JejuSoul
Posts: 1506
Joined: Wed Jul 08, 2015 6:47 am
Location: Jeju
Contact: Website

Re: Battery Ageing Model

Thu Apr 20, 2017 7:52 am

2016Electric wrote:I'm a little concerned about my stats. ..
Yes, that recent reading is bad. But, this data is often extremely varied. We need an average of multiple readings over a few seasons to make a more reliable judgement.

Something else you could try is a calibration check. Here's one I did a year ago. --- Old post from Tue Apr 05, 2016 4:17 am
JejuSoul wrote:... To calibrate your battery pack you need to charge to 100% then drive the pack down to nearly empty and then charge to 100% again. It is on this 2nd 100% charge that calibration occurs.
Results for my calibration test:- charged to 100%, drove 200km, charged to 100%
I stopped driving and started charging at 9% SOC.
GOM before 174km, GOM after 178km. My highest so far this Spring.
Deterioration values did reset. min det reduced from 6.8% to 5.5% at just above 4.08V (some cells were showing 4.10V) It was at 92% SOC, Battery pack =393V
max det reduced from 9.0% to 8.5% at 100% after I turned the car off to unplug the c harger.
I have seen much greater movement in these numbers recently that was not associated with charging to 100%.

Energy charged = 25.0 kWh for 91% (based on the BMS data) => Full pack 25/0.91 = 27.47kWh
Energy discharged = 23.9 kWh for 91% (based on the BMS data) => Full pack 23.9/0.91 = 26.3kWh

The energy charged figure is simple; it is the difference in cumulative energy charged before and after charging.
The energy discharged is the difference in cumulative energy discharged during driving minus the regenerative energy put back into the battery.

I replicated this calibration check over the last 5 days. It is just my normal driving, no special trips or hypermiling involved.

Results for my calibration test:- charged to 100%, drove 210km, charged to 100%
I stopped driving and started charging at 9% SOC.
GOM before 182km, GOM after 187km.
At 92% SOC deterioration values did not reset, but the cell numbers for the min and max did. (from 83, 29 to 1, 2)
max det reduced from 11.5% to 10.5% at 100%.
min det increased from 7.1% to 7.4% at 100%.

Energy charged = 24.5 kWh for 91% (based on the BMS data) => Full pack 24.5/0.91 = 26.92kWh
Energy discharged = 23.6 kWh for 91% (based on the BMS data) => Full pack 23.6/0.91 = 25.9kWh
Regenerated Energy on the trip was 14.5kWh.
C harger efficiency was 24.5 / 26.48 = 92.5%

From this data, even though my car is two years old it still has the same usable capacity.

Image
The codes for using Torque Pro can be found by clicking the link in the website icon under my user name on the left.

JejuSoul
Posts: 1506
Joined: Wed Jul 08, 2015 6:47 am
Location: Jeju
Contact: Website

Re: Battery Ageing Model

Fri Apr 21, 2017 6:37 pm

-
About a year ago Elmil stated that the Ah data is more accurate than the kWh for doing these comparisons. Also the SOC reading is based on the Ah. I started collecting the Ah data in May 2016, so I don't have my original range test Ah data. But here is the same calculations for my recent trip, but done in Ah. Also instead of just using the displayed SOC, I will graph the SOC over the charging period and calculate a more exact starting SOC with one decimal place.

Results for my calibration test:- charged to 100%, drove 210km, charged to 100%
I stopped driving and started charging at 9.5% SOC.

Ah charged = 66.9 Ah for 90.5% (based on the BMS data) => Full pack 66.9/0.905 = 73.9Ah
Ah discharged = 65.4 Ah for 90.5% (based on the BMS data) => Full pack 65.4/0.905 = 72.2Ah
Regenerated Ah on the trip was 42.7Ah.

The rated usable capacity for our battery is 75Ah; if I used this figure it appears my battery has lost 1 or 2 percent. However I would rather use a figure generated by my own car for that comparison.

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

Update: I have been looking more carefully at the data for this charging session in the torque logs. The number 9% I gave for the SOC was that displayed on the dash before I turned off the car, so that I could plug it in. But in the logged data the SOC starts at 11% once torque starts up again after he car has started charging.

Image
The codes for using Torque Pro can be found by clicking the link in the website icon under my user name on the left.

2016Electric
Posts: 137
Joined: Thu Sep 17, 2015 4:05 am
Location: Toronto, Ontario

Re: Battery Ageing Model

Mon Apr 24, 2017 5:40 am

Update...

I performed a 'calibration' of sorts as described by Jeju. I charged to 100%, then used the car over the next couple days until I had the battery down to about 4%, and the Turtle light came on. Immediately thereafter I charged the car to 100% (last night) at a rate of 208-212v/30A (~6 kw).

I checked the battery stats this morning and they have changed. I'm now at around 23000km or so with:
CEC 5765
Ced 5567
Min Det 6%
Max Det 11.5%

Also, I checked the trip log for total distance travelled since the full charge, and it was about 147.5km, with the 4% battery remaining. About 1/3 of that was highway at 100 km/hr, the rest was stop and go traffic in 50-60 km/hr zones. I think one of the highway trips required some heat use as it was night time and cold.

Thats better.. But quite the spread between cells in terms of deterioration.

SoulEV2016
Posts: 572
Joined: Mon Jul 18, 2016 6:26 am

Re: Battery Ageing Model

Tue Apr 25, 2017 11:11 am

Soul EV 27kWh all new (2017) with 700km and 1 time at Chademo 43kW and only others 1,9kW EVSE Kia Official Charge cable.

Image

http://www.automobile-propre.com/forums ... tml#p99735

Deviation Min-MaxDET with Chademo = less than 10%
Deviation with 1kW = less than 1%
Deviation with 6,6kW = less than 5%
Kia Soul EV 2016 Ultimate 27 kWh - Mountain hard driving (fast). Drain 1/3 of Battery after 32km. - Unlimited highway speed (153km/h at regulator).

JejuSoul
Posts: 1506
Joined: Wed Jul 08, 2015 6:47 am
Location: Jeju
Contact: Website

Re: Battery Ageing Model

Wed Apr 26, 2017 6:59 am

-
I just did a second range test, immediately after the first.

Results for my calibration test:- charged to 100%, drove 195km, charged to 100% (Heavy rain reduced range on one of the 4 days.)
I stopped driving and started charging at 12% SOC.

Deterioration numbers for the last week.
max det reduced from 11.5% to 10.5% to 10.0%.
min det increased from 7.1% to 7.4% then decreased to 1.4%

I charge at 32A , that's about 6kWh. It does seem the difference between min and max is greater at higher currents. The average and trend however are the same.

We have seen brand 'new' cars with terrible stats before. Often they adjust back to normal after a few months. 'New' cars that have been sitting on the dealer lot for a long time, or ex-demo cars do not seem to adjust back.
The codes for using Torque Pro can be found by clicking the link in the website icon under my user name on the left.

JejuSoul
Posts: 1506
Joined: Wed Jul 08, 2015 6:47 am
Location: Jeju
Contact: Website

Re: Battery Ageing Model

Tue May 02, 2017 6:35 am

-
A couple of topics recently have me concerned that the model being developed in this thread to determine the ageing of the battery due to deterioration may no longer be applicable. In particular using the internal BMS data.

1. According to these posts by Elmil - SOH change some cars have had a BMS upgrade. The BMS numbers all get reset during the upgrade, and it seems possible that a different algorithm is used to calculate the deterioration. If so comparing these numbers on different cars no longer makes sense.

2. Up until now almost all of us have had SOH =100% or nearly so. We have no idea on how the BMS data changes once the usable capacity begins to shrink. I have been testing a 2012 Kia Ray EV. The results have me worried that all the BMS data ceases to be reliable after a few years.
Am adding a thread to discuss the Ray EV BMS - Comparing BMS data on the Soul EV with the Ray EV
The codes for using Torque Pro can be found by clicking the link in the website icon under my user name on the left.

Elmil
Posts: 142
Joined: Mon Oct 26, 2015 9:46 am
Location: Sweden

Re: Battery Ageing Model

Tue May 02, 2017 11:06 am

.
The CCC/CEC or CCD/CED counters are often counting up a bit just after the car is started. Here's an example:

Normal commute to work, 50km, then parked during the day (no c harger connected), then home 50km.
The graph is computed as ["consumed Ah"] divided by ["consumed SOC%"] = ["Projected capacity at 100% SOC"].

Image

The interesting part here is what happens at 50km when I start the car and drive home. My data shows that the CCC had increased by 0.8Ah and the CEC by 0.3 kWh, that is energy IN to the battery. This is roughly corresponding to 1% SOC, but the SOC is not adjusted, hence the shift in the diagram.

When I check my old logs it seems this is a common behaviour, and the shift can also be in the opposit direction, i.e CCD/CED are counting. I found another case where this leap was 1.5Ah <=> 2% SOC.

Anyone with Torque Pro can easily test this by taking logs just when turning off the car, leave it for an hour or so, and then take logs again when starting up. Check the energy counters before and after.
Kia Soul EV 2015 / 115,543 km Sold
Kia Soul EV 2019

JejuSoul
Posts: 1506
Joined: Wed Jul 08, 2015 6:47 am
Location: Jeju
Contact: Website

Re: Battery Ageing Model

Thu May 04, 2017 10:38 pm

-
Another SOH 100% reading from GDS. This one from Quebec.
MY2015 Soul EV 31,000 km see - https://www.facebook.com/groups/8840565 ... 876501038/

Image
The codes for using Torque Pro can be found by clicking the link in the website icon under my user name on the left.

JejuSoul
Posts: 1506
Joined: Wed Jul 08, 2015 6:47 am
Location: Jeju
Contact: Website

Re: Battery Ageing Model

Tue May 09, 2017 12:50 am

-
Here is some info on a talk by Thierry St-Cyr from Taxelco (Téo Taxi, the EV Taxi company in Montréal — http://teomtl.com). They have L eafs, Soul EVs, and T eslas in their fleet.

They are doing range testing with Téo Taxi and the IVI institute : the put the car on dynamometers and run the car to empty.

They calculate an average degradation of 2.2% every 10,000km driven However, their situation is special because most of the time they charge on DC fast c hargers (77% of the time). This is higher than the degradation we are seeing on most of our cars.

Image

Image

Image

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

Also from Quebec. Battery stats from a MY2015 Soul EV with 90,000km. Calculated SOH is 94.2%. Far better than the taxis above, and in the same climate zone.

Image
The codes for using Torque Pro can be found by clicking the link in the website icon under my user name on the left.

SoulEV2016
Posts: 572
Joined: Mon Jul 18, 2016 6:26 am

Re: Battery Ageing Model

Tue May 09, 2017 7:40 am

The answer is simple for the taxi : BYD use LifePO4 battery.
LifePO4 battery is for heavy duty usage ... like industry.
And this answer is not new.
Kia Soul EV 2016 Ultimate 27 kWh - Mountain hard driving (fast). Drain 1/3 of Battery after 32km. - Unlimited highway speed (153km/h at regulator).

Return to “Technical Discussions”