The 3 phase pump motor needs the voltage and frequency ratio to match as you change speed.  Meaning if you have a 208VAC motor rated at 60Hz and you wanted to run at half speed, you would feed the pump motor 104VAC at 30 Hz.  At 1Hz you would feed the motor 3.47VAC.  

Why not design the wind turbine to provide this exact ratio of voltage to frequency directly to the pump motor.  If the wind speed is slow, you may have problems pumping against such a large head.  However, if the wind was blowing hard the pump would have all the power the turbine could provide with no loss in any voltage conversions.

Because you do not want to "store" water rather than electric power this may be a way to not need the VFD or batteries all together.

Flux is clearly correct on all accounts.  

I would not even consider a centrifugal pump for 170 meters of head.  Positive displacement only makes sense.

Also correct is the problem of directly connecting a 4 meter wind turbine to a 5-7.5HP pump.  (Side question, did you check to be sure 5.5KW is big enough?)
My limited understanding of wind turbines shows that 4 meters is too small for that size pump.  Unless you can increase wind turbine size, I agree with Flux that connecting 2+ turbines to syncronize AC outputs is doubtful.

"Feeding dc to a vfd seems simple in theory but you would need a fair bit of circuitry to hold the volts within limits. You will also need a supply for the drive control circuits."

This is spot on.  The circuitry to regulate the DC Bus voltage is tricky but doable.  The key would be its speed of response.

"It is likely that you will be forced to use a standard unit and modify it. Otherwise starting from scratch I think a vfd could be produced to accept a dc rail that rises with the motor speed just as the motor will accept raw ac as long as it varies in voltage and frequency in the correct ratio ( which it does in this case)."

15 years ago we had VFD's that had a DC Bus voltage which varied in amplitude and we regulated the output frequency to the motor based off this voltage....sadly I filled dumpsters with many of them.  I haven't seen one in 8 years.  They were called PAM (Pulse Amplitude Modulated) drives verses todays PWM's.

My opinion, but if you want to stay simple, and electric, then FLUX's suggestion is the perfect solution "Seriously worth thinking about using more smaller dc pumps"

You could rectify the output of all your wind turbines, connect them in parallel, and feed it down to a Positive Displacement DC well pump(s).  No inverters, no batteries...simple solution.
Remember the wind turbine works off the affinity laws just like the pump does...sounds like a perfect match.

I also agree with Oztules & Dinges posts, water pumper mills are proven, and easier to maintain by most people.

I think the DC pump system would also be easy to maintain.

The question of sustainability is an important one. The wind-electric turbines would be produced in Afghanistan, and the installations would be by Afghans, and backed by private enterprise. I think the wind-electric systems are more versatile, and also seem to be cheaper.

(Let me know if I've gotten something wrong with my numbers.)

One other option for the pumps to simplify things once you have achieved about the 65 Meter depth is a low voltage well pump available from Shurflo.
http://www.northerntool.com/webapp/wcs/stores/servlet/product_6970_200332019_200332019  Is just one vendor, posted because the specs are there.

The aspect of running well pumps from an efficient inverter system is a specific aspect of a book I have been working on that has taken the most time.  the block diagram is easy and simple, but finding a simple low frequency oscilator with a sine wave output that has excellent stabillity....  With the discontinuation of availability of the "MAX038" chip, the solution is not so easy, and as time has been at a premium for me these days....  

BUT the matter is not insurmountable, but would be very useful right now if I had finished this portion of my book.  A solution may be ultimately as simple as/ or to be found in the schematic of one of the newer energy efficient frontload washing machines for concept- they use a 3 phase motor with PWM and a bit of variable frequency to ramp up the speed, and ramp it down.  Having scrapped out one of mine already for mechanical failure, I may take a look at those parts for a better answer.  Hysteresis losses (internal heat) are not as great when doing this with a three phase pump motor- Electrolux is using just PWM square wave as the drive method for their washing machines. I may be giving away part of my book here, but such schematics can be gleaned and adapted from simple motor control devices covered in Industrial electronics books.  

From the sounds of what is available over in Afghanistan from this thread so far, they may have to build the circuitry they need, which is why I raise this.

Maybe fieldlines could get in on the act...set up a donation box on the front page so giving is easy, maybe as part of a new poll...

Some of the other sites I frequent have donation boxes for various things, from "we need a new server" to cancer research. Some sites run multiple drives so if you don't feel like giving to one cause, maybe you'll give to another.

Good point on the motor backup for water pumper sucker rod mills. I lived in a house that had a mill that fed a tank and in no wind times you could disconnect the pump rod arm from the mill and connect it to a "pump jack" that ran off a PTO. Took a few minutes to swap but worked well.

Darned Alshiemers is kicking in, I guess or I would have mentioned it before as a hybrid option.

I am not sure I buy the 300 foot max for sucker rod pumps, either. They are positive displacement so I figured they could be pretty deep. I seem to recall one place a friend had that had a 600 foot deep well but thats long ago and I am not sure.

TomW

Contact: IRC