Battery Bank Configuration for a 48VDC, 2430 AH bank

Tmp4000

Hello Everyone,

I need a little help. I am constructing a solar electrical system with the following specs:

Power Needed:  7000W
# of solar panels:  72 (each 315W)
Inverter:  Outback 8000W -  GS8048A
6 x solar panel arrays:  4 x 3 config (4 panels in series make one string...then 3 such strings in parallel)
Charge Controllers:  6 x MidNite Classic 200
Days of Autonomy: <1 day (13 hours)
Depth of Discharge:  0.7 - 0.8
Battery Bank Needed:  48VDC, 2430 Ah

This will be used to power science equipment.

I need contingency such that even if one battery fails, it won't take out the entire bank...hence the need for parallel strings. So I cannot take 24 x 2V, 2430Ah batteries and put them in series...that would be a bad idea in case one fails.

I was looking at 28 batteries (each 12V, 354Ah). One series string would contain 4 batteries....and then take 7 such strings and put them in parallel. But I read that having so many parallel strings is not advised. Why is that ?

What would be the best battery and battery configuration to create a 48VDC, 2430 Ah bank with contingency in case one battery fails it won't take out the entire bank ? And what would be the best way to connect 6 charge controllers to this battery bank to make sure that all batteries would be charged evenly ?

Thank you very much!

Estragon

A couple of thoughts:

- If one bank failed, wouldn't you still need the 2400ah? If so, you could use 48x2v2400ah cells wired as two series strings of 24. As well as balance problems of 7 strings, you would have 28x6=168 cells to check and water!

- Batteries rarely fail without warning. It's usually more like bankruptcy (it happens slowly, but faster towards the end). With proper routine monitoring and maintenance, they should give plenty of warning if they're starting to fail. The more likely point of failure is the inverter. It may well fail suddenly, so my first choice for redundancy would be there.

For connecting charge controllers and inverters, etc., it's generally easiest to connect to a common positive and a common negative bussbar, and then a heavy short cable to the bank.

mcgivor

@Tmp4000 said
 I need contingency such that even if one battery fails, it won't take out the entire bank...hence the need for parallel strings. So I cannot take 24 x 2V, 2430Ah batteries and put them in series...that would be a bad idea in case one fails.

Using 2 Volt cells would  be the best solution form a ballance perspective, if in the unlikely event  one cell was to instantly fail, it could be removed and have the system  operate at 46V until a replacement is installed. Adjustments to charge parameters  would be needed for the controllers , which must all be exactly the same in parallel on a common bank. Duplication with 2 banks would be better albeit at a cost. Having 7 strings of 12V would be a nightmare from a ballance and maintenance point of view, IMHO

Another observation was a single inverter, perhaps it may be better to divide loads between two inverters, or have a spare on hand for replacement, as an inverter failure, should it happen, usually is without warning.

Marc Kurth

Great advice rendered already!

I would want to hear more specific information about the application before making a recommendation. Having a clear understanding of customer expectations is always step one for me because this is what I do for a living.

Your 13 hours surprises me for a mission critical application - along with the single point failure design with one inverter already mentioned. Tell us more in order to get good info!

Tmp4000

Thank you for the comments thus far.

Okay, here are more specs:

I want to power computers, air conditioning, and equipment for data collection. The power draw will be:

Air Con:  1870W -> round to 1900W
Data Equipment = 3500W
Computers + Peripherals = 1100W

I will be purchasing 315W panels...so I will need about 72 of them.

The above equipment will need to operate 24/7. Once the Sun goes down....we assume that the bank will needs to provide power for 13-16 hours. The actual number of hours of sunlight is about 12 hours (near the Equator). The air conditioners will need to operate 24/7 due to humidity and a temperature differential between inside and outside of 25-30 degrees F.

If I only use one inverter/battery bank for the whole system....the budget only allows for one day of downtime (actually about 16 hours) assuming that I discharge the battery bank 80% (Depth of Discharge = 0.80)... I cannot afford a larger battery bank a.k.a. discharge only 50%. 

My calculated total bank capacity needs to be 2430 Ah assuming a depth of discharge level of 80%...

However, if anyone has any recommendation of having multiple inverters and multiple battery banks (one bank/inverter for computers + peripherals and one bank/inverter for air conditioners) I would be all open to suggestions. 

Thank you again...still learning about solar electrical systems....

Photowhit

So you plan to have a 22680 watt array feeding a 116 Kwh battery bank (48VDC x 2430 Ah)

At 48 volts you could count on the 22680 watt array to produce 22,680 x .75(NOCT value) = 17,010 watts....17,010/58=293amps into a 2430 Ah battery bank or about 12% charge rate. If depleting the battery bank to 20% of capacity you would need 2430 ah x .8 = 1944 amps per day or 1944/293=6.6 hours of direct sunlight each day. This is a down and dirt rough estimate, charging rate will be higher when batteries are at a low voltage in the morning and the batteries will accept less current as they near full capacity.

Several things will be difficult, as the batteries become nearly depleted the voltage will sag. At 20% of capacity they will have about 500 amp hours left in them and you will be drawing energy at 6600 watts / 46 volts = 143 amps or more than 1/4 of the remaining energy. I doubt you will maintain 46 volts or even 44 volts and the system will fail. Indeed the 143 (greater with inverter load) is more than 1/20th of the battery capacity so the battery will effectively be smaller.

I'd guess this is for bitcoin mining? I doubt you will find cheaper energy than the grid. While you can't afford a larger battery bank, this one won't meet your needs. You may have over estimated your needs, Have you run a Kill-A-Watt meter on your loads?

Tmp4000

Photowhit said:

So you plan to have a 22680 watt array feeding a 116 Kwh battery bank (48VDC x 2430 Ah)

At 48 volts you could count on the 22680 watt array to produce 22,680 x .75(NOCT value) = 17,010 watts....17,010/58=293amps into a 2430 Ah battery bank or about 12% charge rate. If depleting the battery bank to 20% of capacity you would need 2430 ah x .8 = 1944 amps per day or 1944/293=6.6 hours of direct sunlight each day. This is a down and dirt rough estimate, charging rate will be higher when batteries are at a low voltage in the morning and the batteries will accept less current as they near full capacity.

Several things will be difficult, as the batteries become nearly depleted the voltage will sag. At 20% of capacity they will have about 500 amp hours left in them and you will be drawing energy at 6600 watts / 46 volts = 143 amps or more than 1/4 of the remaining energy. I doubt you will maintain 46 volts or even 44 volts and the system will fail. Indeed the 143 (greater with inverter load) is more than 1/20th of the battery capacity so the battery will effectively be smaller.

I'd guess this is for bitcoin mining? I doubt you will find cheaper energy than the grid. While you can't afford a larger battery bank, this one won't meet your needs. You may have over estimated your needs, Have you run a Kill-A-Watt meter on your loads?

It is not for bitcoin mining...it is a science project studying the air/atmosphere. Yes, the equipment power usage was measured with a power meter. Alright, so what would be the appropriate bank size (or multiple bank sizes) and how would one go about calculating the appropriate batteries, bank size, etc.... ?

Tmp4000

Photowhit said:

So you plan to have a 22680 watt array feeding a 116 Kwh battery bank (48VDC x 2430 Ah)

At 48 volts you could count on the 22680 watt array to produce 22,680 x .75(NOCT value) = 17,010 watts....17,010/58=293amps into a 2430 Ah battery bank or about 12% charge rate. If depleting the battery bank to 20% of capacity you would need 2430 ah x .8 = 1944 amps per day or 1944/293=6.6 hours of direct sunlight each day. This is a down and dirt rough estimate, charging rate will be higher when batteries are at a low voltage in the morning and the batteries will accept less current as they near full capacity.

Several things will be difficult, as the batteries become nearly depleted the voltage will sag. At 20% of capacity they will have about 500 amp hours left in them and you will be drawing energy at 6600 watts / 46 volts = 143 amps or more than 1/4 of the remaining energy. I doubt you will maintain 46 volts or even 44 volts and the system will fail. Indeed the 143 (greater with inverter load) is more than 1/20th of the battery capacity so the battery will effectively be smaller.

I'd guess this is for bitcoin mining? I doubt you will find cheaper energy than the grid. While you can't afford a larger battery bank, this one won't meet your needs. You may have over estimated your needs, Have you run a Kill-A-Watt meter on your loads?

It is for a science project studying the atmosphere. The power consumptions were measured using a power meter. Ok, so I think I was wrong in sizing my battery bank. What is the proper way to size the battery bank/banks ? Did I at least calculate the correct number of panels (72?) 

Photowhit

Where will it be located? or What are the solar isolation numbers for the area for the test period? You state near the equator, but even then fixed panels will only deliver near full capacity (less 25% NOCT value) for 6 hours and maybe 1-2 addition with all the lesser hours, even on sunny days.

You also state this will run 24/7? So this will cut into the available power to recharge your battery bank as well. 13 hours is also not a realistic idea of the amount of time it will be running primarily from the battery bank, unless you plan to have your array tracking the sun.

Will this be unmanned? Could a generator assisted or just a generator system not be viable!

A flooded battery bank would require some monitoring, at least monthly inspection, AGM might be a viable option, but much more expensive, Marc would have some insight on this. Lithium would allow for faster recharging and less voltage sag, again more expensive. What will you do on cloudy days?

Tmp4000

Photowhit said:

Where will it be located? or What are the solar isolation numbers for the area for the test period? You state near the equator, but even then fixed panels will only deliver near full capacity (less 25% NOCT value) for 6 hours and maybe 1-2 addition with all the lesser hours, even on sunny days.

You also state this will run 24/7? So this will cut into the available power to recharge your battery bank as well. 13 hours is also not a realistic idea of the amount of time it will be running primarily from the battery bank, unless you plan to have your array tracking the sun.

Will this be unmanned? Could a generator assisted or just a generator system not be viable!

A flooded battery bank would require some monitoring, at least monthly inspection, AGM might be a viable option, but much more expensive, Marc would have some insight on this. Lithium would allow for faster recharging and less voltage sag, again more expensive. What will you do on cloudy days?

It will be located along the Equator. Measured actual per day solar insolation is 5.7 - 6.0 kWh/m^2

The Air Conditioning will run 24/7 and one desktop computer. The sensors/data equipment will run during the day...and preferably at night. for several hours...but we would like it 24/7. If not, we can have the sensors on one day and then off the other day.

Yes, it will be unmanned. We expect to visit the site about once or twice a year for maintenance.

mike95490

Maybe after you work the bugs out, you can manage two visits a year.  Till then,   someone may have to camp out at the site,
What's the worse stretch of cloudy weather you can expect, 2,3, 6 days ?  That's how long the batteries have to last to run the air conditioner..

Photowhit

Since no one will be there to maintain the system, I suspect you will want AGM batteries. There would be no out gassing or water to worry about, they could live in the same air conditioning. I think Marc has more experience with these than most here. They have different charging profiles than flooded. 

Wonder if someone hasn't developed a system to record data on an hourly basis and hibernate between readings... Though it might be a continuous reading you need.

At 6 months you will also want to make this a near 'clean room' and perhaps provide a prefilter for the air conditioner.

mcgivor

A system for such an application would need some serious engineering, with completely  integrated components  that communicate with one another, programed to safe guard itself, have remote access  and send messages via email or such to alert the operator, SMA and Schneider are examples of such systems, SMA and Axitec Li-Ion battery systems are designed to work with one another   Mixing components that weren't designed to work together in a situation as you have described would be a mistake, in my opinion. 


http://www.axitecsolar.com/en/storage-systems.html

Estragon

> @Photowhit said:
> So you plan to have a 22680 watt array feeding a 116 Kwh battery bank (48VDC x 2430 Ah)
>
> At 48 volts you could count on the 22680 watt array to produce 22,680 x .75(NOCT value) = 17,010 watts....17,010/58=293amps into a 2430 Ah battery bank or about 12% charge rate. If depleting the battery bank to 20% of capacity you would need 2430 ah x .8 = 1944 amps per day or 1944/293=6.6 hours of direct sunlight each day. This is a down and dirt rough estimate, charging rate will be higher when batteries are at a low voltage in the morning and the batteries will accept less current as they near full capacity.
>


The pv will be supplying ~7kw of loads as well as charging, so it's more like 10kw available. 10kw/58v charging =~170a. 1944/170 =~ 11hrs + a couple to allow for diminishing current in absorb.

A lot of the equator has a rainy season. If this location has anything like that, a generator backup system will be pretty much manditory.

Taking batteries down to 20% SOC will be hard on them, AGM or otherwise IMHO, so twice yearly visits likely won't be enough.

mvas

Tmp4000 said:

Hello Everyone,

I need a little help. I am constructing a solar electrical system with the following specs:

I need contingency such that even if one battery fails, it won't take out the entire bank...hence the need for parallel strings. So I cannot take 24 x 2V, 2430Ah batteries and put them in series...that would be a bad idea in case one fails.

I was looking at 28 batteries (each 12V, 354Ah). One series string would contain 4 batteries....and then take 7 such strings and put them in parallel. But I read that having so many parallel strings is not advised. Why is that ?

What would be the best battery and battery configuration to create a 48VDC, 2430 Ah bank with contingency in case one battery fails it won't take out the entire bank ? And what would be the best way to connect 6 charge controllers to this battery bank to make sure that all batteries would be charged evenly ?

In a remote location, you need redundancy.
Have you considered running multiple Sunny Island Inverters each with its own battery bank?
Below is an example: A 3-Phase system with two separate 48 Volt battery banks.
You can change to 1 Sunny Island per battery bank for Single Phase.
With three battery banks, each with its own 4548 Sunny Island - if one bank fails then system still runs on two banks.
Each Sunny Island manages its own battery bank, fewer strings in parallel.
Also note, power from any functioning Solar Array can charge any functioning battery bank.
Contact an SMA or Outback ( dealer / installer ) for some serious engineering help.

Tmp4000

We are a small team, with a limited budget. We cannot afford to contact a professional solar company and have them design a system for us. That would cost us ~$200,000. We looked into it a while ago. Our current budget for electrical is only 1/3 of that. We were hoping we could put together a basic electrical solar system for our ~7000W instrumentation + air conditioning and have the site unmanned for most of the time. There will be a local person who will go to the site and check things out (make sure nobody stole anything)....but that's only for 2 hours per week....it's not like that person will perform hours worth of maintenance.... 

Dave Angelini

You seem to be pretty far off on your calculations. You need 6 KWH in 24 hours? How much do you need for overnight?
If I am correct above, I would use an AGM 2V in a 48V bank with and XW+ and all the solar you can. Each of the Hvmppt-80-600V controllers can have 6KW of solar on them. If you have an Ethernet connection you can monitor this easily anywhere in the world. Outback has similar equipment but not as easy for installation because of low voltage controllers. I assume that you are using a small 9,000btu mini-split with high SEER. The plot below could have 6kw of solar for each controller, as many as you need. How are you shipping this?

Tmp4000

Dave Angelini said:

You seem to be pretty far off on your calculations. 

Can you please explain what I did wrong so I can learn ? 

I need almost 7000W of electricity generation. That is 7000W*24hrs = 168,000 Wh per day = 168 kWh per day.   How did you get 6 KWH in 24 hours ??

Photowhit

Tmp4000 said;

I need almost 7000W of electricity generation. That is 7000W*24hrs = 168,000 Wh per day = 168 kWh per day.

168Kwh per day with 6 hours of charging = 168kwh/6= 28kw per hour, Solar panels actually produce 75% of panel rating normally. so 28Kw x 1.333= 37.3 Kw array. You have planned for a 22.7 Kwh array.  Even this is a poor calculation as the stored energy will come at an additional cost about 20% for flooded batteries. Less for some Agm's @3% for Lithium.

Also planning for 1 nights autonomy only allows for 1 cloudy day to take your system off line. Also the draw will make a battery bank have a smaller capacity as they are calculated on 1/20th discharge rate. It will also challenge the ability of the charger controller to recharge the battery bank in a short amount of time. Also 1 cloudy day, heck 1 cloudy afternoon, will shut you down running that close to thresholds!

In general we use 3x the amount of production for the same usable energy as a grid connected system. Off grid requires losses to cover to cloudy days and storage. You should get fully charged a couple times a week.

It's just a hard nut to crack. I suspect you have looked at your loads, any way to reduce them near 1/2. That's a lot of energy in air conditioning. Would super insulating, and running highly efficient mini split air condition work?

The big draw is the data collection, are you connected with a university? Can you make it a class project to figure out ways to reduce it's load? This often produces results that you are too close to see. Run it by the 'lab guys' as well.

1100 watts for a server and peripherals isn't too bad having just helped sell off some old servers that ran at 1500 watts by themselves, but it's worth asking around, perhaps some redundant simple computer can collect the data at a minimal load.

BB.

I will humbly suggest that your team review the hardware chosen to run the science side of the project. Using desktop computers vs dedicated low(er) power computers (and/or even laptops, etc.). In general, every watt you put into the computing hardware will have to be removed as heat via the AC. Energy conservation can be a big help in reducing the size/costs of your overall power system and support equipment (like A/C).

A full/remote site (such as a coast guard facility) would have triple redundancy--It sounds like there is no budget for that.

There is always a tension between a very simple/bullet resistant design vs applying full remote sensing/control/automation to a project.

Over the years, I have tended to favor a "simple" basic system with fail over support--Vs a complex system of alternate power paths and such. With complex systems, they always seem to figure out new wild and wonderful ways to fail.

There can be alternatives you can look at too... One is that many AC power supplies can run on high voltage 380 VDC power. In theory, you can get rid of the whole AC inverter side of your system. For telecom systems (at least those in times past), We could buy (for example) PC power supplies with redundant and/or -48 VDC power input.

https://www.google.com/search?q=380+VDC+computer+center+project&ie=utf-8&oe=utf-8&client=ubuntu&channel=fs

Or you can push the redundancy into the compute module itself with (redundant AC input):

https://www.amazon.com/SilverStone-ST55GF-Redundant-Power-Supply/dp/B001O0DOBM

Or redundant 48 VDC (possible both DC and AC input redundant configurations):

http://www.power-con.com.tw/index305a.htm

The downside for high voltage DC systems power systems include there is not any (that I am aware of) direct solar to 380 VDC charging systems. Also, when placing 190x 2 volt cells in series--You now need (IMHO) to have per cell (or per 6 cell) monitoring. In a 190 cell string, a single 2 volt cell failure (short circuit in cell, loss of electrolyte, "poisoning of cell" ,etc.) that are not really "visible" at the 380 VDC level.

If you plan on maintenance visits every 2-4 weeks--Then there is the question of costs/complexity of resilience of your power system... Can your project withstand a loss of power for a week (until the next scheduled visit)? What about on-site/nearby available spares (keep some spare cells/batteries--you may need to add float chargers for those spares, etc.).

-Bill

Dave Angelini

Tmp4000 said:

Dave Angelini said:

You seem to be pretty far off on your calculations. 

Can you please explain what I did wrong so I can learn ? 

I need almost 7000W of electricity generation. That is 7000W*24hrs = 168,000 Wh per day = 168 kWh per day.   How did you get 6 KWH in 24 hours ??

I looked at only one thread here and I mixed up the numbers into KWH. 
"Air Con:  1870W -> round to 1900W
Data Equipment = 3500W
Computers + Peripherals = 1100W"

Your AC requirement does seem really high though. How many square feet? I wish I had more time but as Bill and others have suggested, you need more review and also to state or restate the long term goal. How long will this have to run in years. Is there fuel nearby for a generator. Keep posting and an "aha moment" may occur. Also use only one make of equipment for the power system. A big mistake for gathering data is mixing equipment brands in a power system.

BB.

If you cannot tell us where the system would be installed (security, multiple sites, etc.).... Can you look at one of these sites and get the "hours of sun" per day by month (season). In areas with marine layers, and such--I usually suggest that you design the loads to only use ~65-75% of the "predicted output" of the array on average (i.e., if you can shut down A/C on cloudy days--lower outside temperatures, some equipment can be optionally shut down during "bad weather"), or you have a couple backup gensets to make up for poor production (data is usually based on 20+ year averages--It is very easy to have + 10% year over year, and dark/stormy weather to have as little as 10% down to 1% of clear sky production--Get a couple of weeks of "no sun", and something has to give:

http://www.solarelectricityhandbook.com/solar-irradiance.html
http://pvwatts.nrel.gov/

In general, with flooded cell lead acid batteries (and to a lesser degree with AGM batteries), you cannot recharge the battery bank with ~8 hours of sun per day from 50% depth of discharge. In general, for a long term installation running mostly from solar, I would suggests 2 days of storage and 50% maximum discharge.

So, a "typical" full time off grid power system with 48 volt battery bank would look like:

• 168,000 WH per day * 1/48 volt battery bank * 1/0.85 AC inverter eff * 2 days storage * 1/0.50 maximum discharge = 16,471 AH @ 48 volt battery bank.

Perhaps, you could get away with a bank at 1/2 the AH capacity... But I would suggest somebody like Marc K. here, with lots of AGM battery experience, look at your charging profile.

Typically, 5% to 13% rate of charge for solar--10%+ for full time off grid system (solar panels are "cheap" these days, batteries are not):

• 16,471 AH * 59 volts charging * 1/0.77 panel+controller deratings * 0.05 rate of charge = 63,103 Watt array minimum
  
• 16,471 AH * 59 volts charging * 1/0.77 panel+controller deratings * 0.10 rate of charge = 126,206 Watt array nominal
  
• 16,471 AH * 59 volts charging * 1/0.77 panel+controller deratings * 0.13 rate of charge = 164,068 "cost effective" maximum
  

Also, since you have a minimum 7kWatt load, the solar array has to also add enough panels to keep make up for the loads:

• 7,000 Watt load * 1/0.77 panel+controller derating * 1/0.85 AC inverter eff = 10,695 Watt array to make up for daytime loads (subtacts from charging current available to battery bank.

So that would make 63,103 Watt plus 10,695 Watt as the "minimum" suggested array for the above bank (2 days storage, 50% max discharge).
 
And we have to look at the amount of sun... There is not a lot of "stable" land at the equator for a remote/unattended/very expensive systems.
http://www.solarelectricityhandbook.com/solar-irradiance.html

Palu Indonisa
Average Solar Insolation figures

Measured in kWh/m2/day onto a solar panel set at a 89° angle from vertical:
(For best year-round performance)

Jan	Feb	Mar	Apr	May	Jun
5.26	5.35	5.54	5.76	5.68	5.43
Jul	Aug	Sep	Oct	Nov	Dec
5.51	5.94	5.85	5.53	5.33	5.06

So, taking nominal energy (note, suggest 5 degree minimum tilt to keep array "self cleaning"):

• 168,000 WH per day * 1/0.52 system eff * 1/5.06 hours of sun (December average) = 63,849 Watt array minimum (based on loads and amount of sun) or 202x 315 Watt solar array minimum.

AGM Batteries MAY help reduce the size of your battery bank (perhaps 8,000 AH @ 48 volts)... And looking at your loads and local weather conditions, perhaps some sort of Lithium Ion battery bank (significantly less than 8,000 AH @ 48 volts) would work very well here (Lithium batteries can take much higher levels of charging current, and charge to near 100% SOC much faster than flooded cell lead acid).

Just guessing/making some suggestions here to help understand where we are coming from.

168 kWH per day--That would support two homes (or more) deep in Texas Summer with full A/C and electric appliances (5,000 kWH per month for your installation vs ~2,000-3,000 kWH per month for Tx electric home--guessing).

Or maybe 500 to 1,000 gallons of diesel fuel per month for a non-solar power system? 10-15 kWH genset is not that expensive (two or three for redundancy). Cost of fuel delivered?

-Bill

mvas

Solar Panel Array...
136 KWH Generated daily = 72 panels x 315 watts x 6 hours (equiv Full Sun)
168 KWH Consumed daily = 7,000 watts x 24 hours 
It is worse when AC-to-DC and DC-to-AC conversion efficiency is considered.
One cloudy day, and then what?
Could the Solar Panel Array be too small?

Inverter...
Will the 7,000 Watt inverter support 4,600 watts of equipment and then the A/C Compressor starting?
Could the Inverter be too small?

Battery Bank ...
  93 KWH Battery Capacity    = 2,430 AH x 48 Volts  x 80% DOD
126 KWH Power Consumed  = 7,000 Watts x 18 Hours on battery
6 Hours of "equivalent Full Sun", means 18 hours on Battery
At 80% DOD every day, an AGM battery will be dead in less than 2 years.
Could the Battery Bank be too small?

Your major components:
$22,000 = 72 x 315 PV Panel
  $3,000 = 6 x Midnight Classic 200
  $4,000 = Outback 8000W - GS8048A Inverter
$20,000 = 2 strings in parallel of 24 x 2 Volt 1100AH Batteries 
================================================
$50,000 of the $60,000 budget

Your design of a single DC-to-AC inverter with a single battery bank, gives you no redundancy when a single battery fails.

What about the Balance-of-System?
1) AC Disconnects
2) DC Disconnects
3) AC and/or DC Circuit Breakers
4) AC and/or DC Fuses
5) AC Wire
6) PV Wire
7) DC Battery Cables
8) Conduit, J-Boxes, Connectors, Clamps
9) Roof Mount or Ground Mount
10) Ground Rods
11) Labor to install the PV Array, the Batteries, the Conduit, the Inverter, the Charge Controllers and pull the wire.

Who is going to commission the system?

Lumisol

I would definately hire a caretaker to maintain the power system for you. Someone knowledgeable that likes solitude.
You might end up buying all new equipment a couple times a year when pirates find the spot is unmanned and free for the taking.

Tmp4000

Well, this will be a learning curve for our team. We will be installing the entire system...we won't hire an expert. I asked our hardware folks and they gave us more 'relaxed' numbers:

Air Conditioning (still 24/7 operation):  1400W
Instrumentation + Computers + Peripherals = 4300W

Total = 5700W

Daily energy consumption:  5700W*24hrs = 136,800Wh = 137 kWh

The location will have 7 peaks hours (@ 5.7 - 6 kWh/m^2), so the # of panels needed:

136,8000 Wh / (317W*7h) = 62 panels @ 315W each   Go with 72 panels for connecting each MidNite 200 with 12 panels

Days of redundancy = 1 day, with a battery bank DoD = 0.5

5700W  (assume 7000W w efficiency issues taken into account) 7000W/48Vdc = 146 A. Times 24 hour redundancy:

146A*24h = 3504 Ah

With a Dod=0.5

3504Ah/0.5 = 7008 AH of battery bank needed


Why will this system not be able to give us the 5700W AND also recharge the batteries at the same time ? Even at NOCT, the panels will provide about 16,000W....and at peak they will provide 22,000W...we will be using 5700W. So the rest 16000-5700 = 10,300W

Is 10,300W of power not enough in 7 hours to recharge 3504AH battery bank ?

nickdearing88

Could you push the majority of your computing to the Cloud or servers hosted elsewhere? I assume you will have some kind of remote data connection for monitoring, etc. It would be best to keep your equipment onsite to a minimum.

Tmp4000

nickdearing88 said:

Could you push the majority of your computing to the Cloud or servers hosted elsewhere? I assume you will have some kind of remote data connection for monitoring, etc. It would be best to keep your equipment onsite to a minimum.

There will be only one computer on site...which will control the sensor acquisition stream. We will have a small satellite internet box on site...that will be our remote comm with the site. 

The only major pieces of this system will be the need for cooling the sensors and keep the enclosure cool (it will be on the Equator, where the average outside temp is 95F) and the instrumentation power. 

BB.

At best in a warm climate, you will get an average of 77% or so of the 315 watt rated panels (including roughly 5% controller loss).

Flooded cell batteries are around 80% discharge/charge cycle. AGM run 90% or better efficiency.

You talk about 7 hours of sun, but at roughly 1,000 watts per square meter of std noontime sun, you say 5.7 to 6.0 kw per sqmtr. I.e. 6 hours of sun per day.

And add 85% typical inverter efficiency (include 3% or so wiring loses).

Get some dust on the panels, etc., the loses really grow.

So you have good numbers for the sun by month?

- Bill

Estragon

Assuming you get ~10kw net of daytime loads, 10,000/58a charging gives you ~ 175a. At 50%SOC you need to put back ~1800-2000ah into a 3500ah bank. My arithmetic says you'll need about 9-10 hours of full sun, plus a couple hours of absorb time.

mike95490

Tmp4000 said:

Well, this will be a learning curve for our team. We will be installing the entire system...we won't hire an expert. I asked our hardware folks and they gave us more 'relaxed' numbers:

Air Conditioning (still 24/7 operation):  1400W
Instrumentation + Computers + Peripherals = 4300W   

The location will have 7 peaks hours (@ 5.7 - 6 kWh/m^2), so the # of panels needed:

Wow, the data gear consumes 3x what the computer does ?   Any way to get THAT pared down.

Can you "super insulate" your gear shack, and make it as small as possible, for less surface area to gain heat through?

What month has your WORST sun hours ?  Monsoon month ?

137KWh daily, x5 for nominal foul weather & DoD considerations =  about 700Kwh battery
bank @ 48V = 14,271Ah battery  Wow !!

Estragon

The gear apparently produces roughly 15,000 btu, which has to be moved from inside the enclosure to outside. There will be a point at which reducing the size of the enclosure is counterproductive, depending on the delta from ambient to conditioned space, the r-value of the insulation, and convective vs conductive heating, etc.

Still, 4500w seems like a lot of juice for a remote atmospheric monitoring application, and reducing that would change the air conditioning load as well.. Seems to me there may be ways of rethinking the loads before thinking too hard about accomodating them.