[sciencegeek]

Thanks to ryzvy for motivating this thread with her thread "Energy consumption".

I just did a simple test, where I hammered it on the freeway for 7.2 miles and then drove home via surface streets for 7.2 miles, noting the SoC and the reported kWh/100miles at each step.

I'll spare y'all the gory details but here's the upshot: 14.4 miles used 11% of battery capacity. Assuming 84kWh battery capacity, that's a consumption of 0.11*84/0.144 = 64 kWh/100miles. The reported consumption was 35.4 kWh/100miles.

Yes the numbers are small (few miles, not much SoC delta) but the contrast is too large to be chalked off to some measurement variance. The culprit is the reported kWh/100miles. Its calculation is totally wrong. It's fine if you keep constant speed, but the more you vary your speed the wronger it gets because it's dominated by the time you spend driving efficiently and it incorrectly downweights less efficient driving because its calculation of the average is incorrect.

I'm dumbfounded that JLR could be making such a rookie mistake and I'm wondering if they did that on purpose to make people feel like they have great numbers when in fact their efficiency is far lower. The long term averages, as well as the averages for short trips with highly varied consumption within the trip, are completely totally utterly wrong. Bad JLR.:surprise:

 

[Timbo]

Well, the opposite of that would be driving in my Sprinter. You fill the tank up and it says 800kms range. You get on the highway it drops to 600kms range. You go up a slight climb it drops into the 400’s and then then go downhill and it jumps up to the 700’s. Now THAT is completely useless.

Jag uses an average based on your trips and it doesn’t change to much based on the occasional sprint. Hop out on the highway for 90 miles at 90 miles an hour and the next time you get in the range will be less than you are used to seeing. Piddle around in the city and it will go back up. It is always adjusting and a 14mile run isn’t long enough for it to average out the previous trips. Eventually it will learn your driving style.

 

[epirali]

Honestly all EVs do what we joke as “guess-o-meter” because there is no real answer. We are just used to not having as much of a consequence when you fill up so you don’t notice. But there is no way to look into future. That’s why I for one reset my miles on full charge then compare to SOC along with range.

Tesla seems to lie and always give a consistent range on start but truthfully varies a lot. Leaf would adapt too fast to more recent data. The i3/i8 actually have a cool trick that IF you have a destination the computer not only will optimize EV use but will adopt the range based on topography and road speeds (I’ll see the remaining range increase or decrease when I set destination if it’s a significant portion of remaining charge).

So far in 3 days I am finding that the Jaguar range guesstimate is not bad if I am not making drastic style changes, but then again I don’t live in a very hilly region of country.

 

[sciencegeek]

I'm not talking about the GoM or anyone's driving style. I'm talking about a calculation JLR makes based on measured parameters. Not predictive estimates.

 

[epirali]

Sorry my bad!

 

[PGhiron]

@sciencegeek I think you are looking at it wrong from the following perspective. In range prediction you do not use straight averages but rather a weighted average based on "intent". What you did most recently is weighted heavier than what you did in the past because empirically you are more likely to continue dong what you are currently doing than what you did first. In other words if you had reversed your test you would have far different results.

 

[sciencegeek]

I'm not talking about range prediction. I'm talking about the consumption figure for the completed 'journey' and the historical trip average in kWh/100miles and how it's incorrectly calculated. It's not a prediction.

 

[Sinister_Crayon]

You're also assuming that the percentage used shown in the system is linear but I'm willing to bet it's probably not. Battery charge and discharge rates vary wildly depending on the state of charge as evidenced by the fact that most EV's accelerate faster when they have a full charge as opposed to a partial charge.



You can control for this by trying the same experiment at various charge levels... I'd probably do one at 100, one at 70, one at 45, one at 25 and one at 10 and see what the average looks like... the reason for the curve in the starting charge amounts is to control for the varying discharge rates across the total charge amount. Depends how much free time you have and how much you like driving your I-Pace

 

[ghost]

Actually, I believe the reported average consumption is fairly accurate, but the reported battery percentage is very inaccurate. We already now that the reported percentage (we like to call it SOC) is not linear. If you charge the car and let the battery have time to balance properly, you will notice that you can drive fairly long before the SOC goes down to 99%. Also, there are reports saying you have a similar situation when you reach 0%.

I think we can assume around 3kWh between 100-99 % and the same from 0%-stop. That's leave only around 79 kWh for the remaining 99%-0%.

In addition, it looks like the SOC is not linear in the interval 99-0% either, so any calculation involving a small SOC interval is doomed to be flawed.

If you charge the car completely full and run the battery all the way to 0% is displayed you can calculate the used kWh my multiplying the average consumption with the distance. Add the last 3kWh from 0%-stop and I think we will
be fairly close to the claimed 84.7kWh. At least if you do this when the ambient temperature is not too far away from around 70 degree Fahrenheti (20 degree Celcius).

 

[sciencegeek]

The reported SoC is quite accurate, at least in the midrange. See attached chart of a 7.2kW charging session yesterday. The charge added explains 99.5% of the variance in SoC (that's the R^2 value).

I'm also attaching a spreadsheet where you can see that over a couple of back-to-back ~60 mile drives which center on a SoC of 50% the calculated consumption based on SoC was way higher than that reported by the car.

=> SoC is not the issue here; the issue is with the reported consumption [kWh/100miles].

 

[ghost]

Your second graph is interesting and indicates that the SOC (I assume this was taken from the car and not from the charger) is fairly linear during charging in the 30-65% interval range.

However, even if we assume linearity between 0-99% your calculation in the table will not be correct without accounting for the much larger 100%(full)-99% and 0%-empty slots.

If we account for these 6kWh then each percentage used in the interval 99-0% will have a size of (84.7-6)/99 = 0.795 kWh.
If we do the same calculation as in the table using these numbers we will get:
29*0.795 = 23.06kWh. If we account for roundoff errors in the displayed values for SOC the range of consumed energy will be [22.26kWh-23.85kWh].
In other words, these seemingly 29% of battery actually represents somewhere between 26.3 and 28.2% of the net battery capacity.

I agree there is still a discrepancy in the numbers, but I'm not really willing to conclude that the reported average consumption is the bad guy here.
There are at least two other alternatives.
- Reported SOC could be inaccurate
- There is less energy in the battery than the claimed 84.7kWh. We know that the colder it gets, the less energy can be harvested due to increased internal resistance.
If we assume that the reported average consumption does not include the extra heat generated due to this increased resistance (which I guess is logical) then
we can only make sensible measures at fairly warm battery temperature.

 

[sciencegeek]

The battery is a nominal 90kWh and the usable range (i.e., the 0-100% range) is supposedly 84.7 kWh. I think it's actually less but not 6 kWh less. I agree that there are confounders but they can't explain the results.

Regardless of SoC, I recommend the following to convince yourself that something weird is going on with the consumption calculation: drive hard for a few miles, watch the real time kWh/100miles figure in the Eco Data view. Then get off the freeway and watch the kWh/100miles drop precipitously even though you've only covered a tiny fraction of the total distance driven on this 'journey'. Don't need SoC to see that something is awry.

 

[Time2Roll]

I looked at these screens on my LEAF back in 2011. Lost interest. It is what it is.

 

[ghost]

The battery is a nominal 90kWh and the usable range (i.e., the 0-100% range) is supposedly 84.7 kWh. I think it's actually less but not 6 kWh less. I agree that there are confounders but they can't explain the results.

Regardless of SoC, I recommend the following to convince yourself that something weird is going on with the consumption calculation: drive hard for a few miles, watch the real time kWh/100miles figure in the Eco Data view. Then get off the freeway and watch the kWh/100miles drop precipitously even though you've only covered a tiny fraction of the total distance driven on this 'journey'. Don't need SoC to see that something is awry.

When 0-100% is used like this a lot of people tend to misunderstand. The net capacity of 84.7% is measured from where the BMS stop any further charging (real net 100%) until the BMS will stop any further draw of current (real net 0%). These levels are not the same as the displayed SOC 0 and 100%. There is about 3kWh between a full battery and when the meter starts showing 99%. In other words the car can show 100%, but still lack 3kWh of charge compared to a full battery.

Also, we all know that the car does not come to a halt as soon as the display show 0%. There is more energy there to use.

The usable capacity and the EPA and WLTP ranges are measured between these two extremes.
In addition, there is a brick protection ofr 90-84.7 = 5.3kWh which you cannot use. This buffer is there to prevent damage to the battery.
(disclaimer : It seems to be possible to add some energy through the regen system into this brick protection buffer)

You might be correct that there are weird short term fluctations. I really haven't done much analysis over short distances. I have made a lot of observations over longer trips though, and the average cnsumption displayed has looked fairly reasonable. Also the 100%(full battery) to low number SOC looks OK and do not conflict with average consumption. It's only when I perform measurement in the mid range of SOC I get into this discrepancy.

 

[sciencegeek]

Here's a preview of actual consumption calculation based on battery SoC in kWh reported by the API and made accessible by WattCat.

Prepare to be shocked.

Mostly suburban driving, fairly aggressive because I have zero patience in traffic and make full use of high-regen one pedal driving. A little bit of freeway.

Regardless of driving style, it's crazy and in fact embarrassing that the JLR computation off by this much.

I will keep posting on this until JLR figures this out.

 

[2761]

Here's a preview of actual consumption calculation based on battery SoC in kWh reported by the API and made accessible by WattCat.

Prepare to be shocked.

Mostly suburban driving, fairly aggressive because I have zero patience in traffic and make full use of high-regen one pedal driving. A little bit of freeway.

Regardless of driving style, it's crazy and in fact embarrassing that the JLR computation off by this much.

I will keep posting on this until JLR figures this out.

Thanks, Sciencegeek. I just started using WattCat and only have one trip as a data source but I can confirm Sciencegeek's findings. I-Pace/InControl reported numbers were underestimated by 20% compared to actual consumption.

 

[2761]

At 100% SoC WattCat shows 82.55 kWh available and if we assume 1 or 2 kWh available below 0% SoC to give the driver a safety net then usable battery is around 84.7 kWh as stated by JLR. Now if they can fix the inaccurate consumption numbers we can finally have an accurate range prediction.

I would like to have available battery kWh and miles driven per kWh used added to the trip bank then I could do the math for range myself.

 

[juice]

Here's a preview of actual consumption calculation based on battery SoC in kWh reported by the API and made accessible by WattCat.

Prepare to be shocked.

Mostly suburban driving, fairly aggressive because I have zero patience in traffic and make full use of high-regen one pedal driving. A little bit of freeway.

Regardless of driving style, it's crazy and in fact embarrassing that the JLR computation off by this much.

I will keep posting on this until JLR figures this out.

I'm not shocked. Something doesn't add up, but your trips are way too short and there isn't enough available information to draw any rational conclusions. There is probably a simple non-nefarious explanation for what you're seeing. For example, the car may be warming the battery and JLR may be excluding energy used for that from the consumption numbers it presents. With such short trips, the effect of something like that could be significant. Another possibility is that the JLR numbers are being misinterpreted. Maybe JLR's consumption numbers are computed entirely based on energy used to move the car forward, so if you're using Max AC with seat coolers on full blast and sitting in traffic a lot, a significant portion of your battery energy may not be factored into the JLR numbers. I'm speculating without any facts or evidence.

 

[eJeff]

I also noticed some time ago that the reported efficiency did not seem to match my rough calculations, so I decided to do a data analysis the old fashioned way: measure how much total energy I put into the vehicle over a long period and then divide it into the number of miles I drove over that period. I used the report from my ChargePoint home charger, which is the only charger I have ever used, to determine the total amount of energy I put into the vehicle. Then I simply divided the number of miles driven over the same period by the total energy. In my case, I chose April 16, 2019 as my start date and August 23, 2019 (yesterday) as my end date. During that period, I put a total of 1,361.4 kWh into my I-Pace. Over that period, according to the Jaguar InControl app, I had driven 2,760.9 miles. The simple math is that my efficiency was 2.03 mi/kWh. There will be a slight error in this calculation because my battery was not 100% full on April 16, nor is it full now. The ideal calculation would start and end at 100%, but it is still quite accurate.

So, how does this compare with what the car tells me? Using the InControl app, you can easily calculate the amount of energy reported by the car per trip. I assume the app reports the same number as reported on the display in the car, but I don't know that for sure. If you divide the miles reported for each trip by the mi/kWh reported for each trip, you will get the kWh used on each trip as reported by the app. In my case, over the same period as I used above, the total energy used was 1,103.9 kWh - clearly a lower number than what ChargePoint tells me I put into the battery. In fact, it is off by 23.32%. This is a smaller error than reported by sciencegeek, but still clearly significant.

My typical driving mission, for your comparison purposes, is around the San Francisco Bay area in a combination of city and highway driving. I typically drive 70-80 mph on the freeways and close to the speed limit when possible on surface streets, however quite a bit of time is spent in stop/go traffic on freeways and city streets. I would estimate about 25% of miles are in stop/go traffic and 75% is at high speed on the freeway. However, if you look at time spent, it is probably more like 60% in stop/go traffic and 40% at high speed. Temperature here since April has averaged between 60 and 90 deg F. and I use AC always set at 72 in the car. Some statistics accumulated from InControl are below:

Average trip length: 11.04 mi
Total miles: 2,760.9
Total energy reported by InControl: 1,103.9 kWh
Total energy reported by ChargePoint: 1,361.4 kWh
Total regeneration reported by InControl: 304.7 kWh
Average efficiency reported by InControl: 2.50 mi/kWh
Average efficiency calculated using ChargePoint data: 2.03 mi/kWh

Note that I assume that the InControl app is taking regeneration into account when it reports energy used per trip. If not, then the reported efficiency is off even more than my calculations state. I intend to keep tracking the data over the first year of ownership and then again in a few years to see if there is any degradation in efficiency over time.

 

[eJeff]

For those of you who prefer the kWh/100mi metric, for my numbers above they are:

ChargePoint data: 49.26kWh/100mi
InControl Data: 40.0kWh/100mi