Off-Grid system in Haiti

KeithWHare

Last April, I helped install a solar electric at a clinic about 10 miles east of Port au Prince Haiti.

I'm going to describe what we did, then you all can tell me what we could have done better, what is likely to break first, etc.

12 Kyocera KC130TMPV Modules in three strings
Each String

• Volts 48 nominal
• Amps 7.39
• Watts 520 maximum
• Short circuit 8.02 Amps
• Open Circuit Voltage 87.6

Midnite Solar MNPV6 Array Combiner

• 3 lines in, 10 Amp DC breaker/line
• Delta LA-302DC DC Lightning Arrestor

Power Out of the array combiner

• Volts 48 nominal
• Amps 22.17
• Watts 1560
• Short circuit 24.06 amps
• Open Circuit Voltage 87.6

63 Amp DC Breaker & 63 amp 150VDC DC ground fault protector
(installed in a Midnite Solar E-panel)

Outback Flexmax 60 MPPT Charge controller

63 Amp DC Breaker

Midnite Solar E-panel, with a Delta LA-302 AC Lightning Arrestor

12 Trojan T105 6 volt batteries in three strings
Each string:

• Volts: 24
• Amp Hours: 225 (20 hour discharge)
• Total Amp Hours: 675 (20 hour discharge)

Each battery string is fused with a 110 Amp T-fuse. The strings are combined with Ilsco LDB-26-350s. The battery cables are 2/0 welding cable between the batteries and the combiner block and 4/0 between the combiner blocks and the E-panel

The inverter is a MagnaSine MS4024 (4000 watt, 24 volt)

The AC output side of the inverter is connected into an existing breaker panel. The building is on a large compound that had originally had a fairly large diesel generator (long gone). The nearest utility power pole is a mile up the road. Even if it was possible to get a utility line to the building, the utility power in Port au Prince is intermittent at best. So, this is off-grid.

The load? I wish I knew for sure. Best guess at this point is:

• 10 or so 120 volt compact flourescent bulbs of unknown wattage
• One laptop
• Some sort of wireless internet connection (WiMax, I think)
• Some number of 120 volt fans
• 5-10 cell phone chargers
  
• A 24 volt Sundazer 50L DC refrigerator

Eventually, we will need to power:

• several additional computers
• a small autoclave
• Eye exam instruments
• Who knows what?

The clinic operates about 4 hours a day, five days a week.

We were able to get The Kyocera panels, the MagnaSine inverter, the T105 batteries, wire, and conduit in Port au Prince. We carried in the Outback charge controller, the E-panel, breakers, etc. in our luggage.

Each string of solar panels is wired with #10 stranded and is connected to an Air Conditioner disconnect on the roof. The array combiner is mounted inside the building. We used #8 stranded wire from the air conditioner disconnects to the array combiner. From the array combiner to the e-panel, we used #6 stranded.

The DC refrigerator is approximately 50 feet from the e-panel, wired with a 10 amp DC breaker, and #8 stranded.

The E-panel is grounded to four five-foot ground rods at five-foot intervals in a trench outside the building. We couldn't get solid ground wire in Port-au-Prince, so we used #4 stranded with the insulation removed. The frame grounds from the solar panels are connected through the bus bar in the E-panel.

The solar panels are mounted on the roof of the clinic at approximately 18 degrees, facing slightly east of south.

The Midnite Solar E-panel made things really easy. We really got moving on a Monday afternoon, after a 4x8 sheet of plywood finally showed up, and had everything put together and working by Thursday afternoon. I have one day's worth of data from the Friday we were there. I was really relieved that it all worked as we flew out on Saturday.

In the next week or so, I should have the last 90 days of data from the Outback.

I will post several pictures when I figure out how.


Keith

KeithWHare

Re: Off-Grid system in Haiti

Here are the solar panels mounted in a welded angle-iron frame, with masonry to reduce the chance of movement. There is about a 4 inch gap in the front and at the top in the back to allow for air flow.

KeithWHare

Re: Off-Grid system in Haiti

The E-Panel with the Outback charge controller.

The batteries, in a really nice battery rack.

KeithWHare

Re: Off-Grid system in Haiti

On the first full day (April 23), by 12:38, we had generated 4.4KWH. By the time we left at 4:00, we had generated about 8KWH. The day was mostly sunny, but with some high scattered clouds.

The power output from the panels was pretty close to the panel specifications, which gives me some confidence that the system is going to do ok.

BB.

Re: Off-Grid system in Haiti

Hi Keith,

Boy--not a bad first post from a "Newbie" :D...

Some observations---

KeithWHare wrote: »

12 Kyocera KC130TMPV Modules in three strings
Each String

• Volts 48 nominal
• Amps 7.39
• Watts 520 maximum
• Short circuit 8.02 Amps
• Open Circuit Voltage 87.6

Vmp for the KC130 is 17.6 volts--So Vmp-array would be 4x17.6v=70.4 volts (at STC conditions)

In hot water, the Vmp will fall (upwards of 20% for ~100F no wind day). However, you are running a 24 volt battery bank--so no issues there with "low array voltage on hot days".

Midnite Solar MNPV6 Array Combiner

• 3 lines in, 10 Amp DC breaker/line
• Delta LA-302DC DC Lightning Arrestor

I looked at one of the KC130 data sheets, and did not see a series protection fuse listed... I think the 10 amp breaker is a bit conservative and I would guess the proper fuse/breaker should be 15 amps.

You may get nuisance trips (no system damage from trip--other than missing charging for your battery bank) from a 10 amp breaker on very clear/sunny days (just a guess--not sure). The nearly identical replacement panel calls out 15 amp fuse.

Each battery string is fused with a 110 Amp T-fuse. The strings are combined with Ilsco LDB-26-350s. The battery cables are 2/0 welding cable between the batteries and the combiner block and 4/0 between the combiner blocks and the E-panel

The inverter is a MagnaSine MS4024 (4000 watt, 24 volt)

If you ever pop a battery string fuse, you might want to look at 150 amp or 250 amp fuses per strings... Battery banks do not always share current well and a MS4024 inverter at full power can take:

• 4,000 watts * 1/0.85 eff * 1/21 volt batt cutoff * 1.25 NEC safety margin = 280 amp rating for wiring/fusing.

Towards that end--I would suggest the next installation use a 48 volt battery bank or a maximum of ~2,400-3,000 watt inverter at 24 volts. Large inverters on lower voltage battery banks draw a huge amount of current (heavy wiring, fuses, switches, etc.).

2/0 wiring is good for around 190-280 amps, 4/0 good for 300-380 amps or so.

It would be nice to leave a DMM and/or DC Clamp Amp/Volt Meter behind for maintenance and trouble shooting (if it does not get lost).

Batteries strings can be difficult to get to reliably equally share current flow... Two to three strings in parallels for a battery bank would be my personal suggested limit (also keeps the number of cells to check electrolyte levels reasonable). They do make large 4 and 2 volt cells now so you can get down to one or two parallel strings--but many times you have to make do with what you have...

Did you connect the battery bank following something like this to balance resistance (and balance current flow)?

http://www.smartgauge.co.uk/batt_con.html

What instructions/tools are there to monitor the battery bank state of charge (hydrometer, DMM, Battery Monitor)?

For flooded cell batteries, a hydrometer check once a month to ensure all cells are charging properly and check a "pilot cell" more often to ensure the bank is not discharged below 50% state of charge very often and gets back above 95% state of charge at least once a week or so, etc... All the normal battery recommendations (hard to do if you are not there).

Also, I am a big fan of Battery Monitors (Victron Energy also has a nice battery monitor). Battery Monitors are not perfect--but they do a nice job of displaying battery state of charge as XX% without having to use a hydrometer or measuring resting voltage. I think they are very useful to have non-technical folks monitor battery capacity (and when to cut power and/or start the backup genset to recharge the bank).

Also, you did not say anything--but did you get a Remote Battery Temperature Sensor for the Outback Charge Controller?

The load? I wish I knew for sure. Best guess at this point is:

• 10 or so 120 volt compact fluorescent bulbs of unknown wattage
• One laptop
• Some sort of wireless internet connection (WiMax, I think)
• Some number of 120 volt fans
• 5-10 cell phone chargers
• A 24 volt Sundazer 50L DC refrigerator

If you have 120 VAC 60 Hz--A Kill-a-Watt meter is great for measuring this loads. Cheap and accurate. If you have 220 VAC 50 Hz, then the Euro Version of the Kill-a-Watt may be worth hunting down.

The 120 VAC fans may be a killer for your system... Typically any electro-mechanical device consumes a fair amount of power. You may have to look at the air moving devices and figure out if there is any better way (i.e., use less power) to keep people cool.

The battery monitor will help a lot in estimating the daily power your guys are using and how much charging is being done by the solar system.

Eventually, we will need to power:

• several additional computers
• a small autoclave
• Eye exam instruments
• Who knows what?

Until you get a better idea of how much power you have available--Any additional loads should be planned with great care.

Computers should be laptop or netbook type computers that have been measured with a kill-a-watt meter to ensure low power usage (computer set to power saving mode, etc.).

An autoclave--I would suggest looking at simple non-electric autoclaves that can be heated from a stove (wood, oil, gas, etc.).

The DC refrigerator is approximately 50 feet from the e-panel, wired with a 10 amp DC breaker, and #8 stranded.

That is probably the maximum you would want to run 10 amps on 8 awg wire... 50' (one way voltage drop calculation) gives you about 0.8 volt drop.

Using the PV Watts website to calculate how much power you can generate with your solar system... Assuming Guantanamo Bay Cuba has similar weather, 1,560 watts of solar panels, and the system is around 0.52 efficient (from solar panels to AC power out):

"Station Identification"
"City:","Guantanamo"
"State:","CUB"
"Lat (deg N):", 19.90
"Long (deg W):", 75.22
"Elev (m): ", 17
"Weather Data:","SWERA"

"PV System Specifications"
"DC Rating:"," 1.6 kW"
"DC to AC Derate Factor:"," 0.520"
"AC Rating:"," 0.8 kW"
"Array Type: Fixed Tilt"
"Array Tilt:"," 19.9"
"Array Azimuth:","180.0"

"Energy Specifications"
"Cost of Electricity:"," 2.9 peso/kWh"

"Results"
"Month", "Solar Radiation (kWh/m^2/day)", "AC Energy (kWh)", "Energy Value (peso)"
1, 5.50, 122, 3.54
2, 6.32, 127, 3.68
3, 6.41, 143, 4.15
4, 6.41, 137, 3.97
5, 5.72, 124, 3.60
6, 5.48, 114, 3.31
7, 5.94, 128, 3.71
8, 6.05, 131, 3.80
9, 5.85, 123, 3.57
10, 5.57, 119, 3.45
11, 5.44, 115, 3.33
12, 5.22, 116, 3.36
"Year", 5.82, 1497, 43.41

So, for September, we would expect an average of 123 kWH per month of useful AC power or:

• 123 kWH / 30 days per month = 4.1 kWH per day

Or from your charge controller (panels+charge controller use 0.77 derating):

• 4.1 kWH per day * 0.77/0.52 = 6 kWH per day long term average

Now--This 6 kWH per day from the charge controller assumes that your batteries and loads can consume 100% of the charge controller's output--But in reality, the batteries should be fully charged by the afternoon and therefore the charge controller will be cutting back on charging current as the batteries approach full charge.

So, at this point, it sounds like your solar array + charge controller is operating pretty well... And you need to understand how your batteries are being treated so that they can have a long life instead of an early death.

Impressive job!

-Bill

KeithWHare

Re: Off-Grid system in Haiti

Here's the block we used to combine the battery strings. I did not attempt to balance the connections.

KeithWHare

Re: Off-Grid system in Haiti

BB. wrote: »

Hi Keith,

Boy--not a bad first post from a "Newbie" :D...

Thanks. But can you tell which parts are because I obsessed over the books and specifications for a couple of months, which were because I took my brother-the-electrical-contractor with me, and which worked out by chance or the grace of God?

Vmp for the KC130 is 17.6 volts--So Vmp-array would be 4x17.6v=70.4 volts (at STC conditions)

In hot water, the Vmp will fall (upwards of 20% for ~100F no wind day). However, you are running a 24 volt battery bank--so no issues there with "low array voltage on hot days".

The reason for the 24 volt battery bank was the the MagnaSine MS4024 was available in Port au Prince. I wanted the solar panels to be 48 volts to reduce wire size. About half way down on the airplane, I figured out that this would also allow the system to start charging earlier, charge later, etc.

I looked at one of the KC130 data sheets, and did not see a series protection fuse listed... I think the 10 amp breaker is a bit conservative and I would guess the proper fuse/breaker should be 15 amps.

You may get nuisance trips (no system damage from trip--other than missing charging for your battery bank) from a 10 amp breaker on very clear/sunny days (just a guess--not sure). The nearly identical replacement panel calls out 15 amp fuse.

To get to 10 amp breakers, I used the maximum of 8.02 Amps Isc * 1.25. Since this is really a minimum, 15 amp breakers would be better.

If you ever pop a battery string fuse, you might want to look at 150 amp or 250 amp fuses per strings... Battery banks do not always share current well and a MS4024 inverter at full power can take:

• 4,000 watts * 1/0.85 eff * 1/21 volt batt cutoff * 1.25 NEC safety margin = 280 amp rating for wiring/fusing.

Towards that end--I would suggest the next installation use a 48 volt battery bank or a maximum of ~2,400-3,000 watt inverter at 24 volts. Large inverters on lower voltage battery banks draw a huge amount of current (heavy wiring, fuses, switches, etc.).

2/0 wiring is good for around 190-280 amps, 4/0 good for 300-380 amps or so.

My original intent was to use a 48 volt inverter, but the 24 volt one was available there. This size inverter is 60-65 pounds, which is problematic for the Airlines 50 pound luggage limit, so a reasonable inverter in Haiti was significantly better than a perfect inverter in the US.

But I agree, larger fuses on the battery strings would be better.

In general, the available information on configuring battery banks is a lot fuzzier than the information on the rest of the system.

It would be nice to leave a DMM and/or DC Clamp Amp/Volt Meter behind for maintenance and trouble shooting (if it does not get lost).

I did not leave a DC Amp/Volt meter. I'm still thinking of what to put in trouble-shooting/maintenance documentation for the system. It would be easier if I was actually living with the system.

I did send down a Magnum ME-BMK two weeks ago that someone installed for me. (the shunt was already in the E-panel, so it was an easy install.) I've also been looking at an ARPS do-hickey that works with the MS4024 to upload data. Of course that means running network cables, because the walls are too thick for wireless.

Batteries strings can be difficult to get to reliably equally share current flow... Two to three strings in parallels for a battery bank would be my personal suggested limit (also keeps the number of cells to check electrolyte levels reasonable). They do make large 4 and 2 volt cells now so you can get down to one or two parallel strings--but many times you have to make do with what you have...

I've looked at battery specs and though that a single string of larger batteries would be nice, but the T-105s were available.

Also, take a look at the picture of the battery rack. We found that outside the building embedded in concrete. Someone broke it out of the concrete and we brought it inside. It was a much better rack than the new one we'd purchased, and it exactly fit 12 T-105s.

Did you connect the battery bank following something like this to balance resistance (and balance current flow)?

http://www.smartgauge.co.uk/batt_con.html

See the picture I posted of the combiner block in a previous post. I didn't see an obvious best way to connect the cables.

What instructions/tools are there to monitor the battery bank state of charge (hydrometer, DMM, Battery Monitor)?

For flooded cell batteries, a hydrometer check once a month to ensure all cells are charging properly and check a "pilot cell" more often to ensure the bank is not discharged below 50% state of charge very often and gets back above 95% state of charge at least once a week or so, etc... All the normal battery recommendations (hard to do if you are not there).

Also, I am a big fan of Battery Monitors (Victron Energy also has a nice battery monitor). Battery Monitors are not perfect--but they do a nice job of displaying battery state of charge as XX% without having to use a hydrometer or measuring resting voltage. I think they are very useful to have non-technical folks monitor battery capacity (and when to cut power and/or start the backup genset to recharge the bank).

I took a hydrometer with me, but it got broken.

I'm hoping that the ME-BMK we just added will help.

If I got things correct, I configured the MS4024 to cut out at about 50% charge. The 24 volt refrigerator should cut out at about 10% charge.

Also, you did not say anything--but did you get a Remote Battery Temperature Sensor for the Outback Charge Controller?

Yep, we installed a remote battery temperature sensor for both the Outback and the MS4024.

If you have 120 VAC 60 Hz--A Kill-a-Watt meter is great for measuring this loads. Cheap and accurate. If you have 220 VAC 50 Hz, then the Euro Version of the Kill-a-Watt may be worth hunting down.

This is 120 VAC 60 Hz. I have a Kill-a-Watt meter someplace, here.

The 120 VAC fans may be a killer for your system... Typically any electro-mechanical device consumes a fair amount of power. You may have to look at the air moving devices and figure out if there is any better way (i.e., use less power) to keep people cool.

Fans and computers are the things I worry about the most.

The battery monitor will help a lot in estimating the daily power your guys are using and how much charging is being done by the solar system.

Real data? I salivate over the thought of having real data.

Someone is supposed to be bringing me pictures of the last 100 or say days of data from the Outback charge controller. This will help.

Until you get a better idea of how much power you have available--Any additional loads should be planned with great care.

The doctor who runs the US side of this project is learning that she can make me twitch by suggesting additional loads.

Computers should be laptop or netbook type computers that have been measured with a kill-a-watt meter to ensure low power usage (computer set to power saving mode, etc.).

The trade-offs are low power consumption and reliability in a hot potentially dusty climate. I've been picking at the edges of the computer issue, but don't have any conclusions yet.

An autoclave--I would suggest looking at simple non-electric autoclaves that can be heated from a stove (wood, oil, gas, etc.).

Thanks. I hadn't found that yet. A propane burner would be nice from an electrical point of view.

...
Now--This 6 kWH per day from the charge controller assumes that your batteries and loads can consume 100% of the charge controller's output--But in reality, the batteries should be fully charged by the afternoon and therefore the charge controller will be cutting back on charging current as the batteries approach full charge.

I'll post whatever data I can get when I get it.

So, at this point, it sounds like your solar array + charge controller is operating pretty well... And you need to understand how your batteries are being treated so that they can have a long life instead of an early death.

Impressive job!

Thanks.


Keith

KeithWHare

Re: Off-Grid system in Haiti

The next question for this system is pumping water.

There is an existing well about 200 feet from the corner of the building where the electrical is. There is currently a 1.5hp 240 volt submersible pump in the well at an unknown depth. The well is probably less than 50 feet as there is also a hand pump.

The idea is to pipe the pump directly to the 100 gallon (or so) tank on the roof of the clinic.

The question is, what is the best way to do this?

The options I've thought of are:

1. Add a 120-240 volt transformer and use the existing pump. I'm pretty sure the system will run the pump, but I'm not positive it can start the pump.
   
2. Add a 120-240 volt transformer and replace the existing pump with a smaller 240 volt pump
   
3. Replace the existing pump with a smaller 120 volt pump, but then we have to be more concerned with line loss.
   
4. Something else?

Questions, Comments, Suggestions?

Keith

BB.

Re: Off-Grid system in Haiti

Keith,

Batteries are probably going to be the bane of your existence here... I will post a few links for you to look through:

Deep Cycle Battery FAQ
www.batteryfaq.org

You may want to look at some alternate battery caps like these to reduce battery water usage/acid mist on the battery tops/wiring.

Water Miser Battery Caps

And watch the battery electrolyte levels... If they are darn near exposing the plates every month--too much charging. Over charging will damage the batteries over time.

And not needing to add any water after a month--probably undercharging or "deficit charging"--Also a battery killer.

Do they have distilled water (or filtered rain water) available? Don't use well water or other mineral contaminated water--also a battery killer.

Regarding very deep cycle discharge, don't let them go below 20% state of charge--you run the chance of taking any individual cell to past zero percent discharge and actually reverse charging the cell--another battery killer.

Do you have a back up genset/AC battery charger for the system? You don't want to let the battery bank set below ~75% state of charge for days/weeks/months at a time. Sulfaphation will reduce battery capacity too (is there nothing that does not kill a battery bank--probably not :roll:).

Sizing the Genset would typically start like this... 5%-13% battery rate of charge, and load the genset to ~50% for fuel efficiency. You could load the genset higher, but you need to look at the Power Factor of the battery charger... Many/most of the present charge controllers have a non-linear current profile (sharp current spikes at the peak voltage of the sine wave) which uses more current than would be apparent from the Watts calculations... We use Power=Volts*Amps, but the "real equation" for AC is Power=Volts*Amps*Power-Factor...

Power Factor for motors, battery chargers, and even most CFL lighting can be around 0.5-0.6 or so... 1/0.5 means the current is 2x what would be expected from just a "Watts Calculation"... So, look for the VA (Volt*Amp) ratings of your genset/inverter/etc. AC sources (or kVA or kVAR).

So. for your 24 volt 675 AH battery bank, sizing the genset would be around:

• 675 AH * 29 volts charging * 0.05 rate of charge * 1/0.80 charger efficiency * 1/0.5 genset derating = 1,223 Watts minimum genset
• 675 AH * 29 volts charging * 0.13 rate of charge * 1/0.80 charger efficiency * 1/0.5 genset derating = 6,362 Watts maximum genset

And, for battery charger rating:

• 675 AH * 0.05 = 33.75 amps rough minimum charger
• 675 AH * 0.13 = 87.75 amps rough maximum charger

Also, watch the battery bank temperature. A 10C increase in temperature can cut the bank life by 1/2.

For sizing a battery bank, some handy rules of thumb. Plan on 3 days of "no sun" with standard loads and 50% maximum discharge. This seems to be the "optimum" design for most off-grid systems. Larger batteries banks are more expensive and difficult to charge/care for.

So, ideally (sometimes $$$ limits what you can actually do) is 6x your daily Amp*Hour loading.

Smaller sets will fail sooner from deeper cycles, tend to be less efficient, and you may hit limitations of Amperage from a bank when powering heavy/surging loads (well pump, motors, etc.).

And, many times people don't know their loads, so you can size your battery bank to the solar array or size the array to the available battery bank capacity (we use the 20 Hour Rating for these rules of thumb):

BB. wrote: »

There is a range of current recommended to charge a battery bank... 5-13%

So, if you had a 500 AH 12 volt battery bank, roughly the range of solar panels would be:

• 500 AH * 14.5 volts charging * 0.05 charge rate * 1/0.77 system derating = 470 watts of solar panels
• 500 AH * 14.5 volts charging * 0.13 charge rate * 1/0.77 system derating = 1,224 watts of solar panels

Remember the battery bank is sized on, again roughly, 3 days of no sun and 50% maximum discharge, or ~6x the daily load.

And, lastly, you still need to size the solar array based on what your daily loads are and how many hours of "full noontime equivalent sun" you get per day (by season).

For long battery life, I like these suggestions (and Dave has real experience to back it up):

Dave Sparks wrote: »

I learned this strategy from Dave Surrette (Rolls) in the late 70's. Pretty much the bible on how I design my systems for off-grid.

Assume that the system will never reach more than a 90% state of charge.

Try not to go below 50% SOC, ever! Complete absorption over 90% of the year.

Use the energy stored from 70% to 90% SOC for your daily cycles.
Save the energy from 50% SOC to 70% SOC for aging to get long battery life.

I know Surettes has changed their recommendations over the years but I also know they are in the business of selling batteries! If you do the above you will get 10 to 15 years on their batteries with decent maintenance.

The OP is making it complicated by mixing battery types and not really stating a lot of information that would allow decent specific advice.

Oh yea, I am really happy that Surrette and Trojan are making L16's with 1000 AH capacities @20HR. Been bugging them for many moons to do batteries less than 125LB's!

BB. wrote: »

For a typical flooded cell battery bank (we use the 20 Hour Rate for battery capacity--It seems to best represent the "average" Off-Grid system's average current draw for successful design/operations):

• C/20 (or 5%) of the battery 20 Hour Rate is a good efficient average. Also the minimum charge rate rule of thumb we use for charging a battery bank too.
• C/10 (or 10%) is probably about the most you want to consume on average. Batteries are less efficient at the higher rates and don't deliver as much power (and battery can get warm).
• C/8 (12.5%) Call it the maximum "safe rate" for continuous discharge current. Over that rate, batteries may need forced air cooling (for discharging or charging currents). 13% is also the typical maximum recommended charging current for a typical flooded cell battery bank.
• C/4 (25%) About the maximum current that you can efficiently design for with a standard battery bank.
• C/2.5 (40%) Maximum Surge Current Operation--At this point, you can start to damage/warp the plates, heat up the battery, draw down below operating voltage (even if fully charged)... Drawing this much current or more and you can have flaky inverter operation and such.

Now, the above is just from what I have read trolling around here and the web... And is a good starting point for a a rule of thumb for design.

Obviously, there are batteries designed for all kinds of applications--Car batteries are designed for high surge currents (but not deep discharge).

AGM's have very low internal resistance and at least one vendor (Concorde AGM) claims C*4 (400% current rating) for charging/discharging. Obviously, that amount of current requires a lot of copper bus bars and is very difficult to design for.

By the way, my above calculations are not near as exact as they may seem--I showed a lot of digits so you can check and understand my math...

With solar, if your numbers are within 10-20% of the "math"--you are doing really well.

And the rules of thumbs are just good starting points... You can go higher or lower in many cases--but you want to understand the issues if you go outside the rules of thumb.

For example, to light of loading on a genset will use a lot more fuel since most gensets pretty much use 50% of full load fuel flow with 50% or less loading; too much loading can overheat the motor/alternator--especially with loads like motors and battery chargers that have poor Power Factor; and yet diesel gensets want 50-60% minimum loading for long engine life; etc...

Also, you may want to look at "spares"... So you can "protect" the battery bank from over-discharge while trying to get the system back up and running.

Broken Hydrometers, and battery banks that may last 3 months, 3 years or 10 years depending on who takes care of them--All part of the facts of life.

Any way, enough typing from me right now...

-Bill

BB.

Re: Off-Grid system in Haiti

Keith,

I will also add a link to this thread:

Working Thread for Solar Beginner Post/FAQ

There are a bunch of other "solar projects" suggestions in there.

You might want to look at solar hot water as the next project. Lends itself to DIY type/locally sourced projects better than solar PV systems. And the cost in fuel savings may be a real winner for the clinic.

-Bill

mike95490

Re: Off-Grid system in Haiti

KeithWHare wrote: »

The next question for this system is pumping water.

There is an existing well about 200 feet from the corner of the building where the electrical is. There is currently a 1.5hp 240 volt submersible pump in the well at an unknown depth. The well is probably less than 50 feet as there is also a hand pump.

The idea is to pipe the pump directly to the 100 gallon (or so) tank on the roof of the clinic.

The question is, what is the best way to do this?

The options I've thought of are:

1. Add a 120-240 volt transformer and use the existing pump. I'm pretty sure the system will run the pump, but I'm not positive it can start the pump.
   
2. Add a 120-240 volt transformer and replace the existing pump with a smaller 240 volt pump
   
3. Replace the existing pump with a smaller 120 volt pump, but then we have to be more concerned with line loss.
   
4. Something else?

Questions, Comments, Suggestions?

Keith

Is your inverter 4,000W ? 1.5hp runs at about 2,250w and at least 3x that to start up. Your Transformer NEEDS to be sized for the STARTING surge, or it's magnatic field will fail, and you won't be able to start.

Do you know if you have a 3 wire or 2 wire pump (not counting the green ground wire?) 3 wire starts a bit easier. If you end up replaceing , go with a 240V 3 phase with a VFD controller, then it will be easy starting.

BB.

Re: Off-Grid system in Haiti

If you end up replacing the pump--Look at a smaller pump and a larger cistern...

Either on the roof or with a small DC pump (cheaper, will require repairs every year or three) with ground mounted cistern and/or larger roof tank (or pressurized bladder tank) for pressurization.

You can look at our host's page (NAWS) for some starting points for solar friendly pumping solutions. Grundfos pumps are really nice (and $$$) from a system point of view... They will run from solar panels directly (pump to tank during the day) or from batteries, or AC gensets or utility power.

You could end up with solar powered well pump to cistern and Solar DC powered pump for presurization... Pick the right well pump and you can use solar+battery or even your genset for backup.

Knowing more about the well (how deep, how much sand/grit) and how much water your need (peak flow, average gallons per day, etc.) will help with sizing.

The 3 wire or 3 phase in-well pump with a VFD (variable frequency drive--really a variable frequency inverter) would seem to be a really nice fit.

Finding the right VFD would be a good start (I have no mfg/source information--others may):

stephendv wrote: »

You can use a variable frequency drive (VFD) and a 240V 1.5kW 3 phase pump - this will mean NO inrush current as you can program the VFD to ramp to full speed very slowly. I use one with my 1.5kW well pump and it works great - cost about 200 Euro's.
Plus, you get to use a 3 phase pump which is likely to last longer, especially if it starts up slowly with the VFD.
So if you go this route, then you only need to size the inverter for 1.5kW continuous and don't need to worry about start surge.

mike90045 wrote: »

VFD kits:
http://www.automationdirect.com/adc/Shopping/Catalog/Drives

You will need for your 3 phase pump motor:
Controller
reactors/filters
Fuses & spares

I'd looked into this months ago, and by the time I had all the parts selected, I went back to a conventional 1/2 hp pump and standard controller. Can't even tune it up with a run cap - the start relay won't deal nicely with it. But it works.

stephendv wrote: »

I'm using one of the low cost iDrives: http://www.imopc.com/products/71/idrive.html
The borehole is 110m deep.
Setup is: single phase 230V AC -> 25A breaker -> 2.2kW (max) iDrive -> 1.5kW pump.

Mike, for my application, the only additional part I needed was the breaker or fuse. Didn't need a controller or reactors. The iDrive is a generic VFD to control any 3 phase motor load, so it doesn't include pump-specific controls, but it does have some input sensors which I connected to a water level sensor. When the well is empty, the drive will automatically shut down the pump to prevent dry running.

Not sure if the iDrive is available in the America's but a quick google throws many results for "variable frequency drive".

There are also models built specifically for pumps that might have some useful features:http://www.adlee.com/pump_variable_speed_drive.htm

I used to have one that could control pump speed based on a programmed water pressure threshold, so it could be used to automatically provide a constant pressure source - was triple the price of the iDrive though.

The above are from this well pumping thread...

-Bill

BB.

Re: Off-Grid system in Haiti

Oh, and I forgot to add the Trojan User Guide link (PDF Download).

-Bill

KeithWHare

Re: Off-Grid system in Haiti

mike90045 wrote: »

Is your inverter 4,000W ? 1.5hp runs at about 2,250w and at least 3x that to start up. Your Transformer NEEDS to be sized for the STARTING surge, or it's magnatic field will fail, and you won't be able to start.

The inverter is 4,000 Watts, with support for surges to various levels.

In my calculations, I've been using 746 W/horsepower for electrical motors, but in rereading that's the output equivalent, not the input requirement.

The starting surge is looking uglier.

Do you know if you have a 3 wire or 2 wire pump (not counting the green ground wire?) 3 wire starts a bit easier. If you end up replaceing , go with a 240V 3 phase with a VFD controller, then it will be easy starting.

Looks like two wire + green (which is probably ground).

I've attached a picture. The structure on top is a hand-pump.

Keith

KeithWHare

Re: Off-Grid system in Haiti

BB. wrote: »

Oh, and I forgot to add the Trojan User Guide link (PDF Download).

-Bill

Yep, that's a useful read.

I put together a book of all the documention of the pieces we installed and left it with them. Unfortunately, the documentation and specifications are in English, and some in Spanish. I wasn't able to find anything in French, or in Haitian Creole.

Keith

KeithWHare

Re: Off-Grid system in Haiti

BB. wrote: »

...
Do you have a back up genset/AC battery charger for the system? You don't want to let the battery bank set below ~75% state of charge for days/weeks/months at a time. Sulfaphation will reduce battery capacity too (is there nothing that does not kill a battery bank--probably not :roll:).
...
-Bill

There is a small generator, a 1000 watt Honda. We had it hooked up the charger side of the inverter for a while before we got the solar panels hooked up. On the MS4024, I had to throttle charging to 10% (400 watts, I think) to keep from dragging the generator down.

Charging this battery bank with a 1000 watt generator is a bit better than emptying a swimming pool with a dixie cup, but not by much.

That's why I set the inverter to disconnect when the batteries drop to about 50%. The 24volt DC refrigerator will keep going until the batteries drop to about 10%, but the refrigerator uses approximately 114 Watt-hrs/day so it should be good for a while. Since the refrigerator is for vaccines, insulin, lab reagents, etc. it is more critical than lights.

Keith

KeithWHare

Re: Off-Grid system in Haiti

BB. wrote: »

You might want to look at solar hot water as the next project. Lends itself to DIY type/locally sourced projects better than solar PV systems. And the cost in fuel savings may be a real winner for the clinic.

Any warm water is likely to be incidental to having a black water tank on the roof of the clinic.

Running water is a luxury. Hot running water is way down the list.

But you do have a good point about locally sourced projects.

On this installation, we worked with a group of Haitians including two guys who do welding, metal, and electrical work, and a translator who was a med student until the medical school collapsed. The translator/med student liked running the hammer drill we were happy to let him.

The electricians/welders were pretty sharp. We didn't have to a lot of common verbal communication to communicate the electrical stuff. The goal was to have someone there who had some idea of how the pieces worked together.


Keith

mike95490

Re: Off-Grid system in Haiti

running batteries below 50% will really shorten their life. set the disconenct for about 70%.


With losses, motors consume about 2x theory, several of us have found. My 1/2 hp , 3 wire, single phase, 240V pump @ 165', is right at 1,000W according to the xw6048 inverter readout.

System2

Re: Off-Grid system in Haiti

What my brother was kind enough not say in the post about the hydrometer is that I was the one who broke it. There were not to many supply stores to go out and buy a replacement hydrometer.

BB.

Re: Off-Grid system in Haiti

Bruce,

We have had more than a few threads where broken hydrometers have popped up...

Around 28 hits in Google so far (are you 29?).

Cariboocoot wrote: »

Good!

Here in NA there's lots of those "floating needle" types - although they are more common for anti-freeze - and they aren't much good. The glass tube and bobber type are far more accurate, regardless of make.

If you stick with whatever one you start with you'll be okay, as essentially you're measuring relative SG - relative to your particular 'fully charged with new batteries' mark.

Another care tip: lay them down flat when not in use. Standing them up on the rubber fill tube in a glass jar is a formula for broken hydrometer; the tube will crack. But a lot of people do that. (Okay, it took me three hydrometers before I learned this about 40 years ago. )

-Bill

KeithWHare

Re: Off-Grid system in Haiti

I now have 100 days of data from the Outback mx60 charge controller. The summary is:


Date, Avg Amp Hours, Avg KWH, Total Amp Hours, Total KWH
September 1-6 2010, 198.2, 5.3, 991, 26.4
August 2010, 193.5, 5.1, 5998, 154.1
July 2010, 220.6, 5.9, 6838, 182.0
June 2010, 221.9, 5.9, 6656, 175.9
May 29-31 2010, 211.0, 5.6, 633, 16.8


The peak production was 8.2KWH (June 14) and the low was 3.3KWH (June 3).

I don't have usage data. The best I can say so far is that the maximum Absorb time during the 100 days is 63 minutes.

I don't think the system is being used very heavily yet.

Keith

KeithWHare

Re: Off-Grid system in Haiti

The attached picture is the charge controller statistics from September 1, 2010.

Note the peak output current of 63.0 Amps.

The breaker on the charge controller output is a Midnite Solar 63 Amp DC breaker.

Should I be worried?

Keith

AntronX

Re: Off-Grid system in Haiti

Don't worry, you should be ok. You have 1560W of panels, the peak was 1690W that produced 63A. This was due to a cloud passing by. If breaker never tripped yet, it should not in the future.

BB.

Re: Off-Grid system in Haiti

One of the things to remember is to watch the battery bank capacity... Hydrometer, resting voltage, Battery Monitor, etc...

Many solar charge controllers are known for being 5-10% (or even 30% at times) optimistic regarding their "true" power / amp output...

So, monitoring the battery state of charge is the important thing to mind. Many people "deficit" charge their battery banks (too little charging, too much loading) and can kill their batteries in a couple of years or even a handful of months.

-Bill

KeithWHare

Re: Off-Grid system in Haiti

Google Earth has updated aerial photos of Port au Prince Haiti area.

You can now see the PV panels we installed on the roof of the clinic at:

18°31'0.49"N 72°12'9.04"W

If you use the time-line feature, there is a set of images from either November 7 or November 8, 2010 that show the PV panels better.

Keith

System2

Re: Off-Grid system in Haiti

hi guys. first timer here. wow thats a great job there in haiti you did. i just got back from there yesterday. kinda new at this stuff but i love it. working on powering a 500 watt radio station about 4 hours drive from pap. near henche. 99.5fm. quick question- we have the exact set up as far as batteries. how did you figure the total amps? i know when doubling the volts the amps stay the same. and when doubling the amps the volts stay the same. 24 volts. 6, 6, 6, and 6. doubled. i get 900 amps when i add the 225 for each 6 volt. i know there is a simple thing im not seeing here, im just not so hot at math. is there an online calculator to do this for thinking of different set ups with the t105's? thanks! God bless and merry Christmas

mike95490

Re: Off-Grid system in Haiti

Batteries in parallel, add the amps

Batteries in series, add the volts

icarus

Re: Off-Grid system in Haiti

Just to be clear. If you are measuring loads Volts X Amp =Watts If you double the circuit voltage (with the same load) you reduce the current (amps) by 1/2) but the net watts used stays the sam.

For example, if you are using 1 amp at 120 volts that would be ~ 120 watts. That same 120 watt load would draw ~ 10 amps at 12 volts. Of course the big reason to use higher voltages is that the current carrying capacity of the wire stays the same regardless of voltage so that you can carry more watts so to speak given the size of the wire.

For example, a # 14 wire will carry ~15 amps at 120 volts or 1800 watts.
The same wire, carrying 15 amps at 12 volts will only carry ~180 watts.

In the case of batteries, double the batteries and you double the net amount of power available, sort of no matter how you wire them. For example, 2 6 volt 250 ah batteries wired in series will yield 500 ah at 6 vdc. Wire them in parallel and they will yield 250 ah at 12 vdc, the same amount of power net/net. (not counting any wiring loses.)

Hope this clarifies things.

Tony

KeithWHare

Re: Off-Grid system in Haiti

Helicoptermike,

The battery bank I described in the first post in this thread is 12 T105s, configured in three strings of 4 batteries.

Each string is 4 batteries in series, so the string is 24 volts (4 * 6 volts) and 225 Amp hours (at the 20 hour discharge rate).

The 3 strings are then combined in parallel, so the battery bank is 24 volts and 675 Amp hours (3 * 226 Amp hours).

Earlier in the thread, there's a picture of the strings of batteries, and a separate picture of the combiner block we used.

Keith

FL SUN

Re: Off-Grid system in Haiti

Great thread here with lots of info. I have learned a lot from the details posted by everyone 's participation here. Thanks to all.

The only thing I can add is that I noticed some conduit or pipe sticking out of the ground on the East side of the array. If that is still there, the shadow line will cause great loss of harvest potential as it travels across the array throughout the day.

If it has been eliminated or already mentioned in another post already, please disregard.

Merry Christmas everyone, btw.

KeithWHare

Re: Off-Grid system in Haiti

FL SUN wrote: »

...
The only thing I can add is that I noticed some conduit or pipe sticking out of the ground on the East side of the array. If that is still there, the shadow line will cause great loss of harvest potential as it travels across the array throughout the day.

I completely missed those two pipes when we were placing the frame for the PV panels. It turns out that they run through the concrete wall down to the breaker panel. I didn't want to cut them off short because then if water backed up on the roof, it would have been able to run down into the breaker panel.

The shadow from the taller piece of conduit did not hit the panels in April. I will be there in two weeks and so will be able to tell if it impacts the panel at this time of year. I probably should take a couple of PVC end caps with me.

Keith