Trying to Understand Efficiency

[AlanSubie4Life]

Thus with my math, I used 26 miles to go 18.1. Meaning my efficiency was about 69% (quick math, think I did that right). But my car only indicated 316 Wh/mi. If the baseline efficiency for the car at rated range is 250 Wh/mi, wouldn't 69% efficiency be about 362 Wh/mi?

Getting back to your point:

So using the numbers above, I would say:

26*250 = x * 18.1 => x = 359Wh/mi is what you would expect the trip meter to read.

However, the numbers are wrong. First, use baseline efficiency of 242Wh/mi. Second, let's just say (to get things to align ) that you only used 23.6 rated miles (2.4miles less than what you stated, for fudge factor reasons I stated earlier). Admittedly, I am just fudging your numbers to make it work here.


Then:

23.6rmi*242Wh/rmi = x * 18.1mi => x = 316Wh/mi

23.6 rated miles to go 18.1 (76.7%)

316 Wh/mi vs. rated 242Wh/mi (76.6%)

Can I really burn almost 8kWh by parking overnight in a garage at about 60, and then outside for 8 hours in 30? That would mean if I left my car sitting in an outside lot for a day, it would fully drain.....

Are you sure of what your initial mileage was? In answer to the above question, you certainly can burn a few kWh overnight in the right scenario due to vampire, and if temperature of the pack changes a lot (maybe it was really warm after the supercharge), etc., that can cause the BMS/SoC estimator to change the miles available quite a lot. In short, there isn't really enough information here to tell. We really need to know the starting mileage when you got in the car (when the trip meter actually started - I understand you were using the last charge meter in this case, but for science, let's remove some variables like sitting overnight). 8kWh is a lot though. A lot of it is real unfortunately....

Next time you have an opportunity, post some data (pictures at beginning and end of drive) with a more controlled experiment. And we can go through the numbers again. As part of the experiment, please be sure to turn OFF climate control before exiting the car the night before the experiment. You can turn it on again as soon as you take the car out of park (but be sure to turn it off again every time before going back to park). Also don't spend any time in the car in park. After making these adjustments, let's try calculating again. Longer mileage runs like the one you posted above are ideal.

As I said, I agree the trip meter (and since last charge meter) is garbage. It makes this sort of exercise an exercise in futility unless you tightly control the conditions (which should not be necessary). It's just kind of dumb.

For vampire (not counting any preheating losses, etc.), just add 400kWh/(your miles traveled per year) to your efficiency numbers. So if you're doing 20k miles per year, add 20Wh/mi to your wall-to-wheel energy costs. It's approximate, but close.

 

[Nate977p]

Getting back to your point:

So using the numbers above, I would say:

26*250 = x * 18.1 => x = 359Wh/mi is what you would expect the trip meter to read.

However, the numbers are wrong. First, use baseline efficiency of 242Wh/mi. Second, let's just say (to get things to align ) that you only used 23.6 rated miles (2.4miles less than what you stated, for fudge factor reasons I stated earlier). Admittedly, I am just fudging your numbers to make it work here.


Then:

23.6rmi*242Wh/rmi = x * 18.1mi => x = 316Wh/mi

23.6 rated miles to go 18.1 (76.7%)

316 Wh/mi vs. rated 242Wh/mi (76.6%)



Are you sure of what your initial mileage was? In answer to the above question, you certainly can burn a few kWh overnight in the right scenario due to vampire, and if temperature of the pack changes a lot (maybe it was really warm after the supercharge), etc., that can cause the BMS/SoC estimator to change the miles available quite a lot. In short, there isn't really enough information here to tell. We really need to know the starting mileage when you got in the car (when the trip meter actually started - I understand you were using the last charge meter in this case, but for science, let's remove some variables like sitting overnight). 8kWh is a lot though. A lot of it is real unfortunately....

Next time you have an opportunity, post some data (pictures at beginning and end of drive) with a more controlled experiment. And we can go through the numbers again. As part of the experiment, please be sure to turn OFF climate control before exiting the car the night before the experiment. You can turn it on again as soon as you take the car out of park (but be sure to turn it off again every time before going back to park). Also don't spend any time in the car in park. After making these adjustments, let's try calculating again. Longer mileage runs like the one you posted above are ideal.

As I said, I agree the trip meter (and since last charge meter) is garbage. It makes this sort of exercise an exercise in futility unless you tightly control the conditions (which should not be necessary). It's just kind of dumb.

For vampire (not counting any preheating losses, etc.), just add 400kWh/(your miles traveled per year) to your efficiency numbers. So if you're doing 20k miles per year, add 20Wh/mi to your wall-to-wheel energy costs. It's approximate, but close.

This was pretty controlled. I can post, but I have pictures of it all.

 

[AlanSubie4Life]

This was pretty controlled. I can post, but I have pictures of it all.

Ok I would like to see a picture of the trip meter (not since last charge) and a picture of miles in the morning before you started driving. To take the overnight losses out of it. Those should be dealt with separately. They can be significant of course.

 

[SageBrush]

My WAG is that the starting range number estimate by the car was wrong.

OP: What was the reported range just before you plugged into the L1 ?
How long did you charge ?

 

[dsvick]

My WAG is that the starting range number estimate by the car was wrong.

I'm leaning this way too. I use the percentage display not range and my numbers are fairly in line with what I'd expect them to be. This morning my drive was
Beginning SOC: 75%
Drove 29.6 miles
Ending SOC: 63%
Displayed Wh/mi: 308
Assuming a 75kWh battery: 75 * .12 gives 9kWh used
Going the other way, 29.6 mi * 308 Wh/mi = 9.1 kWh used
Allowing for rounding my Wh/mi multiplied by my distance is always very close to my actual percentage of charge used.

From an efficiency standpoint though I'm not sure how to correlate the two numbers, maybe someone that knows how they correlate can chime in.
9.1kWh / 242 Wh/mi = 37.6 rated miles used
29.6 / 37.6 = 78.7% efficient
(308Wh/mi - 242Wh/mi) / 242Wh/mi = 27% additional Wh/mi used

I'm not an engineer or anything so I'm not sure how the 78.7% efficiency relates to the added 27% used but it sounds about right.

Maybe try your experiment again and write down the beginning number as range, then switch to % display and write that down, then do the same at the end of your trip. I'm betting the % is correct and your miles are off again.

 

[Nate977p]

Ok I would like to see a picture of the trip meter (not since last charge) and a picture of miles in the morning before you started driving. To take the overnight losses out of it. Those should be dealt with separately. They can be significant of course.

My WAG is that the starting range number estimate by the car was wrong.

OP: What was the reported range just before you plugged into the L1 ?
How long did you charge ?

I'm leaning this way too. I use the percentage display not range and my numbers are fairly in line with what I'd expect them to be. This morning my drive was
Beginning SOC: 75%
Drove 29.6 miles
Ending SOC: 63%
Displayed Wh/mi: 308
Assuming a 75kWh battery: 75 * .12 gives 9kWh used
Going the other way, 29.6 mi * 308 Wh/mi = 9.1 kWh used
Allowing for rounding my Wh/mi multiplied by my distance is always very close to my actual percentage of charge used.

From an efficiency standpoint though I'm not sure how to correlate the two numbers, maybe someone that knows how they correlate can chime in.
9.1kWh / 242 Wh/mi = 37.6 rated miles used
29.6 / 37.6 = 78.7% efficient
(308Wh/mi - 242Wh/mi) / 242Wh/mi = 27% additional Wh/mi used

I'm not an engineer or anything so I'm not sure how the 78.7% efficiency relates to the added 27% used but it sounds about right.

Maybe try your experiment again and write down the beginning number as range, then switch to % display and write that down, then do the same at the end of your trip. I'm betting the % is correct and your miles are off again.

Thanks for all the help here, and not going down the other road the thread did earlier

I started the experiment again today. I don't have my last night number as I did get the 110 spot, but have my this am. Will post that, arrival at office, leaving office, and home. That should give good data.

 

[SageBrush]

. I use the percentage display not range

The two numbers should correlate:
100% = 310 miles
==> 1% = 3.1 miles of rated range

My guess about OPs puzzle is that battery energy was expended overnight and not reflected in the start range until he drove off. This effect may be pronounced during L1 charging since the car overhead to stay on while charging is a larger fraction of the power delivered.

The easiest to see hint that this is what is happening is to monitor the rated miles range just as the car turns on and then a few minutes later. I expect a sharp drop as the car takes into account the losses while parked.
I'll go out further on a limb and guess that the discrepancy is only seen when the car does not reach its charge goal.

On 3rd thought: set the trip meter immediately, then the car will log the entire trip's energy consumption.

 

[Nate977p]

The two numbers should correlate:
100% = 310 miles
==> 1% = 3.1 miles of rated range

My guess about OPs puzzle is that battery energy was expended overnight and not reflected in the start range until he drove off. This effect may be pronounced during L1 charging since the car overhead to stay on while charging is a larger fraction of the power delivered.

The easiest to see hint that this is what is happening is to monitor the rated miles range just as the car turns on and then a few minutes later. I expect a drop as the car takes into account the losses while parked.

My original data was from not charging overnight. Went to SC then parked overnight.

 

[SageBrush]

My original data was from not charging overnight. Went to SC then parked overnight.

I think I have only seen the data your posted with an L1 charging.

For a case of first Supercharging then sitting overnight I invoke battery temperature changes. The symptom would be the same: Initial drop in range as the car recalibrates.

Do your calc after 5 minutes of driving.

 

[Nate977p]

I think I have only seen the data your posted with an L1 charging.

For a case of first Supercharging then sitting overnight I invoke battery temperature changes. The symptom would be the same: Initial drop in range as the car recalibrates.

Do your calc after 5 minutes of driving.

The numbers AlanSubie and I did were for the night I charged at SC then went straight home to a heated (60°) garage. But yes tracking data today.

 

[AlanSubie4Life]

Thanks for all the help here, and not going down the other road the thread did earlier

I started the experiment again today. I don't have my last night number as I did get the 110 spot, but have my this am. Will post that, arrival at office, leaving office, and home. That should give good data.

Start picture:
Turn off HVAC BEFORE putting in drive
For picture:
Car in drive but stationary
Climate off (you can turn on after picture)
Trip meter captured 0 miles 0kWh
Capture the rated miles in picture

For end picture:
Turn off HVAC before coming to a stop.
Picture taken while still in drive but stopped with HVAC off
Trip meter pictured
Capture rated miles


Need a pair of pictures for every trip segment.

 

[dhrivnak]

Hello All!

So I have had my P3D for about 13.5k miles now and get efficiency much worse than rated; obviously that is old news.

The part I am confused on is the displayed usage in the car. My understanding is that it does not take into effect heating/cooling when not driving, but does while driving. Take my commute to work this morning.

Starting indicated range: 72.
Ending indicated range: 46.
Drive distance: 18.1 miles (per car).
Temperature: about 25° and sunny.

Commute info: car started 3 stories down in a garage that was 50-60° warm and had been plugged in on 110 (I got the lucky spot!), Thus was not cold soaked. No meaningful elevation change.

Thus with my math, I used 26 miles to go 18.1. Meaning my efficiency was about 69% (quick math, think I did that right). But my car only indicated 316 Wh/mi. If the baseline efficiency for the car at rated range is 250 Wh/mi, wouldn't 69% efficiency be about 362 Wh/mi?

During all this the car did not sit on with the heat on at any point. All this data is from me unplugging the 110, getting right in, driving, and getting out.

Am I messing up the math or missing something?

Thanks!

Two points after driving our 3 18,000 miles. I find short trips are hard to judge as there seems to be start up inefficiencies. My watts/mile almost always drop the longer I drive. Case in point we are at 335 watts/mile after 7 miles but by the end of our trip we will be likely under 240. Now after 22 miles we are at 263. The second is for us to get rated range we need to be below 250 more like 239. But even that does not change your numbers much.

 

[AlanSubie4Life]

Two points after driving our 3 18,000 miles. I find short trips are hard to judge as there seems to be start up inefficiencies. My watts/mile almost always drop the longer I drive. Case in point we are at 335 watts/mile after 7 miles but by the end of our trip we will be likely under 240. Now after 22 miles we are at 263. The second is for us to get rated range we need to be below 250 more like 239. But even that does not change your numbers much.

This is normal; it is primarily due to initial heating, and sometimes possibly due to sitting in drive with the car stationary and the heat on (depending on your exact habits). You’ll see it less if you don’t have HVAC on. And any distances less than 0.5mi to 1 mile are subject to a lot of error in the Wh/mi reading.

Initial heating load can be as high as 12kW, while steady state it could be lower than 2kW, depending on ambient conditions. That’s 300Wh/mi additional vs. 50Wh/mi additional at 40mph. The slower the trip the worse it is. Shorter trips tend to be lower avg speed.

But I am off topic here since this does not affect the trip meter accuracy to my knowledge notwithstanding the “missed” counting previously mentioned.

 

[AlanSubie4Life]

Picture taken while still in drive but stopped with HVAC off

I take this back (though you should turn HVAC off before putting in park):
Actually I guess you might want to put it in park and *immediately* take the picture for the end of each segment. I think this may force an update to the trip odometer (as I said earlier, it does not update continuously so there is a possibility of losing info here unless you get the “complete” summary).

The general picture here is that there are ways the meter is crap, so we are trying to deliberately avoid all those so we can assess its actual accuracy neglecting all those crap factors. This won’t change the fact that there are a lot of idle losses (which the meter misses), which can be quite real. But if you can believe the meter (what we are trying to determine), then you can back calculate the enormous idle losses.

 

[dsvick]

The two numbers should correlate:
100% = 310 miles
==> 1% = 3.1 miles of rated range.

True, but we've seen enough posts from people where the numbers were way off of what they should be. Whether it is from cold, weather, or even a damaged battery.

 

[AlanSubie4Life]

True, but we've seen enough posts from people where the numbers were way off of what they should be. Whether it is from cold, weather, or even a damaged battery.

Actually other than defective batteries we have not (which is why I prefer the miles display to detect degradation). Don't confuse 3.1 miles rated range with actual miles driven. Rated range has nothing to do with range (or miles); it is a measure of available energy. The units are "rated miles," not miles. Rated miles is a unit of energy, not distance. EDIT: (If you want to convert rated miles energy (for the Model 3 LR/AWD) to something more familiar, multiply rated miles by 242Wh/rmi. This may not be a precise conversion as mentioned earlier.)

 

[SageBrush]

Actually other than defective batteries we have not. Don't confuse 3.1 miles rated range with actual miles driven. Rated range has nothing to do with range (or miles); it is a measure of available energy. The units are "rated miles," not miles. Rated miles is a unit of energy, not distance.

That is my take too, but if I was inclined to be picky I would surmise that rated miles are proportional to SoC. SoC is an approximation of energy in the battery affected by temperature, hysteresis, inaccurate coulomb counting ... and who know what else besides our main guess of uncounted vampire losses until the car is driven.

 

[AlanSubie4Life]

SoC is an approximation of energy in the battery affected by temperature, hysteresis, inaccurate coulomb counting

That's definitely true. The coulomb counting may not be as accurate in some cases and the SoC estimation is pretty tricky. How all of that works I do not know and there is no easy way to determine. First step is to make the trip meter as accurate as possible eliminating all the KNOWN things it does not count, and then go from there. It's quite possible that this experiment will show that the meter isn't 100% matched to the rated miles indicator (even after correcting for rounding error on the rated miles display). Or it may imply that the true constant is not 242Wh/rmi (if the error is consistently a fixed % in one direction for many controlled experiments).

 

[animorph]

Try this thread:

Car’s energy consumption (lack of) accuracy


I think this shows losses in the battery itself. The higher the power draw, the less energy you can get out of the battery. And the car can't measure that power lost in the battery. So the battery meter can decrease faster than trip kWh used, with the difference between the two being the battery losses.

You can see this in the number of kWh the car uses to charge the car and the number of kWh the car says it used since last charge if you are careful to avoid non-driving losses.

This must be the hundredth thread trying to reconcile the numbers over the past two years. They don't add up. Don't obsess over it. Search for old threads.

 

[AlanSubie4Life]

I think this shows losses in the battery itself. The higher the power draw, the less energy you can get out of the battery. And the car can't measure that power lost in the battery. So the battery meter can decrease faster than trip kWh used, with the difference between the two being the battery losses.

I've reviewed that thread and don't find that it aligns with my results, so I have tended to disregard its conclusions. Specifically I took a 200-mile trip and the trip odometer showed that the extrapolated full battery drain would have been very close to 75kWh if not a little more. It's also quite a dated thread.

However, I don't doubt that heating of the battery is a thing. But, it is not that hard for Tesla to correct for it. If they know how many watts are being drawn from the battery and being measured (however they measure it), they can approximate internal losses and correct for it. And it would be kind of silly for them not to, as it's a well-known source of inaccuracy in such monitoring - internal losses are real and they are real energy consumption! They just need a model for it and then they can probably guess it quite accurately.

You can see this in the number of kWh the car uses to charge the car and the number of kWh the car says it used since last charge if you are careful to avoid non-driving losses.

This is not quite correct, though it is a related observation. You get about 87% (this is approximate) of the energy added to the car from the wall (AC charging) to the actual battery, as available energy in the pack. The reason is heating of the battery (that's why I say it is related) and charger inefficiency. You should always expect to add a lot more energy from the wall than you used (15% more, 1/0.87 = 1.15); that's just the way these things work. It's also already accounted for in the EPA efficiency estimates (idle losses are not, but this is!). Again, the 87% is an approximation but it's within a couple % in my experience.

It's true this is a tricky topic and to some extent we rely on Tesla being truthful with their in-car meter. However, various battery breakdowns indicate that the Tesla Model 3 battery has a usable capacity in excess of 75kWh (but less than about 78kWh). If you meter the energy used from the wall to fully charge it, I assure you it will take at least 86kWh to do so. So that charging inefficiency is a knowable quantity. I've measured it myself since I can have Chargepoint do metering. (Of course, you depend on that meter being accurate too. But it's free so they don't have an incentive to cheat on that. And people with home metering have seen the same sort of losses.) This is the reason the very best RWD hypermilers get at best about 260Wh/mi wall-to-wheel efficiency with no HVAC use (220Wh/mi driving efficiency).