LFP system design approach - comments welcome

[nebster]

Small comments only:

1) Sounds pretty good, across the board!

2) It's 32kWh, not 32kW. I find it helpful to be precise/pedantic with units, as should you (as a real EE).

3) I'm unconvinced there's much special worth hunting down CALB at this point. Unless maybe you want their CAM aluminum-shell batteries for the added efficiency. (At the same time, Winston's custom yttrium concoction is also un-special.)

4) If you're making room for AGM fallback, it's unclear what you're trading: are you just planning to have spare storage space in the meantime? No problem with that, of course, just wondering.

5) You do need more thermal management (a way to deal with heat, mostly) for LFP longevity than you would for lead acid.

6) You didn't mention charge rates. You may find you need a CV stage to get good top-off depending on how high your rate is under CC/bulk. (I do.)

7) Staying away from balancing/shunting active BMS is smart. Just get good cells, set them up properly up front, and you will be fine.

8) I would engineer for multiple strings on a boat for maintenance and better fault tolerance. As a bonus, if you do have the spare shelving I asked about above in (4), you can expand your system pretty easily if you like.

9) It's not obvious why 48V forces you to all-12V. You can have two converters on board, and in fact the 48V-to-24V supply will probably be more efficient. I think you should make the pack voltage choice based on topology, space and wiring constraints, and aggregate power demand. Moreso if you're going to make a custom monitoring solution (cheap 16s monitors are basically nonexistent).

10) Speaking of that custom PLC, let us know how we can buy one.

[tanglewood]

Originally Posted by CatNewBee

How about a Tesla PowerWall and Inline solar controller? You can then step down to tne 12V for yout electronics by a regular charger and a small buffer battery. Just saying...

I've looked briefly at Tesla, but info is very limited unless you put down a deposit to buy one. That reminds me of buying a car back in the 70s where the dealer wouldn't show any cars until we put down a deposit. Needless to say, we went elsewhere. But based on what I have gathered, here's why I think Tesla isn't a good fit.


1) The expressly exclude mobile apps from allowed/supported use


2) The inverter is integrated into the product, so I would be buying into a whole system of batteries, inverter, BMS. I have always found this to end poorly since you are locked into one vendor, and if anything goes wrong or doesn't perform, you are stuck unless you replace everything.


3) Last I checked, it is designed for grid tied application only. Generator support was "coming soon".


4) There is no provision for direct DC charging, so can't charge directly from solar or alternators. This is part and parcel with the grid-tie design focus, where solar charging comes from a grid tied solar inverter, back feeding the Tesla to charge itself.


I don't begrudge Tesla for their design since the bulk of the market is grid tied, but it's really not a good fit for a boat or any off-grid application. Some enterprizing sole can and probably will hack a way to make it work, but to me the appeal of something like the Tesla would be a plug and play solution like it is intended for grid tie. But it's far from that for off-grid.

[tanglewood]

[QUOTE=nebster;2646033]Small comments only:

1) Sounds pretty good, across the board!

2) It's 32kWh, not 32kW. I find it helpful to be precise/pedantic with units, as should you (as a real EE).
[/quote


Oops. Good catch.

Originally Posted by nebster

3) I'm unconvinced there's much special worth hunting down CALB at this point. Unless maybe you want their CAM aluminum-shell batteries for the added efficiency. (At the same time, Winston's custom yttrium concoction is also un-special.)

I have expanded the search to include Winston, and welcome any other suggestions. BTW, on first inspection, Winston doesn't seem to be much more available that CALB.

Originally Posted by nebster

4) If you're making room for AGM fallback, it's unclear what you're trading: are you just planning to have spare storage space in the meantime? No problem with that, of course, just wondering.

I'll use the space for storage as best I can, but am definitely giving up some of the space-saving advantage of LFP. But I think reversibility - so to speak - is important. If I get stuck somewhere at some point and conclude LFP was a bad move, I want to be able to revert to AGM. I hope not to do that, but want to cover my butt. I also worry that LFP may make the boat unattractive to prospective buyers down the road when I ultimately sell it, so want to be able to revert to conventional power storage if I feel that's important. And in the mean time, you are correct that it provides some expansion space. In fact, it reminds me that I need to plan that expansion in the layout....

Originally Posted by nebster

5) You do need more thermal management (a way to deal with heat, mostly) for LFP longevity than you would for lead acid.

This is one of my big concerns with LFP. Batteries will be in the laz which has power ventilation if needed to cool, and residual heat from boiler etc to stay reasonably warm. I leave the heat on in the boat (but temp set back) when left unattended in cold weather, so the laz stays acceptable warm assuming the heat doesn't fail.


Cooling is a bit more of a concern. With max operating temps in the 30-40C range for most LFP, that could be a challenge in a really hot climate. My thinking so far is to provide good air gaps between cells, and monitor to see how much of an issue is really is.

Originally Posted by nebster

6) You didn't mention charge rates. You may find you need a CV stage to get good top-off depending on how high your rate is under CC/bulk. (I do.)

I only mentioned it indirectly in terms of charge time. Charging will be in the C/2 to C/3 range, with a possibility to go up to C. I have been trying to heed the vendor's recommendation for sub C "normal" charge rates. C/3 to C/4 seems to be the norm, so I'm pushing it a bit with C/2 to C/3, but I think still fine. I actually could get as high as 1.5C if I cranked everything up at once, but would take care not to do that.


I do see that finish charge voltage varies with charge current, so will need to sort that out. The challenge is that charge rate will vary based on which charge sources are active. I have alternators on the main (400A), AC chargers (200A), Inverter chargers (400A), and solar (small A in comparison). This is one good example of how LFP really should be done with a complete redesign of charging with control from the batteries, not from the chargers. But that's more of a challenge to do on a boat where everything is designed for charge control at the charger.


But I have been thinking about a way to address this with the PLC. All PLCs speak ModBus, and almost all of the electrical gear that I'm using speaks Modbus as well. In theory I should be able to adjust the bulk charge voltage in the chargers on the fly based on actual charge current into the batteries. Wouldn't that be slick?

Originally Posted by nebster

7) Staying away from balancing/shunting active BMS is smart. Just get good cells, set them up properly up front, and you will be fine.

8) I would engineer for multiple strings on a boat for maintenance and better fault tolerance. As a bonus, if you do have the spare shelving I asked about above in (4), you can expand your system pretty easily if you like.

I have been planning on dual strings for just this reason, but was originally thinking of a single disconnect contactor and manual on/off switches to isolate the strings. But it has finally dawned on me that separate contactors for each string, and fail safe shutdown on only the offending string would be a much better approach. That should be very doable.

Originally Posted by nebster

9) It's not obvious why 48V forces you to all-12V. You can have two converters on board, and in fact the 48V-to-24V supply will probably be more efficient. I think you should make the pack voltage choice based on topology, space and wiring constraints, and aggregate power demand. Moreso if you're going to make a custom monitoring solution (cheap 16s monitors are basically nonexistent).

You are correct. Unstated was my desire not to create a third DC power system, hence moving 24V loads to either 12V or AC. But keeping a modest 24V system is an option, and significantly reduces the scope of impact if the LFP system shuts down. This is another of my big concerns with LFP. The DC power system on a boat is super reliable, and something we really count on. An LFP shutdown, whether for real or false reasons, would be incredibly disabling on boat. If the LFP were 48V it would probably only power the Inverters, so failure would be limited to that. 24VDC would be set up the same as my 12VDC system, with an AGM and one or more AC or DC/DC chargers.


One remaining challenge is the main engine alternator. 48V alternators are a rare beast, so I would be into exotic parts for that. The up side is that the I2R losses would be a lot less.


There is a lot to like about this and I'll give it more thought. Thanks!

Originally Posted by nebster

10) Speaking of that custom PLC, let us know how we can buy one.

I'm a total nubee with PLCs, but am learning fast. I bought one a few months ago to start experimenting with a handful of possible applications for this boat. The coding of course is custom, but there are lots of industrial automation people around who understand them, and they are used quite a bit on larger vessels so not foreign to marine applications. Just uncommon on smaller boats. What I like is the standardization of components and interfaces, and the bomb proof reliability. I'm just so sick of closed standards, unreliable software, and stuff that needs to be updated (and breaks) every five minutes.


Thanks for the comments. I have enjoyed the "conversation".

[CatNewBee]

dont worry too much about cooling.
I have my winston under the berth in the original battery compartment of my Lagoon,
Winston cells, 4x 1000Ah. Even under heavy load with 7000VA (Victron 5000 + Mastervolt 2000 inverter) they pretty much are unimpressed, stay at 24°C. I have cell sensors that measure the temp, they do not get warm, just +0.1..0.2°C above the surrounding temp. Other point is ambient temp.

If you install them down bellow near the bilge they wont get much warmer than the sea. just keep other heat sources away, dont istall them next to the engines or inverter or generator.