Battery Ageing Model

[JejuSoul]

... I'm a bit confused about the cumulative energy charged/discharged (CED/CEC) counters reported by Torque.
Shouldn't these numbers agree with the trip computer results reported by the car?
....
So why the difference and which number is correct?

They are both correct. They measure different things.

The consumption numbers reported by the car uses only the energy that you pay for. "From the pump".
These figures are equivalent to a gasoline car.
Everyone understands mpg or km/l in this way.
It is these figures that are compared on government websites such as - FuelEconomy.gov

But that is not what we are interested in when we talk about the battery.
The battery is also charged by regeneration.
Battery cycles include the energy paid for, and the regenerated energy.
Battery degradation is proportional to battery cycles.

Example - Here is a screen from my 2015 Soul EV at 57400km. The battery and BMS have never been reset / replaced.



According to the battery data my fuel economy is 57400 / 11394.8 = 5.04 km/kWh.
My reported fuel economy by the car is always around 8.0 km/kWh.
( To understand the difference between fuel economy and consumption see - Fuel economy in automobiles

The large difference of the battery value for fuel economy against the car's value is due to regeneration. I live on the side of a mountain and go up and down hill frequently. Going down hill regenerates energy back into the battery.
In my dataset the person with the least amount of recuperated energy lives in the Netherlands.

[wizziwig]

I'm still not following. In the calculations I quoted, the owner used NET energy consumed from the battery. He subtracted the amount put back by regen.

28 kWh (difference in CED) - 2 kWh (difference in CEC) = 26 kWh = net amount actually lost from the battery.
I'm going to assume here that the trip computer is either reset automatically on each trip or that he reset it manually before the trip.

81.2 miles / 26 kWh = 3.12 mi/kWh

Even ignoring the energy recovered by regen, it still doesn't work:

81.2 miles / 28 kWh = 2.9 mi/kWh.

So how did the car calculate the 2.5 mi/kWh that it reported? Do you know which OBD counters and formula were used?

[ksoul2084]

Hi all, I recently bought a used 2016 Soul EV+ out of Quebec that was a used as a taxi for all of its life until now. My understanding is that it was likely DC Fast charged most of the time as well.

Car has 140k on the odometer. Currently a full 100% charge will report 83km range on the screen. Driving conservatively I'd say I am able to get 100km or so.

Some readings from SoulEVSpy:

Min Cell Det %: 19.1
Max Cell Det %: 36.3
SOH: 82.3%

SOH was 93.4% (Min Det: 0, Max Det: 33.2) 4 days ago before I started to drive it to work each day this week. I assume the min cell % was wrong at that point for some reason.

When I check the battery cell heat map after arriving home from my 35km commute I see:
7 cells @ 3.64v
36 cells @ 3.66v
51 cells @ 3.68v
3 cells @ 3.62v
1 cell @ 3.60
and 1 cell @ 3.58v

Thoughts?

[JejuSoul]

-
I am replying to the ksoul2084 message above and also his later comment in the thread - Re: Thinking of getting a higher mileage 2016

From the comments we see that his used 2016 Soul EV+ has 140k on the odometer.
In the last few weeks SoulEVSpy has reported -

Min Cell Det %: 0, Max Cell Det %: 33.2 SOH: 93.4%
Min Cell Det %: 19.1, Max Cell Det %: 36.3, SOH: 82.3%
Min Cell Det %: 0.7, Max Cell Det %: 33.7, SOH: 92.8%

These are strange readings, and taken together with the voltage deviation of one cell that is behaving differently from the rest leads me to assume that one cell - in this case #95 - is failing. But I don't think it has 'failed' enough for you to get a replacement battery yet. Although it would be worth phoning Kia to ask what criteria they will use to judge whether the replacement is necessary. Maybe they will accept the request based on the low GOM. My guess though is that you will have to wait a few months for the SOH to come down.
There's a more extreme case of cell failure reported here - Math doesn't work - 2015 Kia Souk EV w 36,000 miles - max deterioration is 98% - GOM 65 miles.

To provide some background. I have a 2015 Soul EV with over 58,000km. At 100% SOC my GOM shows 195km; the same as when I bought the car 4 years ago.
Here is the latest reading from Torque, and a plot of the Min and Max deterioration numbers over time. As you can deep spikes lower in the Min do occur fairly regularly. Other cars have seen this to, so I have to conclude it is normal.





I also own a 2011 Hyundai BlueON. That car has a failed cell at #61. I use the voltage heatmap on the SoulEVSpy app to monitor that cell. While driving that cell goes as much as 0.36V LOWER than the rest. When charging starts that cell will instantly switch from being lower to being HIGHER than the rest. Charging will cease when that cell is 0.36V higher than the rest.



Anecdotally all the websites seem to think that once a cell fails it will cause degradation in the rest of the pack to exponentially increase. But I am not seeing this. In fact I'm seeing the opposite. Because the weather is getting warmer, the battery pack in the BlueON seems to be getting better. I don't think I will really see what is really going on until the winter.

[ZuinigeRijder]

My Kia Soul EV from September 2014 has now driven 108251km and the SOH is now 93.55%. Note that the previous owner already got a SOH of approx. 102% after the first year.

Seems that it pays off to mainly charge till 80% SOC (yes, my model has still the option to charge till 80%) and try to keep it above 30%. I also do not drive fast, mostly 90 km/h on highway. My average speed is 57 km/h in the board computer. Also try to park/charge in the shadow. My car has now also the possibility to fastcharge to 91% instead of 86%, but I never used it, because fast charging above 80% takes more time (because it will not charge with 50 kW). I do not use fast charging often.

These are the latest readings of TorquePro, after a charge from 13% till 100% with the home charger:


I measured 23.64 kW with a meter to charge from 13% to 100%. This equals to 27.3 kW when I correct this to charge from 0%.
Torque Pro showed a difference in 23.7 kW CEC and 0.6 kW CED.


My battery cell readings show a Voltage between 4.14 and 4.12 Volt:


And these are the figures over time (you see that the BMS is twice reset due to a software upgrade):


In the winter the SOH seems a little bit lower, I expect because the Voltage is then also lower.



The BMS has been twice reset due to a software upgrade, so the CED and CEC do not show the real values.
I have accounted for the reset in the below figure.



The details of the figures you can find in this spreadsheet:
https://www.dropbox.com/s/h2syw43trwucamo/SOH.xls?raw=1

[JejuSoul]

-
Thanks ZuinigeRijder. That's great data. Your car is doing really well. You have the best battery efficiency figures for any car in this thread - 17.7 kWh/100km. I am assuming that is because you drive on flat terrain - no hills in NL - , and your commute has few traffic lights / speed cameras. True?
Having regen is better than losing all braking energy to heat, But it cannot reclaim all energy lost; it is most efficient not to need to brake at all.

-------------------

Here's some data from my car in Jeju and a second Soul EV also in Jeju. The R squared values are only for my car.
Both cars in the same climate zone, both cars drive about 14,400 km year, both cars always charge at 32A.
My car always charges at night to 80%, His car always charges during the day to 100%.
My overall battery efficiency is 19.9 kWh/100km, whereas his is 23.4 kWh/100km.





My car is doing about 10% better in both the time and distance trend lines.
No way to determine which of the possible causes for this difference may be,
a/ charging to 80%, rather than 100%.
b/ charging at night, rather than during the day. Given that it is cooler at night, battery temperatures will be lower.
c/ more efficient driving.
-------------------

I am only showing the battery cycle data from my car, because I don't have enough data from his.
I calculate the cycle use for each 250km segment based on the SOH at the time, rather than keep a constant 27kWh cycle.

As expected this measure is the best - statistically it has the highest R squared value - of the 3 charts.

[ZuinigeRijder]

-
Thanks ZuinigeRijder. That's great data. Your car is doing really well. You have the best battery efficiency figures for any car in this thread - 17.7 kWh/100km. I am assuming that is because you drive on flat terrain - no hills in NL - , and your commute has few traffic lights / speed cameras. True?
Having regen is better than losing all braking energy to heat, But it cannot reclaim all energy lost; it is most efficient not to need to brake at all.

Yes, I drive on flat terrain in the Netherlands. For my home/work commute I take quiet roads. Indeed without traffic lights/speed cameras.
Actually, taking the highway is a few km longer, but by taking the quiet roads I also drive less (2x37 km) than I would when taking the highway (2x42 km). So the km you do not drive, you do not have to charge. So you have more km left with the same battery pack And the quiet roads are also faster for me and more predictable in time. Consumption is also lower when you drive with a steady speed of around 80 km/h.
On the highway I often keep the same speed as the trucks, they drive in the Netherlands with a speed around 90 km/h.

I drive in D- mode, because this gives less regen and you need to gently reduce speed for a corner. I tried B-mode for some time, but this is more aggressive when you release the gas-pedal fully.

When I first measured CEC and CED, the car was 1 year and 8 months old. And the first year it was driven by the previous owner. Then the measure was 20 kWh/100 km.

In the spreadsheet
https://www.dropbox.com/s/h2syw43trwucamo/SOH.xls?raw=1
I added a column to see the number of Full Charge Cycles (assuming 27 kWh). I have now around 700 full cycles according to my calculations.
And a last column, where you see the kWh/km. You see a gradually improving number, in the winter I consume of course more.

[ksoul2084]

Another update on my 2016 Soul EV+ with 141K mileage.

So this morning for the first time my GOM read over 100km estimated range with 100% charge (101km). Outside temp. is 9c.

I've been adjusting my driving habits to better suit the car and this has been consistently increasing my range over time. My commutes to/from work are generally 15.1 - 15.5 kWh/100km now which seems pretty low to me.

This mornings readings from SoulEVSpy:
Max cell det.: 38.9%
Min cell det.: 18.9%
SOH: 81.01% (this has been dropping slightly each time I check it).

[JejuSoul]

...
Anecdotally all the websites seem to think that once a cell fails it will cause degradation in the rest of the pack to exponentially increase. But I am not seeing this. In fact I'm seeing the opposite. Because the weather is getting warmer, the battery pack in the BlueON seems to be getting better. I don't think I will really see what is really going on until the winter...

Here's the reason why I think the exponential increase in degradation frequently occurs.
If you continue to push the car to perform after a few cells have gone bad, these cells will suffer.
I believe the internal resistance of the cells increases as they degrade. They will lose more energy as heat than before.
This can be seen in a SoulEVSpy voltage heat map from a car in Canada. -



Cells in modules 1 and 3 have begun to go bad. This cause those 2 modules to heat up much more than the rest. 38C compared to 30C.
This extra heat will cause faster degradation.

Before the warranty expires this process will get you a battery pack replacement sooner, if you continue to push the battery hard.
After the warranty expires I suggest you treat the pack more gently, if you want it to last.

Another example of a car (also in Canada) with a cell going bad is here - Soul Spy thread

After discharging to 20% and charging back to 100% cell #1 showing 4.10v the rest showing 3.88-3.90v, so it’s clear that the cell#1 is bad. The temperature above cell#1 was 33C and the rest 28C... So I assume that the cell#1 reaches the max voltage and tell BMS to stop charging to avoid overcharge.

[JejuSoul]

-
The new 64kWh Soul EV uses the same BMS codes as the N iro EV and the K ona EV.
These are only slightly different from our own.

Recently in Canada an owner of a new 64kWh Soul EV with 2000km on the odometer has complained, "It charges at 10kW at 65% SOC. 283km on the GOM fully charged. Dealership doesn't know what is happening on this car. Any idea?"

This error has also been seen on new 64kWh N iro EVs in Korea. "Symptoms were, low GOM & slow charging speed, unexpected turtle sign".

It turns out one of the cells is completely out of balance.



(The Torque screen is not correct for this car - it looks like a 30kWh 2018 screen, that has 100 cells)

Is this a BMS balancing error, or is one cell faulty. Am waiting to hear how this get fixed.