Navitron Renewable Energy and Sustainability Forum

Energy/Electricity Storage and Use/Grid Connection => Off-Grid, Batteries & Inverters => Topic started by: bautsche on October 08, 2013, 12:27:10 PM



Title: Sunny Island 5048 and Battery Fan
Post by: bautsche on October 08, 2013, 12:27:10 PM
Hi.
I've recently added some solar panels to my off-grid setup. The issue being that if the Sunny Island inverter decides to start the generator for a boot run during the morning and it then gets sunny, it doesn't realise that it should turn the generator off because the solar panels (direct attached to the batteries, SI has a current sensor) are charging the batteries on their own. The result is that the batteries overheat and the SI turns itself off (which of course only makes the issue worse as there is no consumers for the power generated by the solar panels now).
So, my idea was to use the battery fan function on the Sunny Island to turn the generator off if the batteries get too warm.
I have set option 241#02 (Rly2Op) to "BatFan" making the relay 2 turn on and off depending on the battery temperature.
However, my manual says that I should have option 221#07 "BatFanTmpStr" (presumably battery fan temperature start) to set the temperature at which relay 2 should activate.
Only, my options under 221# only go as far as 06.
I've got a Sunny Island SI 5048E 2.1 vers. 4.202.

What am I missing? Do I need to update the firmware?

Any pointers greatly appreciated.
Thanks.
Eric


Title: Re: Sunny Island 5048 and Battery Fan
Post by: Tinbum on October 08, 2013, 01:00:44 PM
I think you have some settings wrong somewhere.
The batteries should not get too warm as the sunny island should reduce the charging so they don't get too warm.
I would look at the settings for the generator start and stop.


Title: Re: Sunny Island 5048 and Battery Fan
Post by: stephendv on October 08, 2013, 01:39:24 PM
You shouldn't need to do this.  I remember testing this with generator + midnite classic controller and the sunny island correctly reduced the battery current so that it stays under the max charging amps configured for the batteries.

My guess is that you have the batteries max charging amps set incorrectly. You'll find it under the Battery settings.

Or the current shunt isn't installed properly and the SI isn't reading the current going direct to the batteries correctly.



Title: Re: Sunny Island 5048 and Battery Fan
Post by: offthegridandy on October 08, 2013, 01:54:07 PM
Eric,

I don't know anything about Sunny Island using Outback inverter  and charge controller myself; but I have had the same situation in that the inverter charger will start a charge cycle when the sun is shining.  I have manage to generally avoid this by using the "quiet time settings".  That is to say I set the inverter charger so it is set not to charge during daylight hours.  It will only start the generator during the day for short runs to support a load larger than the batteries will cope with or it can still start a full charge cycle if the battery volts drops below pre set limit for  an  extended period.

I don't know if this set up is possible on SI but it works well for me.   NB If I know I will be making large demands on the batteries ie using the welder I always start the genny manually before I start to avoid the large drain.

Andy


Title: Re: Sunny Island 5048 and Battery Fan
Post by: bautsche on October 08, 2013, 04:40:35 PM
Thanks all for your suggestions.

Unfortunately, the idea of setting the quiet time isn't going to work for me, as I need the generator even on a not-so-bright day at the moment... :-( I'll need more solar, really.

The battery charge current is set to 436A (for a 720AH) battery. That's how the installer has set it, I have no idea if that's reasonable for the battery.
That translates to approx. 21kW, given that the generator will not produce more than 9kW and the solar array is 3.4kWp, the max charge current those can produce is 248A. Looking at a graph of battery temperature and battery current, at the time of crisis (where the temperature kept rising until shutdown) the battery current was approx. 34A and had been for a while. During a normal charge run, the current generally hovers around 65A without causing an issue. I've attached a graph of the fatal day. The outage occurred around 12 o'clock mid-day.

In all fairness, I have no idea why the battery temperature rose as much as it did. The second graph show temperature against battery voltage which was my first suspect, i.e that the voltage rose too far, but again, that's not the case either.

I'm puzzled.
Eric



Title: Re: Sunny Island 5048 and Battery Fan
Post by: stephendv on October 08, 2013, 04:56:41 PM
Mmmm, the charge voltages seem high.  You could try reducing the Boost voltage setting on the SI to 2.4V (the default is 2.5).
The maximum charge current for the battery should be specified in the battery datasheet.  Generally it's about 20% of the capacity, so 144A.

The charging current near the time of crisis shows that the battery is almost fully charged, since it's relatively flat.  You could also try reducing the boost time in the sunny island to 30 mins - 1 hour or so.  So you'll just use the generator to bump the battery state of charge up, but won't use it to fully charge the battery.


Title: Re: Sunny Island 5048 and Battery Fan
Post by: bautsche on October 08, 2013, 05:08:45 PM
OK, so I'm good on the current as it was that day, but I should still adjust it down to avoid too high a current if I add additional solar at some later stage.
The battery Boot Voltage is set to 2.63V and the Full charge Voltage to 2.65 (i.e. 63.12V and 63.6V for the bank respectively). The open circuit voltage for the solar panels directly attached to the battery is 73.6V and the Maximum Power Voltage is 59.2V (or 2.47V per cell).

Eric


Title: Re: Sunny Island 5048 and Battery Fan
Post by: offthegridandy on October 08, 2013, 07:19:50 PM
Eric,

I think stephendre is correct.  Assuming your using lead/acid batteries then the charge rate is generally considered to be around 20% of the sets total ie around 144A.  If you've been charging at 436a I imagine your batteries are boiling.  Have you checked the levels or had to top them up a lot?

Andy


Title: Re: Sunny Island 5048 and Battery Fan
Post by: Tinbum on October 08, 2013, 07:43:37 PM

The battery Boot Voltage is set to 2.63V and the Full charge Voltage to 2.65 (i.e. 63.12V and 63.6V for the bank respectively).

Eric


Those voltages are set very very high for a 48v battery bank. My rolls batteries are 61.92 - 64.08v for an equalise charge!!


Title: Re: Sunny Island 5048 and Battery Fan
Post by: bautsche on October 08, 2013, 08:06:10 PM
Yes, they are lead/acid batteries and yes, it's 48V. The maximum I seem to have ever charged them with is 159.1A. So I do need to set that limit, on that day however, the maximum was around 70A.
I check water levels every weekend and don't refill a huge amount.
I'll double check with my electrician why he set the charge voltage this high then, I guess as the consensus seems to be it's too high for the batteries. I wonder if it might have something to do with the quality of my battery bank (slightly differing cells of differing ages, but who's got 12 grand for a new battery bank.... :-(

Eric


Title: Re: Sunny Island 5048 and Battery Fan
Post by: offthegridandy on October 09, 2013, 08:35:46 AM
Eric,

I think you should be able to buy a new set of cells 800ah for nearer £2000.  I can supply a couple of contacts if required.


Andy


Title: Re: Sunny Island 5048 and Battery Fan
Post by: bautsche on October 09, 2013, 08:51:02 AM
I would be very interested in cheaper battery cells. If you have any contacts, that would be very much appreciated.
Eric


Title: Re: Sunny Island 5048 and Battery Fan
Post by: stephendv on October 09, 2013, 09:01:48 AM
I check water levels every weekend and don't refill a huge amount.
I'll double check with my electrician why he set the charge voltage this high then, I guess as the consensus seems to be it's too high for the batteries.

Good that you check every week, but if you're refilling every week and charging at those voltages then I think you're massively overcharging the batts.  Choosing charging voltages and time for the SI depends to some degree on how much charge you're getting from the solar panels, and how much you're getting from the generator, i.e. how often do you run the gen?

If there is NO charge controller between solar panels and batts, then that could be another source of overcharging, depending on how much solar you have(?).  Would be good to put a cheapo PWM controller in there.

There's some degree of interplay between the absorb voltage and absorb time.  If you run the generator only on occasion and rely mostly on solar, then it's more economical on fuel and better for the batts if you absorb at a slightly higher than normal voltage, e.g. 2.5V for a short period of time, say 1 hour.  The purpose of this is to bump the state of charge of the batts up to more than 80% and to mix the electrolyte, without leaving the gen running unloaded for extended periods of time.  This will not result in a full charge, so you'll be relying on the solar to finish the job- which only works if you have enough solar.

But if you run the generator a lot and don't have enough solar, then you need a lower absorb voltage, but for a longer time to make sure the batteries are getting a proper charge.  Since the SI has both a boost and a full setting, you could then program boost to be 2.5V for 1 hour.  And full charge to be 2.4V for 4 hours, every 2 weeks.  An EQ charge at 2.6V should really only be done once a month or even less, depending on the SG readings from the batts.

Would be good to know how often you run the gen, and how much solar you have, and at what state of charge do you start the gen?


Title: Re: Sunny Island 5048 and Battery Fan
Post by: bautsche on October 09, 2013, 09:21:21 AM
Hi Stephen.

OK, I have a 48V 760Ah battery bank. I also have a standing load of 1kW (24/7) with peaks obviously higher than that.
Before I got the solar, the Sunny Island started the generator about 6 times in each 24 hr period.
I then added 3760Wp of solar, direct attached to the batteries (to save money) and the generator now runs about 3 to 4 times each 24hr period.

Unfortunately, my battery bank is pretty shot, so I also start the generator when the battery voltage drops below 44V and then run it until the voltage reaches 59V.
In all, a boost run doesn't happen more than once every day or two because before the SI decides to start the generator based on state-of-charge (SOC) the voltage has usually dropped below 44V. A boost run tends to happen early in the morning before the farm wakes up when the load on the battery bank is low but the SOC has dropped to the point where the SI starts the generator (the SI starts at 45% SOC). I experimented in the past with starting the generator earlier but that then keeps the generator running for such long periods of time that I can't afford the red diesel.

I'm quite happy that the solar panels are not going to overcharge the batteries in most conditions. I am planning on wiring up a dump load in the form of turning on the immersion heater in the house if the battery voltage rises too high (level still to be determined) but that's a lot of messing around with taking apart kitchen cupboards and digging cables out of plasterboard walls (don't ask, the previous owner basically was clueless what the cable was for and decided to cut it off because it "looked ugly". And the SI, batteries, generator, etc are a quarter of a mile from the house).
I don't have enough solar to charge the batteries in any significant fashion at the moment (see attached graph). On a sunny day (like the 6. Oct.) it puts some charge in around mid-day, but other than that really only supports the site load. I have also attached a few more graphs to show site performance on that fatal day. Ignore the horizontal lines, they are no longer relevant. The SOC crisis at 8:30am that day seems to be some calculation difficulties that the SI tends to have on a regular'ish basis.

It sounds like I ought to set the charge voltage lower for the boost charge anyway. As is, the SI is programmed not to every do an equalising charge. Perhaps I ought to have a fresh look at the charging regime all over.


Title: Re: Sunny Island 5048 and Battery Fan
Post by: DaveSnafu on October 09, 2013, 02:49:19 PM
I would not rely on the MPP voltage printed on the rear of the panels being correct, I have consistently measured 1-2v over panel spec on all the panels that pass through my hands.
For instance sharp solar nd230, Mpp 30.5v, connected directly to a 230ah battery @24v, will raise the volts to over 33v before we cancelled the experiment.


Title: Re: Sunny Island 5048 and Battery Fan
Post by: Tinbum on October 09, 2013, 02:56:53 PM
MPP is maximum power point and not the max voltage of the panels and has no relevance on the max voltage at the batteries. They can go to a higher voltage than the MPP but the current will be less. In theory they could go to the VOC but in practice that won't happen with batteries connected.


Title: Re: Sunny Island 5048 and Battery Fan
Post by: Eleanor on October 10, 2013, 10:34:09 AM
Hi Bautsche, sorry to hear of your troubles. I haven't had time to read all of your posts properly but I think the Sunny Island SOC calculation error may be caused by the fact that battery capacity varies with the size of the load ie the bigger the load the fewer Ah that can be drawn from it (Peukert Effect). If the SI sees a big load it will recalculate the battery capacity and the %SOC will therefore decrease accordingly. Stephen posted the relevant page of the manual a while back and from memory the SI uses just two averaging factors for the capacity in the calculations even though the actual capacities at low currents will vary considerably. This leads to step changes in the %SOC readings which I donít think reflect the real situation. It could also be worth checking that the correct battery capacity and Peukert coefficient are entered into the software.

I wonder if it would be best to AC couple the PV using a grid tie inverter? This way the SI will control all the charging. As Stephen has mentioned earlier presumably all the charging devices go through a common shunt so the SI is seeing all the current going in and out of the battery?


Title: Re: Sunny Island 5048 and Battery Fan
Post by: bautsche on October 10, 2013, 10:44:02 AM
Hi Eleanor.
Yes, the current sensor sees all the power going to the battery. Interesting re how the SOC is calculated. I'm going to upgrade the firmware this weekend anyway and there is some mention about better SOC calculation in the later versions, so I'll see what that does.
The reason I direct attached the solar is quite simply money, i.e. the money I don't have to buy a SunnyBoy. :-(
Looking over the graphs with a friend last night, it seems that this jump in SOC calculation is what caused the SI to start the generator when it wasn't required which in turn lead to the high temperatures....

Thanks you guys for all your help and pointers.
Eric


Title: Re: Sunny Island 5048 and Battery Fan
Post by: stephendv on October 10, 2013, 01:09:55 PM
Unfortunately, my battery bank is pretty shot, so I also start the generator when the battery voltage drops below 44V and then run it until the voltage reaches 59V.

44V!  That's far too low for a normal set of batts.  My guess is that the batteries are so shot that the SoC calculation in the SI is completely thrown off.  For example, the peukert constant would be quite different for new vs. damaged batts, and I think yours are tending towards the latter.

In all, a boost run doesn't happen more than once every day or two because before the SI decides to start the generator based on state-of-charge (SOC) the voltage has usually dropped below 44V. A boost run tends to happen early in the morning before the farm wakes up when the load on the battery bank is low but the SOC has dropped to the point where the SI starts the generator (the SI starts at 45% SOC). I experimented in the past with starting the generator earlier but that then keeps the generator running for such long periods of time that I can't afford the red diesel.

You can control gen runtime through the boost and full time settings.  As I mentioned earlier, I'd reduce the boost and full voltages to something like 2.5V and limit boost time to 30 mins - 1 hour and see how it behaves.

If you have a constant 1kW draw and 3kW of PV, then when the batteries are fully charged they're still being hit with a constant 2kW charge- that could easily overcharge them.  A 60A charge controller is not that pricey: http://www.navitron.org.uk/product_detail.php?proID=8&catID=119  compared to the price of a new set of bats.



Title: Re: Sunny Island 5048 and Battery Fan
Post by: bautsche on October 10, 2013, 01:40:23 PM
For historic reasons, I actually have a Trace 4048 (I think, same as Paul Camilli's anyway) as well as the SI (trace isn't doing anything other than sit there at the moment) and that has a spare relay that can be operated based on high battery voltage, so I plan on using that and a cable that already runs from the inverter shed to the house (to operate a relay to turn on the immersion heater). I just actually have to get around to taking the kitchen cupboards apart to actually get at the cable.... :-(
Eric