Pack Cycle Life - Can you help with the math?

[garolittle]

Question regarding “battery pack cycle-life”. This is a possible battery pack an electric go kart prototype. Here are the assumptions: 81.4V Battery pack consisting of of 22S/20P LG HG2 cells which are 3,000 mAh. Maximum current will be 300A and continuous current will be 200A. Pack will be charged at 15A as needed and the SOC will be between 10% and 90% most of the time during charging and discharging. Battery pack will be balanced with a battery management system while charging and discharging. 

The goal is to achieve a cycle life of at least 750 cycles. Is this possible given the assumptions listed above? We really appreciate any feedback or suggestions we will have. The VenomKarts Team 

 

[john61ct]

Yes it is possible.

But never guaranteed, just too many interacting variables, stuff happens.

A better approach might be to post it this way:

"How can we maximize the odds that the pack will. . ."?

You need a SoH% number as your defined EoL, the cells rarely drop off a cliff, capacity just keeps declining.

Express your current rates relative to capacity, so 0.3C, 4C etc.

SoC% can be a pretty nebulous concept without very precise mapping, for usage guidelines at the top and bottom, resting voltage per cell makes a better lingua franca.

Ignore the maximum stressful vendor specs, not good for lifespan.

So 3.5Vpc at rest or even 3.6 would be a better "working dead flat" definition of 0% SoC.

What precise SoC% that might be compared to 2.5Vpc is IMO irrelevant,

and will differ a fair bit by cell model anyway.

But keeping average DoD% lower can double or triple lifespan.

At a minimum, set your "while discharging" LVC so that your pack - after resting isolated for an hour - is never below that 3.6Vpc point.

That will vary by discharge C-rate at the time.

And also as the pack wears, losing capacity and increasing internal resistance.

The higher the average termination SoC% / voltage - really, the lower the utilization of capacity - the longer the lifespan.

At the high end, again ignore the vendor spec, define your own "working 100%". Lower is better for longevity.

Charge to 4.05Vpc and stop, no holding Absorb/CV is a fair bit lower than holding to say 0.05 trailing current.

4.15V is better for longevity than 4.2V, your judgment call, lower is better but maybe losing too much capacity, forces a higher Ah which makes for too heavy a pack.

Same with C-rates, lower the better but. . .

Temperature of the cells when fast charging is a **huge** life-shortening factor, can outright instantly render the pack scrap. Higher is better, to a point.

While colder is better when the pack is not cycling!

Also keep SoC low as much as possible, do not sit for long at Full, charge just before cycling required.

The actual #cycles impacted by each of these greyscale factors cannot in practice be quantified. No hard rules, only variables to balance against your use case.

Usually not to get too stressed about, best you can do in the circumstances.

 

[garolittle]

I really appreciate the response. That is excellent information. Already implementing some of these ideas on the existing prototype pack. Definitely use “slow charging” and usually stop at 4.1V per cell. Very good points regarding SOC and SOH. Really appreciate the response. Check us out www.venomkarts.com

 

[Okami]

Yes it is possible.

But never guaranteed, just too many interacting variables, stuff happens.

A better approach might be to post it this way:

"How can we maximize the odds that the pack will. . ."?

You need a SoH% number as your defined EoL, the cells rarely drop off a cliff, capacity just keeps declining.

Express your current rates relative to capacity, so 0.3C, 4C etc.

SoC% can be a pretty nebulous concept without very precise mapping, for usage guidelines at the top and bottom, resting voltage per cell makes a better lingua franca.

Ignore the maximum stressful vendor specs, not good for lifespan.

So 3.5Vpc at rest or even 3.6 would be a better "working dead flat" definition of 0% SoC.

What precise SoC% that might be compared to 2.5Vpc is IMO irrelevant,

and will differ a fair bit by cell model anyway.

But keeping average DoD% lower can double or triple lifespan.

At a minimum, set your "while discharging" LVC so that your pack - after resting isolated for an hour - is never below that 3.6Vpc point.

That will vary by discharge C-rate at the time.

And also as the pack wears, losing capacity and increasing internal resistance.

The higher the average termination SoC% / voltage - really, the lower the utilization of capacity - the longer the lifespan.

At the high end, again ignore the vendor spec, define your own "working 100%". Lower is better for longevity.

Charge to 4.05Vpc and stop, no holding Absorb/CV is a fair bit lower than holding to say 0.05 trailing current.

4.15V is better for longevity than 4.2V, your judgment call, lower is better but maybe losing too much capacity, forces a higher Ah which makes for too heavy a pack.

Same with C-rates, lower the better but. . .

Temperature of the cells when fast charging is a **huge** life-shortening factor, can outright instantly render the pack scrap. Higher is better, to a point.

While colder is better when the pack is not cycling!

Also keep SoC low as much as possible, do not sit for long at Full, charge just before cycling required.

The actual #cycles impacted by each of these greyscale factors cannot in practice be quantified. No hard rules, only variables to balance against your use case.

Usually not to get too stressed about, best you can do in the circumstances.

Some good info there.

Wanted to ask about this:

[quote""]
Temperature of the cells when fast charging is a **huge** life-shortening factor, can outright instantly render the pack scrap. Higher is better, to a point.
[/quote]

So how higher do u mean?

From my experience up 1C charge rate should be 'acceptable' for not degrading pack while charging more with more than 0.5C (which has been mentioned in some places as suggested charge rate for longevity.

Though what can be seen in data sheets of batteries, usually 2C charge rate is mentioned.

This, of course, does not take into account Lipo's get charged at 2-5c rates at times.

So yes, of course it is individual chemistry related.

But since we talking about Li-ions most of the time, would be cool to see some more detailed charge rate vs capacity degradation (lifespan) data

Sent from my ALE-L21 using Tapatalk

 

[flat tire]

You need rc lipo not canned cells. Those hg2s will sag and heat like crazy if you use the discharge rates indicated. If they do 750 cycles the capacity will be severely degraded at the end. RC lipo pack of the same capacity on the other hand can handle those discharge rates without a sweat. You can do 750 cycles with good charge discipline too. Turnigy graphene is one option. Yes RC cells have slightly lower energy density than canned cells but not much and the power density blows canned cells out of the water. Much better and more consistent performance.

If you don't care about the weight of the pack and want the best charge, discharge and longevity performance look into LTO. LTO has rather poorer energy density but you can charge it up in 10 minutes.

 

[john61ct]

Temperature of the cells when fast charging is a **huge** life-shortening factor, can outright instantly render the pack scrap Higher is better, to a point.

So how higher do u mean?

From my experience up 1C charge rate should be 'acceptable' for not degrading pack while charging more with more than 0.5C (which has been mentioned in some places as suggested charge rate for longevity.

Though what can be seen in data sheets of batteries, usually 2C charge rate is mentioned.

This, of course, does not take into account Lipo's get charged at 2-5c rates at times.

So yes, of course it is individual chemistry related.

But since we talking about Li-ions most of the time, would be cool to see some more detailed charge rate vs capacity degradation (lifespan) data

First off, trying to get optimized longevity requires the other factors to be also aligned,

if the use case already requires abusive high **discharge** rates, then other factors impact is going to be reduced.

Charging regularly at even 0.2C can really reduce lifetime if temps are anywhere near freezing.

Going over 1C is no problem at all if the cell is at 40°C

But if the use case **requires** fast charging, then you just accept replacing the pack more frequently.

These general relationships between the variables are the same whether LI round cells, hobby LiPo, pouches, EV packs or the big LFP prismatics.

But as I said, every factor is infinitely variable, no black and white hard edges.

And so many factors interrelated, and vary by chemistry, vendor, model even mfg date.

As well as cell age.

So quantification is either impossible, or of very limited use for **your** situation.

 

[garolittle]

You need rc lipo ......

Really appreciate the response. Good info and I will certainly research it more. Thanks

 

[john61ct]

Note that very high **power** density greatly inflates pricing, and is in opposition to **energy** density.

So wrt range, longevity and value for money, a good rule of thumb is,

given equal reputation for "quality" (consistency, reliability, longevity)

do not pay for a higher power density than what you actually need.

If you can expand Ah capacity due to improved energy density, you are reducing the C-rate **and** can leave more capacity in the cell after a given range

which multiplies longevity

further increasing value as in $/Ah **per year**

 

[serious_sam]

You can do 750 cycles with good charge discipline too.

This is one aspect of LiPo cells that I haven't seen any data on. Energy density isn't too far different from li-ion to make any significant real world difference. Definitely they kick arse in power density. IMO, the downsides of LiPo is limited test data (especially on cycle life), and significantly higher cost.

In a race/track situation, you're going to be running high percentage of lap time at WOT, so peak current is going to be occurring for a significant amount of time. There's going to be minimal "continuous" level current usage. Throttle mostly on or off, with a little feathering through corners. You have to pay to play, and LiPo is the logical choice.

 

[garolittle]

.....
If you can expand Ah capacity due to improved energy density, you are reducing the C-rate **and** can leave more capacity in the cell after a given range

which multiplies longevity

further increasing value as in $/Ah **per year**

Now that is an awesome quote I will not forget! Thanks

 

[garolittle]

You can do 750 cycles with good charge discipline too.

In a race/track situation, you're going to be running high percentage of lap time at WOT, so peak current is going to be occurring for a significant amount of time. There's going to be minimal "continuous" level current usage. Throttle mostly on or off, with a little feathering through corners....

CORRECT. Excellent points.

 

[garolittle]

...But if the use case **requires** fast charging, then you just accept replacing the pack more frequently.

These general relationships between the variables are the same whether LI round cells, hobby LiPo, pouches, EV packs or the big LFP prismatics.

But as I said, every factor is infinitely variable, no black and white hard edges.

And so many factors interrelated, and vary by chemistry, vendor, model even mfg date.

As well as cell age.

So quantification is either impossible, or of very limited use for **your** situation.

Good info. Charging would usually occur at about 30 degrees (Celsius) and at at 1C (or less). Cells would be brand new. Really appreciate the response. Good info.

 

[Vbruun]

The current is probably a bit high for that kind of cycle life, but who knows. This is also very much dependant on when you define the pack as "dead". You should probably set some goals for internal resistance and remaining capacity before proceeding any further.

For this scenario, temperature management is also very important, so maybe you should consider that when designing the pack as well.
If you can afford it, I think proper R n D is the way to go with prototype packs and all the jazz. But I don't know if that is an option for you financially? It will probably take a few iterations to reach a pack capable of your goals.


It should not take all that long to reach 750 cycles on a test pack at this current level. 200a continuous on a 60ah battery = 3,3C or just under 20 mins for the discharge. Then for testing you could heat the pack to maybe 20c and charge at 0,5 C for a total cycle time of 2,5 hours.
This gives a total testing time of 72 days if you test non-stop. Maybe not all that fast

A cheaper possibility could be to make a smaller (and cheaper) test-pack or do individual cell tests. These should be scaleable as long as you have good thermal management in the final battery pack.


BTW i strongly disagree that you should use LIPO. That amount of energy stored in LIPOS is just way to dangerous in my opinion. Especially if you intend to have this these cart run by other than yourself.

 

[garolittle]

T

A cheaper possibility could be to make a smaller (and cheaper) test-pack or do individual cell tests. These should be scaleable as long as you have good thermal management in the final battery pack.


BTW i strongly disagree that you should use LIPO. That amount of energy stored in LIPOS is just way to dangerous in my opinion. Especially if you intend to have this these cart run by other than yourself.

I really like the idea of a smaller test pack. These are some great suggestions and I will certainly give all of them some serious consideration. Our first prototype used the Phase Change Composite (PCC) material from Allcelltech.com. Fantastic passive management for heat but a little expensive. I thought one way to go without the PCC is to simply increase the number of cells in parallel. If fact, we could easily incorporate "air cooling" by routing air through the pack as we drive (at high speeds) and we can probably increase the number of cells in parallel to as high as 30 cells.

 

[Vbruun]

T

A cheaper possibility could be to make a smaller (and cheaper) test-pack or do individual cell tests. These should be scaleable as long as you have good thermal management in the final battery pack.


BTW i strongly disagree that you should use LIPO. That amount of energy stored in LIPOS is just way to dangerous in my opinion. Especially if you intend to have this these cart run by other than yourself.

I really like the idea of a smaller test pack. These are some great suggestions and I will certainly give all of them some serious consideration. Our first prototype used the Phase Change Composite (PCC) material from Allcelltech.com. Fantastic passive management for heat but a little expensive. I thought one way to go without the PCC is to simply increase the number of cells in parallel. If fact, we could easily incorporate "air cooling" by routing air through the pack as we drive (at high speeds) and we can probably increase the number of cells in parallel to as high as 30 cells.

I think the key for you is to measure the temperature carefully at different locations in the pack.
Og you can keep these comparable in a smaller pack, your test results should be useful. But I Think that this is crucial to get right if you want to do a scale pack

 

[garolittle]

“I think the key for you is to measure the temperature carefully at different locations in the pack.
Og you can keep these comparable in a smaller pack, your test results should be useful. But I Think that this is crucial to get right if you want to do a scale pack”

Good ideas. Thanks! Gary