Scott Peterson, Jay Apt, and Jay Whitacre are preparing a paper titled
"Lithium-Ion Battery Cell Degradation Resulting from Realistic Vehicle and Vehicle-to-Grid Utilization"

The paper applies some interesting NHTSA driving pattern information to charge/discharge cycles of A123 LiFePo4 cells along with a simulated Vehicle to Grid discharge cycle.

Notably the cells seemed to last LONGER than expected in all test regimes, and MOST notably, depth of discharge (DOD) did NOT have the expected reduction in capacity in extended life cycling. Cells discharged to 95% DOD, instead of fading to 80% capacity in 1500 cycles, seemed destined to hit 5300 cycles before diminishing to that level.

These test results would seem to indicate BETTER life cycle performance that the manufacturers specifications for the cells, and apparently to a pretty dramatic degree.

The paper is athttp://wpweb2.tepper.cmu.edu/ceic/papers/ceic-09-02.asp


Jack Rickard
http://evtv.me

[AmpEater]

I think I speak for everyone when I say; sweet!

LiFePO4 was just what practical EVs needed.

Gone are the days of hauling around 40% more capacity than you need just to keep the cell life reasonable. 5000+ cycles is over 10 years.

[AmpEater]

I can't access the paper without a password...could someone with access relay the LVC and HVC used in their cycle testing? Amperage/C rate? Also, cell temperatures? I'm sure that info has alot to do with their excellent cycle performance

[MN Driver]

As much as I like to be optimistic about things, this in one place where I feel that it might not be quite as rosy as it may seem. The discount chinese LiFePo4 cells that many of us are using or are planning to use are going to be rated as high as they can be by the manufacturer, usually Chinese manufacturers stretch things too. I really think the issue that people will be most concerned about when reaching the 5,000 cycle mark of a LiFePO4 cell will be the increased internal resistence that is characteristic of pretty much all Lithium cells with age, except for Lithium Spinel chemistries(such as LiMnO2) but LiMnO2 has lower densities from the cells I've seen. I don't know about Lithium Titanate though, but it isn't in our price range anyway.

So basically you might have the capacity 5,000 cycles out but you won't have the same performance, you'll have to drive slower to avoid the voltage sag. Take a look at the Thunder Sky LFP cells at 8,000 cycles, the charts shows they can get 70% capacity but if you look at the discharge curve the 0.5C chart looks like the 5C chart of a new cell. Something to consider! Personally I plan on buying a pack that lets me get 2 days of my commute driving(60 miles) and be around 70% SOC(86 miles total range, but really thinking to go for a golden 100 miles if possible), this way, hopefully my cells last as long as possible, provide the best draw as possible in cold weather conditions and if they make it to 5,000 cycles I will have been driving with them for 27 years and would likely bring my batteries with me to a second car. Now, do I think I will get 5,000 cycles out of a cheap LiFePO4 pack? I think I'll be disappointed somewhere in the realm of performance but I'd be thrilled with 3,000 70%(of 100 miles) DoD cycles because that is possible 210,000 miles, of course toward the end each cycle will be closer and closer to 100% DoD but remember, I was planning to have 2 commutes per cycle so if worse comes to worse I could live with half the capacity and likely have good use for the pack. I think that oversizing the pack for cell life is a good idea no matter what. If you thought 10 years(or whatever you want for life) is/was good, oversize it and add more, it's what I plan to do. Right now that looks like a $10k for just the batteries but i'll take the peace of mind plus the ability to pretty much do every trip I want that wouldn't normally require a dedicated roadtrip.

[roflwaffle]

Originally Posted by AmpEater

I can't access the paper without a password...could someone with access relay the LVC and HVC used in their cycle testing? Amperage/C rate? Also, cell temperatures? I'm sure that info has alot to do with their excellent cycle performance

There's a link on the page to request the user/pass.