Featured Tesla has been cheating EPA on mpge and range numbers

https://www.caranddriver.com/features/a15388892/the-truth-about-epa-city-highway-mpg-estimates/https://www.motortrend.com/features/1505-how-the-epa-performs-fuel-economy-testing/

But even with gasoline only cars (hybrids included) the results can be very different than epa in the city or highway. These cars because of the poorness of the test differ widely from epa on highway trips.
Top 10 EPA Fuel-Economy Overachievers

Finally the edmund's test is really even less reliable for important numbers for real world range on long trips since it always goes in moderate temperatures in mainly stop and go LA traffic, their estimates will greatly over estimate range on real trips. Tesla is doing the right thing from doing the 5 cycle test. Still this needs to be improved for range for real world EV road trips. As I said most of us don't really care if we get a little less or more on short milelage days, its the long trip days that matter when choosing a EV.
Electric Vehicles Deserve a New EPA Range Test

 

Being in the EV business longer doesn't give them permission to break the laws of thermodynamics or to redefine the word perfectly.

https://electrek.co/2018/04/24/regenerative-braking-how-it-works/
This link says 60-70% across the industry

This real world test says about 51%
Prius Prime Regeneration Braking Efficiency Experiment | PriusChat

Mike

 

In a thread last year, someone, maybe the person you were responding to here, claimed near-100% efficiency in his own experiments. If that claim comes back up again, I'll revisit a challenge to his methodology and results that I figured but never posted.

 

The problem is what you are measuring and those not about averaging, and there are no bags on a EV for the exhaust.

First problem - some makers - Tesla and Audi do all 5 cycle tests, while most others use a 2 cycle test and the fudge factor from the prius (30%). The model 3 standard range + is very close to that 30% but others are further like the model Y. The model S performance with 5 cycles is only 14% below the 2 cycles.

The Adjustment Factor Tesla Uses to Get Its Big EPA Range Numbers


Motor efficiency and lower road load as well as being able to use a higher percentage of the battery capacity really do give tesla's longer range than similar cars. The more efficient heating only matters on cold days, the way regen works on a tesla doesn't really help on long highway trips, and is exaggerated on epa tests.

IMHO epa should change the high speed test to more accurately reflect long trip driving, but should do it 3 times, cold, hot, and and regular temperatures and report range for each. That would give us each a more realistic look. The city numbers are pretty accurate on my car, and I got worse fuel efficiency than epa in the city significantly on every other car but my gen 3 prius and tesla model 3. It is the highway that many evs don't do as well as epa.

 

Well, you may be right. But the coefficient of drag there doesn't look like a normal coefficient and did you state the actual measured cross secional area? Without that number the coefficient doesn't mean much. Note how the coefficent of internal drag is negative. How can numbers like this be a "obvious" indicator of the point and not just a "possible" indicator.

I'm amazed you admit the Prius Prime may have a "very large internal-drag force" compared to the Teslas. I thought you were saying that the Tesla M3 had worse everything, which if at 80mph the Tesla needs only 2 or 3lbs more force it would seem kind of improbable that every aspect of the M3 were worse, cause that would mean about 1lb more force for air drag, 1lb more force for rolling resistance and 1lb more force for internal/mechanical drag, instead of 5lbs more force for one of those and 2lbs less for another and 1lb less for the third.

But what we can agree on is the Telsa M3 needs a bit more power at the wheel to go down the road and that the EPA numbers favor the Tesla.

 

They did not. If we are accepting the Edmunds test results, only two of the eight tested Teslas returned less than EPA. They were less then 4% different from EPA. Of the other six, one matched the EPA, and the rest did better.

Tesla's published ranges were overestimated, because they use the same thinking as used for the tank ranges for ICE cars. Which assumes you manage to get the full listed amount of fuel into the tank, and you drive until there isn't even fumes.

As for why the others exceeded their listed ranges, I can only say for the Taycan. It has a two speed transmission, and it appears the EPA tests don't let the car make use of it for efficiency. Guesses for the others are they choose the calculation method that ends up more conservative results, don't count the miles after the 'low fuel' light would come on, including early capacity loss, or simple under reporting to not give customers too high of an expectation. Like Toyota has done with the Prius.

They are called bags because they are literal sample bags of exhaust from the cars taken during the test. The gases are analyzed for pollutant and CO2 content. From the CO2 numbers, they calculate fuel used. The EPA test doesn't measure the fuel going into the tank, and what is left over from the test. The bags have a wide difference because they are taken at times when the engine is cold, and when it is hot. The four bags are from one run of the test.

Remember, these tests were developed to test emissions. Fuel economy was just a side benefit.

Since EVs don't have a tail pipe, the energy used during the test is the only data point needed.

40 CFR § 600.114-12 - Vehicle-specific 5-cycle fuel economy and carbon-related exhaust emission calculations. | CFR | US Law | LII / Legal Information Institute
EPA Test Car Database with MPG vs mph charts | PriusChat

 

I found in Tesla's propaganda that they calculated the Model S as having the same CdA as a Prius in 2012. (I hate Tesla's sexual naming in their cars and literature).

the-slipperiest-car-on-the-road.pdf (tesla.com)

But I must correct something I said I read online earlier. According to lots of online referrences, the Model S has the same cross sectional area as the Prius. But Tesla shows that no, it doesn't. It is bigger, but they claim it has the same CdA. But that also mines my confidence in the info out there that the Model 3's CS area is supposedly smaller than a Prius's. Who knows, until someone comes up with an official measurement.

 

you can see the tesla model 3 versus the prius, other tesla's, and other cars. Car and Driver did test an earlier tesla and prius in a wind tunnel and found the tesla had slightly less aerodynamic drag. My Tesla is heavier and has a wider stickier contact patch then a Prius prime. Its pretty clear that the load (look at graphs) of the Prius for rolling resistance is lower, but drag catches up to rolling resistance at high speeds. Do Prius prime drivers go far all electric at 70 mph or more? Not likely, you will go at most 25 miles at that speed all electric, the short range that EVs all can handle, so yes the Prius prime will get more miles for the same electricity if it always switches to gas for the fast miles where it may be at a disadvantage. Does it matter, it is very efficient on gas and electricity. You could make the model 3 more efficient at city driving speeds, lighter, smaller contact patch, less sticky tires, but that is not what it is designed for.

 

That's what it seems like to me. If the Tesla starts out with considerably more friction of some sort (looks like rolling resistance) and ends up almost the same as the Prius Prime, it must have a better something that's good enough to let it catch up. If the Telsa Model 3 were worse in every category, it would start out worse and only get worser instead of nearly catching up to the Prime in total drag and friction.

Another thing to add is that Prime drivers will likely charge on 120V and Tesla on 240V. Usually 240V is more efficient, so from power mains to wheel the Tesla has that advantage. Not that I like Tesla or am defending them. But like any car there are usually both pros and cons.

 

I tested this ~9 years ago on my PIP. For a week I charged only with 120v and measured the wall kwh for every full charge. Then next week I used only 208v ChargePoint chargers at work. The 208v was about 91% efficient and the 120v was about 86 % efficient.

Mike

 

These were the "target" coefficients for the 2019 models:





If i use the "set" coefficients I get:






Negative coefficients violate the 2nd law of thermodynamics and obviously call into question the EPA protocol

 

@ToyXW

My question is target and set coefficients of what exactly?

What car or part of a car has a drag coefficient of about 0.011 to 0.019 (column C)? We were talking about 0.23 to 0.24 earlier for coefficient of drag (Cd) and 0.006 to 0.015 for coefficient of rolling resistance (Crr). Those I understand. Positive coefficients I can understand. I can't understand what I don't understand unless someone explains it to me or points me to the right book.

 

SAE publishes a detailed report on the protocol, but I don't feel like paying $125 to satisfy my curiosity. Best I can figure:




A*cos(atan(Gi-1)) is the primary function dealing with rolling resistance. At a grade of 0, it is simply the "a" coefficent.

B*Vi is the function that seems to deal with bearing grease/fluid resistance. It varies linearly with speed

C*Vi*Vi is the function dealing with aerodynamic drag. It varies with the square of speed.

Me*dv/dt is the function dealing with rotational inertia. Spinning flywheels, axles, wheels & tires all store energy when accelerated and release energy when decelerated. At a constant speed, the initial and final velocities cancel and the entire function becomes zero.

M*ag*sin(atan(G)) is the function dealing with the change in potential energy when you go up or down a grade. If the grade is zero, there is no elevation change and the entire function goes to zero.


So on level ground at a constant speed the equation simplifies to:

A + B*velocity + C*velocity*velocity

A = rolling resistance
B = bearing/grease/fluid resistance
C = aerodynamic resistance



A lot of online formulas and calculators combine A & B into one rolling resistance coefficient that may or may not include the variable of velocity. That should work fine over a narrow range of speeds, but this SAE formula appears more robust

 

... unless the coefficients are just from regression curve fitting of empirical results, not from analytic derivation from first principles.

If the real world operation needs a more complex formula than the results are curve fit to, or the figures include measurement noise or errors or variations, then finding some occasional negative coefficients in there is no surprise.

 

One untested hypothesis is the fixed number may be vehicle overhead ... power to the electronics. We know there are two, identical control computers and they have some pretty strong processors.

As for a negative coefficient, sometimes math and physics works that way. We might project what we think the linear component is but to match the roll-down metrics, a negative coefficient is possible ... curious but not uncommon.

Bob Wilson

 

I'd say coefficients violating laws of thermodynamics prove that the prescribed regression should not be considered valid for those data - even if the regression makes a nice "fit". Garbage in, garbage out...