore, we are also developing "MATLAB" code capable of communicating with "DSpace".
The conversion from xPC Target to Micro-Autobox and DSpace will be completed on the Generation 1 vehicle in 2011. Micro-Autobox will be used for all controls implementation on the Generation 2 vehicle. Installation of the controller on the Generation 2 vehicle will be initiated in 2011.
Input vs. Output Coupled Study: Available power split transmissions can be classified as "input coupled", "output coupled" and compound. An input coupled transmission utilizes a fixed gear ratio between the engine and one pump/motor, while the second pump/motor is coupled to the wheels with a planetary gear train. An output coupled transmission utilizes a fixed gear ratio between the wheels and one pump/motor, while the second is coupled to the engine with a planetary gear train. A compound transmission is one in which both pump/motors are coupled with planetary gear trains.
A study to determine the most efficient powertrain configuration was performed. This was achieved by defining and optimizing a generalized expression that relates the kinematics of the engine, wheels, and pump/motor units. This expression is referred to as Matrix G. Matrix G takes different degenerate forms for input coupled and output coupled drive trains, as seen in Table 1. However, a general Matrix G can represent a combination of the two. The combined configuration is described as a "compound" drive train.
The elements of Matrix G were optimized for a prescribed drive cycle using all three potential architectures. This approach maximizes the opportunity for improving fuel economy. The optimal size of the hydraulic pump/motors is generated as part of the process. Whereas previous approach to determining the optimal power-split configuration explicitly considers and optimizes each discrete physical configuration [4], our approach of the optimization of the kinematic relation (Matrix G) is more efficient since a Matrix G can be realized by multiple physical configurations with the same performance.
Table 1: Preliminary results of the architecture comparison
Preliminary results are shown in Table 1. As seen in the table, the fuel economy of the various architectures appears fairly similar: the optimized results are within about ±5%. The component sizing varies slightly, with the compound architecture requiring the overall smallest pump/motors. However, the input and output architectures are competitive. Refined results will be obtained in 2011.
Hydraulic/Electric Hybrid Comparison: A comparison of the efficiency of hydraulic and electric hybrid vehicles was performed in 2010.The initial results indicate that for the light (1000kg) vehicle that was studied, electric and hydraulic hybrids have comparable fuel economy under standard EPA driving cycle without additional acceleration requirements. It is expected with heavier vehicles, more stringent acceleration requirements, and more efficient pump/motor, the advantages of hydraulic hybrids will be accentuated. Improved analysis will be performed to refine this comparison in 2011.
Achievements and Plans for Generation 1 Vehicle
Work on the Generation 1 vehicle in 2010 has focused on redesigning the drive train. In addition, a fuel sensor has been added. These accomplishments are described below, followed by discussion of additional plans for exploiting the availability of the improved vehicle in 2011.
Figure 1: CAD model of original Generation 1 HHPV powertrain
Drive Train Redesign: The original Generation 1 vehicle drive train, illustrated in Figure 1, suffered from several limitations which restricted its usefulness. The drive train is complicated and it includes several belts and chains. The vehicle's frame would flex enough during driving that the chains would sometimes skip teeth. In addition, the planetary gear trains, which combine power from hydraulic pump/motors with engine power at the rear wheels, were undersized, so they were not capable of carrying the full wheel torque specification.
The drive train was completely redesigned in 2010. A CAD representation of the system is shown in Figure 3 and a schematic of the revised system is provided in Figure 2. All of the problems caused by the belts, chains and frame flexion have been eliminated by using gears. The drive train has been simplified by replacing dual rear wheel pump/motors and planetary gear trains with single units driving a stock automotive rear wheel differential. The original axial piston type pump/motors have been replaced with high efficiency bent axis piston units. Gear ratios and pump/motor sizes are chosen to optimize fuel economy under EPA driving cycles and to satisfy the acceleration requirement.
The redesigned drivetrain can operate in four different modes. The first is "HMT" mode, where the engine is engaged with the rest of the system while both P/M "T" and P/M "S" (see Figure 2) are hydraulically engaged. The second is parallel mode, where pump/motor "T" is coupled to the engine and P/M "S" is locked up. The third is "T-only" mode, which is the same as parallel mode except that the engine clutch is disengaged. The last is "S-only" mode, where P/M "S" alone is used to drive the system. The engine is removed from the system by disengaging the engine clutch and the shaft of P/M "T" is locked up. The last two modes are similar to a series transmission where only the motor is operating and powered only by the accumulator charge.
Transmission components are currently in the machining and assembly stages. A fully assembled transmission expected to be ready for testing in March 2011.
In addition to providing a more robust drivetrain on the vehicle, the transmission is also designed to stand alone. With this new capability, it will be possible to test the transmission on a dynamometer, facilitating efficiency mapping and control development of the vehicle.
Figure 3: CAD model of redesigned Generation 1 HHPV powertrain
Finally, the modular architecture of the redesigned transmission enables the pump/motors to be changed out. We plan on replacing the bent axis pump/motor used as pump/motor "S" with a pulse width modulated fixed displacement pump/motor designed in Project 1E.1 during 2011. The purpose of this test is to compare the efficiencies of the two approaches. This test will provide demonstration of a real world application for the pulse width modulated pump/motor also.
Figure 2: Schematic representation of redesigned Generation 1 HHPV powertrain
Low level Control Refinement: System identification experiments have been performed on the existing Generation 1 vehicle. This together with experimentally derived pump/motor maps in [6] provide improved information for refining the low level control algorithm design. However, the basic control architecture presented in [7] was still followed.
Fuel Sensor: An accurate engine efficiency map is crucial to developing controllers capable of minimizing fuel consumption of a hybrid vehicle. Simulations of the Generation 1 vehicle performed to date have utilized a Willans' line [3] approximated engine efficiency map; the engine in the vehicle may deviate substantially from this approximate map. A fuel flow sensor was calibrated and installed on the Generation 1 vehicle, utilizing its original drive train, during 2010 to enable creating an accurate map.
Normally, the engine would be removed from the vehicle and mounted on a dynamometer to obtain the efficiency map. We have instead utilized pump/motor "T" to load the engine. The accuracy of the results is limited by the accuracy of the efficiency map of the pump/motor, which was created using a test stand designed for that purpose during 2009.
Additional Plans for Generation 1 Vehicle During 2011: Experiments will be performed to operate the Generation 1 transmission as a continuously variable transmission (CVT) rather than a full hydraulic hybrid. These experiments have two purposes. First, operation as a CVT serves to prove the effectiveness of the low level control strategy. Second, the fuel economy obtained from operation as a CVT provides a limit for comparing full hydraulic hybrid modes.
Achievements and Plans for Generation 2 Vehicle
Effort on the Generation 2 vehicle in 2010 has focused on returning the FTI transmission to service, creating a test plan for generating the efficiency map of the transmission, and developing enhanced simulations. Each of these efforts is described in detail below. Continuing plans for 2011 are described at the end of each item.
Returning FTI Transmission to Service: Collaboration with the pickup truck manufacturer and FTI was started in spring 2009. The pickup was donated to the CCEFP. The internals of the FTI transmission to be used in this truck are shown in Figure 4. Testing of the FTI hydro-mechanical transmission was initiated in early winter 2009. The FTI transmission is shown mounted to a 400 HP dynamometer available at the FTI facility in Figure 5.
Problems with the controls on the FTI dynamometer in early 2010 resulted in the transmission being driven at high speed in reverse. Since no lubricant is supplied in this configuration, extensive damage occurred to both mechanical and hydraulic components in the transmission. The pickup truck manufacturer agreed to fabricate many parts to replace the damaged hydraulic components and new planetary gear sets were procured and modified for the transmission rebuild.
|
|
Figures 4 & 5: Internal view of FTI power split transmission (left) and FTI Hydro-Mechanical
Transmission on the dynamometer test-stand at the FTI site (right).
Testing resumed in early autumn 2010 and performance problems, particularly poor efficiency and the inability to generate sufficient torque, were noted immediately. It was initially thought that the inability to produce sufficient torque was due to a new clutch assembly that had been installed in place of a clutch pack in the original assembly. However, substitution of the original clutch pack did not produce a measurable torque increase so it was decided that a complete transmission tear-down was in order to determine the root cause of the problem. Examination of the hardware indicated that the stroke of the pistons on the hydraulic motor unit had exceeded the design limit by a margin sufficient to cause the outer piston ring to catch on the end of the cylinder block. This resulted in breakage of six of the seven outer rings and scoring of the bores in the cylinder block. A thrust bearing is being installed between the front face of the cylinder block and the thrust plate in the yoke to limit axial motion of the cylinder block and thus preclude piston ring damage on the next transmission build.
Another significant failure mode was identified in addition to the broken piston rings: The spherical end of all of the motor pistons and the spherical seat of the retainer plate suffered severe fretting and wear. The retainer plate and torque plate were redesigned to incorporate a counter bore for fitment of a bronze retainer ring, thus eliminating the steel/steel contact surface present in the original design. Again, the truck manufacturer supplied the machining to effect these changes. The transmission is currently being reassembled and is scheduled for resumption of testing in January 2011.
The pickup truck is expected to be delivered to the University of Minnesota with the FTI transmission installed in late Spring 2011.
Efficiency Map Test Plan: FTI has provided a simulated efficiency map with their transmission. However, the transmission was originally intended to be run as a continuously variable transmission rather than a hydraulic hybrid transmission. Therefore, dynamometer tests are being planned to obtain the efficiency map corresponding to hydraulic hybrid operation. This is essential for fuel economy predication and the design the control and energy management system. The test plan, developed by us, will be implemented on the dynamometer available at the FTI site (see Figure 5) prior to shipping the transmission to Minnesota.
The tests must be designed to overcome two unique circumstances. First, the two pump/motors in the FTI transmission are intrinsically coupled; therefore, the mechanical and volumetric efficiency of each pump/motor cannot be obtained individually. Second, a hydraulic power supply cannot be utilized for the dynamometer tests at the FTI site.
Figure 6: Combinations of test conditions for FTI transmission.
Contours represent flowrate through the relief valve.
Both restrictions have been overcome by developing a procedure where flow is measured through a relief valve connected between the high and low pressure ports of the pump/motors. Figure 6 illustrates all combinations in which the combinations of the two pump/motors could operate. Ordinarily, approximately 3600 data points would be required to fully define all portions of this map. In addition, many of these combinations are not feasible with the restriction of the test facility. However, with the assumption that both pump/motor units have similar characteristics, we have devised a means for approximating the entire map by obtaining only 240 data points, which are represented by the stepped red profile in Figure 6. This approach also reduces the risk of operating the relief valve beyond its capacity. Furthermore, regenerative braking scenarios can also be simulated. Data for creating the FTI transmission efficiency map is expected to become available by March 2011.
Development of Enhanced Simulations: Analysis of the FTI transmission utilized in the Generation 2 vehicle was initiated by adapting the "backward facing" simulation tools developed for the Generation 1 vehicle. "Backward facing" means that the drive cycle is known in advance and the transmission components are optimized to provide the prescribed wheel torque while consuming the minimum amount of fuel. Mechanical restrictions imposed by the Generation 2 transmission architecture increase the complexity of the controls strategy development. However, the restrictions appear to have only minor impact on the fuel economy using the backward facing simulation.
In order to further investigate this issue, a "forward facing" dynamic model is being developed and refined. The forward facing model takes driver commands as the input. The forward facing model takes advantage of a MATLAB Simulink model provided by the truck manufacturer, which includes details of the engine dynamics, auxiliary losses of the vehicle, aerodynamics, temperature variation, and the like. We have enhanced the manufacturer model by replacing a model of a conventional automatic transmission with a model of the FTI transmission and adding a model of an accumulator. In addition, the model is also being re-structured into a form where the designed controllers can be directly implemented onto the actual vehicle controls hardware. The forward facing model is expected to predict the fuel economy more accurately due to controlling energy management in real time.
References
[1] Rannow, M., Li, P., Chase, T., Tu, H., and Wang, M., 2010. "Optimal design of a high-speed on/off valve for a hydraulic hybrid vehicle application". Proceedings of the 7th International Fluid Power Conference, Aachen, Germany.
[2] Cheong, K. L., Li, P. Y., Sedler, S. P., and Chase, T. R., "Comparison Between Input Coupled and Output Coupled Power-Split Configurations in Hybrid Vehicles", in press for 52nd National Conference on Fluid Power (Paper 10.2), Las Vegas, NV (March 2011 publication expected).
[3] Richard Stone, "Introduction to Internal Combustion Engines", SAE International; 3rd Edition, 1999.
[4] C. T. Li and H. Peng, "Optimal Configuration Design for Hydraulic Split Hybrid Vehicles", Proceedings of the American Control Conference, Baltimore, MD, 2010.
[5] P. Y. Li and F. Mensing, "Optimization and control of hydro-mechanical transmission based hydraulic hybrid passenger vehicle", Proceedings of the 7th International Fluid Power Conference (IFK), Aachen, Germany, March 2010.
[6] D. R. Grandall, Performance and Efficiency of Hydaulic Pumps and Motors, M.S. Thesis, Department of Mechanical Engineering, University of Minnesota, January, 2010.
[7] T. P. Sim and P. Y. Li, "Analysis and Control Design of a Hydro-Mechanical Hydraulic Hybrid Passenger Vehicle", Proceedings of the ASME 2009 Dynamic Systems and Control Conference #2763, Hollywood, 2009.