e-Niro reference consumption too high?

[ChrSchaefer]

Hi,

for our first long distance trip with our e-Niro 64kWh I got a monthly subscription for ABRP for iOS. I noticed very early during our trip that we consumed _way_ less power than ABRP estimated. First we had a lot of traffic and had to drive slowly. We than changed our plan from four to three stops.

As an example on how much ABRP differed this is our last leg of the trip: https://abetterrouteplanner.com/?plan_uuid=4fa026de-d4f9-485c-b0e4-de89a003ec1e We left with 72% SoC in Schweitenkirchen and continued to Inzell where we wanted to arrive with 20%. Weather and traffic were ideal and I could drive at max allowed speed or 120 km/h (this was our setting in ABRP). We arrived at about the predicted time with 107km/h average and 29% SoC.

If I use ABRP with default settings it shows we would arrive with 13% SoC, which is a very big difference. 

With 170 Wh/km reference consumption the result is somewhat better with 20% SoC at destination. 

To get close to the 29% value I have to enter a reference value as low as 140 Wh/km.

I really wonder why the  difference is so big. We used normal driving mode, air condition, and electronics as usual.

[Lucas]

I have the same issue, i changed the consumtion to 144wh/km (from 159wh/km) on my Ioniq. Katya explained it to me as follows. ABRP lerns with the km you drive. In my case when i drive slow or medium the SOC forecast data are better then if i drive highway (fast), there ABRP is still missing drives. So i need to drive more highway to get the data for ABRP. The avarage consumtion of my car since march +- 4000km is acc. car around 13.2kw/100km

[ChrSchaefer]

Thanks for your reply. I guess ABRP can only learn from your car if you have some data connection established? 

I was under the Impression from the ABRP blog that they got enough data for a generic model if a car is not in alpha or beta anymore. 

I would not mind 5% or even 10% deviation, but in the case I described it was 37% deviation in consumption. I believe this can't be explained by driving style.

[ChrSchaefer]

See 

 

The e-Niro was released with 179Wh/km reference consumption. Currently ABRP sets 193Wh/km by default. I wonder why that has been changed. 

[Soulplayer]

We see the same thing. After 1h30m driving we were at 62% with ABRP estimating 44%... This is a huge difference. 

[Katya-ABRP]

Hi @ChrSchaefer, @Soulplayer,

As Lucas mentioned, the reference consumption is calibrated as you drive and generate data for ABRP to base a new reference consumption value on. However, a live data connection (which supports calibrated ref cons) is needed. You can see which connection options do support it here: https://abetterrouteplanner.com/compare/livedata

The default reference consumption does contain a bit of margin, we add roughly 10% to the originally calculated value just to keep users who don't use live data on the safe side of our predictions.

Currently we set 181 wh/km for the 39kwh model and 193 wh/km for the 64kwh model (margin included in these values). I would assume that the value 179 referred to the 39kwh model but the posts in the thread where this is mentioned isn't clear. I'll have it double checked.

If our predictions are too pessimistic, it could also be due to deviations between settings and reality for things like; speed or weather etc.

/Katya

[ChrSchaefer]

Thanks for the reply. I didn’t notice there could be a difference between the models with different battery sizes. 
 

I don’t have a live data connection. People having this seem to arrive at rather 150Wh/km. 
 

Unfortunately, with the very pessimistic default, new users get the impression that ABRP is not very reliable. Perhaps, if you get more real world data, the model can be improved. 

[Katya-ABRP]

Hi @ChrSchaefer,

Thank you for the feedback!

We do receive questions regarding reference consumption quite regularly and many times users do confuse it for average consumption. There are definitely things we could do to make the concept easier to understand. I'll take note of this.

150 wh/km is indeed much lower than ~175 wh/km which would be the current default for the 64 kWh model without the added margin. This model no longer has any beta label so in fact we do have a good amount of real-world data on it. But, if that data isn't interpreted correctly or matches the driving scenarios your routes includes this might not help us.

Would you say that there could be factors that we might not account well for? Such as elevation / more regeneration downhill, wind influence or similar?

/Katya

[poupoule]

Hi, I use ABRP with torque pro linked to it and what is strange is that the reference consumption keeps on changing depending on the kind of road I'm driving. Then ABRP act just like the guess'o'meter in the E-niro, reporting the remaining range depending on the speed I'm driving now (instead of considering the speed of the trip).
To me, driving at the speed limit (130km/h in France) I have to set reference consumption at 135Wh/km to get reliable informations.

Does it make sense to add +10% margin if the range prediction is reliable ? To me it doesn't, it's like setting my watch 10 minutes in advance to never get late, in the end I would still be late if I rely on my watch not being reliable