Battery Pack Power Density

[Pete9008]

I'm trying to get a handle on the power capability of battery packs, in particular lower capacity packs from Hybrid vehicles. My goal is to come up with a pack configuration that could cope with more or less continuous 75kW power levels while having a weight of 50kg or less.

At the moment I don't have much in the way of data points and am hoping that others here can help fill in the blanks.

Currently my two reference points are:

BMW PHEV packs - The best data I've found is here viewtopic.php?p=34158#p34158 which gives a max short term power of 83kW at 354V so 234A. With a 34Ah capacity that equates to 7C. I'm using one of these in my current conversion and am assuming that this can comfortable be pushed to 10C if monitored/cooling. I'm also assuming a 3C continuous rating, or around 35kW. The pack weight is around 100kg.

Honda Blue Energy packs - The best data I've found is here https://www.gs-yuasa.com/en/newsrelease ... 461410_446 which gives a max current of 300A for a 5Ah cell or 60C. A 96s pack would therefore have a peak power handling of 106kW with a total capacity of 1.7kWHr and a weight of 22kg.

Now the second pack looks like it may be able to get close to my requirement (especially if two packs are used in parallel). The problem I have is that this depends on what the continuous current rating of the cells is and I have no idea what this might be??

I'd also like to better understand why one pack can cope with 7C and the other 60C. I can see that cell technology varies and optimisations for power or energy density will affect things but a factor of x7.5 seems excessive. That suggests either the Blue Energy rating is optimistic or the BMW rating pessimistic.

I know from playing with RC Helis that Lithium batteries can be pushed quite hard, and those batteries are often advertised as 100C rated, but I also know from experience that you can actually measure the cell degradation produced on a charge cycle by charge cycle basis when using them this way. Is cell life part of the explanation?

So my main question - What C rating is reasonable for continuous and peak operation (assuming good cell cooling) and are some cells really that much better than others?

[rstevens81]

There is one thing to remember is to not over think data sheets when we're not designing a 250000mile car, also were not using overly complex thermal heating/cooling, we are only really worried what the cells can do, the rate of expansion of EVs I would consider all cells disposable after 3-5years....therefore any phev pack (BMW,Vw etc al) will suffice, just expect a big voltage sag when it's really cold.

[tom91]

It all depends on what the cells are geared for. Usually this is a balance between power/weight and energy/weight.

Good thing is watching two parameters can easily indicate you are reaching limits for "burst output":
1.Voltage Sag
2.Temperature rise

Of course the cycling of cells all comes down to the chemical degradation this usually is based on a few factors lie:
1.Operating SOC window
2.Stored SOC
3.Operating and storage temps
4.Peak/burst Power draws
5.Peak/burst Charging

Good site for info and benchmarks https://www.batemo.de/products/batemo-c ... chem-e66a/

[Pete9008]

There is one thing to remember is to not over think data sheets when we're not designing a 250000mile car, also were not using overly complex thermal heating/cooling, we are only really worried what the cells can do, the rate of expansion of EVs I would consider all cells disposable after 3-5years....therefore any phev pack (BMW,Vw etc al) will suffice, just expect a big voltage sag when it's really cold.

If it was a factor of 2 or so then I'd agree completely but the difference between 7C and 60C is just a bit too much to ignore. 60C would mean what I would like to achieve might be realistic.

I get the feeling that there are a few people who do know how to do high power density battery packs but they are mostly in the motor-sports world and treat it as a competitive edge that they are not keen to share (understandably).

Would love to know what the compromises/limits are, reduced battery life I can live with, flames not so much!

[tom91]

Limits are the ones you follow as stated:
1.Temps
2.Voltages

The rest can be sent out the window, and getting the propper specs of how long can stay within that window with the draw depends on loads of factors.

[Pete9008]

Good site for info and benchmarks https://www.batemo.de/products/batemo-c ... chem-e66a/

Nice site, thanks.

[tom91]

BMW PHEV gen 1 cell i believe.

https://www.batemo.de/products/batemo-c ... -bmw-530e/

[Pete9008]

BMW PHEV gen 1 cell i believe.

https://www.batemo.de/products/batemo-c ... -bmw-530e/

Thank you!

That must be the 9kWHr version, so assume the 12kWHr would be a bit more?

If so it quotes a continuous current of 216A, that is much more than I was expecting and would mean the pack could in theory sustain a continuous 75kW! Is that realistic?

[tom91]

That is what BMW sized it for.

Again everything is actively managed and heavily load vs time dependant. Plus ambient conditions ofcourse.

What are the specs you want to hit?

75Kw continuous, but it aint as its a discharge over a set time. What energy capacity you looking at.

[Pete9008]

Based on that sites data the 9kW BMW cell has a continuous current rating of 7C and peak of 10C. Assume that would be with perfect cooling which might not be possible when built into a pack, any idea what kind of derating factor is appropriate for cell limits vs pack limits. Is that where the 3C limit comes from?

Got to say that is massively increasing my confidence level that what I'm after may be realistic!

[tom91]

Please specify what you are after.

If you want biggest punch per Kg these are great.
https://www.lonestarevperformance.com/s ... cells.html

[Pete9008]

What are the specs you want to hit?

75kW continuous but alternating between discharge and charge. Pack only needs to be small, say 0.5kWHr to 1.5kWHr, it's for a hybrid boost on a track car, see viewtopic.php?p=54483#p54483 for more details.

[tom91]

Again, thats just one spec.

Voltage constraints? As this is the most important driving factor for cell evaluation.

75Kw at 100V is 750A were at 350V it is "only" 215A

The biggest issue with weight starts going into all the bits that are not the "cells" so there is alot of discussions to be had around voltage.

[tom91]

So lets take some stab at some more spec

Power Rating: 75Kw
Energy Cap: 1kWh (as an example)
Weight: 50Kg

Power Density: 1500 W/kg
Energy Densityy: 20 Wh/kg
Discharge Ratio: 75 C

These are all "obtainable" numbers from my experience.

[Pete9008]

Yep, aware of the issues surrounding voltage, that's why I started with two cases both with 96s. Also why I'm trying to talk in C ratings rather than kW where possible.

Think I've been looking in the wrong places for cells!

That Lonestar cell looks very good but the cycle life of 500 is a bit worrying. I'd kill that in a single track day Kind of explains the trade off though.

Edit - Thanks, good to know it sounds possible, gives me a little more confidence to start looking at the motor package

Again thanks for the help.

[tom91]

Life cycles are always based on full drains, that is why oversizing packs brings down two things:

1.SOC cycle
2.C-rate

[Pete9008]

That still leaves me with one question though -

Are these higher power cells fundamentally different in some way or are they largely being better cooled, pushed harder and a shorter life accepted?

[Pete9008]

Life cycles are always based on full drains, that is why oversizing packs brings down two things:

1.SOC cycle
2.C-rate

Agree, but on a typical PHEV cell, even on full drain, the life would be a lot higher wouldn't it?

So taking the 500 cycle life on the Lonestar cell, if you only use 30% of the capacity would it scale linearly to 1500 or are there other factors coming in there too?

[Pete9008]

I think the conclusion I'm coming to here is that I need to put a pack together and try it!

Given that I am using one of the BMW packs for the Smart Roadster conversion (with a target of 100kW motor power) the best thing to do, once it's running, is take it to a track day and see what it can do while keeping a very close eye on temperatures and voltages (try and get a evening session, the battery will probably only have the capacity for 10-15min on track!). And then compare the capacity before and after.

I think the above has given me enough confidence in the meantime to start thinking about the motor package though.

[tom91]

Motor package will be crucial.

In the end it all comes down to trade offs. Also taken cells beyond their intended scope will have consequences, yet these are hard to define unless they are tested.

Why do you think manufacturers start out with vehicle simulations to define required usage and then feed that into battery models and then into single cell level tests.

Something I created when in Uni in matlab was capable of basic vehicle level simulatins.

[Pete9008]

In the end it all comes down to trade offs.

What doesn't

Also taken cells beyond their intended scope will have consequences, yet these are hard to define unless they are tested.

Have to wonder if that's partly what the Lonestar cells are, the result of pushing the limits and testing the results?

Why do you think manufacturers start out with vehicle simulations to define required usage and then feed that into battery models and then into single cell level tests.

Cost and speed. A simulation is just building the vehicle/battery/cell inside a computer model to answer the same questions that used to need physical testing. Unfortunately it builds on a history of doing the physical tests needed to generate the library of data that is needed for the model to be accurate enough to be of value. If you don't have a good enough model then simulation can be a waste of time. A simulation is a lot quicker to redo when you break something though!

[Pete9008]

Something I created when in Uni in matlab was capable of basic vehicle level simulatins.

Nice! Does it need Matlab or will it run under any of the open source alternatives (Scilab??)

[tom91]

The simulation is so old and mainly developed from a learning point of view.

Would have to create a new one from scratch really. Usually a simple excel sheet gets me most of the headline info anyway so no real incentive to create one.

[Pete9008]

Usually a simple excel sheet gets me most of the headline info anyway so no real incentive to create one.

Agree, was thinking about pareto earlier, I used to strive for 100% all the time. More recently I'm starting to accept that 80% may be good enough an awful lot of the time. Although having said that some things, especially anything safety related, does need to be 100%.

Found a little more detail on the BMW, 50kW continuous so that does sort of tie in with the above (difficult to be sure as the limit could be due to battery, motor or inverter).

Reasonably convinced that the Blue Energy based NSX pack would almost do what I need, two definitely would - difficult to get in the UK though. An i8 pack might but can't find enough data to be sure - easier to get though. Lonestar type cells should although there is the added complexity of building a pack from scratch, possibly also shipping issues. The alternative is to see whether an easily available PHEV pack can be pushed a bit harder (well quite a bit harder!).

The final issue is voltage as you mentioned. The motor gets a lot easier with higher voltages, 350V would be the minimum, 400V would be better, 500V probably about the limit given the inverter I'd likely be using. The problem with most PHEV packs is getting these voltages needs a lot of heavy cells.

Being realistic though it's going to be at least a year till the I even start this project and the motor will probably take a year or two so the battery landscape is likely to have changed considerably. Maybe the simplest thing is to just assume that something is/will become available and start thinking about the other bits. The above has done enough to convince me that it is likely to be the case.

Edit - looking at the temperature rise graph on the Batemo site for the BMW cells also gives a power of 57kW for a 96s pack (for 50C bat temp, convection cooling) so that seems to tie up too. The energy characteristic plot suggests that efficiency drops significantly above 75kW too.

[Pete9008]

Please specify what you are after.

If you want biggest punch per Kg these are great.
https://www.lonestarevperformance.com/s ... cells.html

Just had another look at the data on these cells. The discharge currents look good but the max charge current is only 10A (2C)!

Does that mean they can't handle regen or am I missing something?

Edit - A better way to ask this question may be are cells normally symmetrical for charge/regen and discharge currents or is the always a large asymetry? For a hybrid boost pack for track use, where regen and discharge will be similar for both currents and durations, is it actually the charge/regen current that is the limiting factor?