Solar Electric System for a School

Re: Solar Electric System for a School

As vtMaps says--You cannot ignore losses and "physical" realities... For a pure off grid system, you have to just about double the solar array size because of the losses in the system.

For the AC side, the PF of 0.82 sounds like it is a bit optimistic for "average" induction motor fans and "typical" florescent fixture... Induction motors can run ~0.67 PF and florescent lighting has been as low as ~0.50 PF (cheap Compact Florescent "twisty" bulbs).

But if accurate, the AC inverters and wiring would need to be up sized by:

• 1/0.83 = 1.205 => ~20% larger...

Badar wrote: »

Well BB.

Thanks for your detailed reply. I think your way of calculation is far different from mine
here is my calculation style.

Assumptions:
1: utility company is not disconnected from school. It will be used as "back-up" power in poor weather days.

How many hours do you need to operate from the battery bank per day (worst case)? 2 hours then run generator or you don't usually have power outages on overcast days?

2: Solar panel gives 40% output at 8:00 AM Local and 100% at 10:00 AM Local in clear skies. (I have checked using meter)

So--Just to make things "easy" (I don't know when your solar noon is, and solar noon occurs earlier during winter in the northern hemisphere), lets say the power needs go away at noon--So from 1pm on (say 1pm is ~solar noon during late fall thru early spring), you are recharging the battery bank from solar (~1/2 of the day's worth of solar power)--and you can use the utility for recharging if needed (makes system "smaller/cheaper").

3: All following load will be operated between 8:00 AM to 01:00 PM Local time. Then there will be no load.

5/6/7 days of power usage?

Power Load:
APPLIANCE..................... QTY.....................Watt.....................Total Power
Fans............................ 500.....................100........................50,000 watts or 50kW
Lights........................... 100......................40.........................4000 Watts or 4kW
Total Load.............................................................................54000 Watts or 54 kW

Energy Consumed in one hour:
54,000W

AC Inverter rating:

• 54,000 Watts * 1/0.82 PF = 65,854 VA (volt*amp9) rated inverters/wiring
• 65.9 kVA * 1.25 NEC derating = 82.4 kVA rated wiring/breakers (and inverters?)

In the US/North America, wiring and breakers are designed to not be loaded more than ~80% of rated load (1/0.80 = 1.25). That is why we use the 1.25x derating factor here.

AC Amps to run this load
I=P/V= 54000/220= 245Amp/hours

• I=54,000 Watts / 220 VA = 245 Amps

Note that Amps and Watts are already "Rates" (like gallons per minute)... So, do not add the Amps/Hours or Watts/Hours -- That is something different (first derivative of Amps or Watts from Calculus, and a mix of seconds and hours units--even more confusing--Amps is Coulombs per second, Watts is Joules per second).

DC Amps required to run this load @ 48V
=Ac amps * 5= 245 * 5= 1225 DC amps/hour

Don't ignore the loss factors--In solar they all add up to "kill" your project numbers (like the 52% solar panel to AC power output derating).

• 54,000 Watts * 1/0.85 inverter derating * 1/48 volt battery bank = 1,324 Amps at "nominal" load

Battery bank required for 2 hours back
=DC amp/hours * 2= 1225*2=2450 Amp/hours @ 48v

"Typical" Flooded Cell Lead Acid batteries typically are best operated at ~C/8 rate of discharge--You may push this to C/5 if batteries are kept cool. AGM can handle much higher discharge rates but are also much more expensive. For the moment, to stay "conservative", lets use C/8 rate:

• 1,324 Amps * 1/(C/8 discharge rate) = 10,592 AH "nominal" 48 volt battery bank.

Solar Panels requirement
CS5P-260M (CanadianSolar) gives max 5.27 Amps.
Assuming that it will give 2.5 amps at 9:00 AM.
Panels required=DC AH/Panel output= 1225/2.5=490 Panels
Panels in watt= Qty of panels * power of panels= 490 * 260 = 127,400 Watts of panels required.

For Flooded Cell battery charging, you are looking at a (typically) maximum of C/8 nominal charging rate (our ~13% cost effective maximum rate of charge). Note that a major source of battery losses is they discharge at 48 volts and recharge at 59 volts--P=I*V. Lead acid batteries are near 100% efficient regarding AH of charging/discharging--However, they are around ~80% power efficient.

• 10,592 AH * 59 volts charging * 1/0.77 solar panel+charge controller losses * 0.05 rate of charge = 40,580 Watt array minimum
• 10,592 AH * 59 volts charging * 1/0.77 solar panel+charge controller losses * 0.10 rate of charge = 81,159 Watt array nominal
• 10,592 AH * 59 volts charging * 1/0.77 solar panel+charge controller losses * 0.13 rate of charge = 105,507 Watt array "cost effective maximum"

This is another issue with Lead Acid Battery sizing... They have a limit to the maximum amount of current you can pump down them for multiple hours before they overheat... It takes a larger battery bank to absorb higher charging currents (or active cooling).

A 127,400 Watt array is probably "close enough" that it will not cause major problems. When the loads "go away" at solar noon, the output power begins to fall too.

Costing:
Batteries' cost: 1,440,000/- * 4.3x larger battery = 6,192,000 PKR
Panels' cost:....12,740,000/-
Misc...............5,000,000/-
Total..............19,100,000/- = 22,492,000

Break-even Calculations:
Total load: 54kW/h..............................= 54 units/hour
Unites consumed/day...........................= 54*6=324 units/day (actually 5 hours per day => 270 kWH per day?)
Units consumed /month........................=324*25= 8100 units (=> 6,750 kWH per day--6 school days per week?)
Expected monthly bill on this load...........=Unit consumed * Unit rate = 8100* 22= 178,200/-/Month (6,750 * 22PRK=148,500 per month)
Yealy payment for electricity of this load.=178200*12=2,138,400/- (148,500 per month * 12 moths=1,782,000 PKR/yr)
Keeping in view, rise in prices, expected break-even would be in 4-5 years for all this installation.

22,492,000/1,782,000 PKR per year = 12.6 years

Add new batteries every 5-15 years, new electronics every ~10 years, and your cost of "solar power" is about the same costs as avoided electric costs--With the down side of a 22,492,000 up front cash payment on install... You can use Present Value/Future Worth calculations (based on assumed inflation), but at a first cut, this is not a "risk free" return on investment (replacement batteries, parts also go up in price over time... Training/maintenance for the system are also real costs).

Batteries are a real killer in any off grid power system. That is why "net metering" (using the utility like a giant battery bank) looks really nice. Net metering can also be a huge cost for the utility (they have to maintain the infrastructure and 'buy power' at retail when the sun is up, and sell power at retail when the sun is down)--Not really a good investment for them and why governments have to push utilities to do net metering and why the extra costs are passed on to the rest of the customer base.

The numbers are very fluid at this time--And you can see how one change in assumption makes your payback 2-3 times longer (and approaches never saving money) very quickly.

Making accurate measurements, reducing power usage/increasing efficiency are (generally) your first priorities. In the US, you can get large savings in costs if your loads can be "interrupted"--In your case, you already have the generator. But, they may interrupt your power anyway and not really look at reducing your billing costs (local customs/laws).

In the US, commercial customers with "poor" power factor can get charged a whole lot--More or less 1/PF (in a round about way, commercial customers are charged for kVAH, and not KWH.

In theory, with new fans and lighting (or at least using capacitors to "correct" the motor PF), you could save upwards of 10-20% on your power bill if you have some form of kVAH (or kVR) charging.

If you are not an electrical/power engineer--Perhaps you can partner with a local university electrical/power engineering department and make this a class project/thesis (at least for the initial project study/paper design).

-Bill

Re: Solar Electric System for a School

Badar,

My answer (and it has gotten me into trouble many times before ) to your "problem" is to change the rules.

My guess, if you try to do things the "traditional way" (as documented above with panel+charge controllers+batteries+AC inverter+etc...), you will--At best--Pretty much break even when compared to the cost of power from the utility.

My suggestion, look at each load and see if and how much you can do things non-traditionally.

First the lighting. LED's have made a huge change in the way light is generated and made large changes in the way lamps work (respond to voltage/available power). I would suggest you get a few commercial LED fixtures and do some experiments.

For example (websites are examples, I know nothing about the sites or their products), look at AC LED lighting.

http://www.rabweb.com/product.php?product=FFLED18

Many fixtures are rated from 120 to 277 VAC--But in reality, may work on 120-277 Volts DC or (factor of srt(2)x) of 170 VDC to 391 VDC. You have to contact the mfg or even just buy one or two and take one apart (second for testing/spare) and see what the input circuitry is.

Many modern devices use some form of electronic switch power supply input that do not need AC voltage--But instead will work just fine with DC voltage (perhaps a higher voltage by the square root of two). This even applies to many/most computer power supplies.

In the "old days", everything used a transformer to convert from 120/230 VAC to 12 volt or other low voltage AC before rectifying for use in the device. But transformers use "lots of copper" to make that happen. With the low price of electronics and the high price of copper, AC only transformers are being replaced by AC/DC capable electronics.

Also, with LED lamps, they maintain their (relatively) high efficiency and similar color temperature even at reduced input power (note, LED's optimally are constant current lamps--they need a fixed current regulator, not a fixed voltage regulator/source). Unlike filament bulbs which go towards "red" at lower power levels or florescent lamps which need AC power to operate reliably (they are Arc Lamps and a DC current will usually destroy the electrical elements inside the tube very quickly). While you could make electronics to change DC to AC for florescent, I would choose the LED just to make things "simpler".

Another issue with LED lamps is they generate less heat than filament bulbs, and can be slightly more efficient than florescent lamps--Especially if you can use less lamps for "focused" lighting (i.e., more lights on the black board for the teacher, and less light on the students desks. Or task lighting just lighting desks with less flooding on walkways and storage areas. And if you need occasional lights (library book shelves), you can use motion sensors to only turn on lights/zones when occupied (LED lamps can be turned on/off hundreds of times per day, filament and especially florescent will fail an early life if power cycled a lot).

Another neat thing with LEDs, since they are small--You can sometimes mount them in ways that the heat does not enter the room (install outside and shine through a window, or recessed into a ceiling and avoid heat into the room/living-working space. Room lighting can be a major source of heat into the room, and reducing heat into the room, can reduce heating loads too... However (again using very rough numbers) an adult will add about 100 watts per person to a room's heat load--So a class room with 30-50 students may give you 3,000 to 5,000 watts of heating--More than the heat from lighting--For a school, heating from lighting may be less of an issue (and in many high desert climates, heating from lighting can significantly heat a room on those cold/clear desert mornings (no clouds to reduce amount of radiation of heat back into space--forget "global warming" from CO₂--That does not stop/slow back radiation to space).

So, for lighting, you may be able to (with even commercially produced/lower cost LED Fixtures) simple put 6-12x ~250 Watt panels in series (6x Vmp~30 volts would give you 180 VAC nominal) and enough power for ~60x 22 watt LED florescent fixtures just from sunlight. That would light X number of rooms per array.

Assuming you get direct sunlight when the school day starts, you will have some light when classes start. You could tilt panels more towards morning sun (raise at an angle, face south east) and have a few less lamps because of less sunlight/energy in early mornings.

If you really needed, you could put a high voltage DC Battery bank in there too to supplement solar panel power in the AM/cloudy weather (i.e., ~15x 12 VDC batteries) if needed (I would try my best to avoid batteries--They are a whole pile of issues to get right and you still have expense of replacing batteries very X number of years).

Since you still have AC power (grid/genset), and in many areas of the world, morning utility power cuts are less common, you could perhaps skip batteries completely and just have backup AC power in parallel with the solar arrays (transfer switch AC to DC supply that can parallel with solar array).

There is also a fall back to make a similar system with low voltage DC (12 volts) and standard M16 LED conversion lamps. These LED lamps are defined/designed to work with 12 VDC power for track lighting, but many of these bulbs will work just fine with 12-18 VDC too:

http://www.ledlight.com/12-volt-led-track-lighting.aspx

Same process as above, but use lower voltage solar panels (100-140 watt typically with Vmp~18 volts) and smaller lighting "units" (i.e., only 1-10x 100 watt panels in parallel) per "micro power" installation.

I believe that the above will not be that difficult to implement (getting samples of various LED fixtures/lighting from different sources, taking a few apart to see how they work, and some experimentation with a DC power supply, meter, small oscilloscope, etc.)--And you can work out the details to build small systems, each capable of lighting a few rooms each (rather than building a central solar lighting power station with a large solar array and sending energy to all the rooms). Smaller systems are "less" complex and can be (if done right) have less issues with safety (large/high voltage power systems require professional electricians to install/maintain, large battery bank are heavy/full of acid/and electrical accidents waiting to happen too).

You have not said if you where an electrical/electronics engineer/trained or not. Without some good knowledge of electronics, doing the above may be a bit frustrating/require a few months (or couple of years) of self training. However, if you can get somebody with electrical/electronic education, none of the above is that difficult to do the basic research. And the amount of supply's instrumentation is something that could be done with stuff found in a typical school electronics lab.

Your thoughts?

-Bill

Re: Solar Electric System for a School

Badar,

I like your idea of breaking things into separate projects--At least for analysis, if not actual implementation. It makes the problems clearer and, potentially, easier to solve as each problem may need a different solution.

Fans are the next issue... A typical fan uses an AC Induction Motor. Those types of motors are pretty rugged (and when used in fan type applications) are pretty hard to damage short of hitting with over voltage (a fan, with lower input voltage will simply turn slower--A motor driving a fixed load, like the compressor in a refrigerator, will draw more an more current as the voltage falls--Brownouts will burn out motors used in these types of applications).

So--Again, change the rules. no 230 VAC at 50 Hz (or whatever your power standard is)--Look at different motors and at "non-traditional" power.

The simplest motor for your needs would be a "brushed" or "Universal" motor (such as you see in an electric hand drill, the starter motor in a car, etc.). These types of motors can be directly connected to a solar array, sometimes with electronics for higher efficiency/better air flow in mornings/evenings, very cheaply and easily. However, brushed motors will need brush replacement every ~6-12 months and some other issues which make them less than ideal--So forget these motors for this application.

The next are PM DC motors (permanent magnet) that have electronic commutation (instead of brushes). The are cheap, reliable, and can take variable DC power/voltage from a solar array very nicely. These fans are used in just about every small computer made these days.

The problem may be trying to find this type of motor (PM with electronic commutation) in the size and fitted to fans for air circulation in a class room. If you can find DC Fans like this--Just connect to a solar array + backup battery (or AC to DC power supply) and away you go. China is making a huge amount of "non-traditional" power devices and they may have things like this... I will look later today, but I am running out of time right now.

Another way is with induction motors, but different types than you probably have today. There are three phase induction motors (very common, but perhaps not so easy to find in fractional/sub 1,000 watt ratings). These motors can be driven with a VFD (variable frequency supply) very easily from AC power, and there are VFD's that appear to be relatively easy to convert to DC input). Basically, the VFD is a type of DC to AC inverter, but it outputs 3 phase (rotating field to turn motors) and variable frequency (10-50+ Hz) to adjust motor power RPM needed for application or based on amount of DC power from solar panels available.

And another variation is use the VFD with split phase/capacitor start motors... These motors have an external capacitor that "alters" the single phase input to 2 phase AC for the motor (to provide the "rotating field" needed to start the motor rotation). These types of motors are available in smaller sizes and may be easier/cheaper to make/retrofit into fan applications.

In any case, read about VFD+Solar powered water pumps. Similar solution for a bigger application vs yours.

BB. wrote: »

Some discussions about VFD (Variable Frequency Drives)... Basically a variable frequency inverter with (typically) three phase output. Used to soft start motors (handy for 3 phase well pumps, or pumps with well head starting capacitor) and can also turn an AC motor into a variable speed motor (very handy for pumping applications).

WELL PUMP and Inverter QUESTION
Wind/solar for large scale pumping etc (out of my depth!)
could use knowledge - using Gould jet pump - transfering from 230vAC to ? DC (new link/thread 10/27/2012)
Help required to design off grid system (information on possibilities to connect "standard VFDs direct to solar panels) (new link 1/13/2013)

-Bill

The last link is from user "syedbukhari" -- He is from Pakistan and based on our conversation here has implemented a VFD+solar array water irrigation system, and (I think) is turning this into a business... I would suggest you try contacting him via "Private Message" and see if he can help you with the VFD/Fan questions--And/or even the LED Lighting. If he does not see your PM, let me know and I will send him an email directly. Let us know how it goes with him.

Regarding the computers--I will leave that conversation for another day... There are "interesting" methods to supply solar power--But the basic solar array+battery bank+AC inverter is probably the best solution. Again, conservation is going to be a big help here--Computers are actually quite heavy users of electric power and, obviously, they need stable power to operate reliably.

-Bill