Toyota to pursue plug-ins

Clett,

That's my understanding of what this study was about. I'd characterize it as a worst-case scenario, from which anyone should feel free to modify the numbers to match their own projections of what actually will happen.

This study assumes that:
a) every current gasoline-powered mile would be converted to electricity,
b) there would be no downsizing of vehicles (Hummers become Hummer EVs),
c) in fact, average vehicle weight would go up due to the weight of the batteries,
d) even given their weight, and the extra weight of the batteries, vehicles were not on average particularly efficient at converting electricity into motion. The overall efficiency assumed, before adding the 15% penalty for battery weight, was 71%, whereas the Prius (based on posts here) is already about 80% efficient in converting generated electricity into forward motion.
e) all charging occurs during one eight-hour period, every night.

I don't think the author of the study had an axe to grind, I just think that he made conservative (pessimistic) assumptions to figure out what, in the worst case, we would have to build. Just to get a handle on the order of magnitude. And after checking the numbers, I'd agree that the path he took from assumptions to conclusions was done correctly. In that worst-case scenario (all US transportation is done with large, heavy vehicles that are modestly less efficient in handling electricit than the current Prius, we all plug 'em in at midnight, charge at a stead rate, and unplug at 8 AM) -- then yes, significant new capacity would be required.

The only way I saw, in which the study was not the absolute worst case, is that it didn't take daily and seasonal variation in travel into account. But it also didn't take into acccount a number of plausible charging strategies that would spread the load over more hours. Adding a couple of hours to the period over which EVs charge would offset the effects of the Fridays-in-August peaks. I was not smart enough to figure out whether the issue of four US time zones had any impact on the assumption of the 8 hour charging period.

But, that's not a prediction of the future, that's a calculation. If you want to predict what we actually might need, then you modify the assumptions to match where you think the world will go. In any number of ways. No new capacity would be needed if:
a) only half to maybe two-thirds of vehicle miles were converted to electricity, or
b) average vehicle size were reduced signiificantly (e.g., fewer SUVs) as the fleet was converted, or
c) vehicle electrical efficiency would be around 90% and inefficiencies due to battery weight would be half of what was assumed.

And so on.

Maybe it's because I do work like this all the time (in a different industry), but I don't mind separating the issue of whether the calculation is correct, for what's been assumed, versus making a prediction of the future. The calculation is more-or-less correct, for what's been assumed. That's useful, unless that gets misused as anti-PHEV propaganda. I mean, with very little work, it lets you generate your own estimate of potential capacity needed, and tell within reasonable limit how far the conversion could go with zero capacity expansion required.

So, I'm a two-handed economist. Yes, the calculation is correct. Under those assumptions, capcacity expansion is required. No, the assumptions are probably not the most realistic case. Under plausible assumptions no or fall less capacity would be required.

Any way you shake it, I think the message is that, at worst, we'd need a capacity expansion that is well within historical norms. It would not be a radical change for electrical power generation. Far less radical, in my estimation, than air conditioning was. The only trouble that might occur is of it catches the industry by surprise, so to speak (because even under the worst-case scenario, you wouldn't see nighttime brownouts from EV load until two-thirds or so of the fleet miles had be converted to electricity). In that sense, a serious article of this nature is a useful tool -- it credibly brings the issue to the attention of the people who matter in this, the generators and their regulators.

And at best, we'd need no or little capacity expansion. Also good to know. And now we do know, just by tinkering with the assumptions of that work in a transparent way. And therefore you have a competely credible prediction as well, based on the original scholarly work. So long as everybody can see how your assumption varies from what's in that study. I'd say that's helpful rather than hurtful to have that in hand.

Because -- In other threads I've seen people raging about the huge cost of the expansion of capacity needed to run an EV fleet, and who was going to pay for that. This study, viewed properly, puts that argument to rest. It's not an issue. Under the worst-case-scenario assumptions above, we would end up buying 31% more electricty than now, and have to expand capacity 20% or so. Merely by paying our electric bills, we'd pay for that capacity expansion just fine. And the efficiency of generation would increase as the average load factor would be higher, which means CO2 per KWH would be lowered. And if the conversion takes 30 years, the rate of capacity expansion, even in the worst case, is actually below historical norms for the electrical generation industry.

So, in the worst case, some modest and affordable expansion would be required, with added side-benefits noted above. The only problem you might see is if it catches the utilities unawares and they'd have to play catch-up, as their planning cycles are very long. So this article serves a useful role in puting utilities and their regulators on notice about the issue. And NIMBY, of course - that's always a problem for new power plants.

Then, as you point out, any number of plausible scenarios about the future would result in no or minor capacity expansions required. And in any case, we are talking about decades from now: no expansion would be required until a large fraction of miles had been converted. Projections on that time scale are pretty uncertain anyway.

Anyway, I've made my mind up about the issue. If we have the worst case, and do not plan, and have relatively inefficient charging strategies, then we could have some problems coping with this. No problem paying for it, just problems if we have to get the capacity in place in a hurry. But the likelihood of something near the worst-case scenario is small, I think, the exact requires are essentially impossible to predict, and any problems would occur in the distant future. Under completely plausible alternative scenarios, no or little capacity expansion is required. Except for those literally engaged in long-term planning for the utility companies, I can't see the benefit of thinking about it beyond that. Those guys really do have to think 10 years down the road. But from my perspective, electric generating capacity is not a stumbling block. It may or may not be an issue to be dealt with some decades down the road. It's clearly affordable -- the increment in our electric bills is, in the worst case, far larger than the increment in required capacity. It's more affordable than the capacity expansion we had to do to air condition the country. And it's absolutely not a barrier to pursuing PHEV/EV transportation.

As an afterthought, as ken1784 is a major contributor to this thread, I'm of course speaking from a US perspective. If Tokyo Electric, for example, is far more constrained in its ability to situate and open new power plants, or requires a much longer planning horizon to get a new plant approved and open, then of course they would rationally pay more attention to what I've called the worst-case scenario. Because they would then face far higher risks than we would here, if that worst case occurs.

What I'm saying is that even if we agree on the calculation, we could quite rationaly disagree on what set of assumptions we think we ought to pay attention to.

 

<div class='quotetop'>QUOTE(clett @ Jul 28 2006, 07:06 PM) [snapback]293726[/snapback]</div>

Hi clett, I'm not saying the study is correct or wrong. I just say it is one of assumptions by Dr. Uhrig.

Have you read the paper? The 21.1 kWh assumption was explained.
Where is your 7-9 kWh assumption came from?

It is your own definition. In general, the PHEV-35 is defined as 35 miles EV mode.

Where does the 10.8% come from?
My assumption is 553GWe at 100% peak(refer to my previous post), so even small 90 GWe is more than 16% of peak.

Maybe, you are doing wrong guess. Please show us your certain assumption numbers.

Ken@Japan

 

Ken,

US department of energy lists US operating capacity at roughly 1000 GWe, as shown here.
http://www.eia.doe.gov/cneaf/electricity/epa/epat2p2.html

If it were 553 GWe for the US, that would not square with total US electrical production of 4,038,000 GWH, as 553*24*365 = 4,844,280, implying that we operate at an average of 83% of maximum capacity. That does not seem reasonable given daily and seasonal variation in demand.

On the KWH required for a 35 mile pack, roughly speaking, I think Clett is assuming a Prius-sized car, the study assumes a car 15% heavier than the US average. That's the major difference.

You can build up either estimate starting from the calories in a gallon of gas, and the assumption that the average car is 20% efficient in converting gasoline energy to propulsion.

For a straight gas car getting 20 MPG, if you start with calories in a gallon of gas and do the math, you get .367 KWH per mile. If the car were perfectly efficient (zero losses), that would mean 12.9 KWH for 35 miles. At 61% efficiency, as assumed in the study, it gives (voila) 21 KWH needed for 35 miles.

For a straight gas car getting 50 MPG, ... , you get .147 KWH per mile. If the car were perfectly efficient, that would mean 5.1 KWH for 35 miles. At 80% efficiency (current Prius), that would be 6.4 WKH for 35 miles.

My only comment on the low figure is the phrase "straight gas car getting 50 MPG". I'm not sure such a car exists. Or if it does, it's pretty tiny. The Prius gets 50 MPG, but it's not a straight gas car. That is, I believe the Prius converts about 27% (??) of the energy of the gas into motion.

So, let's just run that to ground. What would a Prius require? Under the assumption that is Prius converts 27% of gasoline calories into motion, and is 80% efficient in translating electricity into motion, and right now gets 50 MPG, the 35 mile pack for a Prius would reuqire 8.7 KWH. Call it 9 KWH for 35 miles.

Which is actually less than Clett assumes in the thread about Toyota bringing forward the PHEV Prius release date (3 KWH for 10 mile PHEV). But is in the same ballpark, meaning, far below 21 KWH for 35 miles.

I hope this clears it up. The main difference is that the study assumes current US vehicle size plus 15%. I'd call that a Popeye assumption (I yam what I yam). With slightly worse electrical efficiency than the current Prius. But most of the reason for the high KWH number there is the large vehicle size. Clett's number seems reasonable for a 35 mile PHEV for a small car with good electrical efficiency, and is just below a fairly reasonable calculation for what a Prius actually would require. Although Clett's other post would put the reuqirements a bit higher than I calculated. (If the Prius actually does better than 27% at converting gasoline energy to motion, let me know, I'll redo the numbers.)

 

<div class='quotetop'>QUOTE(chogan @ Jul 28 2006, 11:25 AM) [snapback]293814[/snapback]</div>

In the thread you refer to, I assumed that Toyota would continue their approach of limiting available SOC. 10 miles range would only require 2-2.5 kWh. The remainder is left unused to protect the battery and extend calender life, like it does today.

 

<div class='quotetop'>QUOTE(ken1784 @ Jul 28 2006, 09:40 AM) [snapback]293770[/snapback]</div>

Hi Ken, I have read the paper. His assumption of 21 kWh for 35 miles is based on completely wrong assumptions. For some reason, he decides to calculate for himself how far an EV “should†be able to go on 1 kWh of charge. Why has he done this? He makes some appalling (from a physics point of view) assumptions about gasoline vehicle mileage and engine thermal efficiency to come up with his estimate.

To “guess†this fundamental part of his argument is crazy when the data already exists to show exactly how far an EV goes per 1 kWh of electrical energy. Why did he not use this? Hundreds of EV-users have been posting real-world kWh usage for years.

For example, a Solectria Force typically manages 5-7 miles per kWh.

http://www.austinev.org/evalbum/734

A T-zero sports-car can go 300 miles on one full charge of 50 kWh (6 miles per kWh).

http://www.acpropulsion.com/LiIon_tzero_release.pdf

GMs EV1, which was a heavy car using primitive, inefficient batteries, managed 75-150 miles range on 26 kWh (3-6 miles per kWh; 6 miles per kWh at 60 mph).

http://en.wikipedia.org/wiki/General_Motors_EV1

The Toyota RAV4 EV, which is a very heavy vehicle (3440 lbs) and perhaps representative of the average weight of vehicle most people in the States drive, typically manages 3.5 – 4 miles per kWh. When Darelldd (on this forum) drives gently he gets 5.2 miles per kWh from his RAV4 EV.

http://www.darelldd.com/ev/rav_data.htm

Now, since we are talking about PHEVs and not BEVs, remember that they will be MUCH LIGHTER than the BEV equivalents. I took a pessimistic approach and assumed that the average PHEV would probably get, on average for all vehicle types across the US, about the same as what the much heavier BEVs get (in reality they would get better mileage because PHEVs are lighter). Say, 4-5 miles per kWh.

35 miles range divided by 4-5 is 7 – 9 kWh required for 35 miles range. Unlike the author, this has been calculated from real-life test data, not assumptions of how an EV “should†operate.


<div class='quotetop'>QUOTE(ken1784 @ Jul 28 2006, 09:40 AM) [snapback]293770[/snapback]</div>

Yes, the first 35 miles per day would be in EV mode. But on a fair proportion of trips the vehicle would go beyond 35 miles in one day and the gasoline engine would have to come on. The author has not accounted for this simple fact at all and appears to be assuming all the PHEVs in his study have magically transformed into pure BEVs that do not use gasoline at all. The reality is that they would use some gasoline. I estimated 40% of total annual miles would use some gasoline. Of course BEVs would be a different story.

<div class='quotetop'>QUOTE(ken1784 @ Jul 28 2006, 09:40 AM) [snapback]293770[/snapback]</div>

I used the figures that the author himself quoted. He suggests 424 GWe would be required to charge up 21 kWh packs. But the packs would really be more like 7-9 kWh. If we recalculate with this adjustment we get 140 – 180 GWe required. (424 / 21 * 7 or 9). But because many miles would still be done on gasoline, assume 90 GWe required. The author states that entire US generating capacity is 850 GWe (page 6). 90 / 850 = 10.5%.

Now back to the other central question of how much spare electrical capacity there is in the US at night. It turns out… a lot. The data for California are presented at this website:

http://www.caiso.com/outlook/SystemStatus.html

You will see that the lowest consumption is at night, with around 25 GW demand (the red curve). However, potential supply is about 50 GW (the upper green curve). There is a HUGE amount of spare capacity – almost 50%.

This website discusses what could be done with all this spare capacity:

http://baltimorechronicle.com/2005/083005Korthof.shtml

There would be no requirement to build any new power stations EVEN if every vehicle in the entire country was a PHEV-35 that charged up from empty to full every night of the year.

 

<div class='quotetop'>QUOTE(chogan @ Jul 29 2006, 12:25 AM) [snapback]293814[/snapback]</div>

The 1000 GWe is an ideal number using every power sources run at the peak.

No. The 553*24*365 = 4,844,280 means 120% work load than actual usage.

Again, my number came from actual in 2005.
553*16*365.2422+276*8*365.2422=4,038,000 GWe.

It is just assumption is different.

It is Dr. Uhrig's assumption.

OK. The 50 MPG is your assumption. But, I don't think the current averages of light vehicles is 50 MPG.
Where does your 80% efficiency come from?
Did you think about all the grid-to-wheel efficiency?
grid -> charger (DC-DC converter) efficiency -> battery charge efficiency -> battery discharge efficiency -> motor efficiency -> drive efficiency(gear/bearing losses) -> wheels

Again, the 50 MPG and 80% efficiency are questionable.

Where does the 3 KWH for 10 mile come from?

No, it is not clear for me.
Your assumptions are questionable.

Ken@Japan

 

press release said a 3 KWH for 10 miles. that is there current working goal as i read it. that would maintain a similiar SOC as the current Prius

 

<div class='quotetop'>QUOTE(DaveinOlyWA @ Jul 29 2006, 01:42 AM) [snapback]293857[/snapback]</div>

Would you please show us the link?
Is the 3 KWh about grid power consumption, or what?

Ken@Japan

 

<div class='quotetop'>QUOTE(clett @ Jul 28 2006, 11:42 AM) [snapback]293824[/snapback]</div>

Clett,

I'll say a few more things then step out of this.

One reason you might calculate from scratch is to avoid relying on anecdote, and to avoid the self-selection that occurs with existing car owners who are willing to share mileage results. As with all things, those who do poorly do not publicize it.

You cite DarellDD, and the data there are a good case study. You may note from other posts on this board that I think DarellDD basically is correct about EVs being a good thing and that he has his facts straight.

So, DarrelDD averages .275 KWH per mile, per his website.

From EPA data, I calculate that the EPA assumes .304 KWH/mile.

Five Cal Edison meter readers averaged .40 KWH/mile. DarellDD provides a link to the Cal Edison 100,000 mile evaluation of the RAV-4 EV where that is published. That's a long term, systematic study using drivers who weren't particularly motivated to drive gently, in a somewhat older RAV4 with NIMH batteries.

But in fact, the RAV4 is a small vehicle by US average standards. The average passenger vehicle on the road in 2003 was about 24% heavier than a (gas) RAV4. You can calculate from the document below. I see the 2003 4x4 RAV4 curb weight listed as either slightly below or slightly above 3,000 lbs. The US average curb weight of 2003 cars was 3170, light trucks at (golly) 4699, from which I calculate an average curb weight for US passenger vehicles at about 3726 lbs, assuming 36% light trucks, which I believe is about right for the current mix.


http://www.nhtsa.dot.gov/cars/rules/CAFE/F...lupdate2003.pdf

DarellDD's EV weighs more because of the batteries. So, 3440 (EV)/3000 (gas) = 15% heavier, which, sorry to say it, is exactly what this study assumed, although they (almost surely incorrectly) assumed that 15% more weight = 15% less efficiency than would otherwise occur.

So, can a car do better? Sure. Can a driver do better in a car? Sure. Can a bunch of enthusiasts average better than this? Sure. Is the average Joe Shmoe going to do that well? I don't think so.

If I had money on the line on this calculation, would I use the data from a driver in a car (e.g., your first link) for the calculation, or a two-seater (EV-1), or advertised performance for a unique sportscar, or even assume the RAV4 was average weight? Nope, I'd go with the blue-collar guys driving around, adjust for the weight relative to the US average, and get on down the highway. Which, in fact, if I did, would give me slightly worse KW/mile than I actually assumed.

Anyway, the study did what it did, and stated it clearly. Convert all existing cars to EV, up their weight by 15%, assume fairly low efficiency, and ask how much more capacity that would require, based on roughly 2002 battery technology. I think the were somewhat pessimistic in their assumptions. Is that a good projection of what is expected in the future? No.

If you want to ask a different question and in my opionion far more reasonable question -- downsize the fleet and convert it to PHEV 35 --- of course you'll get a different answer. But you'll get the same answer whether or not you assume efficiency based on self-reported mileage from motivated drivers, just grind it out from the underlying physics, or even make some pessimistic assumptions such as that study made. See my earlier post -- PHEV 35, at the current US trip mix, would replace just over half the gasoline use in the US, not all of it. There should be no problem accommodating that with nighttime charging using existing capacity using your assumptions (for sure), my assumptions (for sure), or this study's assumptions (at the borderline). Only if we went beyond an all-PHEV-35 fleet would there even be the potential for disagreement.

Anyway, I agree so much with PHEV that I donated $1K to CalCars last week. Just something to help move the process along a bit. And I've been thinking about how much I value not voiding the warranty on my Prius with a PHEV mod, and have pretty much decided that in another year, I won't care. Let's get on with this and make this happen.

 

<div class='quotetop'>QUOTE(ken1784 @ Jul 28 2006, 12:33 PM) [snapback]293851[/snapback]</div>

There are threads elswhere dicussing optimizing Prius mileage. My recollection is that using the engine to charge the battery, then using the battery to run the electric motors, is calculated to waste 20% of the energy. In the Prius. Not having any better information, that's what I assumed.

The efficiency number does not account for transmission losses in the grid, which I understand are roughly 10%, but have not checked that number against anything.

As noted elsewhere, a 50MPG car is not average. Even the RAV4 is (by my calculation) 24% below the weight of the average US passenger car.

I just plain didn't understand your calculation of electrical generating capacity. Given the choice I would rely on the published data.

I'm now bowing out of this thread.

 

<div class='quotetop'>QUOTE(Cheap! @ Jul 28 2006, 05:58 AM) [snapback]293737[/snapback]</div>

Don't do that. They'll just turn it into oil via Fischer-Tropsche!