System Design and Component Selection

joeedens · April 2011

I'm looking for feedback from all of our resident experts, and anyone with an opinion, on my system design and component selection. At this point I haven't placed any orders so nothing is set in stone yet.
I am designing an off-grid system, with the potential to use grid power for backup battery charging (instead of a generator) or to allow Grid pass through in the event of hitting a low battery set point. I am NOT interested in being on a grid tie system. I am shooting for 1 day of autonomy and 50% depth of discharge.
I have done a load analysis and have come up with the following:
Average Watt-hours per day: 16600. There is some room for improvement, but I have a solid need for 2 refrigerators and 2 large freezers. Am also on a rural water well system with 3/4 HP 240V pump and need to water a large garden. I'm not using electricity for any type of heating or AC loads.
Potential simultaneous loading 4750 watts.

Here are the components that I'm leaning towards:

HUP Solar One 48V 950 AH Battery Bank
2 each Midnite Solar Classic 250 Charge Controllers
24 Evergreen ES-A-215-fa3 215W Panels. 3 Strings of 8
1 Outback Flex power Two FP2-10 VFX 3648 (Dual Inverters)
1 MNPV 6 Midnite Solar Combiner Box

My Questions:
1. Am I missing anything besides mounting, breakers, and wiring and system monitors?
2. What would you change if you were implementing it, and Why?
3. What type of system monitors would you recommend? I am a MAC user and I know that many of the data logging packages are Windows only. All of my components will be located approximately 150 away in my barn/workshop.

mike95490 · April 2011

Re: System Design and Component Selection

I am shooting for 1 day of autonomy and 50% depth of discharge.

That's going to greatly effect your battery life (in a bad way). 1 day of autonomy is ok, but that one day still should only take your batteries down to 70%-80% of full. Taking the batteries down all the way to 50% will greatly shorten their life.

What I don't understand, is why, if you have grid power, you are choosing not to use it? If you have $$ to burn - send some my way.

48V 950 AH Battery Bank

needs 95A to recharge and stir the electrolyte, your array is only forecast to produce 4128 w and at 54V, that's 77A. A bit short, but I don't know how much stratification abuse the batteries can take.

Do you have existing well pump, or is it on shopping list. 3/4 hp is a lot of pump unless you have a really deep well. 2 wire or 3 wire ? 3 wire with the remote capacitor is supposed to be an easier starting load.

2 fridges & 2 freezers ! Set the freezers a bit colder and put them on a timer to only run daytime, and not off the batteries at night. ON time should be after sunrise, and the PV's are powered up.

mike95490 · April 2011

Re: System Design and Component Selection

Water - do you have elevated storage, or pressure tank ? larger pressure tanks = less start/stop on the pump. Lots' more hints from others here I'm sure

Cariboocoot · April 2011

Re: System Design and Component Selection

Must have passed out when I read "16.6 kW hours per day" consumption.

Seriously, the sizing seems fine except that the battery bank is a tad large. At 50% DOD you're looking at about 692 Amp hours, which would fit better with the panels too. But heed Mike's warning; the DOD is a trade-off with life expectancy of the batteries.

His other suggestions are absolutely spot on too, in my opinion.

icarus · April 2011

Re: System Design and Component Selection

I'm sure you know, and I tr to mention it once, but off grid, compared to grid tie comes at about twice the cost net/net per watt, at about half the efficiency, resulting in power tha is about 4times as expensive. (Not mention! The replacement cost of the battery is fairly significant )

The grid is PV's best fiend.

Tony

joeedens · April 2011

Re: System Design and Component Selection

Mike, thank you for your response. This is exactly the feedback that I was hoping for.
Mike said: "needs 95A to recharge and stir the electrolyte, your array is only forecast to produce 4128 w and at 54V, that's 77A. A bit short, but I don't know how much stratification abuse the batteries can take."

What type of derate factors do you use to come up with the forecasted 4128 w from the 5160 w STC? Is this discounting the losses of the charge controller?
I'm still learning about batteries, so is it a rule of thumb that you need 10% of the battery bank AH rating in order to properly "stimulate" the batteries? Is this what is needed during the "bulk" stage? Is this true regardless of the depth of discharge?

Mike said: "Do you have existing well pump, or is it on shopping list. 3/4 hp is a lot of pump unless you have a really deep well. 2 wire or 3 wire ? 3 wire with the remote capacitor is supposed to be an easier starting load."

Yes, I already have the 3/4 HP well pump. It was spec'd out by the well installer when installed. It is a 3 wire start. It has a control box with 2 wires (240V) input and inside is a circuit with a large capacitor and 3 wires to the pump. I've got a large pressure tank. I think it was called 40 gallon, but somehow I think I remember that meant that I've got something like 15 to 20 gallons of storage in it.

Mike said: "2 fridges & 2 freezers ! Set the freezers a bit colder and put them on a timer to only run daytime, and not off the batteries at night. ON time should be after sunrise, and the PV's are powered up."

I tried experimenting with a timer on a freezer once and a week after I started using the timer (on-off 4 times per day) I ran into problems. I think it burned out a sensor used for self defrost and was defrosting constantly. I live 65 miles from the nearest repair technician that could respond in one day before I lost a whole beef, and it cost me big time for his services. I stopped using the timer.

Thanks again for the info.

joeedens · April 2011

Re: System Design and Component Selection

Cariboo,
Thanks for your response.

Cariboo said: "Seriously, the sizing seems fine except that the battery bank is a tad large. At 50% DOD you're looking at about 692 Amp hours, which would fit better with the panels too. But heed Mike's warning; the DOD is a trade-off with life expectancy of the batteries."

To calculate my AH required, adjusted the numbers for 90% inverter efficiency and used a battery temp derate factor of 1.11. So my calculation was as follows: 16600/48=346 then 346*2=692 (for 50% DOD) then 692/.9=769 (for 90% inverter efficiency) then 769*1.11=854 (battery temp compensation)
Then I found the battery I liked and the capacity that met my needs was 950AH. So, I take it that this is a conservative calculation? If so, then doesn't that just work in my favor as far as depth of discharge is concerned? Maybe my conservative calculations would just allow a shallower depth of discharge and that would be a good thing, right?

Cariboocoot · April 2011

Re: System Design and Component Selection

Some of the basic rule-of-thumb info we use around here includes a derating of 77% for panels * charge controller. Like this:

4128 Watts @ 77% efficiency = 3178 Watts. Divide by charging Voltage to get potential peak charge current. In this case that's about 57 Volts for 55 Amps. Now before anyone thinks I'm saying Mike is wrong, keep in mind that 77% is not written in stone nor is 57 Volts charging.

Another rule of thumb says to shoot for 5% to 13% of the batteries' Amp hour rating for peak charge current. The battery FAQ's references this and other important information: http://www.windsun.com/Batteries/Battery_FAQ.htm

And yes, for refrigeration it's best to have the power available whenever it needs it. Older refrigerators would work well under the timer, but modern ones have more electronic controls. Also, the warmer the unit is allowed to get, the more power that will be used to cool it down again. It's best to get freezers cold and keep them there. You can reduce power use if you can keep the outside temp low, as in locating it in a cooler spot, and add insulation without blocking the function.

Cariboocoot · April 2011

Re: System Design and Component Selection

joeedens wrote: »

Cariboo,
Thanks for your response.

Cariboo said: "Seriously, the sizing seems fine except that the battery bank is a tad large. At 50% DOD you're looking at about 692 Amp hours, which would fit better with the panels too. But heed Mike's warning; the DOD is a trade-off with life expectancy of the batteries."

To calculate my AH required, adjusted the numbers for 90% inverter efficiency and used a battery temp derate factor of 1.11. So my calculation was as follows: 16600/48=346 then 346*2=692 (for 50% DOD) then 692/.9=769 (for 90% inverter efficiency) then 769*1.11=854 (battery temp compensation)
Then I found the battery I liked and the capacity that met my needs was 950AH. So, I take it that this is a conservative calculation? If so, then doesn't that just work in my favor as far as depth of discharge is concerned? Maybe my conservative calculations would just allow a shallower depth of discharge and that would be a good thing, right?

Yes it works in your favour as far as DOD; the lower discharge the better. But it "works against you" in terms of not having sufficient array to do a good job of recharging. Not saying it won't work; at my 55 Amp estimate it's still 5.7% potential peak charge current.

mike95490 · April 2011

Re: System Design and Component Selection

joeedens wrote: »

What type of derate factors do you use to come up with the forecasted 4128 w from the 5160 w STC? .

I simply used 80% of the STC spec. Since the STC spec is only valid in the lab, at 70F @ 1 Sun, with normal atmospheric condition, and hot panels (130F) you get about 20% loss off the spec. = 4128

Now I stray to possibly faulty memory:

The scary part is when you use batteries, someone here has a very simple calculation: PV spec * 4hours / 2 = what goes into the batteries. More in summer sometimes. (did I get that right, or blow it ?)

5160 * 4 = 20640 / 2 = 10,320 watts stored. your 24 hour load should not be above that, and if you aim for 25% discharge (75% remain) is a battery of 41,280 watts ~ 800ah @ 48V

--

Water - using a pressure tank, you need to get a couple more, and they stack in parallel. Gives you more stored for nighttime use, and lesser starting cycles on your pump, if your well can keep up.
Water the garden after NOON ( when the batteries are out of BULK), so the sun can power the pump directly, then you get 95% efficiency of your solar harvest, instead of 50% when pumping from the batteries. This varys with clouds and such.

stephendv · April 2011

Re: System Design and Component Selection

I can't see a reason to choose 2 x classic 250 controllers with those panels. The 250 only has a maximum peak charging current of 63A, whereas the 150 has a peak of 96A. Seems if you rewired your array to fit within the 150's limits you'd only need 1 classic instead of 2.

dwh · April 2011

Re: System Design and Component Selection

mike90045 wrote: »

The scary part is when you use batteries, someone here has a very simple calculation: PV spec * 4hours / 2 = what goes into the batteries. More in summer sometimes. (did I get that right, or blow it ?)

I believe that's Tony's rule of thumb of what comes out of the inverter, rather than what goes into the batteries.

icarus · April 2011

Re: System Design and Component Selection

My name plate/2*4 rule of thumb assumes the inclusion of all loses up to and including the inverter. Of course not a hard and fast rule, but one that has proved itself over time to me at least.

T

BB. · April 2011

Re: System Design and Component Selection

Technically, the rule of thumb numbers I use:

81% efficiency for Solar Panel "marketing rating" for warm day/hot sun
95% efficiency for controller efficiency
80% efficiency for battery (flooded cell)
90% efficiency for battery (AGM)
85% efficiency for Inverter

So, for a system with solar panels + charge controller + flooded cell battery bank + AC inverter, the end to end efficiency would be:

0.81 * 0.95 * 0.80 * 0.85 = 0.52 = 52% end to end efficiency

Just a rule of thumb... There are changes a person can do to make efficiency higher or lower:

Cycle a battery mostly between 50 and 85% state of charge--more efficient
Cycle a battery mostly 85-100% state of charge, less efficient
Charge during the day, use power at night, less efficient
Charge during the day and use power during sunlight hours, more efficient.
Use an AGM battery instead of a Flooded Cell, more efficient
Leave AC inverter on 24x7, less efficient
Use a large inverter with small loads, less efficient
Use a small inverter sized to the loads, more efficient
Use an inverter with "search mode", more efficient
Turn on an inverter only when AC loads needed, more efficient
Don't assume that you can use the "predicted" available power every day--Assume that you can use 50-75% of your planned power every day and leave some leftover for growing loads, visitors, lights left on, etc...

With so many variables, it is not worth trying to come up with a much more accurate rule of thumb other than 1/2 * solar panel * hours of sun per day = Maximum/Average AC power available per day.

It is why we say that monitoring your off-grid power system is so important. And why two people with identical systems can have such different stories of successes and failures.

-Bill

Cariboocoot · April 2011

Re: System Design and Component Selection

Well I'll take issue with the 85% factor for inverters as the good ones are over 90% efficient.

But by and large, that's about it. Horrible though the results are. Keep in mind that solar installs are highly site-specific and greatly affected by wire runs, ambient temperatures, elevation, local weather, et cetera. Some people do better than "50% out" (like me) but on average it's what you can expect. If you get more than that you'll be happy.

If you expect more and get less you'll be back here crying about it.

icarus · April 2011

Re: System Design and Component Selection

The previous two posts highlight the difference between a engineer and a seat if the pants tinkerer. The engineer wants (needs?) a gnats ass answer to a problem, while the tinkerer is usually satisfied with "close enough".

There is no implied criticism of either here, as I think there is a place for both, just illustrating the difference!

Tony

Have
To edit that however since 'coot got in between. The posts I was referring to was Bill's "calc" and my napkin math.

icarus · April 2011

Re: System Design and Component Selection

It should also be noted, that while the fifty percent rule of thumb is ballpark, in the real world off grid performance is often even worse. This is simply do to the fact that some production is lost when the charge controller dials down current due to batteries becoming fully charged. In an ideal case the batteries would become 100% full, the minute the sun goes off the panels.

T

BB. · April 2011

Re: System Design and Component Selection

Regarding 85% vs 90%--Again, it depends on how you load the inverter... Only run it at mid-range power and you can see >90% efficiency... Operate it above 80% load or under 10%, and you will see <85% efficiency.

And Tony highlights why people get frustrated when asking an engineer for help... 52% vs 1/2 -- Both answers are virtually the same.

If you get within +/- 20% of your predictions with solar using generic rules of thumb and estimates--Count yourself as being dead-on.

If you get within +/- 20% of using PV Watts, calculations to 4 decimal places, accounting for 3% initial output drop from a solar panel, cleaning your panels weekly, and 1% every year there after--Count yourself as being dead-on.

In most regions, you can see a 10-20% variation in solar output when looking at monthly/yearly data (that has been constructed from 20 year or longer averages) just from normal variations in weather patterns.

The only reasons I keep 3-4 significant figures in the math is so people can reproduce my numbers and to avoid adding even more round-off errors.

-Bill

tonygcan · April 2011

Re: System Design and Component Selection

This is a good thread. Is it safe to assume then that for a system with a solar array of 2220 watts one should discount this down to 1110 watts effectively?

mike95490 · April 2011

Re: System Design and Component Selection

tonygcan wrote: »

This is a good thread. Is it safe to assume then that for a system with a solar array of 2220 watts one should discount this down to 1110 watts effectively?

Only if you only use power at night. Daytime usage (pumps, vacuum cleaners, dishwasher, clothes washer) comes off the PV, and not out of the battery, so that's consumed at 80% efficiency.

icarus · April 2011

Re: System Design and Component Selection

tonygcan wrote: »

This is a good thread. Is it safe to assume then that for a system with a solar array of 2220 watts one should discount this down to 1110 watts effectively?

A typical battery based PV system will ( using my rule of thumb) deliver the following. Take the name plate rating of the PV, divide by two to account for all cumulative system loses, then multiply that number by four to represent the average number of hours of good sun a d PV system might get on a per day basis, averaged lever the course of the year.

100watts fir example.

100/2=50*4=200 watt hours per day.

Now if you are powering loads directly from PV, one would derate perhaps 25% and then (other loses additional ) multiply by 4, or what ever number of hours of good sun you might expect on average.

The one hard and fast rule of solar, is that people almost always over estimate the amount of power they can actually harvest, while at the same time they underestimate their loads.

Tony

System Design and Component Selection

Comments

Categories