Comments on Darryl Siry's Blog: The Problem with EV Range Figures

Hi Darryl, These three points are crucial, but ano...

2010-01-24T00:36:38.539-05:00

Hi Darryl, These three points are crucial, but another danger looms in terms of public perception. That is the extractive nature of the lithium mining that is rapidly increasing in countries like Chile and Bolivia. Consumers will exert evermore pressure on firms to reduce supply chain impacts through time. Considering that lithium is the 31st most abundant element on earth, I think it's a race for the private sector to figure out more economic ways of harvesting it that also have less environmental impact. I wrote an article covering the life cycle impacts of electric car batteries you might find interesting here: http://carbonpig.com/article/electric-cars-dont-use-fossil-fuel-whats-environmental-impact-and-life-cycle-batteries
Hope this adds to the dialogue,
CarbonPig

Companies have been overestimating battery life in...

2009-07-09T01:30:16.960-04:00

Companies have been overestimating battery life in portable computers and cellphones for 15+ years, in the face of endless complaints by reviewers and customers. They do rundown tests at the lowest possible power consumption, when real life usage is easily twice that. So perhaps customers are already aware they need to cut the marketed battery life of any product in half to get a realistic figure. Or they know they can rely on reviewers and bloggers to test products and report something closer to the truth.

It seems easy enough to specify both the maximum and average range, but you have to be aware that customers in the mass market don't spend any effort at all trying to understand what the numbers mean. They are compared directly between vendors as one of the few measured differences in products bought for emotional reasons such as aesthetics and reputation.

For poorly understood products, adding any metrics at all is as dangerous to your sales as it is helpful, because your competitors can top you simply by being a little more deceptive. They can redefine your standard to something looser that looks the same to uneducated buyers. Witness "gigabytes" that went from a power of two to a power of ten - hard drive companies were sued, paid a petty settlement, and continue the practice today.

The way customers shop then is to try comparing apples to apples if they can. If every car maker overestimates range in about the same way, you can still see that an electric Mini gets twice the range of an electric Hummer, whatever the numbers are.

Lots of great thoughtful comments here - I guess t...

2009-06-25T20:50:41.497-04:00

Lots of great thoughtful comments here - I guess that happens when you have a great article to comment on!

a couple of thoughts from a HEV engineer:

1) Battery degradation is extremely complex to work with in an engineering context. The mechanisms are not yet fully understood, and are highly dependent on the chemistry within the li-ion space (phosphate vs. cobalt), and there still doesn't not exist a decent (if any) model that can be plugged into our simulation tools in order to estimate durability as a function of real world duty cycle. The cobalt understanding is better of course since it's been around for so much longer, but trying to predict the life of a battery undergoing various SOC swings in various, unpredictable temperature conditions is pretty much impossible still.

2) Prius and other hybrid batteries are lasting the full 8-10 years and 300000 km because those manufacturers oversized the batteries and limited the SOC swing. With a brand new prius you are using (If I remember properly) around 20% of the pack capacity. By the time end of life rolls around 10 years later, you are still getting the same battery performance, but since the chemical capacity has degraded, the SOC swing is getting up to 100%. NiMH's SUCK for battery degradation compared to Li-ion (in general). We've toasted packs in a matter of months by driving a hybrid like EV.

3) EPA looked at modifying the test cycles back in '93 to better approximate real world fuel consumption. At that time, their focus was on regulated emissions - and ultimately, the decided the benefit of improved fuel consumption estimates was not worth the massive cost of changing the compliance testing procedures (which is extremely valid!). They did change the procedures last year however to newer 5 cycle test (which is when the hybrid mileage dropped dramatically). I've had a chance to look through the proposed cycles for PHEV's from ANL and they are COMPLICATED! They are talking about 5 day tests - can you imagine what compliance testing will cost the car companies? I wouldn't be surprised if EV testing procedures are also under review, although I'm not sure who, if anyone, is working on them. Darryl's points are extremely valid in this area that the procedures need to change in a big way specifically for EV's. Never mind that the EV companies don't do themselves any favours by publishing ridiculous projections - at one point Aptera had numbers that were clearly breaking the laws of energy conservation (although I didn't read a single blog post about that one for some reason).

Yes, "environmental factors" play a part for sure....

2009-05-18T13:57:00.000-04:00

Yes, "environmental factors" play a part for sure. If it is too hot or too cold then you (and your battery control computers) may be turning the heating or cooling system on which uses up some of your traction pack energy that would otherwise have been available for more range.

Like with the Prius NiMH, on my NiMH RangerEV the range is noticeably better in the summer. I still don't fully understand why. Some have told me it is because warmer air is less dense so the vehicle doesn't have to push as hard to get through the air. Others have said that the lubricants are more viscous when cold which also hurts range. Perhaps the batteries don't chemically work as well in the winter. In any case, EVs can and do have different range capability depending on the weather. But that rolls back to "your mileage may vary"!

Thanks for writing this post. I touched on the con...

2009-05-15T17:39:00.000-04:00

Thanks for writing this post. I touched on the concept of range in one of my posts a few weeks ago. You've expanded one of my bullet points to a degree far more detailed than I could hope to write.

Thank you.

Don't we need a fourth asterik? I may have this w...

2009-05-15T12:18:00.000-04:00

Don't we need a fourth asterik? I may have this wrong, but aren't environmental factors involved as well?

Driving a Prius in 70°F weather yields me 52 MPG overall, but when its cold, the mileage drops. My friends in snow country lose 20% of their mileage gains when the temperature is below freezing. Running the heater causes the Prius to run the gas engine more. On an EV, how do you heat the interior? Electric resistance heating coils would use some of the charge, I'll bet. And increased rolling resistance from snow, etc. on the roads will also be a factor.

A third take may be in order. Electric range, lik...

2009-05-15T12:00:00.000-04:00

A third take may be in order.

Electric range, like conventional range, can be enhanced through software, instrumentation, and driver education.

I test drive a different vehicle every week to check against the EPA estimates and I've never missed the numbers (yet) ... quite often, my results greatly exceed the government numbers. I don't play games or use tricks ... I just drive sensibly (and yeah, I do mash the throttle when necessary).

If proper instrumentation is used (look at the crazy cool gauges in the 2010 Ford Fusion Hybrid as a starting point), drivers will develop a better feel for the energy flowing through their feet.

Thus the need for 20 minute charging and a 200 mil...

2009-05-14T21:24:00.000-04:00

Thus the need for 20 minute charging and a 200 mile range. To me, that's the perfect balance. I will gladly wait 20 minutes after driving 200 miles on a roadtrip. A washroom and snack break is needed anyway.

The "Your mileage may vary" is an understatement. ...

2009-05-14T21:18:00.000-04:00

The "Your mileage may vary" is an understatement. It isn't just about driving aggressively versus sedately. In a gas car with a large tank you can waste fuel and not notice it so directly. With the limited energy storage of an EV you become much more aware of the range penalty you pay for various energy usage differences. For instance, drafting behind a big rig on a highway makes an obvious improvement you might not have noticed in a gas car. On the flip side you can notice a range degradation from driving into a headwind or up a long hill. Switching to different tires can also impact your range in a noticeable way. You start to appreciate every little efficiency improvement you can find.

(Disclosure: I drive an EV with much shorter range than the Tesla Roadster so my experience is more extreme than most.)

@Nick I agree with you that the cycles should be ...

2009-05-14T20:32:00.000-04:00

@Nick

I agree with you that the cycles should be more accurate for both electric and gas cars. Lord knows my EVO never comes close to 24 mpg on the highway (but that might be driver related)

The problem is that since it takes 4 minutes to fill up a gas tank and several hours to charge a battery pack, the "penalty" for a EV driver is overall usability of the car while the penalty for the gas driver is just a slight inconvenience and more cash.

Darryl, Thank you for the clarification of the ex...

2009-05-14T20:07:00.000-04:00

Darryl,

Thank you for the clarification of the extended range mode. That said, my point stands about such deep cycles rarely being needed. The only time would be an long trip. Day to day driving will never touch this. This is the advantage to a huge battery.... 40 miles of driving won't have much effect on it at all.

As for the EPA stuff, it's one thing for the EPA to change its standards. But the onus should not be on EV makers to develop their own standard test. That would only serve to make EVs appear less attractive than gas counterparts. If the EPA changes its methodology then it would be ok, because gas cars would be held to the same standard.

Not sure about you, but I think Toshiba's SCIB is going to be fantastic for EVs. These batteries will be good for well over 10,000 deep cycles. They're also rapid charge capable -- 20 min for a full charge.

In my view, 20 minute rapid charging combined with a long deep cycle life will make EVs perfect for almost all consumers. And it's right around the corner.

I personally want an EV with a 200 mile range, 20 min charging, and a long life expectancy. SCIB delivers on all fronts, though A123's Nanophosphate is also very interesting.

I am surprised Tesla isn't using Toshiba's battery for the Model S. I'd rather have a 230 mile range via the lower density SCIB batteries than a 300 mile range via lithium cobalt.

In response to Todd Lassa - the EV1 used NiMH batt...

2009-05-14T19:29:00.000-04:00

In response to Todd Lassa - the EV1 used NiMH batteries which do not suffer as much from this problem. (So GM doesn't have that as an excuse ;-).

There are many Toyota RAV EV drivers who have 100K plus miles on their original batteries and still get about the same range. Also there are many Priuses in taxi fleets with over 300K miles and still getting the same capacity on the original NiMH battery.

Darryl, This was a good post and I think I can ad...

2009-05-14T19:26:00.000-04:00

Darryl,

This was a good post and I think I can add another asterisk to your list. It turns out that it's very difficult to measure the state of charge (SOC) of a Lithium-ion battery which compounds the EV Range problem you describe. To the driver this translates to having an undependable and/or non-linear "gas" gauge. You never really know how much charge is left in the battery at any given time. We've all become used to having computers in our ICE-powered cars that give us an accurate "distance to empty" based on remaining fuel in the tank and current driving conditions. Even if you're driving hard you still have a good idea of when you'll need gas. It just isn't so in an EV because the SOC measurement depends on things like voltage, temperature, state of battery health (age), how you've been driving, and a lot of other factors. Fortunately, it looks like this is an area where technology can be applied and we should see constant improvement in the accuracy of our SOC gauges and DTE computers in electric cars as the battery pack manufacturers continue to improve their SOC algorithms.

Having performed official EPA testing on productio...

2009-05-14T19:23:00.000-04:00

Having performed official EPA testing on production EVs myself, I think Darryl's comments are very poignant and his estimates are pretty good too. Range estimation and communication is a crucial aspect of marketing EVs, and we need a better approach.

@Nick Aziz I think the analysis is pretty accurat...

2009-05-14T18:21:00.000-04:00

@Nick Aziz

I think the analysis is pretty accurate, unfortunately. And it applies to everyone, not just Tesla.

A 20% decline in capacity over 5 years for a cobalt oxide pack would be miraculous. That's going to happen even if you drive zero miles due to calendar year effect alone.

Also - 1000 cycles a cobalt oxide pack is way too high. Can you share where you got that info? Have you seen data to that effect? You are right that the active cooling in the Tesla pack helps mitigate this effect but it isn't a panacea. Without cooling, the effect would be totally unacceptable I think.

Last, the point about the EPA cycle being too optimistic is more about customer expectations vs. reality. I agree with you that the the dilemma manufacturers face is that the EPA cycle has become the standard of comparison, so you can't avoid communicating it. That just makes the problem worse, which is why I think the EPA has to change the cycle.

The Tesla Roadster is set up to allow you charge to "extended range" where 100% charge is possible. The "standard range" is 85% charge. The EPA range numbers come from the former. You should check with them on this though.

I think there are some serious flaws in this analy...

2009-05-14T18:04:00.000-04:00

I think there are some serious flaws in this analysis, and I'm a bit surprised since you work in the field.

This comment about expecting a 20% decline in range after 5 years is particularly odd. Since you've been talking a lot about the tesla roadster, let's use that as an example. Let's assume its real world range is just 200 miles. Using it as a daily driver, one might go 12,000 miles per year. So that's 60,000 miles in the 5 year timeframe you're talking about. Even if you run the battery pack from full to zero every time, you're only talking about 300 deep cycles. The Roadster's lithium cobalt pack is good for 1000 deep cycles before it's left with a 20% reduction in capacity. So how does this make any sense? On top of that, who is gonna deep cycle their battery constantly like that? It would only happen for long trips. The daily commute followed up by overnight trickle charging should not result in any major capacity loss.

As for your EPA comments, one has to compare apples to apples. Gas cars are rated in this way, both range and milage. A BMW 550i for example has a 312 mile EPA range. So advertising the Tesla Model S with a 300 mile range is totally fair in the context of competition.

I am also confused by your claim that EPA tests involve fully charging and discharging the pack beyond its "allowed" range. My understanding of the Tesla battery pack is it always keeps between a certain range (perhaps 80% of theoretical capacity of the cells) and it won't let any driver exceed this. Furthermore it's my understanding the EPA test is done using the car in a standard operating mode and 244 miles reflects this.

I think we also need to consider the effect of the liquid cooling. By preventing batteries from getting hot you greatly extend the life span.

That said, I am a big fan of lithium iron phosphate and lithium titanate chemistries. Toshiba's SCIB is especially interesting due to good density and very agressive anticipated pricing.

Daryl: Good post. It seems the right time for the ...

2009-05-14T17:57:00.000-04:00

Daryl: Good post. It seems the right time for the EPA to implement a test cycle based on the one it uses for conventional internal combustion-powered cars and trucks. The test cycle should take into account a variety of driving conditions and varied speeds. And your asterisk #3 serves as a good explanation of why GM didn't release its EV-1 to the public after the lease-test period ended.
-- Todd Lassa
Detroit Editor
Motor Trend