Battery Charge efficiency and battery monitors

samuel

I just read through a brief article/study on battery charge efficiency (for a C/33 rate). Here is the link

If I'm reading this right there appears to be a giant drop off in charge efficiency once the battery reaches about 80% SOC. From 0-80% SOC the battery readily accepts about 90% of the charge current, but then as it begins to top off a battery will only accept something like half of the current being put out by the charger. This seems like a big deal! In winter when there is no sun for a stretch of several days and I use a generator for my power needs It seems very un-economical to charge the batteries up past about 80% (since the batts are cold sulfating isn't a huge concern, but the loss of 20% SOC off the top is).

Next, what about my battery monitor. If my PV/battery set up is typically operated at 70-100% SOC what on earth should I set my "charge efficiency" number at? 55%? 60%? The monitor will reset to 100% once the batts are fully charged so my main concern is an extended stay when the batts don't reach 100%.

Any thoughts on the validity of this article and the best "charge efficiency factor" to program into my battery monitor (Victron BMV-600s, CEF=.85 currently, C/33-C/50 is the usual charge rate from PV for my set-up).

Blackcherry04

Re: Charge efficiency and battery monitors

I'v been doing it for 30 Years on Batteries, on a Boat using a generator ( 20kw ) for charging and using 50 % DOD after you leave the dock at 100% you only have the 35% of useful Battery power to play with. Charging with a large generator becomes a fuel use problem. I'v done it long enough on a 12V bank ( 1300 amp hrs ) , I let it drop to around 12.2 v and Bulk back to 13.6 V and then cut it off. I try to fire up my Honda EU 2000 and top them up once a week or so. I like to run the generator around our TOU and charge during the late evening and the bank is sized to give 22 hrs of use time. We have a 3 ton Heat Pump and a 5000 btu water cooled heat pump.

I use GC2's , but I consider them as expendable and a 5 year life is the cost of having them.

niel

Re: Charge efficiency and battery monitors

samuel,
that was what turned me away from using my trimetric 2020 as it is a guessing game on charging efficiency. even if somebody gave me a good figure that works for them with the same battery, the dod and battery age would be different and change the efficiency figure. guessing is not a good thing to do as too high or low could be detrimental to watching usage and battery soc and can cause wider false figures down the road.

of course it is certainly better than nothing, but one does have to keep in mind the accuracy of such a meter isn't exacting and even if by chance you got the right number it will most certainly change with age or different dod points so don't swear by it.

Vic

Re: Charge efficiency and battery monitors

samuel,

Did not read the article, but sounds about right to me.

BUT, this situation goes with the territory. And, with your Honda EU3000, this in not such a big problem.

Your batteries need to be fully charged often. Opinions differ on just how often, but better to err on the side of a bit too often, vs not enough.

RE batt monitors, could not agree more. Setting the efficency is a problem, and this value will change over time. But, at least with FLAs, one can acutally measure the SOC, which is a good thing.

JMHO, Vic

westbranch

Re: Charge efficiency and battery monitors

Samuel, have you considered a E1000i to use rather than the E3000, when the 80% SoC is reached? they will loaf all day on 2 litres of gas. this seemed to be one of your concerns.

samuel

Re: Charge efficiency and battery monitors

Usually the generator will take the batts up to about 80% in winter before leaving the cabin. Then the hope is that the PV panels will top things off while we're gone.

The % SOC readout on the Batt monitor seems to be super accurate ONLY if I'm starting from a full charge on the batts otherwise it's not worth beans for accuracy - way too much guessing over charge efficiency. At least now I'm beginning to understand why that is.

I think I'll set the charge efficiency around 55% for now and see if things start to make sense. On the upside, the Victron batt monitor reports volts to the hundredth (eg. 12.45V). Because of that accuracy, I put together a SOC/voltage chart based off of a 20 amp DC load and Specific Gravity readings to use until I figured out how to get the batt monitor to work the way I want to. I attached it below.

Attachment not found.

Looks like the chart might be here to stay.

westbranch

Re: Charge efficiency and battery monitors

Samuel, good link, finally read to the bottom, where I found the clincher:

Charge efficiencies at 90% SOC and greater were measured at less than 50% for the battery tested here,
requiring a PV array that supplies more than twice the energy that the load consumes for a full recovery charge.
Many batteries in PV systems never reach a full state of charge, resulting in a slow battery capacity loss from
stratification and sulfation over the life of the battery.

So the rule of thumb: ~120% input, relative to the output, is needed to achieve a full charge; is not even close, particularly with light discharges.

Windsun

Re: Charge efficiency and battery monitors

This is an important but often overllooked topic so I have stickied the thread. Please keep on topic, off topic replies will be moved or removed.

SolaRevolution

Re: Battery Charge efficiency and battery monitors

Some battery monitor manufacturer's recommendations for setting charge efficiency:

The Magnum ME-BMK manual (2010) Page 24. The last 2 paragraphs of section 5.2
The Trimetric 2020 manual. Bottom paragraph of page 6.

The charging efficiency is a moving target due to many factors. You need do the initial setup and guess at a charge efficiency % to begin with, then after a few cycles check the AH I/O meter in the BMK or the AMP-HR "AH" for the T-2020. If the reading is negative, increase the efficiency. If the reading is positive decrease the efficiency. This should be repeated from time to time.

This will not make the SOC perfectly accurate. It will still be thrown off by imperfect "charged set point" volts and amps and by the effects of tempurature variations. The T-2020 does not have tempurature compensation. Hmm. I don't know about the Magnum BMK or the Outback Flexnet DC. I'll have to look that up.


SOC monitors are to be eyed with suspicion. They are making a suggestion for what the SOC% may be. The operator should understand the relationship of amps, volts and tempurature and be able to tell if the SOC% reading seems appropriate. Unfortunately not everyone who owns a battery powerd system can or will grasp it.

petertearai

Re: Battery Charge efficiency and battery monitors

If that is the case , then would not the battery moniter algorithim take the state of charge into account? ( maybe Victron could answer that one?) Still the moniter is useful guide, especially when friends or kids go up to the holiday home.
I suppose as my battery's age I should reduce the efficiency setting.

vtmaps

Re: Battery Charge efficiency and battery monitors

In flooded batteries gassing is a good and necessary thing, but it costs energy and that contributes to the low charge efficiency.

Here is a link to an interesting article in Home Power about the research at Sandia Labs:
homepower.com/view/?file=HP89_pg120_IPP
The author of the HP article recommends that you undercharge your batteries to save money. What he is saying is that because of the expense of achieving full charge with a generator (low charge efficiency during absorb charge) it is cost effective to just let your batteries remain a bit undercharged.

The state of charge (SOC) of a battery is most accurately measured with a hydrometer, and is indicated as specific gravity (SG). Most RE users rely on amp-hour meters to provide convenient (although slightly less accurate) battery SOC information. During the Sandia tests, full batteries had a SG in the range of 1.290. The long, engine generator run times needed to achieve this SG translate into dollars and pollution (both audio and atmospheric). Perhaps there is a “middle way” that preserves the lifetime of the batteries while reducing the time and cost of engine generator finish charging.
<snip>
Perhaps batteries should be rated based on their application. For instance, a battery used in a standby application (such as utility backup system with grid recharging) might specify a full charge SG of 1.290. The same battery used in an application that regularly cycles the batteries (such as a PV system with engine generator backup) might have a recommended SG of 1.250 to be considered full.

--vtMaps

samuel

Re: Battery Charge efficiency and battery monitors

Looks like I might just bulk in winter and kill the generator when absorb is reached.

It's a bit amazing that after two years of research I'm just now learning about this phenomenon. Thinking about it, I'm not sure how a battery monitor could account for this… there is Peukert's constant for discharging, but I'm not aware of a constant that describes charging, especially since charging rate and battery SOC would both play a part. Then add in the variability of solar. Best to set a conservative charge efficiency based on how far down the batteries are commonly drawn down to and try to get them to full charge (from PV power) when possible to re-zero the battery monitor.

Windsun

Re: Charge efficiency and battery monitors

The usual and/or typical way is to charge with a generator to 90 or 95% or so, then let the panels finish the final charge. The obvious catch is that you need the panels to be producing enough excess power to both charge the batteries and supply and power usage going on at the same time. Getting that last 5-10% from a generator can take as long as the first 90%.

nsaspook

Re: Battery Charge efficiency and battery monitors

samuel wrote: »

Looks like I might just bulk in winter and kill the generator when absorb is reached.

It's a bit amazing that after two years of research I'm just now learning about this phenomenon. Thinking about it, I'm not sure how a battery monitor could account for this… there is Peukert's constant for discharging, but I'm not aware of a constant that describes charging, especially since charging rate and battery SOC would both play a part. Then add in the variability of solar. Best to set a conservative charge efficiency based on how far down the batteries are commonly drawn down to and try to get them to full charge (from PV power) when possible to re-zero the battery monitor.

You can't use a set value of CEF that's an average of the changes that happens during the entire cycle and still track partial recharges. It's possible to do but you have to use the information from several cycles to train an algorithm that models the battery using a simple expert system (hard, because I'm not an expert at battery physics but I am learning). The expert system has to know the important properties of battery physics and monitor important environmental factors like temperature. I use a Ah/Peukert/dynamic SOC/CEF factor expert system for my DIY monitor. The dynamic SOC/CEF factors are driven by data collected every 30 seconds and are used by rules that also create a weighted number that blends all the data into a overall quality number that on my system decides what battery from 4 banks needs the PV power most.

I can see why most commerical battery monitor system depend on full charge resets because with so many things to track and calculate it's super easy to drift off by huge amounts quickly.
System data plot from my small off-grid developement system.
Debug screen system data: http://farm8.staticflickr.com/7141/6848168603_751c114d7b_b.jpg

SolaRevolution

Re: Charge efficiency and battery monitors

Windsun wrote: »

The usual and/or typical way is to charge with a generator to 90 or 95% or so, then let the panels finish the final charge.

Be careful to watch your charge controllers when doing this. Running the generator in the AM can make your charge controllers go into float for the rest of the day. You may need to "force bulk" or restart the charge controllers. A lot depends on your array/battery relative size, charge controller settings etc. I've seen Outbacks read "BatFull" after a half hour of generator runtime with batteries still in the 85% range.

telljf

Re: Charge efficiency and battery monitors

The battery monitor needs a good reference to start with, and it needs to be revised often. What I do is remove the load and charge the battery till it reaches float, keep it in that state for 2 days, verify SG of all cells, ensure that they are all at the same state specified by the manufacturer for a fully charged battery. Then reset the net AH reading of the battery monitor to zero (I have a Nasa BM1 battery monitor). So that is the new reference. The PV should be able to replace and add an additional 20% of what is drawn every night so that the net AH reading of the monitor is always in the positive side at the end of the day. Even with all these, the reference will slip for a host of reasons, and hence needs to be checked/updated whenever possible.

nsaspook

Re: Charge efficiency and battery monitors

telljf wrote: »

The battery monitor needs a good reference to start with, and it needs to be revised often. What I do is remove the load and charge the battery till it reaches float, keep it in that state for 2 days, verify SG of all cells, ensure that they are all at the same state specified by the manufacturer for a fully charged battery. Then reset the net AH reading of the battery monitor to zero (I have a Nasa BM1 battery monitor). So that is the new reference. The PV should be able to replace and add an additional 20% of what is drawn every night so that the net AH reading of the monitor is always in the positive side at the end of the day. Even with all these, the reference will slip for a host of reasons, and hence needs to be checked/updated whenever possible.

A major problem with most battery monitors is they assume that "Charge" is conserved under charging and discharging like in normal EM theory with a "Capacity model" and "Voltage model" that works well for constant charge and discharge with static correction factors. The intermittent charge/discharge of PV systems causes a complex relationship with the "rate capacity effect" and "recovery effect" that makes the L = C/I for lifetime invalid for "Electrical-Circuit" models. (lifetime (L),capacity (C),discharge current (I))

With simple "Analytical" models the "rate capacity effect" from "average" discharge current is handled mainly by adjustment from the Peukert's factor in most good monitors but few take into account the "recovery effect" during battery idle times. The errors from simple Peukert's calculations can be huge (greater than 25%) with non-linear loads.

Advanced models like the "Kinetic Battery Model" mimic chemical processes that models a battery as two internal energy sources that flow between each other to model discharging and the recovery effects. (available-charge and bound-charge in large lead-acid storage batteries)

Most of technical documentation is from this.
http://doc.utwente.nl/64556/1/BatteryRep4.pdf

boB

Re: Charge efficiency and battery monitors

There are two (2) efficiencies to deal with when charging batteries. Amp-Hour efficiency and Energy efficiency.
Amp-Hour efficiency is pretty high, say, in the 90% to 95% range normally and Energy efficiency is much lower.

It takes more energy to get that last 10% or so of charge in Absorb and especially inefficient if you are running your
generator (if you can use renewables instead because of the lower power requirement for the last part), because
the generator is lightly loaded and very inefficient at running just lightly loaded.

I think that your 55% efficiency number is energy efficiency. Amp-Hour efficiency should never be quite that low
I don't think and BMK should not have efficiency number in that range if I am remember its setup correctly.

boB

nsaspook

Re: Charge efficiency and battery monitors

boB wrote: »

There are two (2) efficiencies to deal with when charging batteries. Amp-Hour efficiency and Energy efficiency.
Amp-Hour efficiency is pretty high, say, in the 90% to 95% range normally and Energy efficiency is much lower.

It takes more energy to get that last 10% or so of charge in Absorb and especially inefficient if you are running your
generator (if you can use renewables instead because of the lower power requirement for the last part), because
the generator is lightly loaded and very inefficient at running just lightly loaded.

I think that your 55% efficiency number is energy efficiency. Amp-Hour efficiency should never be quite that low
I don't think and BMK should not have efficiency number in that range if I am remember its setup correctly.

boB

The actual Ah efficiency goes down to that low number as more energy is needed to complete the conversion from electrical energy to chemical compounds via the movement of ions. It becomes less efficient when most of the (easy) surface area has already been converted and the reaction needs to happen deeper in the metal electrodes to complete the charge cycle.

The "Kinetic Battery Model" explains it like this.

In the model the battery charge is distributed
over two wells: the available-charge well and the bound-charge well (cf. Figure 6). The
available charge well supplies electrons directly to the load, whereas the bound-charge
well supplies electrons only to the available-charge well.

When you charge the battery at some point the first "well" is full so you have to push charge through it to get charge to the second "well". Most of the lost energy from pushing charge past the first well is turned to heat and gas from electrolysis.

BillBlake

Re: Charge efficiency and battery monitors

SolaRevolution wrote: »

Be careful to watch your charge controllers when doing this. Running the generator in the AM can make your charge controllers go into float for the rest of the day. You may need to "force bulk" or restart the charge controllers. A lot depends on your array/battery relative size, charge controller settings etc. I've seen Outbacks read "BatFull" after a half hour of generator runtime with batteries still in the 85% range.

The information you supply is some Key stuff. Thanks.
Also thanks to --vtMaps for pointing out this thread.

The Price of Meat (fuel) Keeps Going Up

but Your (Solar Panels) Keeps Going Down :-)

What are we going to do about it ??


Bill Blake