[PStechPaul]

Generally the Peukert effect is considered during discharge of lead-acid batteries, but for charging there is a similar effect. I found a detailed explanation of this phenomenon here:
http://www.smartgauge.co.uk/peukert_chg.html

and in more detail:
http://www.smartgauge.co.uk/peukert_depth.html

Since I plan to use SLAs for my tractor project(s), I want to understand fully what to expect during discharge as well as charge cycles, and to be able to calculate accurately the state of charge. I also want to be able to extend the life of my batteries as much as possible. I would consider using one of the smartgauge products, but the cost seems prohibitive at 150 to 300 British Pounds which is about $225 to $450, and I would need one for each of my proposed 20+ batteries or a way to switch from one to another.
http://www.power-store.com/?id=273

This may be of less interest to those who use LiFePO4 cells which have little to no Peukert effect, but they still have significant internal resistance which may affect the accuracy of SOC gauges using coulomb counting or A-h or W-h measurement, especially under conditions of high current charge and discharge. So I will try to propose and discuss a design which may be used for either lead-acid or lithium technologies as well as perhaps NiMH. Since it will be microcomputer based it should readily adapt to other technologies.

As discussed in the articles linked above, a simple model for a battery consists of a "pure" energy storage element and a series resistance. More accurately, the series resistance changes according to the current, similar to the tungsten filament of an incandescent lamp, and thus the Peukert effect is exponential rather than linear. The actual Peukert constant may be accurate for a fresh battery, but it will change as it ages. And the state of charge or battery health may also be determined by performing a load test and measuring voltage and current under various conditions.

So, a comprehensive and accurate means of determining SOC and end-of-life may involve accurate voltage, current, and time measurement, as well as probably temperature, and a history of charge/discharge cycles. It would also probably need a way to perform certain tests on the battery from time to time, and be able to control charging and discharging. Thus I would design this as a 12V BMS which would be semi-permanently attached to each battery (over its lifetime), and perform its functions as part of a series string of 20 to 50 batteries (240 to 600 VDC).

I'll add further details as this project proceeds, and I welcome comments and discussion so as to make this a useful product, especially for anyone who still uses or plans to use lead-acid batteries.

Thanks!

 

[EVfun]

There is not really any Peukert's effect when charging. You have to replace the actual amp hours used plus 5 to 10 percent to make up for the charge efficiency being less than 100% for lead acid. If you amp hour counter is displaying a Peukert's adjusted amp hours available or percentage SOC then it is indicating you have removed more amp hours than you have actually removed. You will need some type of correcting factor to scale up the amp hours returned during charge.

I feel it would be better to determine the actual amp hours you have available at EV discharge rates and enter than for the battery capacity and set the Peukert's exponent to 1.0 (no Peukert correction.) This will typically be only 50 to 60 percent of the nameplate capacity. The meter will scale your SOC based on that part of the capacity that is actually available. Since it will be showing actual amp hours you don't have to do any number fudging when you charge.

 

[Siwastaja]

they [LiFePO4 cells] still have significant internal resistance which may affect the accuracy of SOC gauges using coulomb counting

Why would any internal (series) resistance affect coulomb counting?

 

[PStechPaul]

Maybe LiFePO4 cells are different from other chemistries in this regard, but the portion of current that results in heating does not really contribute to the actual ampere-hours stored in the battery. I know that NiMH cells continue to conduct current at full charge and the result is a significant rise of temperature, and in lead-acid cells it results in heat and gassing.

The effect is probably much less for lithium cells but I think a 10 Ah cell charged at 100 amps for 0.1 hour probably gets fairly hot and will have less capacity than the same cell charged for 10 hours at 1 ampere. If the internal resistance is 0.01 ohms (and no temperature coefficient), there will be 100 watts of heat at 100 amps, while only 10 mW of heat at 1 amp. And the energy loss will be 10 watt-hours compared to 0.1 watt-hours at the slower rate.

Rate of charge may only affect efficiency, which may not be a big deal, but there is definitely an effect on the life of the cells comparing charging rates of 10C, 1C, and 0.1C. The actual state of charge may not change much in these cases unless the cell is overcharged, in which case I think it will overheat and deteriorate rapidly, and at that extreme, coulomb counting will certainly be inaccurate.