Review of the Research Program of the Partnership for a New Generation of Vehicles: Third Report (1997)

A key measure of the ultimate success of the PNGV will be its ability to integrate R&D programs that collectively improve the efficiency of converting fuel into motive power for automobiles. This improvement must take place within the very stringent boundary conditions of size, reliability, durability, safety, and affordability of today's cars. At the same time, it must meet even more stringent emissions and recyclability levels and use components capable of being mass produced and maintained in a manner similar to current powertrains.

In order to achieve the Goal 3 fuel economy target (up to three times the fuel efficiency of today's comparable vehicles), the efficiency of the combustion engine (e.g., a CIDI, Stirling engine, or gas-turbine engine) or fuel cell, averaged over a driving cycle, will have to be approximately double today's efficiency to achieve at least 40 percent thermal efficiency (see Appendix E). This is a very challenging goal, considering all of the constraints noted above.

The candidate systems and subsystems have not changed during the past year. They are as follows:

• four-stroke CIDI engines

• gas turbines

• Stirling engines

• fuel cells

• reversible energy-storage devices¹

• electrical and electronic power-conversion devices

HEVs use an energy storage device to modify the fluctuating demands on the primary power plant. This modification allows the engine or fuel-cell peak power output to be reduced and provides an opportunity for improved efficiency by both restricting the power fluctuations and recovering some of the vehicle's kinetic energy during braking operations. The PNGV is sponsoring research on batteries, flywheels, and ultracapacitors for this purpose.

Committee members reviewed each of these R&D programs in considerable depth to assess the status of each, the progress that has been made, and the developments required for the future. The PNGV Technical Roadmap has been updated for most of these technologies, and it provides a good summary of the program goals. However, following in-depth reviews, the committee members almost all reported that the people in charge of performing the technical work had very limited information about the detailed requirements that would be imposed by a vehicle installation. This lack of information reflected a major concern for all of the candidate technologies, with the possible exception of the CIDI engine; namely, the systems analysis work and packaging studies that would provide this information have fallen significantly behind schedule. This lack of direction to the individual technical teams makes it difficult for them to focus their efforts. This basic flaw in the program cannot help but reduce R&D effectiveness and efficiency.

Traditional passenger car diesel engines exhibit 15 percent to 30 percent better fuel economy, 10 percent to 20 percent lower carbon dioxide (CO₂) emissions, nearly zero evaporative emissions, and very low cold-start emissions when compared with similar gasoline engines. However, CIDI engines also suffer from size, weight, noise, and cost penalties that have limited their market acceptance in passenger cars unless their purchase is encouraged by a substantial fuel cost differential. Eliminating these disadvantages, while retaining or increasing the superior fuel economy of these engines, represents major challenges on several fronts.

During the past year, the CIDI engine has been selected as the most promising of the four-stroke, direct-injection PNGV engine candidates for either stand-alone or hybrid vehicle use, and a technology road map applicable to CIDI engines has been completed. Research priorities have been established, and five high priority areas are being addressed through dedicated research programs. A workshop was held in which individuals from industry, academia, and government discussed and prioritized the critical technologies necessary for CIDI engines to meet the PNGV objectives. Also, the PNGV USCAR participants and

the U.S. heavy-duty diesel-engine industry have established communications and agree that there are several areas of common interest including:

• understanding combustion fundamentals

• combustion gas after-treatment

• fuel-composition effects on engine performance

The five high priority CIDI research areas identified by the CIDI technical group are:

• lightweight engine architectures

• dimethyl ether (DME) as an alternate fuel for CIDI engines

• combustion-related processes

• lean NO_x catalysis

• alternative CIDI fuels assessment (including diesel reformulation)

The focus on combustion and DME is related to concerns about particulate emissions from CIDI engines. New cooperative programs, initiated in 1996, address critical aspects of each area. Other programs, most notably those involving fuel injection technology and lean NO_x catalysis, were already in place. The OEMs are supporting several in-house CIDI development programs, primarily through their European subsidiaries, but also in the United States. These projects involve both development of subsystems (fuel delivery, turbochargers, and engine- and emission-control strategies) and prototype vehicles through the HEV programs, which will serve as a test bed for the new technologies. CIDI hybrids are under contract at both Ford and Chrysler, with different delivery schedules. Because the HEV programs preceded PNGV, these hybrid vehicles are required to double fuel economy rather than meeting the PNGV Goal 3 vehicle target of tripling fuel economy.

The critical characteristics of a CIDI engine, suitable for application to a Goal 3 vehicle, are shown as a function of PNGV milestone targets in Table 4-1. The last column shows that the 1995 targets have been met or exceeded in both test engines and some production engines. Passenger car diesel engines typically weigh 20 percent to 40 percent more than their gasoline counterparts. Table 4-1 shows a need to reduce the engine specific weight (usually referred to as specific power) by 26 percent and increase the displacement specific power (power density) by 29 percent in the 1995 to 2004 time frame. A production Volkswagen engine, which includes a variable geometry turbocharger and intercooler, is within 5 percent of the specific power goal but has made no progress toward the power specific weight goal. This indicates the need for radically new materials or construction techniques that can reduce the weight of CIDI engines. Using a turbo-charger and intercooler also adds to the packaging challenge. Compared to the

gasoline engine, these components suggest an initial cost penalty of perhaps $800 to $1,200 in addition to the undetermined added cost of a sophisticated electronic fuel-injection system. The need for these added components makes keeping the cost to $30/kW (shown in Table 4-1) a major challenge.

Fuel economy targets for the CIDI engine have been moderated in recognition of the likelihood of having to meet the increasingly stringent emissions standards shown in Table 4-2. The CIDI team believes that the most technically challenging aspect of the CIDI program will be meeting the NO_x emission standards; however, possible stringent fine-particulate standards could also become a significant barrier for diesels. (Note that recently EPA issued a proposed rule for an atmospheric air-quality standard for fine particulates.) The PNGV should discuss with the EPA the likelihood of more restrictive particulate emission standards and their potential effect on the PNGV program. Intensive development of both in-cylinder combustion control and exhaust-gas after-treatment will be required

reduce emissions from the CIDI engine, including reducing NO_x emissions. In-cylinder combustion control will require a sophisticated, high-pressure, fuel-injection system that is matched to the engine's in-cylinder flow field over the entire operating range of the engine. Exhaust-gas after-treatment will require the development of lean NO_x catalysts, NO_x traps, or plasma-aided catalysts.

Two types of reductants have been considered for lean catalysis: urea (or ammonia) and hydrocarbons. Urea reduction of NO_x is a well-known technology developed for treatment of exhaust from stationary sources. However, the technology has not been tested under highly variable vehicle operating conditions, which would require a sophisticated control system to avoid ammonia in the exhaust as a result of incomplete conversion of injected urea. The possible formation of ammonium sulfate particulates and the need to refill a reservoir tank of urea solution also may hamper consumer acceptance of this technology.

Diesel fuel can also be used as a hydrocarbon reductant for NO_x; it can be injected into the exhaust stream. Like the use of urea, the injection technology must be developed for optimal operation with minimal fuel consumption. At present, noble metal (e.g., platinum [Pt], and rhodium [Rh] catalysts are the most promising, but control of nitrous oxide (N₂O) emissions, formation of metal sulfates, and prevention of emission of unburned hydrocarbons are still significant issues. Laboratory data suggest that the Tier II target could be met by matching the operating temperature window of the catalyst and the exhaust-gas temperature, but at a 1 percent fuel penalty.

NO_x trap technology uses an alkaline earth oxide to trap NO_x by reacting

with it to form metal nitrate when the engine is running lean. Periodically, the engine runs fuel-rich, and the temperature of the exhaust increases. Then the metal nitrate is decomposed to release NO_x, which is treated by a catalyst, such as the proven three-way catalyst. The long-term stability of such a system has yet to be demonstrated. The formation of stable metal sulfate and carbonate could cause severe degradation of the performance of such a device.

Plasma-assisted catalytic removal of NO_x is a new technology that has the potential to remove NO_x and particulate matter (PM) simultaneously. The technology has not been tested commercially. In the PNGV CIDI Workshop of May 22–23, 1996, in Detroit, Michigan, it was reported that for this technology to be successful the required power must be reduced. It was also reported that the NO_x removal rate needs to be increased while restraining the formation of nitric acid, ozone, and other atmospheric contaminants. Many of the research results have not been reported; therefore, the current status of and progress made with this technology is not clear.

Laboratory test data presented to the committee by USCAR for the newest Volkswagen engine indicate that the Tier 0, and possibly the low-emission vehicle (LEV) PM standard can be met without exhaust after-treatment. However, the more stringent ultra-low-emission vehicle (ULEV), or Euro III standard, will probably require an oxidation catalyst. Because of the trade-off between PM and NO_x, it is unlikely that these standards can be met by engine modification alone; an after-treatment device will be necessary. Catalytic oxidation devices to reduce PM emissions are currently used successfully on some heavy-duty trucks.

A lean NO_x cooperative research and development agreement (CRADA) was formed in 1994 involving the USCAR partners and five government laboratories, supported by the Department of Defense (DOD) and the U.S. Department of Energy (DOE). The government budget for Fiscal Year 1996 was $293,000 from the DOE Office of Defense Programs technology transfer initiative, and $950,000 from DOE. Unfortunately, because of late appropriation of the DOE funds, there was a nine-month stoppage of research at Sandia National Laboratories and Lawrence Livermore National Laboratory. The Fiscal Year 1997 request from DOE is for $840,000 and the emphasis of the research will be on methods to coat catalysts onto a monolith support and on production of full-size converters for testing and optimizing the reductant. No PNGV funds have been allocated for catalyst R&D because there are strong business incentives for companies to pursue proprietary research in this area. Catalyst suppliers worldwide, as well as Ford, General Motors, Cummins, and other engine manufacturers, both in the United States and abroad, are supporting intensive R&D of lean NO_x reduction technology.

Significant advancements in the fuel-injection system's ability to control the fuel injection rate to match it to in-cylinder combustion processes are required for

the CIDI power plant to be successful in the PNGV concept car and prototype. The concept engine will probably use more sophisticated electronically controlled injectors than are currently used. Two high-pressure fuel-injection configurations are under consideration: a hydraulically actuated common rail and a hydraulically intensified fuel injection system. They are referred to as the common rail and the hydraulic electronic unit injector (HEUI) types. Both are capable of providing an injection pressure independent of engine speed. This characteristic yields engines with relatively high low-end torque, which results in very good driving characteristics. Using a very high pressure common-rail injection system and large amounts (35 percent) of exhaust-gas recirculation (EGR), led a European contractor, AVL, to conclude that engine emissions eventually may be able to achieve the ULEV NO_x standard (Meurer, 1996).

The injection system must be capable of optimizing the injection rate shape and increasing the tolerance for EGR. A small pilot injection at varying times prior to the main injection has shown promise both for controlling combustion and reducing noise. However, as the EGR tolerance of the engine is increased, through increased injection-pressure and injection-rate-shape control, the carbon monoxide (CO) and unburned hydrocarbon emissions are also increased. These emissions are not a problem in current CIDI designs, but they may become a problem as regulations become more stringent and engine operation is extended into the operating range projected for the PNGV concept vehicles. The development of these injection systems for a concept engine is being pursued cooperatively between the USCAR participants and injection system manufacturers. Both types of injection system are being aggressively developed, and the prognosis is good for successful deployment to an engine in the required time frame.

In 1996 the PNGV program established a 4-year combustion CRADA, with an initial allocation of $700,000/yr, plus matching funds from industry, to address the in-cylinder combustion processes of fuel injection, air-fuel mixing, combustion, and emission formation. The CRADA objective is to provide ''the technological understanding required to develop a new CIDI diesel engine which meets the efficiency and emissions standards of a PNGV vehicle." Participants are Sandia National Laboratories, Wayne State University, and the Engine Research Center of the University of Wisconsin, Madison. Optical diagnostics will be performed on a single-cylinder engine at Sandia; engine bench testing will be performed at Wayne State University; and modeling will be done at the Engine Research Center at the University of Wisconsin, Madison. The CRADA is just starting, and initial results are expected by the second half of the 1998 calendar year.

The Tank Automotive Command (TACOM) of DOD and USCAR initiated a program to investigate alternative engine architectures and lightweight materials

leading to procurement, development, and pre-prototype engine testing in 1997. The requested funding of $2.44 million for 1997 is not yet firm, and if it is held at a carryover level of $450,000, such lack of funding would seriously hamper progress. Concept designs being developed by Ricardo Engineering Ltd. are now focused on the possible use of magnesium for the engine block (Talwar, 1996).

A specific manufacturing issue has surfaced related to the need for extremely close tolerances within the injection system; namely, the capability to make 0.10 mm diameter fuel injector holes consistently, compared to today's limit of 0.15 mm. A number of design and manufacturing approaches, for both affordability and weight reduction, are suggested throughout the program status documents. However, the benefits projected for these actions are not quantified with respect to their contribution to the cost or weight objectives, nor are these elements visible in the PNGV work plans.

The relatively advanced small-displacement CIDI engines in production in Europe, together with programs put in place through the PNGV and by individual companies to improve these engines, make the PNGV CIDI performance goals seem potentially achievable.² The highest technology risk appears to be the ability to develop a combustion system and after-treatment system to meet an uncertain set of exhaust emission requirements in 2004. The PNGV program has put in place a modest effort to address both engine-out emissions reduction and after-treatment, but the funds are minimal, and results to date are insufficient to assess a rate of progress. However, substantial commercial programs are under way to address these issues.

The cost and weight objectives of a CIDI engine that will meet the challenging goals for 2004 appear to be major obstacles. Ideas have been proposed for novel engine architectures, systems integration, reduced parts count, and manufacturing improvements to address these problems. However, it seems likely that the sophisticated high pressure injection system, variable geometry turbocharger, and high cylinder pressures, combined with the need for lightweight, high strength materials will continue to bring a substantial cost premium for this engine. It also

appears that the proof-of-concept demonstration of lightweight structures and cost-reduction features for the CIDI are lagging the year 2000 concept vehicle schedule by a year or more. Obviously, further delays will be incurred unless funding is significantly increased from the 1996 level.

In spite of the significant risks of not meeting the PNGV goals as noted above, it is clear that CIDI technology is, by far, the best understood and most highly developed of any power plant technology being considered in the program. Because it is almost certain to be one of the final candidates for use in the year 2000 concept vehicles, the CIDI technology program needs to be more focused and needs a substantial increase in financial support.

Based on its review of the program status and progress for CIDI engines, the committee makes the following recommendations.

Recommendation. The PNGV should expand efforts to devise lightweight, low-cost alternative CIDI engine structures, and additional resources should be made available.

Recommendation. The PNGV should immediately assess the possible effect of regulatory actions aimed at reducing the atmospheric levels of fine PM on the viability of passenger car CIDI engines, and the research and development program should be modified, if necessary. To help the EPA make decisions based on the best possible information, the EPA should be continually informed of decisions made by the PNGV during the downselect process. Furthermore, the PNGV and EPA should work together to determine the trade-offs between vehicle performance and environmental standards and associated impacts on social benefits and costs.

Gas turbine engines have some attributes that make them potentially successful as a Goal 3 vehicle engine and other attributes that will make realizing this potential very difficult. The low-pressure, excess-air, continuous-flow combustion provides very low levels of untreated emissions and a broad multifuel capability. The all-rotating machinery leads to low levels of vibrations and noise. The continuous flow, annular flow path, and high rotational speeds result in high power-to-weight and power-to-volume ratios for the "core" of the engine. Furthermore, the dominance of the gas turbine engine in commercial and military aircraft has generated an enormous amount of R&D, which has resulted in significant technological advances. However, for more than 40 years, substantial efforts to develop the gas turbine for automotive applications have been unsuccessful for several reasons. Most notable are (1) achieving sufficiently efficient

aerodynamic turbomachinery components, which becomes more difficult as their size is reduced; (2) attaining high enough turbine inlet temperatures; (3) developing efficient means to recover a portion of the turbine exit thermal energy; and (4) developing materials, manufacturing techniques, and designs capable of meeting cost requirements.

The "core" of the gas turbine engine generates hot, high-pressure gas that is expanded through one or more turbines to produce work. These engines lose efficiency very rapidly if the main shaft speed is allowed to decrease in partial-power operations. A single-shaft turbine engine can attain a relatively high efficiency over a very limited speed range. Therefore, automotive gas-turbine engines have generally required at least two turbines operating on separate shafts, which make the engine bulkier and more expensive. However, for a series hybrid vehicle, the engine speed is independent of the vehicle speed, the engine-speed range can be reduced, and the single-shaft arrangement may be adequate, depending on the hybrid-vehicle control strategy. Idle fuel consumption of gas-turbine engines is much higher than for similar-output reciprocating engines, so systems optimization is critical.

High temperature gas at the turbine exit in an automotive gas turbine represents lost thermal energy, making regenerators or recuperators essential for efficient operation. These devices are heat exchangers that transfer some of the turbine exit thermal energy to the compressor discharge flow. However, they are large and heavy and thus detract from the power-to-weight and power-to-volume advantages of the engine "core."

Turbine-engine size can be reduced and thermal efficiency improved dramatically by increasing turbine inlet temperatures. Cycle studies have shown that to approach the thermal efficiencies required for a Goal 3 vehicle it is necessary for this temperature to approach 2,500¹F, although there is some indication that it may have to be limited to less than 2,300¹F to limit NO_x emissions. These temperatures exceed, by several hundred degrees, those possible with even the most exotic uncooled metal turbines. Even if they did not, the cost of these exotic materials and their fabrication would prohibit automotive usage. The alternative is to develop a ceramic material with the potential to meet both the cost and high temperature objectives.

In its second report, the committee reported that gas turbines represent a promising technology for hybrid vehicles and that considerable progress had been made in the previous year, especially in turbo-alternator design, bearings, combustors, heat recovery, and controls. However, the committee also reported that progress was behind schedule and there were major technical roadblocks (NRC,

1996). Foremost among the roadblocks was the lack of a turbine material that could withstand high temperatures, that was capable of being precision-formed to provide efficient aerodynamics, and that could be mass produced at low cost. It also appeared that little had been accomplished with respect to systems studies at the vehicle or power plant levels. To some extent, this clouded the actual magnitude of the technical barriers because the requirements and trade-offs were largely unknown.

Some completed systems studies give better indications of the more desirable configurations and the projected performance parameters. The single shaft, turbo-generator configuration has been chosen, thereby limiting the vehicle application to a series hybrid. This configuration uses an all-electric drive system, in contrast to the parallel hybrid, where there is a direct mechanical link between the engine and the drive wheels. Some systems studies have projected better drivetrain efficiency for the parallel hybrid. If this is confirmed, the challenge of enhancing gas turbine efficiency will be increased. The systems studies presented to the committee were very preliminary, and there was no evidence that systems trade-offs had been made.

The Allison engine design is complete, and the Teledyne Ryan design is well under way. Both designs appeared reasonable, although probably optimistic, in their projected goals. The designs have different target characteristics based upon two different hybrid-vehicle performance and control strategies. The Teledyne design power is 55 kW to 60 kW and its efficiency at one-third power, required to meet the Goal 3 vehicle fuel economy target, is 43 percent. Comparable values for the Allison design are 40 kW and 38 percent. The most optimistic projections would put attainable efficiencies at least several points lower than these values. Although many component tests have been run supporting the plausibility of the projections, the necessary resources have not been applied to iterative component and subsystem developments to indicate that these levels of performance can be maintained for the complete engine system.

The single most critical component yet to be demonstrated is a suitably sized, efficient ceramic turbine, capable of being mass produced. The committee found the significant progress has been made by Kyocera-Vancouver and AlliedSignal Ceramics in making high-quality ceramic components with complex shapes and using processes with high potential for scale-up to automotive volumes and cost. Extruded regenerator rotors, integral turbine rotors, and scrolls have all shown new capabilities which, although beset by tooling and startup yield problems, have produced quality engine parts and have supported successful component

durability demonstrations. Previously, precision silicon-nitride parts were made by hot isostatic pressing, which is impractical for large-volume production. Today, laboratory-scale parts can be produced in 10 minutes using a gelcast process, a much better time frame than the 8 hours required by the previous procedure. However, yield and consistency for this process have yet to be established. Some silicon-nitride components have been manufactured using this process and have been installed on an experimental basis in engines running at metal temperatures (approximately 2,000¹F), primarily to improve wear. A silicon-nitride diesel engine turbocharger rotor for use in automotive engines has been in large-scale production by Kyocera for several years.

Two turbine engine manufacturers reported combustor tests results confirming an ability to meet California ULEV standards, and two other manufacturers are pursuing catalytic combustors with still lower emission potential. Thus it appears that emissions will probably not represent a significant barrier. Three turbine engine manufacturers developed air bearings. They reported engine or dynamic rig testing results that meet engine requirements. This indicates that the cost and parasitic losses associated with conventional oil lubrication systems can be eliminated. Although air bearings are not an enabling technology, they do represent a very desirable feature for minimizing engine cost and complexity.

Heat-recovery devices have also shown considerable progress. In particular, the Corning extruded design appears to offer a low-cost alternative, and its 500-hour test has begun to demonstrate the required durability and performance capability. An effective, low-cost, heat-recovery device is an enabling technology for the turbine, and progress in this area is impressive. However, high-volume cost, long-term durability, and performance characteristics still must be demonstrated. Adapting low-cost, automotive-type electronic control systems that can meet gas turbine requirements is also meeting with success.

Considerable progress in the last year has moved the gas turbine closer to being a successful PNGV engine candidate. Gas-turbine manufacturers project that they can meet or approach the PNGV weight, volume, and efficiency goals. They also believe turbo-generators can meet other constraints, including operability, emissions, noise and vibration, and durability. These projections have not yet been validated, but given sufficient time and resources, they may be realized.

A ceramic turbine is necessary to meet or approach the efficiency goals. Even though impressive progress has been documented with the silicon-nitride gelcast approach, it is still not clear that this process can meet automotive parts-manufacturing requirements and costs. Considerable time will be required to

demonstrate durability and repeatability of manufacture. Also, the efficiency and durability of such turbines at the required temperatures have yet to be demonstrated. Another major factor relative to the PNGV time frame is that there are no testbed engines for testing and integrating into the vehicle. A workable design and workable components are not enough to demonstrate effective system integration in a vehicle.

In summary, much progress has been made, but even the most optimistic projections for the required development time puts the gas turbine well beyond the time frame for PNGV concept vehicle demonstrations.

Based on its review of the program status and progress for gas turbines, the committee makes the following recommendations.

Recommendation. Development of the gas turbine engine should continue, and acceleration of the basic technology and systems and design optimization efforts should be considered. Some of the deficiencies and inconsistencies among the developers indicate that a better technical focus in areas such as bearings, overall system optimization, and manufacturing goals should be defined.

Recommendation. A testbed for the gas-turbine engine should be made available for vehicle integration studies.

Recommendation. Potential goal 1 and 2 applications of some of the gas-turbine engine technologies, such as oil-less bearings, gelcast parts, and extruded ceramics, should be explored for other automotive applications.

The Stirling engine is an external-combustion, reciprocating-piston engine. This power plant has been under development for more than 40 years and has been used successfully for stationary solar to electricity energy conversion, as well as for some submarine and satellite applications. It was selected by General Motors for development as part of the DOE HEV program. This engine is expected to have performance, size, weight, and emission parameters similar to those projected for the PNGV ceramic gas turbine. Stirling engines are typically very quiet and are relatively vibration free. The engine being used by General Motors was built by Stirling Thermal Motors, a company with extensive experience with engines of similar size for use in solar electricity generation. The uncertainty about having ceramic turbine components in time to meet deadlines of the DOE HEV program, together with the higher predictability that the Stirling Thermal Motors design can be successfully adapted, were the primary factors in General Motors' decision to continue engine development.

A 40-kW, four-cylinder, swashplate Stirling engine and alternator unit was delivered to General Motors in September 1996, and preliminary testing has begun. Packaging has not been optimized, and this initial unit does not meet specific power targets for a PNGV vehicle. Although steady-state emissions are expected to be very low (the burner is very similar to that for a gas turbine), little is known about emissions during start-up. Thermal efficiency is expected to meet the DOE HEV requirements, but the engine-head temperature would have to be increased or the heat-rejection temperature decreased to approach the requirements for a PNGV vehicle. The Stirling engine cycle efficiency is very sensitive to the temperature available for heat rejection, making radiator effectiveness a key system design parameter.

Stirling engines typically operate with either hydrogen or helium as the working fluid that is sealed within the engine. Performance is substantially better with hydrogen, which is currently being used in the General Motors engine; however, the use of hydrogen increases the difficulty of eliminating leakage past seals and from permeation through hot metal surfaces.

This program is at a relatively early stage of development. Much will be learned about the potential of the Stirling engine for possible PNGV application through the current hybrid program. Cost, manufacturing hurdles, and durability for automotive applications are not well understood at this time. Because this is an external-combustion engine, the performance of the five heat exchangers (heater head, regenerator, air preheater, gas cooler, and radiator) creates the most uncertainty. Cold-start emissions also need to be determined. The HEV program will provide a benchmark for the performance of these components, as well as an assessment of the critical issue of working-fluid containment. The potential of this power plant in the PNGV program will be much easier to assess when this information is available.

Based on the review of the program status and progress for Stirling engines, the committee makes the following recommendation.

Recommendation. The PNGV should review results from the General Motors/DOE hybrid vehicle program on an ongoing basis, and the potential of using a Stirling engine in a PNGV prototype vehicle should be assessed through appropriate vehicle systems, modeling, and packaging studies.

Fuel cells hold the promise of converting fuel to power for vehicle propulsion much more efficiently and with lower emissions than combustion engines. The proton-exchange-membrane fuel cell (PEMFC) is the best candidate among established fuel cell technologies for a hybrid vehicle. It can operate at about 80¹C and its efficiency, specific power, and power density (operating on a hydrogen fuel) approach the goals of the PNGV program. Like other fuel cell technologies, it operates best on hydrogen as a fuel. However, because an infrastructure for the production, storage, and distribution of competitively priced hydrogen is not likely to be developed in the foreseeable future, the strategy of the PNGV program is to convert hydrogen on board the vehicle to hydrogen for use by the fuel cell. Pollutant emissions from PEMFCs should be extremely low because high-temperature combustion processes are not involved and the fuel (e.g., gasoline) must be processed into a hydrogen-rich fuel stream with extremely low levels of carbonaceous compounds (except carbon dioxide) to be compatible with PEMFC electrocatalysts. However, the requirement that fuel be converted completely and with very high thermal efficiency poses difficult technical challenges for the fuel-processing subsystem, especially if gasoline is the primary fuel. Incompatibility of fuel cell electrocatalysts with CO in the output fuel stream from the fuel processor remains an important issue. An electrocatalyst that yields a higher fuel cell efficiency with lower noble metal loadings but retains high activity in the presence of CO from the reformer would remove a significant barrier to progress. Subsystem integration and plant engineering (i.e., water, thermal, and air management, and operation under pressure) also require extensive development efforts. Major advances are needed in manufacturing at the component, subsystem, and system levels to reduce greatly the weight, volume, and especially the cost of fuel-cell power plants if the goals of the PNGV program are to be met.

The PNGV targets for 2004 include the following: energy efficiency of 80 percent for the fuel processor and 57 percent for the fuel-cell stack; power densities of 0.75 and 0.50 kW/L for the fuel processor and stack, respectively; and costs of $10 and $35 per net peak kW, respectively. These goals do not include the volume or costs of energy storage or the power-conditioning subsystems. In addition, a durability of 5,000 h is sought, along with a warm-up time of about 1 min and a transient response to changes in load within 10 s for the fuel processor and 1 s for the stack. To meet durability and efficiency goals with currently available electrocatalysts, the CO content of the fuel stream passing from the fuel processor to the stack is targeted to be less than 10 ppm at steady state and less than 100 ppm for transients.

The achievement of efficient, compact on-board fuel processing is one of the most challenging problems in developing practical automotive fuel-cell power plants. Successful development requires low CO content (preferably below 1 ppm) to avoid poisoning the anode catalyst of the fuel-cell stack; low processing temperatures to minimize warm-up time, thermal-insulation requirements, and thermal inefficiencies that arise when fuel and hydrogen-rich product streams are heated and cooled; high degrees of thermal integration and fuel utilization to avoid compromising the overall energy efficiency of the fuel-cell system; and reduced capital costs. These challenges are greater if gasoline is the fuel. Although fuel processing of methanol requires temperatures in the range of 200¹C to 300¹C, gasoline reformers will need to operate at about 700¹C to achieve practical conversion rates while avoiding deleterious deposition of carbon on the catalyst.

Up to now, the PNGV fuel-processor development efforts have taken advantage of the earlier development of reformers and fuel processors for alcohol fuels under the DOE HEV program. Now part of the PNGV program, these developments include a successful laboratory prototype (6-kW equivalent) of a single reactor combining partial oxidation and steam reforming of methanol (at Argonne National Laboratory) and a complete 50-kW equivalent ethanol fuel processor comprising reactors for partial oxidation, steam reforming, and CO shift to hydrogen and carbon dioxide (at Arthur D. Little, Inc.). These fuel processors appear capable of meeting (or even exceeding) PNGV goals for specific power and power density, and they are projected to meet cost goals when operating on alcohol fuels. However, infrastructure constraints have led the PNGV to stipulate gasoline as the primary fuel. Gasoline is more difficult to process than alcohol fuels because of its more complex composition and chemistry and because it contains potential catalyst poisons such as sulfur. Efforts at Argonne National Laboratory and Arthur D. Little to process gasoline are now beginning. The efficacy, efficiency, and durability of these fuel-processor technologies when operated with gasoline are yet to be established, as is their effectiveness in reducing CO to the required levels, and the dependence of the fuel processor efficiency on part-load and transient conditions.

In programs funded under the DOE HEV program, now part of PNGV, and by the OEMs outside of the PNGV program, several fuel-cell power plants in the 10-kW to 30-kW range have been developed, built by subcontractors to Chrysler, General Motors, and Ford, and operated with methanol or hydrogen fuels.

Typically, these prototype units fell about 40 percent short of the PNGV goal for power density and a few percentage points short on efficiency. Work at Los Alamos National Laboratory (LANL) on the important electrocatalyst poisoning issue shows that small amounts of air (about 2 percent) added to the fuel stream fed to the electrochemical cell can raise the CO tolerance level to about 100 ppm. Although it compromises the overall efficiency somewhat, this approach may make the requirements of the fuel-cell stack compatible with the capabilities of the fuel processor. Because CO tolerance is an electrocatalyst property, exploration and development of improved fuel electrode electrocatalysts for CO tolerance should continue to be pursued albeit at a low level given the high technical risk. This work should proceed in parallel with alternative approaches, such as the LANL preferential-oxidation technique, to achieve CO levels at the electrocatalyst surface sufficiently low to be compatible with the limitations of present materials and technology.

Currently used proton-conducting membranes perform adequately, and thinner membranes of lower resistance are becoming available. Membrane costs are high, but this is not considered as critical as achieving an acceptable overall efficiency and producing a system that can meet performance targets. The need to develop lower-cost materials and fabrication techniques for bipolar plates is being addressed.

In addition to the fuel processor and the cell stack, auxiliary systems are needed for water management, thermal management, fluid-flow control, and safety. Appropriate water-transport management and hydration control are critical to membrane performance and life. Even in a fuel cell power plant, more than 50 percent of the fuel energy is released and must be removed as heat. These needs translate largely into engineering and design problems. These should be put on a schedule and priority commensurate with the time needed to deal with the most critical issues, which include catalyst poisoning, fuel-processor performance on gasoline, and overall system efficiency.

The PNGV cost goal for the entire fuel-cell power plant is $50/kW. Because commercial phosphoric acid fuel cells presently cost more than $1,000/kW on a larger, less constrained scale, major efforts will have to be made to find low-cost component materials and to develop low-cost manufacturing methods for PEMFCs. Substantial cost reduction can be anticipated to accrue from high-volume production using automated processes (for fabricating key components like bipolar plates, membranes, and electrodes, and for assembly of cell stacks), but reliable overall cost estimates must await the resolution of the costs of

materials for the aforementioned components necessary to achieve efficiency and durability for operation with gasoline as the fuel, and the conceptualization and cost estimation of manufacturing practices for all key components and subsystems. It is not clear to the committee whether and how the cost goals can be achieved because no analyses of materials, components, or manufacturing costs were presented to the committee.

Ballard Power Systems, Inc., in Canada, is a leader in developing PEMFC technology for vehicles. Ballard is collaborating with General Motors on the PNGV program and with Daimler Benz (in Germany) to develop power plants for minivans and passenger cars. Ballard also has a program to develop power plants for buses. All of these systems use either pure hydrogen or hydrogen produced from the on-board processing of methanol.

The Toyota Motor Company recently demonstrated a PEMFC-powered vehicle using hydrogen, stored onboard as a hydride, as the fuel. Toyota has also investigated electrocatalysts (Pt-Ru) for CO tolerance. Siemens in Germany and De Nora in Italy have also designed and constructed PEMFC power plants (rated at 1 kW to 10 kW) for defense and vehicle applications. The existence of fuel-cell development and demonstration programs does not mean that fundamental issues—from gasoline fuel-processor performance to electrocatalyst and membrane performance to overall system efficiency and cost—have been resolved. At the present time, U.S. research laboratories are in the forefront of fundamental and applied research investigations to advance PEMFC technology, and International Fuel Cells, a subsidiary of United Technologies, appears to be only behind Ballard in technology development and demonstration.

PNGV and worldwide progress in automotive fuel-cell technology have been impressive in a number of important areas, especially in improving the power density of the fuel-cell stack and in lowering the projected costs. Nevertheless, PNGV's PEMFC technology targets are unlikely to be technically achievable by 2004 and cannot be expected to be cost-competitive with state-of-the-art automotive power source technology before 2010 to 2015.

By 2004, the PNGV technical goals for fuel-cell-stack power density and specific power could be met by using thinner membranes and bipolar plates and by paying careful attention to stack design and construction. No fundamental scientific breakthroughs will be needed. A substantial breakthrough will be required to obtain electrocatalysts with long-term stability, low noble-metal content, and high CO tolerance. However, this must be recognized as a very high-risk endeavor. In the absence of electrocatalyst breakthroughs and considerable

reduction of the production costs for some key components (e.g., bipolar plates), the inherent cost of the system will remain high, and the overall efficiency may not meet PNGV goals. Durability of key components and subsystems has not been demonstrated under conditions of anticipated use, including, most importantly, use of gasoline as fuel. Refinement and demonstration of gasoline fuel processor technology are required.

A high degree of system integration will be needed to realize the fuel cell's potential for higher efficiency compared to heat engines. For example, if the fuel processor has an 85 percent thermal efficiency, the fuel stack reaches 55 percent efficiency, the energy-storage (battery) system returns 95 percent of its input energy, and conversion of electrical to mechanical energy reaches 81 percent efficiency, the fuel-cell system would have an overall efficiency of only 36 percent. This is below the projected values of 37 percent for gas turbines and 43 percent for CIDI engines. Therefore, one important conclusion is that fuel-cell stack and electrical/mechanical conversion efficiencies need to be targeted for significant improvements if fuel cells are to offer meaningful efficiency advances.

Based on its review of the fuel-cell program and status, the committee makes the following recommendations:

Recommendation. Development of automotive fuel-cell technology should be continued because of the inherent potential of fuel-cell power sources for very high efficiency and near-zero emissions. However, PNGV program emphasis in the coming years should be on achieving major advances and/or breakthroughs in areas critical to achieving high efficiency, long life, and low manufacturing costs rather than on prematurely demonstrating prototypical power sources with less-than-competitive performance.

Recommendation. A consistent systems analysis should be conducted that (1) determines the overall efficiency of vehicles using a fuel cell with a gasoline reformer and compares it to vehicle power sources using other advanced energy-conversion systems, (2) clearly identifies areas where significant improvements are needed, and (3) establishes realistic targets and the most promising technical approaches to attain these targets.

Recommendation. Because a high-performance, compact, efficient, low-cost gasoline fuel processor is the key to automotive fuel-cell applications in a gasoline-fuel scenario, high priority should be given to advancing fuel-processor technology, validating it in an engineering prototype (e.g., 50-kW), and integrating it with a prototype fuel-cell stack.

Recommendation. Electrocatalyst, membrane, and bipolar-plate durability tests should be run on small cells with simulated gasoline-reformate fuel to determine

the prospects of retaining high efficiency, electrocatalyst specific activity, and tolerance to poisons and degradation processes over the longer term.

Recommendation. Efforts to develop more cost-effective materials and high technology automated techniques for manufacturing and assembling key fuel-cell components, such as bipolar plates, membranes, electrodes, end-plates, and sealants, should have higher priority than costly and possibly premature efforts to demonstrate integrated fuel-cell power plants in parallel efforts.

Batteries, especially some of the advanced batteries currently under development for electric vehicle propulsion, have the potential to meet the energy storage requirements of HEVs. The PNGV Technical Roadmap compares HEV requirements to those set for electric vehicles (see Table 4-3).

The HEV targets for battery specific energy (60 Wh/kg) and energy density (75 Wh/l) exclude lead-acid and nickel-cadmium batteries from consideration. Advanced batteries are being developed in the U.S. Advanced Battery Consortium (USABC) program and in European and Japanese industrial and government

efforts, but only the nickel-metal hydride and lithium-ion systems appear to have potential for attaining hybrid targets and meeting the PNGV time table. However, neither the capabilities of these battery systems for meeting the specific power and power density requirements nor the number of shallow cycles delivered had been established when the PNGV high-power-battery development program began.

The PNGV Technical Roadmap correctly identifies achievement of high power-to-energy ratios, long cycle life, retention of cell balance in a series-connected multiple battery, very high charge-discharge efficiency, and affordability (low cost) as the major development challenges for the PNGV battery program. The road map also indicates that developing a battery technology to meet these objectives will require significant, and possibly major, modifications in current battery designs and manufacturing methods.

Because of delays in developing the PNGV systems analysis and modeling programs (see Chapter 3), definitive energy-storage subsystem requirements for hybrid batteries based on appropriate trade-off studies have not been established. The road map does, however, recognize the differences in battery requirements imposed by power plants with significantly slower response times to power demands than conventional automobile engines. Batteries designed and developed for ''slow-response" power systems may not be able to meet the "fast-response" requirements and vice versa. Table 4-4 shows these requirements. This distinction expands the range of capabilities for HEV batteries and increases the number of battery types that might qualify for at least one of the hybrid applications. Even lead-acid and nickel-cadmium batteries might eventually be able to meet the "fast-response" requirements.

Between January and May 1996, the PNGV awarded Phase 1 contracts through USABC for initial exploration of three different battery types. Four organizations received contracts: SAFT (conventional lithium-ion); Yardney and VARTA (nickel metal hydride); and SRI International (novel lithium-ion concept). All of the technologies except the SRI lithium-ion technology are well developed and involve established developers and manufacturers.

The SAFT technology meets all major HEV battery target performance goals except for cost, including goals for energy density, power density, and cycle life, at the 0.85-Ah cell level. The project is on schedule and within budget. Capability was demonstrated for 120,000 shallow cycles with less than 20 percent capacity loss and, more importantly, with only a small increase in cell impedance. Also, basic feasibility was established for scaling the technology up to the 10-Ah cell size envisioned for HEV applications. Refinements in electrical and thermal design and in charging control are still needed, and safety under all conceivable operating conditions must be demonstrated. Safety requirements will be the focus of a follow-on program to develop a 50-volt module. Priority objectives include analyzing failure modes and effects and developing a potentially low-cost electric/electronic battery-control system from existing technology. A key issue is whether cell-level current control is likely to be required (as for consumer-product and electric vehicle batteries) or whether a combination of quality control in cell manufacturing and the shallow charge/discharge hybrid-battery duty cycle (around intermediate states of charge) will permit less elaborate and, thus, less expensive control strategies and equipment.

Achieving the PNGV production cost goal of $150/kWh for a 3-kWh battery will be extremely difficult. Several developers project that electric-vehicle versions of lithium-ion batteries (less expensive than HEV versions because they are larger in scale and lower in specific power) will cost more than the PNGV target even if produced in volume (1 to 2 million kWh/year). SAFT currently projects a cost well in excess of the target for 3-kWh lithium-ion hybrid systems that include adequate provisions for electric and thermal management. The main uncertainties and risks regarding the timely availability of prototype 50-volt modules arise from the required combination of high levels of safety and the potential for low production costs.

Initial explorations of the potential of nickel metal hydride battery technology for hybrid duty are nearing completion, but only very limited performance and cycle life data are available. To date, the Yardney program (started in April 1996) has demonstrated that small (0.44 Ah) test cells can deliver 10,000, one-minute pulse discharges and that the cell columbic efficiency promises to meet the PNGV target at low states of charge. However, the power density projected for a battery using this technology is only 0.33 kW/kg. The VARTA program, begun in May 1996, has demonstrated capability for more than 60,000 shallow cycles, but other performance results were not available for committee review.

The nickel metal hydride system should be a good candidate for developing high-power batteries because electric-vehicle versions have already demonstrated energy densities approaching those of lithium-ion as well as comparable peak-power capability and deep-cycle life. Cell and battery safety are more easily achieved with nickel metal hydride than with lithium batteries because of less reactive battery materials and the ability of cells to sustain overcharge safety and without damage. Meeting efficiency and cost goals will be more difficult, however, because of the nature of the electrochemistry (including the relatively low cell voltage) and the materials involved.

It is not clear whether the current programs will be able to demonstrate battery performance and cycle life close to those needed to meet PNGV goals. It is also unclear whether the program's nickel metal hydride batteries will approach, much less exceed, the very promising performance reported by DAUG (a battery-development organization owned jointly by Daimler-Benz and Volkswagen in Germany) for prototype, full-scale nickel metal hydride cells developed for hybrid applications. Cost remains a major concern. The inherently less expensive electric-vehicle nickel metal hydride batteries are projected to cost at least $250/kWh to $300/kWh unless there is a materials breakthrough. Therefore, meeting the PNGV goal of $150/kWh seems quite unlikely.

The SRI concept differs from the concept of other lithium-ion cells. SRI uses a noncarbon anode (negative electrode) host material and a nonflammable solvent base for the electrolyte. These features and the nearly 2 V less-negative

anode potential should considerably reduce safety concerns although at the expense of reduced energy-density and power-density potential. Since June 1996, SRI has been fabricating and testing small laboratory cells. The ability of these cells to deliver and absorb 10 s current pulses at the nominal 15 C rate (discharge of the nominal capacity in 1/15 of an hour) has been established, but the battery power density projected from cell data appears limited to about 0.4 kW/kg. No cycle life information was made available.

Even after the current SRI program has been completed, this battery technology will be in a much earlier development phase than other systems. Key cell and battery materials, design, engineering, manufacturing, safety, and system aspects are not yet developed to the point where a prototype program could be initiated with a definable probability of success. If results of the current program indicate that performance, cycle life, and cost requirements might ultimately be met, the next step should be designing, fabricating, evaluating small engineered cells. Availability of SRI production prototype battery modules in time for the Goal 3 prototype vehicle appears unlikely. Also, because larger amounts of active materials and larger electrode areas are required to achieve the same energy and power levels as more conventional lithium-ion batteries, the SRI technology is likely to have higher production costs.

The PNGV battery program has benefited from DOE's long-term support of battery R&D and from the more recent, major commitments by DOE and its industrial partners to the USABC electric-vehicle battery program. As a result, the PNGV can draw on the services of a substantial number of capable industry and government people with relevant technical and management experience to define and direct its high-power battery development program. In addition, the PNGV had a "running start" because the program was based on battery technologies with proven promise and battery developers who performed well in the USABC program. Together with the battery R&D and testing capabilities established by DOE at several national laboratories, this combination of government and industrial capabilities is more than adequate to define, manage, and technically support the PNGV battery program.

The PNGV criteria provide a reasonable guide for identifying candidate HEV battery systems. With the identification of "slow" and "fast" primary vehicle powertrain response modes, a useful refinement of the criteria has been made. However, the impact of this refinement is not yet reflected in the program's technical strategy. Given that the vehicle systems analysis and its link to the storage subsystem model are not yet fully developed, the PNGV storage system criteria

appear to be based on somewhat arbitrary and restrictive assumptions for performance requirements. Therefore, these criteria should be used as guidelines rather than as rules for selecting candidates.

The importance of PNGV's modeling and systems analysis and its impact on battery/storage system selection, development, and integration, argue for a continuing, systematic refinement of modeling and analysis of batteries as integral parts of hybrid vehicle power systems. However, the PNGV Phase 2 battery program plan does not appear to include a corresponding line item, and it is unclear whether adequate resources will be available for modeling, systems analysis, and high-power battery testing. Adequate resources for these critically important activities are essential.

Testing developmental and prototype batteries in ways that realistically simulate the operating conditions of batteries in hybrid power systems will become an essential part of the PNGV battery program when prototype cells and modules are available not only from the PNGV but also from other development programs. The testing methods and plans presented during the committee's review seem appropriate for this task, but no corresponding task and budget line item was shown in the program plan.

The extensive PNGV compilation of worldwide advanced battery development organizations and technologies does not identify any programs aimed specifically at hybrid applications. Leading international battery developers such as SAFT (France), VARTA (Germany), and Matsushita/Panasonic and Sony (Japan) undoubtedly are exploring high-power versions of their systems. However, no major program or product has emerged. One possible exception is DAUG's successful development of a 14-Ah nickel metal hydride high power cell technology, but DAUG does not appear to have immediate plans for commercializing this technology. Thus, the hybrid/high power battery developments funded under the DOE HEV program and PNGV represent leading-edge efforts to achieve the challenging technical targets for power density, shallow cycle life, and high "round-trip" efficiency with potentially affordable battery systems. If the PNGV battery program in Phase 2 results in prototype modules that meet the proposed technical targets, the United States almost certainly will have the lead in hybrid/high power batteries.

PNGV funds for the Phase 1 efforts undertaken in 1996 appear to have been adequate, although relying on a single contractor to explore conventional lithium-ion technology is somewhat risky. Funding for the SAFT, Phase 2 lithium-ion module development effort appears adequate, mostly because of SAFT's large (50 percent) cost share. It is not clear whether or not sufficient funding will be available to fund appropriate follow-on development efforts at Yardney, VARTA, or alternative developers. Testing and critical evaluation of several nickel metal hydride technologies against refined PNGV targets probably could be accomplished relatively quickly and with a modest budget. Similarly, an effort to establish the advantages and limitations of the SRI technology on a

meaningful scale of technology could be carried out with a modest budget. These actions should considerably enhance the program focus and the probability of success in Phase 2.

Assuring a high probability that the most promising battery systems are selected and production prototypes are developed that meet the PNGV performance, cycle life, safety, and cost goals will require substantially higher funding levels in the future. It seems likely that a meaningful share of the funding required for each system (on the order of $60 million to $150 million for the critical prototype and pilot-plant phases) will have to be provided by PNGV over the next five to seven years.

Based on its review of battery-development programs, the committee makes the following recommendations.

Recommendation. Development of the high-power lithium-ion battery should continue until the prototype module level is reached, with early emphasis on ensuring safety under all foreseeable conditions. The control requirements for safe operation of modules and batteries in the hybrid mode should be determined, and development of potentially low-cost electric and thermal control systems should be initiated.

Recommendation. The ongoing exploration of high-power nickel metal hydride batteries should be completed. Based on data and test data from promising nickel metal hydride batteries available from other sources, the PNGV should determine whether this technology offers advantages over lithium-ion batteries in hybrid applications and how these advantages might be captured for the PNGV Goal 3 vehicle.

Recommendation. Modest efforts should be supported to explore the potential of other battery systems (such as the SRI concept) that show promise of providing superior performance, safety, enhanced cycle life, and/or lower cost. This support should be provided at least until the capabilities of the lithium-ion and/or nickel metal hydride batteries to meet PNGV goals can be predicted with confidence.

Recommendation. Continued efforts should be supported to (1) develop storage subsystem and vehicle subsystem models that provide realistic performance, cycle life, and cost targets that establish a capability for conducting trade-off analyses; and (2) establish a testing methodology and a physical capability for evaluating high power batteries for hybrid electric vehicles both from within the PNGV program and externally.

Flywheels have very attractive power-to-weight and power-to-volume characteristics of the "core" system (compared to batteries) both in delivering power and in accepting the high power developed during vehicle braking. However, serious problems related to safety, cost, and size must still be solved.

A detailed vehicle systems analysis, with appropriate vehicle subsystem trade-offs, has not been performed to define flywheel requirements for a PNGV hybrid vehicle. Nevertheless, the PNGV flywheel technical team has assumed parameters in order to specify its application to vehicles designed with either a fast-response or slow-response power plant. The fast-response power plant is assumed to react very much like a conventional automotive engine, responding very quickly to vehicle power demands. This type of power plant places the least demand on the flywheel system. The slow-response engine puts a much greater demand on the flywheel for the delivery of instant high power. This demand requires a much larger and more costly flywheel system. The flywheel objectives metrics for both types of power plants are shown in Table 4-4.

There is a high probability that the design performance objectives shown in Table 4-4 can be achieved for the basic flywheel system because material specifications and performance characteristics are well known. However, because of the potential for catastrophic flywheel failure, safety considerations require that a containment structure be included. Requirements for this structure have not yet been determined; therefore, the ultimate cost, volume, and overall system weight cannot be determined at this time.

The PNGV flywheel technical team has put together a mission statement and a technical plan. Objectives of the technical plan include developing automotive-applications requirement guidelines and action plans for testing, identifying high-leverage issues, developing a scalable model, and reviewing the feasibility of other mechanical energy storage systems (e.g., hydraulic) and progress documentation. The team is also monitoring flywheel design and testing efforts being pursued for other applications. For example, the University of Texas Center for Electromechanics at Austin is developing a flywheel system for a bus with AlliedSignal and another project, funded by the Advanced Research Projects

Agency (ARPA) and the Houston Metropolitan Transit Authority, to develop a safety containment system.

Program guidelines and vehicle test data are not expected to be available until the end of 1997. The action plan for the test protocol is scheduled for completion by May 1997. The University of Texas Center for Electromechanics plans to mount a flywheel system on an advanced technology transit bus by the end of 1997, and they will complete a substantial number of flywheel burst-containment tests by May of 1997. The Federal Railroad Administration is testing composite flywheels that are one-third of the full design size making them comparable to the bus requirements. The ARPA flywheel safety program will provide a fundamental understanding of composite flywheel behavior and will develop rational means for developing safe designs. Containment of failures will be pursued through (1) careful design to avoid flywheel failure, (2) detection of initial failure potential and shutdown of the flywheel system to minimize failure loads, and (3) barriers to contain and mitigate potential damage.

The PNGV flywheel technical team believes it is unlikely that flywheel subsystems will be able to support the first PNGV concept vehicles; but, if estimated funding of $1.3 million for Fiscal Year 1997 is made available, they should have a laboratory-scale subsystem by the end of 1997. The committee concurs with the team's assessment, while noting that other activities have produced significant data on flywheels systems that can assist in the downselect decision. These activities include the Trinity model flywheel from Lawrence Livermore National Laboratory, the Unique Mobility flywheel contracted by Ford Motor Company, and the Satcon work on the flywheel for the Patriot Vehicle at Chrysler. Containment cost and weight remain significant issues that are, as yet, undefined in terms of total vehicle system requirements.

Based on its review of the program and progress for flywheel development, the committee makes the following recommendations.

Recommendation. After appropriate vehicle system trade-off studies have been conducted, performance objectives should be created that satisfy the requirements of the fast-response power plant vehicle system, and a plan should be developed for evaluating and integrating a flywheel subsystem in post-2000 concept vehicles.

Recommendation. The ARPA comprehensive flywheel failure containment plan should be pursued, including collecting burst/collision failure test data from all available sources.

Ultracapacitors are in the same category as flywheels—they are compatible only with the power-assist (fast-response engine) type of HEV configuration. The technical targets for ultracapacitors are essentially the same as those of the flywheel. The goal of the projects, sponsored mainly by DOE, is to develop high-power ultracapacitors that meet or exceed the energy storage requirements for fast-response engines.

Several laboratories in the United States and other countries are examining the prospects for using ultracapacitors to provide pulse power for military applications, but there is little evidence of R&D abroad of ultracapacitors for use in the HEV. However, carbon-based, aqueous electrolyte ultracapacitors have been developed and commercialized in Japan for other applications, such as providing instantaneous power to high-speed computers during power failure.

The ultracapacitors being developed for the HEV are of the electrochemical type, the active material being high-surface-area carbon, a noble metal oxide, or a conducting polymer. Of these types, ultracapacitors using carbon in an aqueous electrolyte are the most highly developed. However, the cell potential for this type of device is only 1 V, compared with 3 V for a cell with an organic electrolyte. General Electric Company is developing ultracapacitors of the latter type for the Ford HEV program. Based on the testing of unpackaged single cells, they project the following performance characteristics at 2.75 V: specific energy of 4.1 Wh/kg, energy density of 5 Wh/L to 7 Wh/L, specific power (to half maximum voltage) of 1,100 W/kg, and power density of 1,500 W/L. Maxwell Laboratory, Inc. is developing this same type of cell and has built 24-V modules and determined temperature effects and cycle life performance. At a constant power (1,100 W/kg) the performance was found to be better at a higher temperature, but there was a problem with an increased self-discharge rate. There is also some loss in energy after about 60,000 charge/discharge cycles. Sixteen cell modules (electrode area 200 cm²) are being built.

Ultracapacitors using noble metal oxides and conducting polymers have the advantage of achieving a high specific power (2,000 W/kg to 4,000 W/kg); their drawback is having lower values of specific energy. Ultracapacitors with p-doped and n-doped conducting polymers, have demonstrated a value of 10 Wh/kg (the

PNGV goal for 2004 is 15 Wh/kg) for the specific energy. Los Alamos National Laboratory has demonstrated a very high specific power (about 10 kW/kg) with devices of this type. The Army Research Laboratory at Fort Monmouth, New Jersey, has obtained promising results with a new electrolyte, which is benign and exhibits a high specific conductivity, high solubility, high electrochemical stability, and the ability to work satisfactorily over a wide range of temperature and Small (-40¹C to 70¹C). In addition, Small Business Technology Transfer and Small Business Innovation Research Phase I awards were made in 1996 for innovative research on electrode structures, "wet" electrochemical ultracapacitors, and solid-state ultracapacitors, including thin-film dielectrics.

Ultracapacitor projects are in an early stage of research and development. Their only potential applicability in the PNGV program is in conjunction with fast-response engines. The current programs are behind schedule, and the milestones for 1996 have not been reached. Developing ultracapacitors for energy storage in an HEV is a high risk, high payoff project. It is unlikely that the ongoing projects will be able to meet the PNGV timing, performance, and cost goals for the following reasons:

• Even though the specific power and power density could reach high values, the energy density is likely to remain low.

• For this application, the ultracapacitors must have a time constant of less than 100 ms. The high capacity needed for a high specific power and a relatively high energy density makes it very difficult to have such a low time constant because the equivalent series resistance will have to be of the order of a few ohms.

• The current costs of materials and fabrication processes are very high; to meet the PNGV cost goal for this energy storage device ($150/kWh), these costs will have to be reduced by two orders of magnitude.

• Self-discharge rates, particularly of the carbon ultracapacitors, are relatively high; although 100,000 to 200,000 cycles have been demonstrated in small cells, there is a significant loss in capacity after about 50,000 cycles.

The prospects for developing batteries (particularly the high performance nickel metal hydride and lithium-ion batteries) that meet the PNGV technical goals for energy storage, both for the slow-response and fast-response engines, are excellent. The ultracapacitor technology will require at least 10 times the current level of financial support for a period of 10 to 15 years to reach the same state of development. A considerable amount of research is needed just to identify the type of ultracapacitor that might attain the performance level required for an energy-storage device for the HEV. It is premature to carry out scale-up and

cell-stacking tasks for the HEV application. Because ultracapacitors may be used in other applications, for which there are ongoing investigations, the PNGV should follow the progress of these programs and then analyze the needed research, development, and demonstration for the HEV at a later date.

Based on its review of ultracapacitor programs and progress, the committee makes the following recommendations.

Recommendation. PNGV should conduct appropriate systems studies to determine the prospects for ultracapacitors in HEV applications in comparison with high power batteries and other energy storage devices, such as flywheels.

Recommendation. Work on ultracapacitors for HEV applications should be limited to basic and applied research studies at universities, national laboratories, and industrial R&D centers and should be directed toward making fundamental advances and breakthroughs.

All of the PNGV vehicle configurations involving electric motor drives, energy recovery, flywheels, or fuel cells require electric motors/alternators, power electronic inverters, and sophisticated electronic controllers. There are other vehicle subsystems and accessories, such as power steering and heating, ventilation, and air conditioning (HVAC) systems, that also require electric power control. During its second review, the committee noted that the PNGV Technical Roadmap demonstrated a good awareness of the state of the art for these devices and had established appropriate and challenging targets for performance, efficiency, weight, and cost. The milestones in the PNGV Technical Roadmap show the overall electric driveline efficiency improving from today's estimated 70 percent to 80 percent in 10 years, and the weight decreasing by 47 percent. The cost of the electronic converter/controller must be reduced by 85 percent, and the cost of the electric motor must be reduced by 80 percent. The committee commented that these are very ambitious goals. The committee did not conduct an in-depth review of the technologies during its second review and did not make specific recommendations.

The following assessment is based on presentations on October 10, 1996, by the PNGV electrical and electronics power conversion devices team to the committee

subgroup on electrical and systems analysis. The PNGV-USCAR partners have varying degrees of experience in developing and producing power electronic converters and controls, and electrical motors and alternators of the type necessary for a Goal 3 vehicle. They are actively exploring various types of electric motor configurations, including induction, permanent magnet, switched reluctance, and synchronous reluctance. It was stated to the committee that power electronic converters available for demonstration today are approximately twice the weight required to meet the needs of a Goal 3 vehicle. Members of the team have some recent experience working with qualified semiconductor manufacturers in the development and fabrication of insulated-gate bipolar transistors (IGBTs) for high-power applications and a very preliminary understanding of the producibility of IGBT power converters and their associated production costs.

PNGV-USCAR recognizes the importance of reducing vehicle accessory loads, such as HVAC loads, to reduce overall vehicle electrical loads. Moreover, they believe that the trend toward higher accessory loads in response to consumer demand will continue. This will make the challenge of reducing accessory loads even greater by the year 2004. PNGV needs to develop plans and allocate resources to address this issue.

The PNGV-USCAR partners have made progress in understanding the effect of system requirements on the system architecture and primary components. The PNGV vehicle-defined need for 0.5-g vehicle acceleration will determine the configuration and size of the electric drive motors.

The vehicle requirements for this subsystem have not been adequately defined for all vehicle configurations under consideration. The interface with the vehicle engineering team and the system analysis team has not been effectively established. It appears to the committee that this team is relying heavily on other electric-vehicle and HEV programs to guide their efforts in developing technology for the Goal 3 Vehicle. Chrysler is using its experience with the Patriot concept HEV; General Motors is using its experience from the Impact concept electric vehicle and the production of the EV-1; and Ford is using experience from the Ecostar and other advanced vehicle studies. Chrysler, Ford, and General Motors also have government-sponsored contracts to demonstrate HEV technology in concept vehicles. These programs provide the technology base for collaboration on electric motors, power converters, and other related components. It must be pointed out that these applications can provide background, but the PNGV requirements demand very specific design and development considerations in order to meet the performance targets.

The DOE-funded HEV programs require the three OEMs to explore alternative HEV architectures. These include both parallel and series hybrid vehicles. Two of the partners (Ford and General Motors) selected parallel architectures for their HEV assumptions; the other partner (Chrysler) selected a series configuration.

Chrysler received its HEV contract during the past year while General Motors and Ford have had contracts for a longer period. The partners have organized

technical steering groups that share information generated from the individual HEV programs. DOE and PNGV should make sure that detailed data developed in the HEV program is universally and uniformly shared with USCAR partners on a periodic basis. The technical steering groups have identified responsibilities, and members have collaborated on identifying cost, weight, and size targets for components. The steering groups have also developed various technical performance measures for the high-priority components, and they are assessing the state of existing technology. Analytic results are emerging, and limited test data are available from the partners' electric-vehicle and HEV efforts.

The Chrysler Patriot program was targeted at a racing-technology vehicle; therefore, its requirements were quite different from those of a passenger vehicle. What was learned during development of the Patriot vehicle is useful to the PNGV program, especially in the design of the power controller. It must be noted, however, that manufacturability and cost were not major considerations in the Patriot vehicle program.

The committee also reviewed an Office of Naval Research and DOE program, known as the Power Electronic Building Block (PEBB). This program is dedicated to developing new technologies to advance the state of the art in power electronics control. The goal of the program is to reduce the cost and weight of electric-propulsion systems while achieving a high degree of functional integration, efficiency, and reliability. Semiconductor development is a major part of this program. The committee questions the direct applicability of this program to PNGV requirements. The focus is on large packages containing multiple die, which are very expensive from both a device and system-manufacturing point of view. The cost constraints for automotive design will require optimum design for manufacturing and assembly where both weight and space must be minimized.

In the course of the committee's review of this program, very little data on component and subsystem performance relative to PNGV requirements were presented. Both Ford and General Motors have made good progress in developing components that could contribute to the PNGV in their HEV programs. The new General Motors drive-motor design is an advance over currently available components, and the packaging density is impressive. These advances should lead to some design baselines and should provide needed cost and reliability data. General Motors' Impact vehicle could also provide basic cost data for a production vehicle.

The PNGV-USCAR partners identified and ranked the highest priority areas for component technology development. These include electric motors and alternators, power control electronics, HVAC, and electric/electrohydraulic power steering. High performance efficiency for these subsystems and components is vital and a high priority because they greatly affect the overall efficiency of an HEV configuration. If the HEV configuration is selected for a Goal 3 vehicle, overall system performance will be significantly affected by the design and integration of the power electronics and electrical subsystems.

Typically, accessory loads, such as HVAC, power steering, wipers, cooling pumps, instrumentation, and the antilock braking system, can increase the demand for engine power by approximately 30 percent (Malcolm, 1996). Although the team presented limited descriptive material on these systems and their impact on fuel economy, it is clear that the power required by these loads must be reduced, and this will require a well-directed active effort by the PNGV team to achieve maximum efficiency. An aggressive cost-reduction program is also mandatory. For example, the electric or electrohydraulic steering subsystem cost must be reduced by a factor of three to meet the vehicle cost objectives (Piccinato, 1996).

The group working on regenerative braking has identified a number of barriers to the development of a practical system. Of these, the efficiency of the energy storage device is critical. The battery technology now available cannot accept a high charging current efficiently. Questions concerning over-charging and depleting batteries as a storage device remain unaddressed. The efficiency goal for regenerative braking is 75 percent, without considering energy storage efficiency; current efficiency is estimated at 35 percent. There are many constraints to efficient braking energy regeneration, and recovering the energy at low vehicle speeds is a significant challenge.

The team did not review or discuss the awareness of global technology. In Japan and Europe, for example there are several electric-vehicle and HEV programs in various stages of development. The team should make an effort to capitalize on all applicable technology.

The committee concluded that very little was accomplished overall by the electrical and electronics power conversion devices team during the past year. The report given to the committee subgroup in October 1996 was virtually the same as the status reported at Vice President Gore's symposium in the spring of 1996. It should be noted that the team did not have an appointed leader for most of the year, and the committee believes the team lacks on overall sense of urgency. The PNGV Technical Roadman details four PEBB milestones. The first PEBB milestone, scheduled for 1995, has been missed. At the current rate of accomplishment, the second milestone, established for 1997, probably will be missed. The third and fourth milestone are scheduled for 2000 and 2002.

Many tasks described in the PNGV Technical Roadmap are not being addressed. For example, no trade-off format with well defined requirements has been established to support the downselect process in 1997. The committee was informed at its November 11–13, 1996, meeting that a part-time team leader had been selected. The question of adequate funding was not addressed by the team. The committee had requested that this be reviewed and can only assume that it is not an issue.

Although the team did not specifically identify barriers to progress, the committee concluded that the following major issues could constitute barriers:

• Overall driveline efficiency is currently estimated to be 70 percent. The PNGV goal is 80 percent. No data were presented to establish a confidence factor in achieving the goal.

• Today's electric/electronic subsystems, such as power steering and HVAC, and accessories, such as wipers, cooling pumps, instrumentation, and antilock braking systems, require electric power equal to approximately 30 percent of the PNGV vehicle drive power. This level of power must be reduced. Sufficient, directed technical effort is necessary to achieve the required efficiency.

• Costs of motors and alternators, the power controller, vehicle electrical subsystems, and accessories are estimated at from 50 percent to 300 percent above the individual subsystem targets. No plan of action to achieve these targets was presented.

• The committee has a major concern about the overall process of assigning attribute targets to individual components without the benefit of a top-down, overall system definition supported by a systems analysis. Optimization of overall vehicle performance can be achieved only by this approach. Evaluation of the differences among the power plants is especially important.

Based on its review of electrical and electronic power conversion device programs and progress, the committee makes the following recommendations:

Recommendation. The new electrical and electronic power conversion devices team leader should function full time and should determine how to make up for lost time and establish a schedule for the team. The team leader should identify the staff necessary for effective team performance and should commit them to the team's activities. One of the highest priorities for the new team leader should be developing interfaces with the vehicle engineering and the systems analysis teams.

Recommendation. The impact of the team's schedule slippage on the technology downselect process should be reviewed immediately by the PNGV, and plans should be made to meet schedules that support the overall PNGV effort.

Herzog, P. Future High Speed Diesel Engines for Passenger Cars. Presentation to the Standing Committee to Review the Research Program of the PNGV, National Academy of Sciences, Washington, D.C., November 12, 1996.

Kizer, T. 1996. Performance and Cost Metrics. Presentation to the Committee Subgroup on Non-electrochemical Storage Devices, University of Texas Center for Electromechanics, Austin, Texas, October 7, 1996.

Malcolm, R. 1996. HVAC Status and General Requirements. Presentation to the Committee Subgroup on Electrical Systems and Systems Analysis, the Chrysler Technology Center, Auburn Hills, Michigan, October 10, 1996.

Meurer, P. 1996. Fuel Injection Systems. Presentation to the Committee Subgroup on CIDI Engines, Ford Scientific Research Laboratory, October 22, 1996.

NRC (National Research Council). 1996. Review of the Research Program of the Partnership for a New Generation of Vehicles, Second Report. Board on Energy and Environmental Systems and the Transportation Research Board. Washington, D.C.: National Academy of Sciences.

Piccinato, P. 1996. Electrical Steering Review. Presentation to the Committee Subgroup on Electrical Systems and Systems Analysis, the Chrysler Technology Center, Auburn Hills, Michigan, October 10, 1996.

PNGV (Partnership for a New Generation of Vehicles). 1996. Technical Roadmap (draft). Dearborn, Michigan: PNGV.

Talwar, C. 1996. Lightweight Engine Structures. Presentation to the Committee Subgroup on CIDI Engines, Ford Scientific Research Laboratory, October 22, 1996.