Setting up a Battery Monitor - Peukert's Law?

wildebus

Full Member

Messages
6,217
For non-Lithium batteries, it is generally understood and accepted that the higher a current you draw for a given battery, the lower the available capacity is.
This is why you will usually see different capacities quoted e.g. For my Leoch Extreme AGMs, the C100 Capacity is 110Ah, but the C20 Capacity is 95Ah (C100 means the battery will be fully discharged in 100 Hour at a constant current - that current being 1.1A (1.1A x 100Hrs = 110Ah), but 95Ah if fully discharged in 20 hours - so a current of 4.75A (4.75A x 20Hrs = 95Ah). And if you discharge at higher rates the capacity reduction continues (some battery info includes capacity values for C10, C5 and even C2).
This 'effect' is first noted by a scientist called Wilhelm Peukert over 100 years ago. A starting point to read more about this could be seen here - https://en.wikipedia.org/wiki/Peukert's_law
When setting up a monitor like a Victron BMV, you have a entry setting to enter a Peukert Factor to accurately calculate the Battery SOC (State of Charge) to recognise the reduced capacity encountered when drawing a high current.

Now I have been doing some testing of battery capacities against usage to get a better practical understanding of Peukert's Law.
And what I am going to say will tend to go against the commonly understood principle of Peukert's Law and I am coming to the conclusion that it is not being understood and applied correctly in the SOC algorithms. And the following is the reason ...


So my Battery Bank is 4 x Leoch Extreme AGM Batteries. Capacity at C100 is quoted as 110Ah each; at C20 they are 95Ah each. It is the C20 Rating that Victron advise you to enter when setting up the monitor (so my BMV-712 is setup with a Battery Bank of 380Ah). I left the Peukert Factor as the default setting.
At a C100 Rating the battery bank is nominally 5280Wh Capacity (4 x 110Ah x 12V) and at C20 it is 4560Wh.

I fully charged the battery bank and then set the water heater on. When it is on, it draws around 190A through the Inverter - that is equal to a nominal C2 rating on my Battery Bank.
I have it setup to be a cycle of 1 minute on, 2 minutes off. This chart shows the on-off cycle and how the SOC is reducing as power is being drawn from the battery.

Extreme SOC Drop
by David, on Flickr
The Heater was on for 1 hour - this works out to be a power usage of ~700W. The SOC has gone from 100% down to about 65%.
Doing the maths, drawing 700W in one with a 35% battery capacity use means the battery bank capacity would be just 2000W - or 167Ah! That is a massive drop and taken on face value would mean using a high power inverter would be super-inefficient.
And now that high power draw has been removed, the battery (on the C20 Setting the BMV requests) the remaining capacity is 2964Wh (4560Wh*65%) giving a total capacity of 3664Wh rather than 4560Wh.

But is that right? where has the rest of the capacity (900Wh, or 75Ah) gone? that is a lot of power to just disappear in 1 hour isn't it?!

I don't actually think it is gone anywhere and the way Peukert's Law is applied in Battery Monitors is flawed. You will see fom the chart above the heater came on for a couple more minutes twice more, plus there were other small general loads. So around lunchtime today when I looked at the Battery Charge I got this from the Victron Monitor:

Screenshot_20190101-114556
by David, on Flickr
So I am below 50% and in the "danger" level!
I happen to have another SOC Monitor installed that I am testing. On this one I set the same Initial Battery Bank size (380Ah) but there is no Peukert Factor applied in this monitor. This is what this other monitor is saying

IMG_20190101_114514
by David, on Flickr

The Monitor with the Peukert Factor is 49% and the one without is 62% - which is correct?
Well, we can use a fallback method to check .... As can be seen on the monitors, there is a current draw of 4.1-4.5A (depending on which monitor), and the voltage is sitting at 12.2-2.28V (again depending on Monitor).
Using a traditional Battery State Chart (ideally when there is no load or charge on the battery for a few hours) you can correlate the Voltage Reading to the Charge left.
There are a few variations on these, but they are all pretty similar. Below is an pretty good example.

Battery Voltage Chart
by David, on Flickr
Using the Voltages shown in the monitors and applying them to this chart, the SOC is really between 60-65%, which makes a lot more sense then the 49% the BMV tells us and the Monitor that ignores the Peukert effect would seem to be the most accurate!

There is a guy on Youtube that makes some good videos about Batteries and Inverters (he's a bit of a geek but for geeks like me, it is interesting anyway :) ) and he has similar thoughts about Peukert and has done some testing for his own curiosity.


I think I will be removing the Peukert Factor from my BMV settings (or more precisely just making it 1.0) as - IMO - it is giving misleading information as regards the SOC, at least for my style of use that involves very short bursts of very high current. I will keep with the C20 Rated entry (380Ah) rather than the larger C100 (440Ah) value (the difference actually being more than the usuable capacity of an entire battery!) as a conservative value and as my usual current draw is in the C20 range.

I would be interested in others opinions, thoughts and experiences (but based on practical data rather than just a link to a internet resource that just says "this is what it is" without providing empirical data).
 
Y'know, there are times when I really feel blessed to be a thicko !!!!!!!!!!!!!!!!! A battery either works or it doesn't, and if it's the latter, mend it with a new one
 
just depends on what interests you and 'floats your boat' :)
Ask me about Motorhome gas setups and you will get what I know on the back of a postage stamp :D
 
I'd agree that Peukert applied in a barttery monitor while on load seems a flaky concept. You'd have to remove the load and then check rest voltage after a few hours.

I've also found that trying to calculate what time I can get under high loads like 500 to 750w even using Peukert estimations always seem to be innacurate and on the high side. For example I'd calculate I could use a 500w load for half an hour yet the inverter would alarm after 15 or 20 mins.

I generally attributed this to reduced capacity on batteries due to cycling and them not being charged to the capacity they were capable of.

For a wet battery that is not in optimum condition or has a slightly dodgy cell, the Peukert factor seems to multiply. There's not much research on this, but intuitively it makes sense. And, let's face it, the batteries most of us own have been cycled 100 times, maybe sulphated to an unknown degree and in unknown places.
 
The idea of the Peukert Factor on a Monitor is that the SOC is dynamically calculated BASED on the load the battery is on. You couldn't use Peukert in any other way.
Removing all loads (and charges) for a few hours and then checking the voltage will give you a good idea of the capacity remaining in the battery (but of course, it won't tell you what that capacity actually is). That is how you would use a Voltmeter of course, but that is not a battery Monitor.

Your inverter alarm will likely not be a battery low warning but a VOLTAGE low warning - very different things.
 
For non-Lithium batteries, it is generally understood and accepted that the higher a current you draw for a given battery, the lower the available capacity is.
This is why you will usually see different capacities quoted e.g. For my Leoch Extreme AGMs, the C100 Capacity is 110Ah, but the C20 Capacity is 95Ah (C100 means the battery will be fully discharged in 100 Hour at a constant current - that current being 1.1A (1.1A x 100Hrs = 110Ah), but 95Ah if fully discharged in 20 hours - so a current of 4.75A (4.75A x 20Hrs = 95Ah). And if you discharge at higher rates the capacity reduction continues (some battery info includes capacity values for C10, C5 and even C2).
This 'effect' is first noted by a scientist called Wilhelm Peukert over 100 years ago. A starting point to read more about this could be seen here - https://en.wikipedia.org/wiki/Peukert's_law
When setting up a monitor like a Victron BMV, you have a entry setting to enter a Peukert Factor to accurately calculate the Battery SOC (State of Charge) to recognise the reduced capacity encountered when drawing a high current.

Now I have been doing some testing of battery capacities against usage to get a better practical understanding of Peukert's Law.
And what I am going to say will tend to go against the commonly understood principle of Peukert's Law and I am coming to the conclusion that it is not being understood and applied correctly in the SOC algorithms. And the following is the reason ...


So my Battery Bank is 4 x Leoch Extreme AGM Batteries. Capacity at C100 is quoted as 110Ah each; at C20 they are 95Ah each. It is the C20 Rating that Victron advise you to enter when setting up the monitor (so my BMV-712 is setup with a Battery Bank of 380Ah). I left the Peukert Factor as the default setting.
At a C100 Rating the battery bank is nominally 5280Wh Capacity (4 x 110Ah x 12V) and at C20 it is 4560Wh.

I fully charged the battery bank and then set the water heater on. When it is on, it draws around 190A through the Inverter - that is equal to a nominal C2 rating on my Battery Bank.
I have it setup to be a cycle of 1 minute on, 2 minutes off. This chart shows the on-off cycle and how the SOC is reducing as power is being drawn from the battery.

Extreme SOC Drop
by David, on Flickr
The Heater was on for 1 hour - this works out to be a power usage of ~700W. The SOC has gone from 100% down to about 65%.
Doing the maths, drawing 700W in one with a 35% battery capacity use means the battery bank capacity would be just 2000W - or 167Ah! That is a massive drop and taken on face value would mean using a high power inverter would be super-inefficient.
And now that high power draw has been removed, the battery (on the C20 Setting the BMV requests) the remaining capacity is 2964Wh (4560Wh*65%) giving a total capacity of 3664Wh rather than 4560Wh.

But is that right? where has the rest of the capacity (900Wh, or 75Ah) gone? that is a lot of power to just disappear in 1 hour isn't it?!

I don't actually think it is gone anywhere and the way Peukert's Law is applied in Battery Monitors is flawed. You will see fom the chart above the heater came on for a couple more minutes twice more, plus there were other small general loads. So around lunchtime today when I looked at the Battery Charge I got this from the Victron Monitor:

Screenshot_20190101-114556
by David, on Flickr
So I am below 50% and in the "danger" level!
I happen to have another SOC Monitor installed that I am testing. On this one I set the same Initial Battery Bank size (380Ah) but there is no Peukert Factor applied in this monitor. This is what this other monitor is saying

IMG_20190101_114514
by David, on Flickr

The Monitor with the Peukert Factor is 49% and the one without is 62% - which is correct?
Well, we can use a fallback method to check .... As can be seen on the monitors, there is a current draw of 4.1-4.5A (depending on which monitor), and the voltage is sitting at 12.2-2.28V (again depending on Monitor).
Using a traditional Battery State Chart (ideally when there is no load or charge on the battery for a few hours) you can correlate the Voltage Reading to the Charge left.
There are a few variations on these, but they are all pretty similar. Below is an pretty good example.

Battery Voltage Chart
by David, on Flickr
Using the Voltages shown in the monitors and applying them to this chart, the SOC is really between 60-65%, which makes a lot more sense then the 49% the BMV tells us and the Monitor that ignores the Peukert effect would seem to be the most accurate!

There is a guy on Youtube that makes some good videos about Batteries and Inverters (he's a bit of a geek but for geeks like me, it is interesting anyway :) ) and he has similar thoughts about Peukert and has done some testing for his own curiosity.


I think I will be removing the Peukert Factor from my BMV settings (or more precisely just making it 1.0) as - IMO - it is giving misleading information as regards the SOC, at least for my style of use that involves very short bursts of very high current. I will keep with the C20 Rated entry (380Ah) rather than the larger C100 (440Ah) value (the difference actually being more than the usuable capacity of an entire battery!) as a conservative value and as my usual current draw is in the C20 range.

I would be interested in others opinions, thoughts and experiences (but based on practical data rather than just a link to a internet resource that just says "this is what it is" without providing empirical data).

You must really have a sad life to post this on New Year's day. I agree with Simon, but we are also sad for reading it. My excuse is that I'm waiting for my venison. Moira's out with the gun!
All the best for 2019, and respect for knowing what Airfix used to do.
See you at Moffat.
Gordon
 
I'm watching "Stalag 17" as well on Retro Movies - super sad life :D
 
just depends on what interests you and 'floats your boat' :)
Ask me about Motorhome gas setups and you will get what I know on the back of a postage stamp :D
Mind you thats probably twice what I know.
Honestly though, it's thanks to folks like you that the askers of truly stoopid questions are really grateful, and as long as your shoulders are broad enough to avoid folks like me taking the wee, just keep on being you.
Thanks from a thicko
 
I see you have assigned the value of "1.0" to the troublesome variable. Something like Einstein's problem with the Cosmological Constant?
 
Sorry Dave,
I’m surprised nobody has said this before. Coming from me, you’ll know there’s no spite intended..
I read it and thought, oh! Peukert it. ?

Have a great year my friend.
 
I see you have assigned the value of "1.0" to the troublesome variable. Something like Einstein's problem with the Cosmological Constant?
I had to look that one up :)
If Einstein was willing to change his mind about something, surely anyone else should be willing to as well? (and he changed his view due to a new discovery which itself was superceeded 70 years later - and this was on just a simple little subject like the Universe, not complex stuff like campervan battery technology :LOL: )

Peukert came up with his observation in the 19th century, and technology has come a long way since then of course, so maybe his effect is temporary, for example, on current batteries? :geek:

Victron have a Peukert Calculator - https://www.victronenergy.com/upload/software/PeukertCalculator.exe - which can be used to work out what to enter in their BMV setup. I didn't change from the default of 1.25. Their Calculator works out the exponent to be 1.1 for my particular battery (so much closer to the 1.0 'no effect' value) so to be fair to the BMV 'incorrect' SOC reading, my setting was off (although doesn't change my doubts about how Peukert is generally understood and applied).
I changed it to 1.0 yesterday. maybe I will set it to 1.1 and repeat the test and see how the SOC changes? That might be interesting :unsure:
The slightly annoying thing is that when I changed it yesterday, the SOC resets to zero and so the "full cycle" counter clicks up - so it looks like the batteries have been totally and fully drained once (not very good for them if it were true). Chances are same will happen again when I change it again :cautious:


Of course, it should be said that all this playing around with the settings in the monitor makes no difference to how the batteries are actually working or charging, so is not affecting them in anyway (but of course it does help the user decide if they should carry on using the batteries without recharging them) :)
 
Interesting test, thanks for sharing it. However, I am not sure I agree with your conclusion. The battery state of charge table you are using appears to be for standard lead acid batteries not AGM, which have higher voltages. I have seen a table for AGM batteries that shows 50% discharge when the voltage drops to 12.3V, in which case the the 12.28V and 49% state of charge shown by the Victron look pretty much spot on. If it were me I would be very much inclined to trust the Victron with the Peukerts exponent set at 1.25, Victron kit is not cheap but it is well engineered. Also I think battery voltage is at best a rough guide to state of charge because battery chemistries and condition vary, even within the same basic types.
 
I had missed your more recent post and can see that 1.25 might be high, I will be interested to see your results at 1.1.
 
I do not wish to be rude..
But this is totally OTT for most of us !
I know you are "Electric only" so of real interest to you.
I monitor my LB voltage and late evening it some times gets as low as 12.7
I am happy with this..
If I turn heating off when it is dull at night the Voltage may go up to 12.8 as the battery recovers
Good enough for me...why overegg it ?

Winter use is different !
 
Interesting test, thanks for sharing it. However, I am not sure I agree with your conclusion. The battery state of charge table you are using appears to be for standard lead acid batteries not AGM, which have higher voltages. I have seen a table for AGM batteries that shows 50% discharge when the voltage drops to 12.3V, in which case the the 12.28V and 49% state of charge shown by the Victron look pretty much spot on. If it were me I would be very much inclined to trust the Victron with the Peukerts exponent set at 1.25, Victron kit is not cheap but it is well engineered. Also I think battery voltage is at best a rough guide to state of charge because battery chemistries and condition vary, even within the same basic types.
I did see charts for AGM and they didn't really vary that significantly (except I did see one that had 12.8V = 90% which I thought rather dubious?). There are so many variations of these charts who really has a definitive one? I tend to eliminate the extremes and go with the average
I am not saying Peukert is wrong, far from it, but that how what he discovered - the faster you discharge a battery, the less available capacity you have - is not really being applied right as the capacity (energy) cannot just disappear.
(There are some from heat, yes, and also losses in cables etc, (but that is taken account of, for example this "2000 W" heater I am using actually draws more due to those losses, but that is included within the calculations and the monitor sees a draw of 2200W or so))
It is just that it is not available at that time. What the SOC in the Meter is claiming is the capacity is just gone though and does not allow for any recovery and reaction time. A bit like if you took a voltage reading when the battery was under load and used that in a voltage chart and didn't let the battery time to recover to check.
A battery monitor is still an essential tool for a setup that is very battery-dependant, but like most things, it needs to be viewed with some 'adaptability'
 
I do not wish to be rude..
But this is totally OTT for most of us !
I know you are "Electric only" so of real interest to you.
I monitor my LB voltage and late evening it some times gets as low as 12.7
I am happy with this..
If I turn heating off when it is dull at night the Voltage may go up to 12.8 as the battery recovers
Good enough for me...why overegg it ?

Winter use is different !
As low as 12.7V? battery recovers to 12.8V when the heating is off? I would be inclined to check your voltmeter against another one!

As far as OTT for most of us, maybe, but more and more people are looking at going Electric (just look at the 'Electric Car' topic!) and to get the optimum use for a Battery Bank, you have to know what is available.
This is posted in the Forums '... Technical Section' also.
Simple answer ... Not Interested, Don't Read? (I take that approach in the "Jokes" section as I am a boring old fart)
 
I had missed your more recent post and can see that 1.25 might be high, I will be interested to see your results at 1.1.
I am doing that right now :)
 
As low as 12.7V? battery recovers to 12.8V when the heating is off? I would be inclined to check your voltmeter against another one!

As far as OTT for most of us, maybe, but more and more people are looking at going Electric (just look at the 'Electric Car' topic!) and to get the optimum use for a Battery Bank, you have to know what is available.
This is posted in the Forums '... Technical Section' also.
Simple answer ... Not Interested, Don't Read? (I take that approach in the "Jokes" section as I am a boring old fart)

I posted that I did not wish to be rude... Your reply borders on that !
I have read you other posts with interest....( and this one ) it is a shame you do not appreciate other opinions !
I did not contradict anything you posted I just offered my take on it
I have used this system to monitor my LB for the lats 3 years without problem
Ok I have not posted detailed voltages/monitoring etc but IT WORKS !.
Why do you suggest the reading is inaccurate when you have NO knowledge of my setup.

PS My on-board system allows me to view the status of both my batteries with a "Lozenge type display"
This ranges from Red (=dead) to orange then amber and then green and super green.
My additional "Digital voltmeter" simply confirms the on-board display regarding to my LB
 
Last edited:

Users who viewed this discussion (Total:0)

Back
Top