It's All About the Delta (and Temperature, but the BMS will Protect You) -- Test: Effect of temperature on battery cells life

[Maxx]

The x axis covers 20 cycles of the RPT testing.

How many actual cycles does that translate to? In the graph you posted, that is a sharp drop at 10 cycles. That seems awfully early even at 90% SOC to be regular cycle.

he warned us a few months ago about the the low thermal stability point of the cells shown in one chart. He stated our cells have a thermal stability point that is 95 degrees F lower than what he has seen in similar cells of the same chemistry.

Can you expand on this? What do you mean by lower thermal stability? Does that mean our cells do better with hard acceleration and max regen at lower temperatures but not so much at high temps comparing to other EVs with similar battery chemistry? Does it mean they lose capacity at high DOD and high temp faster than they do at low temp comparing to others. Or it means something else?

Sponsored

 

[Maxx]

Does anyone understand what is up with the green line in the left? Why is 30% is dropping faster than 90%?

 

[Scorpio3d]

TLDR: 1. The most important issue and key takeaway is to avoid large temperature swings. The BMS should take care of this, and seems to do a fine job given the battery temp is always in the middle on my dash display. The original leafs with air cooled batteries degraded so quickly because a lack of a BMS and temperature regulation.

2. Avoid large charge/discharge cycles. It seems not as important as 1 according to Ford, since they say you can charge to 90% every day. These batteries love to be closest to 50%. You can take the one way % consumption of your commute and add it to 50%. If you use 15% to get to work, charge to 65%. By the time you get home, you should be at 35%. This keeps you closest to 50%. IMO, I just charge to 80 every night and 100 for roadtrips.

I read this twice as well, even though this is well out of my wheelhouse (so far out that I am in a rowboat and you are in a rocket)and I think you overall nailed it except for maybe this one thing he said
@MickeyAO said:
“On the side, I ran my own test cycling between 80% (actual so 85% SOC on the display) and 50% SOC. Results closely tracked the cycle 30% delta at 25C as I suspected, but I did not graph it.”
so my thinking is that even with a higher state of charge at 77°F it is basically the same??

 

[Grumpy2]

I am not an expert only reading and learning along with everyone else. I just like to know how things work.

He said he only provided "normalized graphs, so I assume this is the end result of a large amount of tables, charts, and graphs. This testing took many months, so you can count on a box full of test results behind this summary lines.

What do you mean by lower thermal stability?

Others have used this term to describe the loss of stability, i.e. thermal runaway... This temperature limit is almost impossible to obtain unless your in an extremely hot environment situation.

As the OP advised, you now have a 7 year warranty that has few limits on how your treat your truck, so just enjoy it. We will probably learn more in the years ahead of how these batteries survive.

 

[Scorpio3d]

How many actual cycles does that translate to? In the graph you posted, that is a sharp drop at 10 cycles. That seems awfully early even at 90% SOC to be regular cycle.



Can you expand on this? What do you mean by lower thermal stability? Does that mean our cells do better with hard acceleration and max regen at lower temperatures but not so much at high temps comparing to other EVs with similar battery chemistry? Does it mean they lose capacity at high DOD and high temp faster than they do at low temp comparing to others. Or it means something else?

I think lower thermal stability comment had more to do with the way our battery packs are made or the difference between them and the battery cells/packs from other manufacturers. I.e. Ford packs versus GM packs or Tesla, etc.

 

[MickeyAO]

Can you expand on this? What do you mean by lower thermal stability? Does that mean our cells do better with hard acceleration and max regen at lower temperatures but not so much at high temps comparing to other EVs with similar battery chemistry? Does it mean they lose capacity at high DOD and high temp faster than they do at low temp comparing to others. Or it means something else?

Thermal Stability is the highest temperature a cell can sit at without an effect to the cell. Once you go above this temperature, the cell will start going through an a thermal reaction and will start increasing the temperature of the cell without other inputs. It will steadily increase in temperature until it hits thermal runaway temperature.

 

[MickeyAO]

Does anyone understand what is up with the green line in the left? Why is 30% is dropping faster than 90%?

The static capacity of that particulate sample is dropping faster than the other sample. I think I called it out this out in my surprises.

 

[MickeyAO]

I read this twice as well, even though this is well out of my wheelhouse (so far out that I am in a rowboat and you are in a rocket)and I think you overall nailed it except for maybe this one thing he said
@MickeyAO said:
“On the side, I ran my own test cycling between 80% (actual so 85% SOC on the display) and 50% SOC. Results closely tracked the cycle 30% delta at 25C as I suspected, but I did not graph it.”
so my thinking is that even with a higher state of charge at 77°F it is basically the same??

Yes. I duplicated my normal drive cycle on the side. 30% delta with the upper limit at 85% SOC (displayed, 80% actual)) closely tracked closely with the 30% delta at 25C at 50% SOC as the center point.

 

[lightspeed]

This doesn't say much really and the information is presented in an intentionally cryptic way.

A bunch of graphs with no markings isn't very informative. They did up to 20 cycles of something to the pouch cells. Apparently, without temperature management that the truck would do via liquid heating/cooling. Some of the graphs show apparently large drop-offs, except there are no units so impossible to really interpret.

Two years in so far, and my BMS reports 1.5% "health" loss and I see no difference in 100% range.

 

[Kit2874]

This doesn't say much really and the information is presented in an intentionally cryptic way.

A bunch of graphs with no markings isn't very informative. They did up to 20 cycles of something to the pouch cells. Apparently, without temperature management that the truck would do via liquid heating/cooling. Some of the graphs show apparently large drop-offs, except there are no units so impossible to really interpret.

Two years in so far, and my BMS reports 1.5% "health" loss and I see no difference in 100% range.

Lol was thinking the same thing..
For the people that like these graphs and everything, hey cool!....

But 8 years 100k Is the battery warranty If it fails before that, they can fix it. And if I still like the truck, most likely the batteries will even be cheaper than they are now. I'll put a new battery in it.

 

[lightspeed]

Lol was thinking the same thing..
For the people that like these graphs and everything, hey cool!....

But 8 years 100k Is the battery warranty If it fails before that, they can fix it. And if I still like the truck, most likely the batteries will even be cheaper than they are now. I'll put a new battery in it.

Yeah I'm hoping in 10 years that a double-capacity battery at roughly the same weight will cost less than $20K. Given current trends, should be very possible. The only question is the difficulty of integrating with the truck. Might just be more cost effective to buy the updated trucks.

 

[ctuan13]

TL;DR you are missing out on actual data that is already summarized as much as I can while presenting the facts about the cells in your truck. I'm not going to summarize it further for you

So, I recently retired from the Institute (and received my final paycheck with my 8 weeks of vacation being paid out) and the exit documents I signed only talked about trade secrets, so I think it is safe to share some actual (normalized) data explaining what I've been talking about over the last couple of years on this forum.

For the last 14 years I have been testing Li-ion cells from all sorts of chemistries (LTO, LCO, LMO, LFP, NCA, NMC, and others that I can’t talk about, but hopefully you should be hearing about in 3 years if all goes well). My lab was more automotive centric, but I have certified camera battery packs for the space station, tested cells for medical implant devices, and even worked on the project for a 3mW pack for a mining company. I actually got my start in energy storage by working on the control system for a flow battery 16 years ago (my degree is in Computer Science, and what I was originally hired at the Institute for). Over the years I have learned how to kill cells (some clients wanted cells to fail) and how to make them last (virtually) forever, and purposely making them catch fire through abuse testing. I say this to show I’ve been at this awhile and know what I’m doing by observing real time data (and maybe a little patting myself on the back after 2 successful 20 years careers and retirements from the USAF and the Institute ).

You all have seen my advice on what is happening with the cells and how to make them last (virtually) forever. I'm sure you have seen all the naysayers who drove EVs for xx number of years and have seen data from the OBD port saying it doesn't matter. Here is some hard actual data.

About the methodology, a Taguchi L9 matrix with 3 charge currents, 3 discharge currents, 3 temperatures, and 3 delta SOC for cycling, with 2 samples per test. The delta is centered on 50% SOC. That means a 30% delta is between 35% and 65% SOC. Despite the varying conditions, all cells end up cycling just about the same kWh of energy throughput. It makes it very easy to see which of the conditions is most dominant in degradation. We are then able to build a mathematical formula that will calculate the lifespan for any type of duty cycle we feed into it. On the side, I ran my own test cycling between 80% (actual so 85% SOC on the display) and 50% SOC. Results closely tracked the cycle 30% delta at 25C as I suspected, but I did not graph it.

Calendar life is 3 temperatures and 3 storage SOCs. When you see 80% SOC on calendar life, it means it sits for 4 weeks at 80% SOC (actual) at the listed temperature.

Cycle life Reference Performance Test (RPT) is performed every week at 25C so we can compare the results across all test coditions. Calendar every 4 weeks at 25C. While we take pulse power and resistance for charge and discharge at 10% SOC increments (90% to 10%), the graphs we produce typically only show discharge pulse power and resistance at 50% SOC. I'm not going to bother showing those graphs. I will show capacity loss based on the test conditions. RPT0 (the first dot) was done before any of our testing started.

We do 10 RPTs for calendar, and 20 for cycle life. We were only at RPT8 for calendar life (RPT9 started on my last week) but have completed cycle life. If you see a test that suddenly has no more dots, then it had to be pulled from testing. In most cases, it was because the cell started to swell (produced off gassing that was trapped in the cell skin). In a couple of cases, the resistance just got too high for us to conduct the RPT…the voltage would shoot past Vmax at the first application of the pulse.

I've normalized the graphs, so you can't see the actual values. I will tell you that RPT0 was just under the rated capacity for the cell (we bought modules from a 3rd party that did some light vehicle testing before pulling the modules to sell to us), and the top of the graph is around 10 Ah above the rated capacity. I'll leave it to the reader to find the rated capacity for the SKE805A... hint, it's in the nomenclature of the name of the cell. BTW, I have 1st hand knowledge that you cannot replace a single cell in a module.

Calendar life started at 25C, 45C, and 55C. The early results from 55C and the previous results of the thermal stability test forced us to use a different temperature for cycle life (I threatened my colleagues that I would stop cycling anytime the cell hit 60C and let the cell cool back down). 10C was chosen as it helped the spread for the model, but I was arguing for 35C as a more real-world condition.

All testing was at actual SOC of the cell without a BMS, so you will have to convert with what the truck reports on the screen. This is accelerated testing, but I have not seen the model (it will be produced in about 10 weeks, and I won't have access to it), but this falls somewhat in line with what I have seen over the last 14 years of cell testing.

For cycling, the 2 dominate conditions on degradation were temperature (very dominate but the BMS should protect you from this) and delta SOC (almost as dominate). Minor degradation conditions were charge and discharge currents (high being the worst for both). Nothing here surprised me EXCEPT for the VERY low thermal stability point (about 35C or more below what I’ve seen in other NMC cells), the 1 calendar condition at 45C (I have never seen where 80% SOC outperformed lower SOCs and made a post about it), and 30% delta at 10C (Li-ion cells don't like being cold, but 30% delta should have been better on sample 2). As my mentor Bapi liked to tell tours, Li-ion cells like the same temperatures humans do.

All this said, your battery has a warranty for 8 years down to 70% capacity loss. This information is only a concern if you plan on keeping the truck for the long term or don't want to go through the hassle of a warranty replacement on the battery.

Thank you so much for sharing your data with us. Also you once again demonstrate why you are one of our most knowledgeable members here on the forum.

While I by no means am perfect with regards to how I charge and drive my truck, this data does offer me good reassurance that my general reluctance to ever allow my truck to dip below 40% even on a cross country towing road trip has always been good practice. It also confirmed for me that my top end charging limit of between 85-90% (SOCdisplay) (80-85% actual), while not perfect, is certainly better than consistently and persistently charging to 100%.

 

[VTbuckeye]

Thank you so much for sharing your data with us. Also you once again demonstrate why you are one of our most knowledgeable members here on the forum.

While I by no means am perfect with regards to how I charge and drive my truck, this data does offer me good reassurance that my general reluctance to ever allow my truck to dip below 40% even on a cross country towing road trip has always been good practice. It also confirmed for me that my top end charging limit of between 85-90% (SOCdisplay) (80-85% actual), while not perfect, is certainly better than consistently and persistently charging to 100%.

It provides me assurance as well. I usually charge to 60 percent and my typical day uses less than 10 percent. If I charge to 70 or 80 it shouldn't be any more challenging to the battery. I think I've been down to the low teens once and up to 100 percent maybe 5 times (prior to longer drives/towing). I will probably charge more often and keep my soc above the 30s when traveling in the future.

 

[MickeyAO]

This doesn't say much really and the information is presented in an intentionally cryptic way.

A bunch of graphs with no markings isn't very informative. They did up to 20 cycles of something to the pouch cells. Apparently, without temperature management that the truck would do via liquid heating/cooling. Some of the graphs show apparently large drop-offs, except there are no units so impossible to really interpret.

I take it that you didn't read the write up and skipped to the graphs. I do state where RPT0 started, and what the top of the graph is. From that you should be able to calculate each of the hash marks.

 

[lightspeed]

I take it that you didn't read the write up and skipped to the graphs. I do state where RPT0 started, and what the top of the graph is. From that you should be able to calculate each of the hash marks.

Why don't you just summarize key results like any research paper would do instead of making us decode the graphs?

For example, in the charge graphs, some of the cells drop off a cliff at a low number of charge cycles. But is each test a single charge cycle or does it cover a group of charge cycles? Are some of your test batteries just bad?

The 45c graph shows a cell apparently dying after only 8 charge cycles. That seems extreme.

What do these results mean in the context of the BMS and its active cooling/heating?

Sponsored