How to predict Cycle life

[walter-]

For a given cell, is there any way one can predict the cycle life?

I see manufacturers state numbers like 3.000 or 6.000 or whatever... But is there a test one can do to predict this number?

The question also relates to the meaning of grade B; we all test the capacity of cells and are happy if it reaches a certain number. But I believe the real test is how the cell holds after 1.000ths of iteration. What is the value of a cell if it indeed has now a capacity but if it fails in a few years.

Walter

 

[sunshine_eggo]

No.

 

[toms]

Find someone with similar use to yours that has the same type of battery, and see how long they have lasted. The more systems you compare - the more accurate your prediction will be.

eg: Winstons have proven plenty of times that if you use sensible parameters they will last at least ten years. (i think in another 5 years i’ll revise this to 15 - but we’ll see).

If the cell type you are looking at doesn’t have thousands of systems in place that have exceeded their cycle specs, then spec is for speculation - not specification.

 

[walter-]

In my case, looking for the 280Ah EVE cells; but I don't think anyone has experience with these cells over 10 years or like 5.000+ cycles.
So it is a wild guess if EVE can hold to its promises. Or worse how EVE resellers "grade B" batteries will do.

 

[walter-]

...

eg: Winstons have proven plenty of times that if you use sensible parameters they will last at least ten years. (i think in another 5 years i’ll revise this to 15 - but we’ll see).

...

Do you have a reference to one of these plenty times?
I think we should rather speak of cycles rather then years.

 

[walter-]

Note; the question also as I want to build a solution that next to store energy from solar panels, i want a battery system that can make use of a dynamic tarrif; this could mean going to 300+ cycles per year versus only 100 when only using solar energy to be stored in the batteries.

 

[toms]

Do you have a reference to one of these plenty times?
I think we should rather speak of cycles rather then years.

What makes you think cycles will kill the battery before time?

 

[walter-]

What makes you think cycles will kill the battery before time?

The specs of the (lifepo4-cells-) manufacturers who only talk about cycles. Or do you have a reference of aging specs based on time?

 

[SeaGal]

Note; the question also as I want to build a solution that next to store energy from solar panels, i want a battery system that can make use of a dynamic tarrif; this could mean going to 300+ cycles per year versus only 100 when only using solar energy to be stored in the batteries.

There are too many variables to predict. Cycles are ill-defined for a start. Then, just because you charge from grid doesn't mean you are going from 5% to 100% and back to 5% again, you may only go from 25% to 90%. Then there is the rate of charge and discharge current (greater rate = more degradation) and then there is temperature of cells to put into the guesstimate.

So, if you charge discharge from 5 to 100, at a rate of >0.5C at 5 degrees C your cells will probably last(*) only a quarter(**) of the time they would if you charge from 25% to 90% at a rate of <0.25C at 25 degrees C.

(*) by last I mean only hold 80% of their original charge
(**) the bit where I say "quarter" is just a random made up figure - I have no idea, but gut feeling and lots of reading suggests the variance could be that large

 

[walter-]

I would disagree that cycles are bad defined; at least each supplier will define a specific cycle their batteries will support x-times before reaching 80% (or whatever value)... so that is well defined.
But indeed; how this specified cycle translate into actual use and 'real' life cycles is another question.

My question comes from the grade B type of cells; I would state the cycle live is more important then the starting capacity. I would pay a lot extra for a battery I test at 270Ah but that can do 6.000 cycles then a battery that now rates 280Ah but will only do 3.000 cycles.
All these people happy with their capacity tests might completely miss the point if they got batteries with clearly lower life cycles.

And if grade B cells also have less cycles ahead of them and we cannot verify this, we are just stuck....

Note; when charging from net as well;
- just charging from sun; 2.000kwh would go in in and out of the batteries maximizing solar usage
- charging from net during winter would do another 4.000kwh minimizing energy cost
These 2 combined and with about 2.000kwh directly used solar energy would cover the entire household energy (heating, warm water, electricity).

 

[toms]

The specs of the (lifepo4-cells-) manufacturers who only talk about cycles. Or do you have a reference of aging specs based on time?

LiFePO4 cells started becoming commercially available in 2007. The cells i use became available in 2010. I have yet to see a Winston cell that has been treated conservatively actually fall below 80% initial capacity.

Nobody can really tell you the failure mode at sub 0.5C rates, they are just guessing. (Including the manufacturers).

The outlook on the lifespan of a “B” grade EVE cell isn’t known, they haven’t been in use long enough.

The experience I had with “B” grade CALB SE200 cells suggests they start to fall over around the 6 year mark (typical use meaning partial daily cycle).

I think you will have a clear picture on the quality of the EVE (etc) cells on another 5 years.

By the way, if you insist on asking for references - you really need to apply that theory to the information you are receiving from manufacturers. They publish specs with no reference to the testing procedure or analytics behind them, and people take it as gospel.

My information comes from personally working with two independent LiFePO4 chemists at different times, as well as the operation of the dozens of systems I’ve installed over the last decade, and the correspondence of hundreds of people I have helped setup systems over that period.

LiFePO4 isn’t a new thing, but the nature of the chemistry is such that high C rate simulations don’t correlate to low C rate simulations, so testing for the latter needs to be done in real time. Hence 6000 cycles takes 16 years to test. Hasn’t happened yet..

 

[OffGridInTheCity]

Agree with all said above. However, one could look at it as glass half full in that you may very well get 5,000 or 7,000 cycles or more.

My 121kwh powerwall is 18650 lithium-ion and these are reported to have limited cycle life... however, we know there are some 10yr old 18650 powered EVs where the battery is still functional. We believe that low DOD / low stress gives a significantly better chance at long cycle life.

In my case, I'm averaging <36% DOD, low stress <0.1C, good ambient temps, and my oldest battery has 1,600 cycles with no degradation I can tell. On another forum a guy did continuous 100% DOD charge/discharge up to 2,000cycles on several different 18650 cells and a couple of brand name cells held up that long.

Eventually we will be able to share some meaningful data but meanwhile, there's hope

 

[scubadoo]

Do you have a reference to one of these plenty times?
I think we should rather speak of cycles rather then years.

Our 4 cell 300Ah Sinopoly LiFePO4 battery has survived 9 years of full time travel.
It also starts the motorhome 3.9l turbo diesel engine often multiple times a day.
Average daily draw from the battery is c120Ah.
My last annual November capacity test shows just a 3% capacity loss but it still exceeds 300Ah.
A month ago I added a 4 cell LF280K pack in parallel to the mix and will report longevity in another 9 years.?
What is a cycle related to RV use? I have no idea and would have to take a guess.
My Victron SmartShunt reports just 54 cycles over the last 3 years or so. Meaningless to me.

 

[timselectric]

For a given cell, is there any way one can predict the cycle life?

Time machine

 

[50ShadesOfDirt]

Interesting question(s)

There seems to be a good amount of research data out there, on just about any particular interest you have.

For example, if we are worried about cycles *or* calendar aging, some studies/thesis, etc. suggest that both are something to consider (worry about):

"Over time ... cells degrade and lose capacity in accordance with two different aging
phenomena: cycling and calendar aging."
"Initial storage history [storing a cell in full state, etc] ... leads to shortened cycle life."
uwspace.uwaterloo.ca/bitstream/handle/10012/12177/Catton_John.pdf (a thesis)

Other sites discuss "cycle testing":

"Cycle Testing: This is perhaps the most important of the qualification tests. Cells are subjected to repeated charge - discharge cycles to verify that the cells meet or exceed the manufacturer's claimed cycle life. Cycle life is usually defined as the number of charge - discharge cycles a battery can perform before its nominal capacity falls below 80% of its initial rated capacity. These tests are needed to verify that the battery performance is in line with the end product reliability and lifetime expectations and will not result in excessive guarantee or warranty claims."
www.mpoweruk.com/testing.htm (it's not a guesstimate, it's a process)

YMMV as you do the googling, but it seems that there's lots of available research material out there ...

Hope this helps ...

NOTE: I didn't answer the OP question, but maybe we can conclude that if the manufacturer is reputable, and/or you can get at their testing data (you may have to ask for this), you'd have better confidence that their "cycle life" claim is achievable and is itself a prediction vs a guess, assuming that calendar aging doesn't get you first.

 

[SeaGal]

I would disagree that cycles are bad defined; at least each supplier will define a specific cycle their batteries will support x-times before reaching 80% (or whatever value)... so that is well defined.
But indeed; how this specified cycle translate into actual use and 'real' life cycles is another question.

No doubt each cell maufacturer will have their own definition. But, unless all cells manufactures adhere to the same measurement and every BMS / shunt measures the same, then the cycle count that you will receive from the BMS/shunt won't necessarily match the supplier's concept of a cycle.

... My question comes from the grade B type of cells; I would state the cycle live is more important then the starting capacity....
...And if grade B cells also have less cycles ahead of them and we cannot verify this, we are just stuck....

Is there any research to show that Grade B cells have a reduced cycle/longevity? I _thought_ (but happy to be corrected) that they just had a reduced and not well-matched capacity (i.e. Ah) that their grade A cousins? Surely they use the same physical and chemical construction, so why would they degrade quicker?

 

[upnorthandpersonal]

Is there any research to show that Grade B cells have a reduced cycle/longevity?

No. The term hasn't even been around for more than three years and is not used in the industry. The reasons why a cell is 'grade B' is because they didn't pass internal QA at the manufacturer, for whatever reasons.
Personally, I just put them in a pack and put them to use. I'm three years in with mine (definitely 'grade B' cells), powering my house 24/7, no issues.

 

[walter-]

...

...I _thought_ (but happy to be corrected) that they just had a reduced and not well-matched capacity (i.e. Ah) that their grade A cousins? Surely they use the same physical and chemical construction, so why would they degrade quicker?

But the funny thing is all people testing them confirm they do have the capacity ;-). I have never seen anyone saying his grade B cells don't match. So capacity probable is not the main factor.
I would rather think it is more related to higher internal resistance (which could also result in lower capacity) or less efficient (how much energy you put i versus how much comes out, which is also related to the internal resistance). And higher internal resistance means more heat, faster degradation in every device I know. I don't often see people test the resistance at high loads or high charging currents and check that against specs.

Things being same "chemical construction" doesn't mean equal. The impurities in each cells can be different. This is the case in many industries. For example in IC chip construction. They are in theory all the same, but testing the chip will say which one can be sold at the highest Ghz speed, which one lower. And yes you often can run the low quality chip also at higher speeds, but at risk of an occasional error or at fast degradation or you need extra cooling...
But a main difference is that Intel or AMD will sell the chips on their own account. They remain liable. Here you talk to sellers where it is very unclear how much responsibility they want to take in case of issues. And you still don't know what the quality of your product is (why did they fail QC in the first place?).

 

[houseofancients]

LiFePO4 cells started becoming commercially available in 2007. The cells i use became available in 2010. I have yet to see a Winston cell that has been treated conservatively actually fall below 80% initial capacity.

Nobody can really tell you the failure mode at sub 0.5C rates, they are just guessing. (Including the manufacturers).

The outlook on the lifespan of a “B” grade EVE cell isn’t known, they haven’t been in use long enough.

The experience I had with “B” grade CALB SE200 cells suggests they start to fall over around the 6 year mark (typical use meaning partial daily cycle).

I think you will have a clear picture on the quality of the EVE (etc) cells on another 5 years.

By the way, if you insist on asking for references - you really need to apply that theory to the information you are receiving from manufacturers. They publish specs with no reference to the testing procedure or analytics behind them, and people take it as gospel.

My information comes from personally working with two independent LiFePO4 chemists at different times, as well as the operation of the dozens of systems I’ve installed over the last decade, and the correspondence of hundreds of people I have helped setup systems over that period.

LiFePO4 isn’t a new thing, but the nature of the chemistry is such that high C rate simulations don’t correlate to low C rate simulations, so testing for the latter needs to be done in real time. Hence 6000 cycles takes 16 years to test. Hasn’t happened yet..

i share the same experience with my plastic encased CALB 150 cells

 

[houseofancients]

i share the same experience with my plastic encased CALB 150 cells

my first batch was produced in 2015 and are still well above 80% , used daily