AC propulsion’s used a 500cc Kawasaki engine rated at 20kW output to maintain their tZero’s battery at around 80% SOC. Without any battery supplementation, the device could achieve between 32-38 mpg. Though an interesting experiment, I think that this study was flawed in a few ways – or at least not completely applicable to future efforts.

1) The “trailer” concept-

The range extender was carried within a separate trailer that added its own aerodynamic and rolling resistance. AC Prop themselves noted that changing the tires of the trailer increased efficiency by 2%. By contrast, a self-contained genset would probably reduce weight, aero, and friction resistance.

2) The engine-

AC Prop chose their motorcycle engine based on compactness and power density, rather than on strict thermodynamic efficiency. Their tank-> battery efficiency was around 22%. Considering that most petrol ICE’s top out at around 25% efficiency under the very best circumstances, that would indicate an alternator efficiency of around 88%, which is fairly reasonable.

0.25 X 0.88=0.22

If an ICE designed from the ground up to be a genset could be used, one would probably expect a base thermodynamic efficiency of around 35-40%. For example, Argonne National Laboratories found the Toy prius’s engine achieves around 37% at its most efficient point of operation. Diesels regularly achieve efficiencies up to 40%, and alternative external combustion engines such as the Cyclone are in this general range as well. Use of a such a device in conjunction with an optimized generator would probably see much better results in a test similar to what AC prop devised.

3) “Transmission” efficiency-

Generator and motor efficiency has come a long way in recent years – optimized permanent magnet generators achieve very high efficiency (90-95%). Compact traction motors such as that employed in the FCX clarity or Equinox Equinox fuel cell have almost no gear reduction at all apart from the the differential and CV joints. A dramatic example of efficiency is that used by the Nuna 3 solar racer, which achieves 98%.

By contrast the AC prop system – despite its sophisticated power electronics – seems to use a greater degree of gear reduction – e.g. even their recent eBox project necessitates the use of the Scion xB’s manual transmission components.

The ultimate goal for the range extender system is for the genset and traction motor to get the energy from the tank to the wheels with an efficiency comparable to that of a mechanical connection. That’s the main reason why series hybrids haven’t been used outside of diesel-electric trains or old submarines.

4) Speed-

AC Prop has noted that efficiency dropped if the ICE was heavily throttled – not surprising at all. Future gensets should be kept at maximum efficiency at all times, while using the battery to cope with greater power demands as well as regen.

In summary, I applaud AC Prop for their pioneering study, but the disappointing mileage of their genset trailer should in no way be regarded of a harbinger of problems to come.

However, a news report shot by a local Detroit news station got inside the design center and got to see a lot more in detail. Here are some decent-sized images of what they saw:

And here are my observations:

1) On a physical model of the front end, the sunken grill, headlights, and foglights have been smoothed out. Interestingly for aero, there is a “splitter” element on the bottom of the front end, rather similar to what the EV1 had. That feature might have something to do with cutting up the airflow and smoothing it out along the whole body.

2) We can now see from disguised clay renderings what the general flow of the roof line will be like (higher). The rounded front end is still thrust far forward in a very un-Prius like fashion. A sketch of the rear profile on a whiteboard looks slightly less angular.

3) A foam and wood model might give some hint of the C-pillars.

4) Of great interest is the dummy interior – there is an arm rest with drink holders in between the two rear passengers. I don’t know if this is a fold-down, but it portends the possibility that the Volt will be a four-seater.

Overall, although the baby-Camaro look has been replaced with more aero-sensible lines, I still think it looks quite nice – European with a dash of American. Contrary to fears that it would basically imitate the lines of a Prius – there’s not a grain of rice in the new Volt.

I look forward to seeing more, as well as what the competition has to offer in the next few years.

I already took Irv to task in a prior post. However, here’s some additional evidence for why we can comfortably discount Irv as an authority on anything engineering-related. The Dept. of Energy and GM run an event called Challenge X, where university students are given a Chevy Equinox, some money, and told to go wild on improving the “green-ness” of the vehicle. One of those teams, UC Davis, has taken their Equinox, added a 1.5L Atkinson-cycle engine (pulled from a Prius, admittedly), beefed it up with bigger electric motors at each axle, and fitted some lithium ion batteries.

The result? Well, it’s a plug-in hybrid SUV that’s capable of going 40 miles on a charge!

Sorry, Irv.

Source: LA times.

Additionally, sharp eyes will notice one frame from the video where engineers appear to be fitting electric drive components under the hood of a Chevrolet Cobalt:

The Volt concept was on an altar right at the front of the section, where two very attractive young ladies were giving enthusiastic lectures to convention-goers. I talked with them in person after the presentations, and learned some new things. Most significantly, “Melanie” explained that there was a “Quick Charge” function being planned, whereby the ICE could spin at “higher rpm” to fully recharge the battery with “18-20 minutes”, and using “3/4 gallons of gasoline” in the process (that’s ~53mpg at the end of the exercise). At first, they didn’t explicitly say whether this could also be done off the power grid, but they later hypothesized that it “was certainly possible”, given that the batteries could absorb so much energy so rapidly.

One could easily argue against an ICE-based quick charge, given that it might cause the engine to deviate outside its ideal efficiency range. But for now I’ll just wait and see what happens. If rapid-charge algorithms are being planned PERIOD, then that’s good news in of itself. It would certainly be harsher on the batteries, but would definitely help people avoid gasoline use more often.

 Update 11/28/07- Video interview with Denise Gray makes another hint at faster charging options.

Elsewhere was the new Malibu. This car looks a LOT better (lighter and sleeker) in person than it does in photos. There was also an E85-capable Corvette. There were two-mode hybrids from both Chevy and Dodge, and Volkswagon had a surprisingly large display dedicated to high temperature fuel cells and lithium ion batteries. They also had a nice display for next-generation clean diesel technology. The Germans seem to do diesels better than anyone else, and I hope to see more of this technology in the future, especially with range-extenders and commercial vehicles.

The Nissan GTR was the obligate automotive porn for the south hall. It’s an amazing-looking, purpose-built super car. I was interested to learn that it has a coefficient of drag of 0.27, which is only 0.02 above a Honda Insight! So there’s proof that aerodynamic vehicles don’t have to look as ugly as a PRius.

Speaking of which, Yokohama had two interesting exhibits – one of which was a Tesla Roadster (!) fitted with speciality tires. In person, this car is VERY small – then again, it IS based on a Lotus. The battery takes up the vast majority of the space behind the seats, although there is a small space for a few golf clubs.

The other Yokohama exhibit was a Prius modified with special tires and enchanced aerodyamics – such as a greatly reduced front grill area, and rear wheel covers, a la the EV1 and Insight.

The Volvo ReCharge concept was on display. Not a lot of technical detail, but the car itself is rather shapely and handsome, having been based on the new C70. I grew up driving Volvos, so I hope they follow through with this. The limiting factors seem to be PMLFlightlink’s expensive-ass motors and slow production schedule. That plus the lithium polymer batteries – I hope they don’t even touch the Kokam LiPos that PML used in the concept. Those things faired miserably in cold weather duty when they powered New York city electric taxis.

Porsche’s parallel hybrid technology was on display, as was the Lohner Porsche hybrid.

A few more pictures of my visit below:

I apologize for the horrible video quality since it was captured on a mobile phone.

After they set the new record, the bike’s owner let me onto some interesting possibilities for the future:

-A123 is going to be providing the team with a new, 1,000hp battery pack using ultra high power density cells.

-In order to deal with this power, the motor will need to be upgraded to AC instead of the traditional, “stone age” brushed DC motors.

-A new bike frame will need to be built to accommodate the different battery pack, the new motor and controllers, as well as a much bigger and better rear tire.

The KillaCycle folks want to outrace the top fuel (nitromethane+methanol) motorcycles, and with the right amount of sponsorship, money, and time, they don’t anticipate having ANY problems doing this.

The test subject: an A123 M1 cell. It was at about 30-40% state of charge before the test, unlike some other lithium battery tests where the cells have been basically dead. That’s about the lower limit of SOC that will be encountered in production vehicles.

The test: a fine point drill bit sent straight into the side of the battery, followed by being cut in half with a hacksaw.

The results:

After drilling a few millimeters into the battery, the drilling point suddenly glowed BRIGHT red hot for about 5 seconds, and there was a strong smell(stink!) of electrolyte. The heat lit a small spot of the battery’s cardboard sleeve on fire, and it continued to burn a very small, calm, steady flame, probably fueled by the electrolyte as well as the cardboard. However, there was no continuous, escalating thermal reaction or any more glowing from battery itself. I blew the flame out after about a minute. The battery case was hot to the touch, although not uncomfortably so. A slight, almost unnoticable “crinkling” sound could be heard from the battery.

The dark gray plastic cap on the anode end turned a much lighter gray – almost white.

Drilling into the battery again – both from the original point, and a new one – produced no noticeable reaction whatsoever.

Cutting the battery in half with a hacksaw across the original drill point did not produce any subsequent reaction either. The smell of electrolyte did increase somewhat. The condition of the battery material at the point of short circuit was not noticeably different from other points, apart from having a hole in it.

My interpretation:
-Drilling into the battery created a brief but intense short circuit with highly localized effects (apart from the whole cell getting warm). This would have probably happened with any battery.
-A small area on the cardboard/paper sleeve (directly adjacent to the drill point) caught fire, indicating the temperature was at least 451º F (~233ºC).
-Most importantly, the cell itself did not explode, rupture, or provide fuel for the initial fire, although the electrolyte (which is flammable) may have contributed to the burning of the sleeve.

I plan on repeating this experiment at some point without the sleeve present, and also with a fully charged battery.

I got to take a spin in the AC Propulsion eBox, as well as a small truck NEV built by Roush. They were both very smooth and pleasant to drive – although the eBox was much MUCH quicker, and its regenerative braking was very powerful (takes some getting used to). After flooring it, it’s tempting to let off the gas and coast for a bit, like with an ICE. With regen on max, that made for some jerky driving initially. Thankfully, it has a tiny gauge which adjusts the regen force very easily, although I eventually got the hang of modulating both power and braking with my right foot alone. I found this car particularly interesting because its motor is VERY similar in power and design to what’s planned for the Chevy Volt. Hopefully the Volt will also have the regen adjust feature. A new feature was a trip computer that calculated energy expenditure among other things – the car had averaged 327 Wh/mile.

The eBox had its air-cooled, 600+pounds of batteries mounted in an “I” shape under the front and back seats, and along the central tunnel – very reminiscent of EV1, and definitely good for crash protection. Despite using 18650′s, they don’t seem to have had the packaging and cooling nightmares that Tesla has, probably due in part to the centralized position as well as using a “different chemistry”. They wouldn’t reveal which chemistry, however.

I didn’t get a chance to ride in the Honda or Mercedes-Benz fuel cell vehicles – there was a really long waiting list, and you couldn’t actually drive it yourself. I did notice that every time they went by, there was a lot of fan and compressor noise coming from them.

Some interesting news that was circulating around the event – apparently AltairNano will be selling its batteries to the public by 2009, although I didn’t hear about any sort of cost estimate.

Unsurprisingly, A123System’s representatives wouldn’t comment about their new automotive cell type.

Oh, and there was a mint condition EV1 there. Ghostly…

But how do the contractors stack up against one another, and what are the inherent compromises of each approach? Each company is understandably reticent to reveal their exact specifications, as it is competitive information. However, with a little clever research into published white papers and presentations, you can uncover their approximate progress relative to one another.

IMPORTANT NOTE: This is a work in progress that only uses publicly available information. Please submit rebuttals, corrections, etc, to ensure that this comparison is accurate and up to date. I have no commercial interest or financial connection to any of these companies, other than as an American consumer.

Observations for each contractor:

A123: Their M1 technology is widely available, routinely reconfigured by the RC scene, and has a reputation for very high power and good safety even under abuse. Thus, the fundamental properties of this technology are well understood. Their more advanced batteries are based on the M1, and will supposedly feature thickened electrodes that will sacrifice some power density but increase overall energy storage. Only official word so far is “greater volumetric energy density”, but no word about gravimetric. To give an idea of where this could lead, maximum theoretical energy storage for LiFePO4 is variably quoted at 130-140Wh/kg. This puts on a damper on A123′s ability to meet the USABC’s long term energy storage goal of 200Wh/kg, but they still have a very competitive technology at the moment. Their choice of a cylindrical cell will make pack design slightly complicated due to poor heat dissipation and space management (compared to stack formats), but cylindrical formats tend to have the very highest energy densities, so as long as the amount of cooling and packaging equipment isn’t excessive, the tradeoff will be worthwhile.

Update 11/28/07- Confirmed from multiple sources at this point – A123 is developing a “flat” cell specifically for General Motors.

Additionally, 32 series energy density (volumetric) is confirmed at 260Wh/L, versus M1 cell at ~220Wh/L.

CPI/LGChem: These lithium polymer cells have been under development for some time – three different prototype generations have been described. The white paper indicating 95 Wh/kg is from 2003 however, so it’s expected that this has been improved upon since then. One apocryphal reference is http://www.aevehicles.com, which built the electric Pike’s Peak hill climb dragster using LGChem batteries. They claim energy densities of 160 Wh/kg for the cells that they use. Regardless, one source of concern is that as thermal stability of prototypes has improved, the energy actually went DOWN from 118 Wh/kg to 95Wh/kg. Their 95% capacity at 200 100%DOD cycles also does not compare favorably with the other contractors, but this was for the first generation, and is actually somewhat typical of conventional lithium polymers. This has apparently been amended using “improved fabrication technique”, but by how much, we do not know. Thus, 95%@200 cycles is the very minimal cycle life we can assume.

EnerDel: Lithium titanate is a much newer technology that has a reputation for very good safety as well as very good lifespan. Indeed, the cells are claimed to achieve 1000+ cycles even at elevated temperatures. However, EnerDel has explicitly confessed that LTO has much lower voltage than graphite or even hard carbon electrodes, and this makes for a very heavy energy penalty.

Nonetheless, they claim to have met the USABC mid term criteria. This means that a minimum of 80Wh/kg is assured, as well as a power density of at least 150W/kg. Similar LTO technology made by AltairNano has only about 78Wh/kg and 1600W/kg, so EnerDel probably falls within this range. We need more information before passing further judgment.

Whether this will be used in HEVs or PHEVs really remains to be seen, but for the moment other technologies seem to have an edge as far as versatility. EnerDel also has a hard carbon anode chemistry that also claims good safety, but better capacity. Specifications on this are unknown at the moment, however.

For pack construction, EnerDel has claimed that air cooling is sufficient for their cells. However, it’s known that chilling lithium batteries to below ambient temperatures extends their calendar life, and as far as I know it’s not possible to chill packs below ambient using air cooling.

Update 11/28/07- The folks from EnerDel got ahold of this comparison through the grapevine, and stressed  that their batteries’ good safety characteristics were a missing factor in the comparison. I don’t have access to safety metrics (if there are any), but when I do I’ll try to incorporate them.

JCI/Saft: Saft batteries are formidable because they are used in high-demanding military applications. The cell featured here is the VLE lithium ion, and has the highest energy density of the group as well as a very long cycle life – although it’s not published what %capacity remains at 1500 cycles. As far as safety and thermal stability, however, it’s worth noting that Saft provided the batteries for GM’s Sequel fuel cell concept, and this concept had to stop twice during its record-setting 300-mile test run because the battery pack had overheated. When assembled into OEM modules, the VLE packs hold about 110Wh/kg.

Not mentioned is 3M, which is not developing full size batteries, but is instead researching fundamental improvements in electrolyte chemistry and electrode formulation.

Where do things go from here?

Each contractor’s individual cells appear to have met the basic, mid-term USABC targets. Perhaps the biggest factor for success, however, is going to be how effectively the individual cells can be incorporated into large modules and packs that make good use of space, and can be efficiently heated, cooled, and protected. Some of this is going to be chemistry dependent – both the fundamental battery chemistry as well as the electrolyte- and some of it is also going to be dependent on cell construction. Cylindrical cells, for example, are physically robust and energy dense due to their tightly wound electrode material and strong metal can. But, they are less efficiently packaged than stack/laminate formats, and tend to build up heat very quickly. Laminate cells, on the other hand, are easier to assemble into large packs, and dissipate heat much more effectively. They are much more fragile, however, and do not store as much energy. One rather significant problem facing them is that lithium ion cells tend to have gas evolve during heavy charging; while cylindrical cells are able to vent this pressure quite easily, laminate cells are an inherently closed system, and must be designed to be able to flex and expand slightly. Proper, slow charging should not induce this gas formation – but in the rapid recharge/regen and discharge environment of an EV or PHEV, this may not be possible.

Cylindrical, stack, and prismatic formats are all undergoing continual refinement. One novel cell format, however, is a “button”-shaped cell from Illinois-based startup Inventek, which achieves good volumetric energy density through the use of wound, ribbon-like electrode material. “Rolled Ribbon” also advertises very high power (>2000W/kg) and good heat dissipation due to large, high surface area current collection and heat dissipation. As with all the other cell formats, putting many cells close together in a module essentially creates a coupled thermal mass, and thus proper spacing and active cooling are going to be necessary. Nonetheless, as the USABC has obviously recognized, the large diversity of technological solutions is a good thing – the more competition the better!

References:
Specific company references:
A123:
-http://www.A123systems.com

http://www.a123systems.com/images/charts/techCompare.jpg

http://www.greencarcongress.com/2006/02/a123systems_rec.html

EnerDel
-http://www.enerdel.com
-http://www.ener1.com
-http://www.ases.org/solar2007/presentations/tuesday/400pm/forums/4-battery/4-ota.pdf
JCI-Saft

-http://www.saftbatteries.com
-http://www.gm.com/explore/fuel_economy/news/2007/hybrids/lithium-battery-010407.jsp
-http://www.autobloggreen.com/2007/05/16/two-chevy-sequels-go-over-300-miles-on-real-roads-with-hydrogen/
http://www.rathboneenergy.com/batteries/battery_cells_by_mfg/saft/saft_lion_pdfs/VLE.pdf
CPI/LG-Chem

http://www.compactpower.com/

http://205.168.79.26/vehiclesandfuels/energystorage/pdfs/aabc_poster.pdf
http://www.aevehicles.com

Other:
http:www.inventekcorp.com
General References
http://www.uscar.org/guest/article_view.php?articles_id=74

http://www.uscar.org/commands/files_download.php?files_id=73
http://www.batteryuniversity.com

(click to enlarge image)

Data efficiency for several of the motors is not publicly listed. However, power density is fairly easy to calculate, provided one includes the mass of the motor controller, whose size varies along with the motor. Some caveats: for the Raser Symetron, their controller modules are only rated for 180 kW, so I played it safe, and included the mass for two controllers to roughly equal the max output of their most powerful motor.

Tesla also does not list the weight of their PEM (power electronics module), so I could not include them on the graph.

That said, the claimed power density of the Raser and PML Flightlink motors is staggeringly above that of more conventional systems. What’s more, the output has been achieved using two different approaches – brushless DC, and AC induction.

Some important considerations, however:

1) PMLFlightlink’s technology is intended for in-wheel motor use. Assuming PML’s claims are true, the fact that they are accomplished using permanent magnet technology probably means that the motors aren’t scalable beyond their current size, lest they suffer from the efficiency losses that come about through hysterisis and eddy losses. As engineers from GM, Aerovironment, and Tesla have frequently pointed out, AC induction motors are best suited for larger, more powerful motors. Then again, many cars won’t require power needs beyond 120kW.

2) The deal with Raser Technologies is essentially one of “put up or shut up”. The company is legitimate enough to have marketed their alternator technology to Wilson, but we’ve yet to see any substantial independent verification of their claims.