## Progress in Developing SOFC APUs for Heavy-Duty Trucks

##### 13 June 2008
 Projected system efficiency of the Delphi prototype SOFC APU compared to that of a conventional diesel APU. Click to enlarge. Source: Delphi

Two US DOE (Department of Energy)-funded projects to develop a SOFC (solid oxide fuel cell)-based APU (auxiliary power unit) system for heavy-duty trucks reported on their progress this week during the DOE’s Hydrogen Program annual merit review in Washington, DC.

One team is led by Cummins Power Generation and includes Protonex LLC and International Truck and Engine. The other is led by Delphi—which is leveraging its SOFC work in the Solid State Energy Conversion Alliance (SECA) (earlier post)—and includes OEMs PACCAR Incorporated (producer of Kenworth, Peterbilt, DAF and Foden Trucks) and Volvo Trucks North America.

 Diagram of a solid oxide fuel cell. Click to enlarge.

The goal of each project is to build and demonstrate a diesel-fueled SOFC truck APU. APUs are increasingly being turned to as a solution to reduce the fuel consumption drain of idling and hotel loads on heavy-duty long-haul diesels. While current APUs can use a small diesel-fueled genset, a diesel-fueled SOFC APU would ideally deliver improved efficiency and thus lower fuel consumption.

SOFCs use a hard, ceramic compound of metal oxides as an electrolyte, rather than the thin, permeable polymer electrolyte sheet in a PEM. In a PEM fuel cell, hydrogen ions cross the membrane; in an SOFC, oxygen ions cross the electrolyte. SOFCs operate well on hydrogen and mixtures of hydrogen and carbon monoxide, among other fuels.

Benefits of a SOFC APU for a truck application include:

• Hydrocarbon fuel reformation requirements for SOFCs are greatly simplified (they are thermally matched, the CO output is a fuel constituent, and there is some sulfur tolerance);

• There is no internal water management issue;

• SOFCs are a lower cost fuel cell option with no or low requirements for precious metals;

• No external cooling is required; and

• There is a high quality, high temperature single waste heat stream that can be used in the fuel reforming process as well as for vehicle heating.

On the other hand, there are thermal management issues, as well as issues with startup time, stack degradation, and the OEM requirement for zero net water diesel fuel reforming.

 The Cummins SOFC APU system. Click to enlarge. Source: Cummins

Cummins Power Generation. The Cummins program began in 2004, but was placed on hold due to budget issues, before being restarted in 2007 (the Delphi project has a similar history.) DOE’s share of the project is $3,225,611, with the industry partners providing$1,732,938.

The Cummins system delivers 2 kW of steady state power from the Protonex SOFC stacks, and about 3 kW of intermittent power from a battery bank. A bi-directional inverter and control (which Cummins leveraged from its work in the recreational vehicle industry) provides up to 5 kW of intermittent power or 2 kW of continuous power to the truck systems.

The stack consists of 4 modules, each comprising 66 tubular fuel cells, which are currently delivering 12.8 W each. The initial target was 10 W each. Each module will include a CPOX (catalytic partial oxidation) reformer, fuel cells, recuperator, tail-gas combustor, and insulation. Cummins says that the SOFC system fuel to electric efficiency is 21% gross, 17% net.

The partners have demonstrated both atomization and vaporization of the fuel. They will use vaporization for the initial units, but will move, in the longer-term, to the atomizer, which requires less start-up energy and has extended maintenance intervals.

Protonex has screened four catalysts and selected one that is capable of more than 93% carbon and H2 selectivity. The team has demonstrated steady-state operation of the tubes for more than 500 hours with the selected catalyst, without carbon formation.

Later this year, the team will build single modules and 4-module sets, and begin testing. Performance optimization is targeting for Q4 2008, along with fuel feed system improvements.

Road testing of the units is due to begin in 2009.

Delphi. DOE is contributing $3,000,000 to the Delphi project—which is now 50% complete—while Delphi is contributing$1,750,000. Delphi’s system has a rated net power output of 3.5 kW, with a target fuel to electric efficiency at rated power of 25%.

The Delphi system is designed with a cycle life—i.e., to go from ambient temperature to the operating temperature of around 750°C and back down to ambient—of 250 cycles. Delphi anticipate one cycle/week, 50 weeks per year, for a 5 year lifetime. (In other words, Delphi sees the system powering up at the beginning of the truck work week and staying on until the end of the week.)

Life the Cummins system, the Delphi system is packaged into the form factor of an existing diesel APU unit.

Delphi is developing reforming technology for diesel/JP-8 SOFC applications by modifying its existing natural gas reformer. Two main designs are being developed:

• A CPOX reformer; and

• A recycle based (endothermic) reformer.

The CPOX reformer offers moderate efficiency and simplicity of design. The endothermic reformer offers higher efficiencies through anode tail-gas recycling. With an SOFC, water is created on the anode side of the fuel cell. Delphi is looking at taking that water and bringing it back into the reforming process to accommodate autothermal or steam reforming capability.

The current version of the platform only uses CPOX; the next generation will use the endothermic capability.

Through the rest of this year, the Delphi team will complete the SOFC APU hardware design and build and begin subsystem testing and development iterations. The following year will see complete system module testing and the beginning of full SOFC APU testing, with road tests scheduled for 2010.

Delphi sees the mobile SOFC market expanding beyond heavy-duty diesel trucks to recreational vehicles (diesel and LPG), truck and trailer refrigeration (diesel), and military (JP-8) applications.

Perhaps I don't understand the hotel load problem; but, I thought Firefly had this problem handled with it's huge new batteries. Why use all this complexity to solve a simple power storage problem?

The are already using the generator that would recharge that battery to power the truck as such it wouldnt be any better to add a battery to the generator.

Instead they want to replace the generator with something better.

I got a better idea: if it take more than overnight, send it by rail.

The trouble with SOFC is that they cost a fortune. A 5kW SOFC cost about $175.000 see http://www.acumentrics.com/tools-faq-fuel-cell-faq.htm The following text will determine the economic viability of using a battery EV long-distance truck at 80000 lb when compared to a similar diesel powered truck. You will be surprised big time I promise. First some assumptions regarding diesel heavy trucks: 1) A heavy truck is operated 60 hours a week, 50 weeks a year = 3000 hours a year. 2) It drives on average 35 mph * 3000 hours = 105.000 miles per year 3) It spend 1 gallon of diesel to get 5 miles or 21.000 gallons per year. 4) At$4.7 per gallon the annual diesel budget is about $100.000 per year. Reference for these assumptions http://www.tc.gc.ca/pol/EN/Report/truckGrainCosting1999/C4.htm#4.2.2 Then some assumptions regarding a heavy duty EV truck: 1) 35kW is enough to propel a 80000 lb truck at no elevation by 40 mph. 2) The truck will need a 350 kWh battery to do the typical daily drive (max 10 hours). 3) The battery cost$500 per kWh or (350*$500 ) =$175.000.
4) The battery can do 2000 cycles so that it will last about (2000/350 days) = 5.71 years.
5) The battery will be leased for (175000/5.71) = $30.000 per year. 6) The EV truck with a leased battery cost the same or less than a diesel truck or about$120.000.

Now the EV truck need to lease a battery for $30.000 per year and it need to buy electricity for$0.1*35kWh*3000 hours a year = $10.500 a year. This is$30.000 + $10.000 =$40.000 for ‘fuel’ per year.

In other word, the EV truck will save $100.000 -$40.000 = $60.000 per year on fuel versus a diesel truck and it will cost no more to buy!!!! Proposal: Instead of 3 months of patrolling in Iraq the US could spend the$50 billion dollars saved to build a charging infrastructure that covered every corner of Mexico, Canada and USA with charging outlets for truckers and people with EVs. Make it a national priority and it could be built in less than 36 months for less than $50 billion I am most certain. Chargers are not rocket science. It needs to be build ahead of EVs actually being made available because that would remove the only obstacle for people to buy pure EV vehicles. Pure battery powered EVs in all vehicle classes where the battery is leased are economically viable right now. What are we waiting for???? "What are we waiting for????" erm, an economically viable (without subsidies), renewable and sustainable way to generate the massive expansion of electricity generation capacity that you propose??? Well, wind power cost about 7 cents kWh without subsidies at an average on-shore location. That is 30% less that the 10 cent per kWh I assumed in my case above. Coal is about 5 cents per kWh but unlike wind power it uses water resources, it emits CO2 and it pollutes the air with particles that cause widespread diseases among people. According to Bush’s top science advisor professor John Marburger the planet will become unlivable (his word) unless manmade CO2 emissions are stopped. See http://news.bbc.co.uk/2/hi/science/nature/6994760.stm EVs are not coming overnight to everybody. We need to build battery factories and so forth so the global wind turbine industry can easily supply all the needed extra power as it is demanded. Did you know that this year global wind power will provide 10.79% of the total global growth in electricity produced and that wind turbine manufacturing expands by 30% annually? If this growth rate can be maintained, it will only take until 2018 for the global wind power industry to supply 107% of the entire extra needed electricity globally in that year. There are no problems of any importance that remains. All we need is to recognize we have the opportunity and then act on it accordingly. Hwneic.. a typical heavy truck eats 200 gallons of fuel.. converting to electric energy and fiddling for egg of engine vs battery and motor.. We will need 2 megawatt hours of power on boatd to replace the MEDIUM haul trucks and twice that to replace a long haul truck. Average car goes 17 mph and drives 30 miles a day.. but no one is average. Thats why the interest and money in fuel cells.. a 50 kg insutrial h2 tank can handle medium trucking needs and a few of em ganged together can handle long haul.. But even 20 years from now a 2 mwg battery will be MASSIVE.. likelt weighing 20 tons with all its stuff... For BIG thinds h2 makes massive sense as it weighs very little and packs a huge punch. And the fuel cell needed isnt much bigger then a cars fuel cell and industrail h2 tanks are cheaper. And h2 can SUPPOSEDLY be made cheaper then diesel... As for the sofc.. its 175k because the company makes 20 of em a year and needs x million a year to run.. when they are making 100000 of em they expect a 5 kw model to be CHEAPER then the current genset apu to last longer then them to use less fuel to be quiter and perform better and be safer and easier to maintain... Oh and be smaller and lighter... So either a fast charging battery that can handle ..1000 miles or a slow one that There are times when we really need the energy density of liquid fuels and long/medium range trucking is one of them. We are struggling with battery supply and specifications. So lets use the batteries we have for stop/start motoring first - city buses, delivery trucks, taxis, then cars for personal use. Alternately, you could move more freight by rail, which has been electrified successfully in many cases. or just use a small apu or plug it in in a truck park. The comments have gotten way off the track of the subject, which is about APUs. While I think it is an encouraging thing for the government to be working with industry to produce a solution, one that might actually be built by the companies chosen, the most important issue is... What is the original question? And why are we trying to answer it? APUs are there primarily to make a habitable sleeping/relaxing area for long haul truckers. Right now, the truckers run their diesels continuously to generate the electricity and heat for their quarters, even when not moving. This is wasteful of fuel to use a large diesel merely to spin the generator. There is a secondary issue not recognized as well; in cold climes, the truckers are fearful that they may not be able to restart their trucks after several hours of "cold-soak" in freezing temperatures. In defense they run their diesel just to keep them warm and running. This "high tech solution" is but a partial Rube Goldbergish answer to but a portion of a question that never needed to be asked, in the first place. Surely the only thing needed is the installation of electrical power taps, for rent, at truck stops. The APU can then be nothing but an electrical connector, to provide all electric service for HVAC and lighting on a truck, with perhaps a voltage and AC/DC converter. Had the government encouraged the truck stops to install such electrical service, The truck stops gets a way to charge for another service; and attract customers. It may also provide the assurance for a truck re-start, with a jump starting service, too. Don't discount that issue; many truckers would refuse to shut down their diesel even if all other issues are addressed. You can find many truckers taking a motel room service, and still idling their diesel all night. Theoretically, the simple engine heating coils, long provided in cold areas, these coils could be energized off the electrical tap to warm the oil, aiding a cold start. There are additional benefits that are not even addressed by the proposed APU solution. A fourth service provided by the truck stop answer, is that resting at a truck stop allows, some personal security to the trucker, versus resting at rest stops in high crime areas. A fifth service the truck stops provides, is an opportunity to obtain luxury food, snacks, and drinks along with entertainment, music, DVDs, books etcetera to the trucker. For that matter, the simple luxury of being able to mix and mingle with people in the same business, and meet and converse, is of an inestimable value to a lonely trucker, as well. There is another implicit benefit as well. Why carry around the weight and material of the APU? It diminishes fuel economy. Why pay for the overly complex APU? It just increases the capital cost of a trucker's equipment. To answer the critics who say APUs exist for a reason, that is true; but truck stops have historically not provided electrical taps for rent, forcing the APU creation. When they do, and appropriate electrical APU taps are available, then for the very occasional circumstance when a rest stop is all that is available, why not rely on just letting the diesel run? That was the original APU. Letting the truck stops solve the "problem" is a way to do it more cheaply, more efficiently, and better all around. "Instead of 3 months of patrolling in Iraq..." Iraq poor and oppressed want democracy to make them free to buy Hummer and big Tundra heavy duty vehicles. No in fact thats ANOUTHER reason they want duel cell apus.. a genset does not like to not be run a duel cell doesnt care.. and no matter what a truck needs its own onboatf power.. And it helps bootstrap fuel cells and gey em ready to take over for engines BEFORE oil goes to heck in a handbasket. As the TRUCKING industry is very keen on the idea I assume they know what they need and why.. so ill leave it up to em. Stan, you got some good points as usual.. but truckers do not like to stop their engines when they are sleeping.. it may not start in the morning and now they are stuck in the middle of nowhere with a perishable load waiting days for repairs. Several thousand$$down the drain. Now a diesel 35kw APU and a battery pack, that could be enough for a plain serial hybrid truck.. I really doubt 35kw will be enough for a large 18 wheeler. Do these APU have an "idle" mode?, to keep them at operating temperature at night and ready to go in the morning? Mahoni touches on reality about using the best solution to each requirement. Right now we use liquid fuel for the shortest trips and the longest. For the largest vehicles and the smallest. That universal mode is what no longer makes economic sense. We need to first utilize each replacing technology where it is best suited. I strongly suspect urban delivery, service, and transportation vehicles will be where we gain the most from fuel cells, hybrids, and EVs. Stan pointed out that lugging an APU, of whatever kind, for hundreds of thousands of miles is not free. Weight burns fuel. Period. End of story. It is also unlikely that APU power will ever be cheaper than power from the plug at a truck stop. Cost-wise the APU is beat. Its merit lies elsewhere, in convenience and certainty. So perhaps we should look at the convenience and certainly problem. These surely cannot be the same for all long-haul truckers. Could it be done with a modular APU? Standardized, easily attached or removed. Truckers would use them on some routes, perhaps with some cargoes, and not use them at other times. Wintermane suggests the trucking industry can figure out what they need. I somewhat agree but notice these two companies axed similar work in 2003-4 and their funding today remains modest by any measure. @ Henrik ~> Electric trucking is possible even for long haul, (with battery swapping or a significant recharge improvment) but this is hardly the "low hanging fruit" of the BEV truck market. Inter-city delivery and pick up are a far easier truck market to open to the new technology. @stas peterson ~> I'm actually agreeing with you here. The better solution is to have engine block hearers in trucks, plugs installed at truck stops, and ensure that all make "restart-service" available.$3,225,611, & $3,000,000 =$6,225,611 which would go a long way to installing these services.

Extra money spent to increase the reliance on fuels we know to be in ever shorter supply is irrational.

The reason they arnt spending much money is there realy isnt much THEY are doing here. A thousanf small steps taken quickly get ya there.

As for the industry... all the trycj companies are doing is look at gizmo... is it USEFUL... not yet... WHEN? Doon... Ok we will look at it again in... 2 years.. LUNCH!

In this case it looks like its getting very close so likely things are going to heat up as they go from wait and see to plan and build mode.

Oh and again it does depend on the load.. plug ib will only work for loads under 240v x 60 amp as the cord and pluf for bigger power needs is amazingly enough heavier then an apu;/ and bloody hard to use too... Niot to mention most truck stops cant get that much power for that many trucks.

Wintermane

You say the EV truck I suggests need a 2000 kWh battery in order to use (2000/350)*35kw = 200kW to propel a 80.000 lb truck at no elevation at 40mph. If you are correct on this then my business case collapse I fully admit.

I am also not sure that my own guess of 35kW is the correct power and so far I have not succeeded to find any source that can confirm it. I would really appreciate to get some web sources that can confirm how much power we need to propel a 80.000 lb truck at no elevation at 40mph. Remember an EV truck with a large lithium battery has no loss to transmission (could use in wheel electric motors), electric motors almost work at about 90% efficiency all of the time and nearly all braking energy can be recovered, for instance, when hitting the brakes going downhill.

My business case would break even if 70kW is needed to propel a 80.000 lb truck at no elevation at 40mph. The lease of the 2*350kWh battery would then cost $60.000 a year and the needed electricity a year would cost$10.500 *2 = $21000. Total cost is then$81000 and the annual saving would be $100.000 -$81.000 = $19.000 per year compared to diesel. However, this money is likely about lost in less space/weight for cargo since we need to carry a 700*20 lb = 14.000 lb around. Another issue that will improve my business case is that the cost of batteries is set at$500 per kWh which is likely to be the price you need to pay today if you buy them wholesale. That price should drop in the coming years and with it so does the cost of the battery lease.

I found a source for the power needed to propel an 11 tons (22.000 lb) EV bus at 85 km/h. It is 600Wh/km.

See http://www.e-traction.com/presentation.htm

I assumed 35kWh to get 40 miles with a 80.000 lb truck running at 40 mph or 64 km/h. For comparison this is 547 Wh /km. The bus drives faster so it will use more power but then again my truck is 4 times as heavy. I don’t know how to adjust for that.

Henril its easy to see how much total energy is needed.. Just take the fuel tank size of a truck and add a zero.

So if your truck commonly comes with 2 50 gallon tanks and NEEDS them then you need 1000 kwh at very high dod.

To figure out MAX power needed you look at the engine currently used. They are BIG. So your talking 3-400 kw max and alot less then that for most of the time.

But the real issue is even your 350 kwh pack will weigh an enormous amount and take up a huge space. Unless they manage to use the same bat as in the volt.. NOT CHEAP thats gona be about 50 times the size of the volt battery and weigh about 10000lb And that means the truck just LOST a huge amount of cargo cap because it cant weigh over x lb.

I wouldnt worry tho as batteries get better longer haul trucks will be made.. so we dont have to stress about the math.. as people can buy a truck that goes as far as needed as long as needed as they need it to do so... they likely will buy em.

An d as giel cells improver they will replace whatever they are better then at that time.

As long as we get good batteries and good fuel cells we are better off then if we only have one or the other.

Wintermane

You have me convinced that my 35kW estimate is too little to propel a 80.000 lb truck driving 40mph at no elevation. I checked the Smith electric truck. This is a 25.000 lb truck and they say its range is 130-150 miles using a 80kWh battery pack (4*20kWh zebra batteries). Their truck run 50 mph and uses a 120kW drive train. See http://www.greencarcongress.com/2007/01/enova_to_supply.html

The implication is that the truck use between (80kWh/130miles)= 0.615 kWh/miles or (80kWh/150 miles)= 0.533 kWh/miles. In order to propel a truck that is 80.000 lb it should be able to do it using maximum as much extra energy as its weight increase imply (it should be maximum because the air drag is certainly not going up proportionally and neither may friction). In other words, a 80.000 lb truck need max (80000/25000)*0.615 kWh/per miles = 2 kWh/miles. For comparison my business case above assumed you could travel 40 miles on 35kWh or 0.875kWh/miles. In other words, you need (2/0.875)*35kW = 80 kW to propel the 80.000 lb truck but the speed is now 50 mph.

Now I got an idea that will save my business case from falling apart. A 350 kWh battery should give the 80.000 lb truck a range of 350/2 = 175 miles at minimum 50 mph. Now if these batteries can quick charge (as we know the LiFePO4 are capable of) the driver can simply change his drive cycle resting 30 minutes for every 3 hours and at the same time charging the trucks battery.

My business case is now still $30.000 for the battery lease but we need more electricity. We need$0.1*80kWh*3000 hours a year = $24.000 a year. This is$30.000 + $24.000 =$54.000 for ‘fuel’ per year.

In other word, the EV truck will save $100.000 -$54.000 = $46.000 per year on fuel versus a diesel truck and it will cost no more to buy!!!! I am therefore still convinced batteries are economically viable right now for heavy duty long-distance trucks but we need an electric infrastructure which is far less costly and quick to establish than liquid hydrogen. Also Wintermane I know you know that hydrogen will need twice as much electricity as a battery because we lose about 0.75*0.65 = 0.4875 for respectively electrolyzing and in the fuel cell. Liquid hydrogen will lose even more (perhaps another 30%). This is important because the electric bill will at least double to$50000 a year in my case and we still have not included cost for expensive electrolyzes.

I just realized that the batteries will be charged about 80/35 = 2.3 times more often so they will wear out faster. The battery lease will therefore be $68500 and not$30000. So this is $68.500 +$24.000 = $92.500 for ‘fuel’ per year. In other words, the EV truck will save$100.000 -$92.500 =$7.500 per year on fuel versus a diesel truck and it will cost no more to buy. It is still economically viable but not fantastic anymore. The case will improve if batteries can be cycled more than 2000 times or cost less than $500/kWh or if diesel goes further up from$4.7 gallon all of which is likely to happen in the coming years.

I am sorry but there are other errors in my calculation so I shall restate the whole case.

The following will determine the economic viability of using a battery EV long-distance truck at 80000 lb when compared to a similar diesel powered truck.

First some assumptions regarding diesel heavy trucks:
1) A heavy truck is operated 60 hours a week, 50 weeks a year = 3000 hours a year.
2) It drives on average 35 mph * 3000 hours = 105.000 miles per year
3) It spend 1 gallon of diesel to get 5 miles or 21.000 gallons per year.
4) At $4.7 per gallon the annual diesel budget is about 4.7*21000 =$100.000 per year.

Reference for these assumptions
http://www.tc.gc.ca/pol/EN/Report/truckGrainCosting1999/C4.htm#4.2.2

Then some assumptions regarding a heavy duty EV truck:
1) The Smith electric truck 25000 lb uses (80kWh/130miles) = 0.615 kWh/miles running at 50 mph. In order to propel a truck that is 80.000 lb it should be able to do it using maximum as much extra energy as its weight increase implies (it should be maximum because the air drag is certainly not going up proportionally and neither may mechanical friction). In other words, a 80.000 lb truck need max (80000/25000)*0.615 kWh/per miles = 2 kWh/miles.
2) The truck go 50 miles in 1 hour using 2kWh/miles*50 miles = 100kWh or 100kW in continuous power.
2) The truck has a 350 kWh battery (350*20lb=7000 lb) in order to give it a 350/2 = 175 miles range.
3) The battery cost $500 per kWh or (350*$500 ) = $175.000. 4) The battery can do 2000 cycles (both Valence and A123 warrant 1000 times in heavy duty use, normally that implies it lasts 2000 times on average) so that it can do 2000*350kWh = 700.000 kWh. This is enough to propel the truck 700.000 kWh/(2 kWh/miles) =350.000 miles. 5) The truck need to do 105.000 miles per year just as the diesel truck so the battery will last (350.000 miles/105.000) = 3.33 year. 6) Therefore the battery will be leased for (175000/3.33) =$52.500 per year.
7) The EV truck with a leased battery cost the same or less than a diesel truck or about $120.000. Now the EV truck needs to lease a battery for$52.500 per year and it needs to buy electricity for $0.1*2kWh*105.000 miles =$21.000 a year. This is $52.500 +$21.000 = $73.500 for ‘fuel’ per year. In other word, the EV truck will save$100.000 -$73500 =$26.500 per year on fuel versus a diesel truck and it will cost no more to buy. So my conclusion is unchanged that this business case is economically viable. Of cause we still need an electric fast charge infrastructure which is rather inexpensive and fast to build.

Henrik:  You're proposing to do things the hard way.  Long-haul vehicles are the most difficult application for batteries, and vehicles with weight limits will suffer the greatest payload impact.

The solution is to minimize the required battery range, by supplying the bulk of the electric power via wires.  This can be accomplished by putting the freight on rails, either as electrified trains or electrified dual-mode trucks.  Roadways can be equipped with rails similarly to city streetcar lines, so dual-mode traffic would not be limited to dedicated rail routes.

How much battery energy would you need?  It depends partly on the electric range (trucks could operate tri-mode with small diesel sustainers until the network was built out).  Suppose that you need to run up to 20 miles from the nearest rail terminus to the loading dock or other place to park and recharge.  A typical heavy truck gets in the neighborhood of 6 MPG.  At 140,000 BTU/gallon and roughly 40% engine efficiency, this is about 2.7 kWh/mile at the crankshaft.  A 20-mile electric jaunt would use about 54 kWh, so call it about 80 kWh of battery capacity required to have adequate margins.

It turns out that the Smith Newton is available today with an 84 kWh Zebra battery, so it appears that the necessary hardware can be purchased off the shelf!

This might even make the truck lighter.  The 200 HP motor in the Tesla roadster is descended from the unit in the tzero, which weighed only 70 pounds; call it 140 kg for the motor.  5 16.2 kWh Zebra battery units (1 hour rate) @ 195 kg, 990 kg.  3 liter 150 kW diesel sustainer engine (optional), guesstimate 150 kg.  The motor can operate without gears, so the total weight would be about 1280 kg.  A Caterpillar C9 engine is about 3/4 ton by itself, so subtracting the transmission and maybe 150 out of 200 gallons of fuel (1050 pounds less) would yield a truck with about the same payload as the combustion-engine version even for the unit with the ICE sustainer; without, 700 kg engine + 200 kg transmission + 630 kg fuel is 1530 kg, about 400 kg more than 990 kg battery plus 140 kg motor.

Getting back to APUs, the 17% efficiency figure stinks on ice.  SOFCs can hit 60%, which leads me to wonder why the performance of these units is so bad.  However, the SOFC is quiet, can be used to keep engine coolant warm and fuel un-gelled, and probably needs closer to 1 quart per hour compared to a gallon or so for an idling engine.  The APU would save its weight in fuel every few days, making it pay its way and then some.

@EP, @mahony and @no name

You have all correctly pointed out that I am going after the non-obvious, high hanging fruit of the possible EV market. This is quite deliberate. If a good business case can be established for the higher hanging fruit then we know the more obvious applications (city busses, garbage trucks, and fixed short distance deliveries) will also work and that there is even more reason to start spending money building a charging infrastructure. The weakest characteristic of an EV vehicle is its limited range and that can be dealt with by focusing on batteries with fast charge capability and by building a tight network of high power capable charging outlets. I doubt that mechanical replacement of batteries is the way forward as suggested among other ideas by project better place and I fear it can be used to establish monopolies for battery supplies.

Using railways for long-haul traffic is benefitted by increasing diesel prices. It will certainly rise in the coming years but its potential is limited by the fact that it is not nearly as fast and flexible as trucks that can deliver gate to gate.

Got a little bit of luck. Found a source that comes close to confirm the latest case calculation. Turn out a firm called www.balqon.com builds 60000 lbs EV trucks for harbor container transport. It goes 60 miles using 140 kWh running at 40 mph. That is 2.33 kWh per miles a little more than the 2 kWh that my long-haul truck uses. However, this thing must have lots of start stops during the day and although EVs can use the braking energy there must be some loss.