Battery Aging Model

JPWhite said:

Put an Excel spreadsheet here
http://dl.dropbox.com/u/2881567/Degradation Model Pub.xlsx

Click to expand...

Let's see...

Ft. Lauderdale...
10,000 miles a year...

10% loss the first year. First capacity bar falls a little after year 2. 20% down by the time I turn the car in at the end of year 3.

Looks accurate.

Should be interesting if maintaining a relatively low, average SOC allows me to beat the odds.

Weatherman said:

Let's see...

Ft. Lauderdale...
10,000 miles a year...

10% loss the first year. First capacity bar falls a little after year 2. 20% down by the time I turn the car in at the end of year 3.

Looks accurate.

Should be interesting if maintaining a relatively low, average SOC allows me to beat the odds.

Click to expand...

OK, I have implemented a rough version of the prediction portion based on the model:

Fort Lauderdale
10,000 miles per year
Average efficiency 4 miles/kwh from dash

Predicted Cumulative Losses:

Year 1 - 11.09%
Year 2 - 16.00%
Year 3 - 20.11%

Looks like a pretty good match with the other spreadsheet (which I haven't looked at). Still need to clean things up, automate lookups of aging factors by City, add average speed to calculate miles/kwh, and then try for the harder stuff like high or low SOC, solar loading, etc.

PS I now need some data to compare with my model from the range test or Casa Grande to see how it is coming so far.

OK, the Prediction Model (TM) gives the following numbers for Phoenix at 12500 miles per year (so far, more factors need to be added):

Miles per kwh = 4, cumulative loss:

Year 1 - 12.11
Year 2 - 17.79
Year 5 - 31.11
Year 10 - 49.43

EOL - 4.7 years

Miles per kwh = 3.5, cumulative loss:

Year 1 - 12.45
Year 2 - 18.46
Year 5 - 32.8
Year 10 - 52.81

EOL - 4.3 years

Model does not yet account for solar loading, percent of time at high SOC, etc.

LEAFfan said:

ENIAC said:

I found this study which looked at parking lot temperatures in Phoenix at various times of the year. They measured temperatures in June of 149 deg F (figure 2a). BTW, measurements are taken at 8 feet above the surface.

Click to expand...

Remember, a lot of us almost always park in shaded parking spots so those temps are only valid for those that usually park in unshaded ones.

Click to expand...

True, however I suspect many owners aren't aware of the issue and don't park in the shade. Also, often shaded parking is simply unavailable. Do Nissan dealerships in Phoenix park their LEAFs in the shade? At 149F, how many hours before permanent battery damage occurs? I was just reading another study which said that parking lot temperatures in Atlanta GA in the summer are nearly 50F higher than the ambient temperature. With that sort of temperature variability, feels to me like an owners parking habits should be one of the dominate variables in the battery aging model.

LEAFfan said:

ENIAC said:

I found this study which looked at parking lot temperatures in Phoenix at various times of the year. They measured temperatures in June of 149 deg F (figure 2a). BTW, measurements are taken at 8 feet above the surface.

Click to expand...

Remember, a lot of us almost always park in shaded parking spots so those temps are only valid for those that usually park in unshaded ones.

Click to expand...

at 147º i am gonna have to guess that unshaded parking is pretty rare. those temps are more than enough to cause 3rd degree burns

DaveinOlyWA said:

LEAFfan said:

ENIAC said:

I found this study which looked at parking lot temperatures in Phoenix at various times of the year. They measured temperatures in June of 149 deg F (figure 2a). BTW, measurements are taken at 8 feet above the surface.

Click to expand...

Remember, a lot of us almost always park in shaded parking spots so those temps are only valid for those that usually park in unshaded ones.

Click to expand...

at 147º i am gonna have to guess that unshaded parking is pretty rare. those temps are more than enough to cause 3rd degree burns

Click to expand...

140F+ water would cause a 3rd degree burning in a matter of seconds, we're talking air temp here. Also, in my experience, the Phoenix valley has a lot of unshaded parking.

ENIAC said:

140F+ water would cause a 3rd degree burning in a matter of seconds, we're talking air temp here. Also, in my experience, the Phoenix valley has a lot of unshaded parking.

Click to expand...

ooooh we can play that game!!

step one
bake a car at 147º for 2 hours in the middle of the day

step two; toss the hot pads

step three; open door and grab steering wheel tightly with both hands for 60 seconds...

DaveinOlyWA said:

ENIAC said:

140F+ water would cause a 3rd degree burning in a matter of seconds, we're talking air temp here. Also, in my experience, the Phoenix valley has a lot of unshaded parking.

Click to expand...

ooooh we can play that game!!

step one
bake a car at 147º for 2 hours in the middle of the day

step two; toss the hot pads

step three; open door and grab steering wheel tightly with both hands for 60 seconds...

Click to expand...

Again, we're talking air temperature at 8 feet above the surface of a parking lot. Are you disputing the parking lot temperature readings in the Phoenix study? I'm grasping to understand your point.

surfingslovak said:

I left my Leaf out in the sun on a hot day once. The sheet metal on the roof was at about 130 F. The battery container was a few degrees warmer compared to what I would see when the Leaf was parked in a shade. Additionally, the battery was about two degrees warmer on side facing the sun.

Click to expand...

How many degrees? How hot was it? What time of year? The number is important in deciding how to account for solar loading, and another data point would be very valuable.

Remember that the study of the Prius found the average annual battery temperature to be 1.3-3.1 degrees C. higher (depending on geographic location) with solar loading (parked every day in the sun vs in the shade) than without. That isn't a huge amount, but adds 10-24% to battery aging for the year (24% was for the Phoenix location) based on doubling of degradation for every 10 degrees C. increase in temperature. Presumably I should add that factor in to both the calendar loss and cycling loss. I might scale the numbers back a bit to account for the fact that the aging factors we calculated at for the hottest climates (e.g., Phoenix) were about 20% higher than the empirically derived values in the model.

Other thoughts welcome. Those super high numbers measured in Phoenix parking lots do make me wonder if some batteries may actually be damaged rather than just suffering extreme heat degradation.

I found this resource to help estimate solar loading by geographic location:

http://rredc.nrel.gov/solar/old_data/nsrdb/1961-1990/redbook/atlas/" onclick="window.open(this.href);return false;

I selected average and annual, but couldn't decide which of the 14 instrument orientations would be the best proxy. Anyone have thoughts about this?

149° is hot enough to burn you.

Here are couple of photos taken in Phoenix around 3 pm on September 15. One was taken in a covered parking lot, and the other when stopped at a traffic light. Infer from it what you will, hope it helps with the discussion.

Stoaty said:

surfingslovak said:

I left my Leaf out in the sun on a hot day once. The sheet metal on the roof was at about 130 F. The battery container was a few degrees warmer compared to what I would see when the Leaf was parked in a shade. Additionally, the battery was about two degrees warmer on side facing the sun.

Click to expand...

How many degrees? How hot was it? What time of year? The number is important in deciding how to account for solar loading, and another data point would be very valuable.

Click to expand...

It was a hot day. Ambient temps were between 80 and 90 in the afternoon. Battery enclosure temperature was about 80, and 82 F on the side facing the sun. If I parked the car in the shade in similar conditions, the pack would not heat up as much during the day, and the container wall temperature would not exceed 76 F. The sheet metal on the roof of the car reached 130 degrees. I saw a report somewhere, which estimated the effect of solar loading at about 3-4%, but I forget which locales they were considering. I will try to find it.

FWIW, I totally believe the study. You can literally fry eggs on your hood during the summer. Locals learn very fast to run from shade to shade like Van Diesel on Crematoria (Chronicles of Riddick) .

I settled on this graphic as the best proxy to estimate average annual solar radiation, since it "represents the maximum solar radiation at a site available to a collector". As expected, Arizona, New Mexico, and parts of Texas, California, Nevada and Colorado are really getting cooked!

I added the solar loading (kwh/square meter) for each city to the spreadsheet, and also added it to the table of battery aging factors on the Wiki:

http://www.mynissanleaf.com/wiki/index.php?title=Battery_Capacity_Loss#Factors_Affecting_Battery_Capacity_Loss" onclick="window.open(this.href);return false;

Solar loading varies between 4.5 and 9 kwh/square meter depending on geographic location, although it may well be somewhat higher in places like Phoenix. The data I can see on the map isn't granular enough to make finer distinctions; perhaps I can find the data for individual cities from which the colored map is drawn.

I haven't figured out yet how many degrees C. increase in battery temperature on an annual basis to expect for various degrees of solar loading (besides the Prius model), but these solar loading numbers should allow appropriate scaling for each city. I will probably start with the Prius numbers, include an adjustment factor if I want to change it later, and put in number of days per week parked in the sun as input for the prediction model.

Further thoughts welcome.

Solar loading has been added to the Battery Prediction Model, but I need more data to tune the numbers. Currently the model uses 3.0 degrees C. increase in average annual battery temperature for Arizona and related areas, and 1.5 degrees C. for areas like Northwestern Washington state. This is for parking the Leaf in the sun every day for a year. For fewer days parked in the sun (e.g., only 5 days a week at work, or only 2 days a week at home), the numbers are scaled accordingly. See this thread to add a data point:

http://www.mynissanleaf.com/viewtopic.php?f=27&t=10397" onclick="window.open(this.href);return false;

The solar loading factor certainly looks reasonable (but that doesn't mean it is correct, of course). I tested my predicted remaining capacity with and without solar loading:

16 months
11500 miles
Battery aging factor adjusted for model - 0.9
Solar loading factor 1.19
Days/week parked in the sun - 4 (no covered parking where I work!!! :twisted: )
Miles/kwh - 5.7

Without solar loading - 8.64% predicted capacity loss
With solar loading - 9.57% predicted capacity loss

I don't have a good handle on my actual capacity loss, except that an incomplete range test to just past LBW gave an extrapolated 106 miles, and theoretically I should get 119.7 miles with a new pack assuming 21 kwh usable, which calculates to a 11.4% loss, for a difference of 1.83%. However, I never tested my Leaf when new so don't have a good baseline.

An additional factor that my model doesn't take account of is the extra heat of 1.6 summers in 16 months, and 1.6 summers worth of solar loading (I have been leaving the Leaf at home this past summer when predicted high is greater than 95 degrees). As with ambient temperature, solar loading is not easy to model for fractional years.

For 5 years, the numbers predicted are 18.27% (no loading) and 20.26% (solar loading)

For 10 years, the numbers predicted are 27.58% (no loading) and 30.57% (solar loading)

I took the data on Azdre/Opossum and plugged it into my model:

Phoenix, AZ
21,834 miles at 17.25 months
Assumed 4 miles/kwh
Assumed no parking in the sun
Assumed no significant time spent at 100% charge (but haven't incorporated that into the model yet anyway)
Scaled mileage to 16.25 months

Model predicts 85.12% capacity at time of Casa Grande test

Bench test at Casa Grande (after 16.25 months of ownership) showed capacity of 85%

I re-tested TickTock's data and got 86.94% predicted capacity remaining at the time of Casa Grande test (mileage and time weren't properly scaled before; I used 13.5 months at time of Casa Grande testing and scaled mileage to 16536 at time of Casa Grande testing). Casa Grande testing showed 87% per the Wiki, but 85% per TickTock's own graph. Not sure which is correct.

Validation page added to Battery Aging Model spreadsheet. I am collecting all of the data in one place so I can see how well it fits the model, and whether the match gets better or worse if I make changes to the model. Model version number added, currently at version 0.81. As soon as I improve the model a bit more I am going to upload it to the Wiki so others can play with it. Just requires OpenOffice installed on your computer.

Didn't both TickTock's and Azdre/Opossum's LEAFs actually suffer two, rather than 1.6, Phoenix "Summers", the ~3 months with the highest ambient temperatures?

So aren't you actual "tuning" your model to match an incorrect value, a world where Phoenix has two summers in only ~16 months?

Stoaty said:

...An additional factor that my model doesn't take account of is the extra heat of 1.6 Summers in 16 months, and 1.6 summers worth of solar loading...

I took the data on Azdre/Opossum and plugged it into my model:

Phoenix, AZ
21,834 miles at 17.25 months
Assumed 4 miles/kwh
Assumed no parking in the sun
Assumed no significant time spent at 100% charge (but haven't incorporated that into the model yet anyway)
Scaled mileage to 16.25 months

Model predicts 85.12% capacity at time of Casa Grande test

Bench test at Casa Grande (after 16.25 months of ownership) showed capacity of 85%...

Click to expand...

Do you not consider variables in Battery care may be very large factors in capacity loss?

azdre
* We always charge to 100% (we drive it a lot, 17,000 miles in 14 months of driving).
* We do 'top it off' a lot. That's the one item we got 4/5 stars on the battery check in March 2012.
* The car sat at 100% for a month in May, 2011, not plugged in...

Click to expand...

http://www.mynissanleaf.com/viewtopic.php?f=30&t=8802&start=20" onclick="window.open(this.href);return false;