Battery Ageing Model

[Birkeland]

I will do a calibration next week and keep you posted. I am currently on paternity leave with twin boys, so getting in my car and drive for 200km in one go is a luxury I don't get to do everyday these days

To monitor the Cumulative Energy Charged and Discharged counters, do I have to reset the counters (they are currently showing about 4000kWh as seen in my former post), or do I write down the values before the test, and compare it to the new cumulative numbers at the end?

 

[JejuSoul]

A calibration check doesn't need to be done in one drive. Take as long as you want to drive the 200km. I did mine over 3 days. I timed the start and end points so that the car would not sit very long at either zero or 100% SOC. It is only a Tony Williams style range check that would require a single constant speed journey.

To reset the trip meter and energy consumption meter use the button on the steering wheel. Although we don't really need these. The BMS cannot be reset unless you happen to be Matilda's Dad and are selling used EV's as new. You need to screenshot the energy counters from Torque 3 times. At 100%, zero, and 100% again in order to calculate the energy used while driving and charging.

And congrats on the twins. Hope you are enjoying being a father.

 

[ZuinigeRijder]

These are my detoriation numbers over time. Did not have all data.

Date	km	Max Det	Min Det	Max Cell	Min Cell	BMS	SOC	CEC	CED
2016-01-17	18709	11.6	10.0	3	15	55.00	57.00		
2016-01-27	?????	11.1	9.6						
2016-01-??	?????	13.4	11.1	43	51				
2016-02-01	?????	10.2	8.2						
2016-02-??	?????	10.0	8.4						
2016-02-??	?????	10.4	8.4						
2016-03-04	?????	11.6	10.4	3	8	32.00	32.50		
2016-03-07	?????	11.6	10.4	3	8	18.00	17.50		
2016-03-07	?????	11.1	9.6	3	83	95.00	100.00		
2016-03-11	?????	11.8	10.3	9	15	46.50	48.00		
2016-03-14	?????	11.1	9.6	9	8	95.00	100.00		
2016-03-21	?????	11.0	9.1	43	49	95.00	100.00		
2016-03-25	?????	12.6	10.6	42	59	95.00	100.00		
2016-04-01	?????	13.7	12.1	42	15	95.00	100.00		
2016-04-08	24454	11.1	9.6	9	1	95.00	100.00	5062	4951

Looks like the detoriation numbers change a lot, dependent if battery calibration has been done or not.
The first and last row are close, while 5745 km driven.
From now on I will also measure CEC and CED when at 100% and just before and after calibration. Calibration seems to be triggered when SOC is below 20% and charging to 100%.

 

[JejuSoul]

The deterioration numbers on my car have just improved another 1%. That all four of us have all seen improved numbers in the last few weeks is interesting. The average mileage on our cars is 22,382km. Our average deterioration numbers have just improved from 9.4% to 7.5%. (Note - these numbers are not SOH, our cars still have 100% of usable capacity)

Calibration can go either up or down so why are we all seeing improvement. This could be:-
a/ random, the same odds as tossing a coin and getting 4 heads in a row.
b/ self selection, only people with improvements are posting the data
c/ a spring effect. Deterioration numbers are artificially decreased in cold weather and will adjust back to normal in the spring as the weather warms.

I have been looking on the mynissanl eaf.com forum at recent battery data postings which seem to be commenting on the same effect. More data would help to show if this effect is real or not. Am hoping the OVMS users could post their latest data.

Update: A week later and the deterioration numbers on my car have just improved again by 1%.

 

[AndY1]

What is the difference between deterioration numbers and battery SOH?

I thought, that the 100% - deterioration number in % = SOH in %.

 

[JejuSoul]

SOH = State of Health = percentage of usable capacity remaining. We don't know how to find this value from the BMS data yet. Maybe soon.
The deterioration numbers we do have are the max and min values which show the percentage of total capacity lost.
There is a buffer at the top of the battery which shrinks first before any usable capacity is lost. We don't know how big that is. If I had to guess I would say about 7.5% which is about where we are at the moment.

 

[AndY1]

Aha, I understand now!

It's like in Ampera, where the SOH windows ends at 85.9% SOC (and starts at 22%). The 85.9% SOC -> 100% SOC (14% battery degradation window buffer) is the SOH buffer for battery degradation.

 

[Birkeland]

Then I finally completed the test of the batteries. No change in the deterioration values.

Start:

Before Charge:

Finished:

 

[JejuSoul]

Birkeland: Thanks for the data. So how many km did you drive in that test. All 3 cars in this recent test have similar values. All 3 cars are a year old and seem not to have lost any usable capacity.

JejuSoul's Car:
Energy discharged = 23.9 kWh => Full pack 23.9/0.91 = 26.3kWh
Energy charged = 25.0 kWh => Full pack 25/0.91 = 27.47kWh
Regen = 11.3kWh

Birkeland's Car:
Energy discharged = 24.5 kWh => Full pack 24.5/0.93 = 26.3kWh
Energy charged = 24.9 kWh => Full pack 24.9/0.93 = 26.77kWh
Regen = 3.4kWh

Elmil's Car:
Energy discharged = 25.1 kWh => Full pack 25.1/0.97 = 25.9kWh
Energy charged = 26.1 kWh => Full pack 26.1/0.97 = 26.9kWh

 

[Birkeland]

I only drove 127,8km. I drove really hard at very high speed on the freeway to run down the battery as fast as possible. I also let the car stand still with climate control at 30 degrees Celcius at the end with the doors open to deplete the battery further. I actually dont think I have ever gotten this short range on a full charge ever at theese temperatures. I will do the test over in a while, taking it more easy when the opportunity comes. It was interesting to see no change in the deterioration values. I thought it would be. To me, it seems the calculation of the min/max deterioration values are pretty accurate since there were no change. Further testing in the future will tell.

 

[JejuSoul]

Wow - driving so fast gives really bad efficiency = 5.2km/kWh.
I drove slowly and got 8.4km/kWh.
The fastest speed limit here on the island is 80km/h and there are frequent speed cameras. Even in the winter I never got as low as 5.2km/kWh
Although the only 2 days the temps got below zero I didn't drive.

In Norway which way do you use the efficiency numbers. 5.2km/kWh, or 19.2kWh/100km, or 192Wh/km?

 

[Birkeland]

I actually got about the same range or a little better in the winter time when it was -24 degrees Celcius. The freeway here permits 100km/h, but I drove about 120 and avoided the cameras :lol: I also drove alot uphill, and coasted in Neutral downhill to avoid regen in order to run down the battery.

By the way, the car performs really good during the winter, with low temperatures and snow.

Usually in Norway, when we talk about fuel consumption, we say liters/mile (1 mile being 10 km) so my other car, a Toyota Avensis uses about 0.74l/mile of gasoline.

When it comes to EVs I really don't know, but I tend to use kWh/10km. But the display shows kWh/100km

 

[ralyon]

Usually in Norway, when we talk about fuel consumption, we say liters/mile (1 mile being 10 km) so my other car, a Toyota Avensis uses about 0.74l/mile of gasoline.

Wow, I didn't realize you guys used a different mile than us. Here a mile is a little over 1.6 km.

 

[Elmil]

Usually in Norway, when we talk about fuel consumption, we say liters/mile (1 mile being 10 km) so my other car, a Toyota Avensis uses about 0.74l/mile of gasoline.

Wow, I didn't realize you guys used a different mile than us. Here a mile is a little over 1.6 km.

It's a Scandinavian thing, actually it's a 'MIL' =10 km.

I think the EU standard is '19.2kWh/100km'

 

[ZuinigeRijder]

I also measured start at 100% charged, discharge till 13.5% and full charge on trickle homecharger (220V, 10A, approx 2000 Watt, took approximate 12 hours):

Data:

Date	km	Max Det	Min Det	Max Cell	Min Cell	BMS	SOC	CEC	CED
4/13/2016	24722	11.1	9.6	9	1	95.00	100.00	5105	4994
4/14/2016	24876	11.1	9.6	9	1	14.00	13.50	5108	5020
4/15/2016	24876	10.2	8.7	43	15	95.00	100.00	5132	5020

Looking at the computation of JejuSoul, I think this gives:

Energy discharged = 23.0 kWh => Full pack 23.0/0.865 = 26.59kWh
Energy charged = 24.0 kWh => Full pack 24.0/0.865 = 27.75kWh
Regen = 3 kWh
Driven = 154 km
Boardcomputer consumption shown: 13.9 kWh/100 km (reset at start of trip)

I also have measured with a flow meter how much actually was taken from my household electricity, and that was 24.7 kWh.
The Max and Min deterioration numbers improved.

Unfortunately I do not have the 2 digits for CEC and CED, so there might be rounding errors. Something for the next time :twisted:

 

[JejuSoul]

ZuinigeRijder: that data looks good. It's amazing how similar all our cars are behaving.

 

[JejuSoul]

The AVT website has updated their deterioration data. - https://avt.inl.gov/vehicle-button/2015-kia-soul
I have added the latest data into an Excel spreadsheet. The Kia Soul EV is behaving completely normally compared to all the other EVs being tested in Arizona. Much better than the L eaf and iMiev. Exactly the same as the Ford Fusion. Similar to the BMW i3, Mercedes B class, eGolf and Spark.

Click image for bigger graphic.

 

[Elmil]

Borrowed the dealers 2016 Demo Soul EV while having mine at 2:nd service. It had only done 950km.
I connected my "Soul Spy" logger and found some interesting stuff.

Stopped at a Chademo QC on the way to work, and noticed a peculiar behaviour of the SOC from 81 to 84%:

Timestamp
08:50:06 SOC= 81.0 U= 390.2V I= -34.1A P= -13.3kW Ei= 337.0 Ai= 914.3
08:50:10 SOC= 81.5 U= 390.2V I= -33.7A P= -13.1kW Ei= 337.1 Ai= 914.6
08:50:14 SOC= 81.5 U= 390.2V I= -33.6A P= -13.1kW Ei= 337.2 Ai= 914.7
08:50:18 SOC= 82.0 U= 390.2V I= -34.2A P= -13.3kW Ei= 337.2 Ai= 915.0
08:50:22 SOC= 82.0 U= 390.3V I= -33.6A P= -13.1kW Ei= 337.3 Ai= 915.1
08:50:26 SOC= 82.5 U= 390.3V I= -34.0A P= -13.3kW Ei= 337.4 Ai= 915.4
08:50:30 SOC= 83.0 U= 390.2V I= -32.1A P= -12.5kW Ei= 337.5 Ai= 915.5
08:50:34 SOC= 83.0 U= 390.2V I= -32.0A P= -12.5kW Ei= 337.5 Ai= 915.8
08:50:38 SOC= 83.5 U= 390.2V I= -32.1A P= -12.5kW Ei= 337.6 Ai= 916.0
08:50:42 SOC= 83.5 U= 390.3V I= -31.6A P= -12.3kW Ei= 337.7 Ai= 916.2
08:50:46 SOC= 84.0 U= 388.7V I=   0.0A P=   0.0kW Ei= 337.8 Ai= 916.4

As you see from the time stamp, it took only 40 seconds from 81% to 84%. What is more interesting is the energy counters Ei (Cumulated Energy In; kWh), and Ai (Cumulated Amperage In; Ah). Ei counts 0.8kWh during 40 seconds, and Ai counts 2.1Ah during the same time. If these counters were reflecting the actual realtime energy transfer into the battery, we would see numbers like -190A and -72kW...
This looks like the same type of behaviour as I have noticed on my car, but sort of 'inverted'; This car rapidely counts up the SOC and Energy In counters, while mine slows down the SOC counting and increases the Energy Out counters. Probably calibration effects and this probably also explains why some cars show larger difference, between cumulative energy in and out, than others.

I suspect that this demo car has very rarely, if not never, been charged to 100%. It was down to 29.5% before QC, and at work I slow charged to 100%. The deterioration values changed from 0.0/0.0 to Max=0.5%/Min=1.4%. Didn't we see the same reversed Max/Min from another car?

 

[2016Electric]

The AVT website has updated their deterioration data. - https://avt.inl.gov/vehicle-button/2015-kia-soul
I have added the latest data into an Excel spreadsheet. The Kia Soul EV is behaving completely normally compared to all the other EVs being tested in Arizona. Much better than the Leaf and iMiev. Exactly the same as the Ford Fusion. Similar to the BMW i3, Mercedes B class, eGolf and Spark.

Click image for bigger graphic.

Thanks for the info! I've checked back there and tried to find updated info (it wasn't there last time I looked).
I'm thinking (and hoping) the battery deterioration is not linear. From the literature and research I've seen, Lithium NCM batteries deteriorate quickly to around 90% (lose their first 10% capacity), and then the deterioration slows down somewhat substantially.
I'm hoping that's the case here too... and that would explain the motivation to over-size the battery by 3 kw/hr or so.. masking that first 10% faster loss.

If that isn't the case.. the first 10% loss according to AVT happened over around 11k miles... projecting that linearly would yield around 90% capacity loss by 100,000 miles (the end of the warranty period). There's no way that could be right.

Thoughts?

 

[JejuSoul]

... this probably also explains why some cars show larger difference, between cumulative energy in and out, than others.

Can you explain why you say this. It doesn't match the data I have collected for our cars. What I see is that initially the cumulative discharge is higher because the battery is manufactured fully charged, but after that the cumulative charge is higher by about 2%. This is consistent with all the cars we have measured so far.

There's no way that could be right.

The AVT data is from the Intertek site in downtown Phoenix. Summer in Phoenix is very hot and sunny. You can look at in Google Earth. The cars sit on the tarmac and charge during the day without any shade. This is the worst possible way to treat a Lithium Ion Battery without dropping it into a volcano. I used to think that having the EV testing center in Phoenix Arizona was a bad idea because it gives the worst possible results. Looking at the above data however I realise that it is good to have worst case scenario data. It is the reason my priority topic for the next few months will be how to keep the battery cool during a hot summer.