Single Phase Inrush & Generator? 1

[sjohns4]

I'm working on a water supply project in Haiti. The pipes, pump, and tank are the easy parts.
Our power source is going to be limited to a portable generator that operates 2-3 hours every couple of days.
The submersible well pump I've selected is a of European manufacturer with approximately a 3 HP motor.
From the spec sheet: 3450 RPM, 230V, 1-phase, 17 FLA, 5.6A at operating condition, 610% FLA inrush.
The 610% inrush is the problem given the generator limitations.

First, I'm getting beyond my knowledge base as most of my experience has been with 3-phase motors.
My basic question is: to overcome the starting current that may stall the generator, is it possible to utilize a bank of capacitors? If so could someone direct me to a resource that would provide some guidance in sizing such capacitors?

Otherwise, are there any other options short of a larger generator? Given that it's a submersible motor I don't know we have access to winding connections beyond the L-L-N conductors in the submersible cable. Also, given where we're at, I'm trying to keep the components simple - converting to 3 phase with a VFD or similar would be a last resort.

Thanks for any info!

Mike

 

[itsmoked]

Hmmm. I did exactly the VFD>3ph for my well which allowed for a simpler less expensive, and considerably more reliable, down-hole 3ph pump. It also allowed for a smooth startup and shutdown. It completely eliminated the hit on the small generator I was using.

But to answer your cap question. No that won't help you as it doesn't work that way. Use a larger generator for the most reliable and unfortunately least efficient setup. With a smaller generator your system will see less of an inrush than expected but this is at the expense of a larger voltage sag at pump motor-start. If the generator is not running other things this usually isn't a big concern.

You'd likely want about a 10HP generator for this but wait for waross to chime in as he's the dirt-lot generator guru around here.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

The absolute minimum that I would instal would be a 7.5 HP equivalent generator capacity. I have never specified a generator at 2.5 times the motor load but I have been stuck with several of them when someone else undersized the generator. We found by trial and error that a generator must be a minimum of two and a half times the capacity of the motor it is starting.
At 250% the voltage drop is such that this configuration is not suitable for any uther loads beside the motor.
For a new instal, 3 times the motor capacity is about the minimum. You can use motor full load current times three and look for a generator rated at three times that current.
I have no problem at all with Keith's 10 HP solution.
Most sizing software will tell you that even three times is too little, but I have a couple of dozen sets that have been starting A/C units for about 20 years with no issues. After about 15 years or more of successful service I ran some sizing software on the installation and found that none of the installations should be working. But the people who supply the software also sell gen sets and bigger is more profit.
I like Keith's VFD solution.
For single phase, about 50% more capacity on the starting capacitor will aid starting.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[sjohns4]

Thanks guys. I figured capacitors would be too simple. I suppose a VFD of this size isn't terribly expensive and practically a commodity item now anyway?
Any recommendations on the drive? I see a lot of AB and ABB in the water wastewater world. One of my concerns from the get go with a VFD is that there will be no local support or service. Who ever is going to operate the system will have no training beyond what a volunteer from the group (likely myself) can provide. Given those conditions I think I'd like to select a fairly basic, "dumb", and robust drive. We're on a Caribbean island within sight of the coast. I'm thinking a fiberglass NEMA 3R enclosure mounted indoors (should keep the hurricanes out of the drive), but humidity will be terrible.

Thanks,

Mike

 

[itsmoked]

Have I got a drive for you!
Check this baby out:

Screw NEMA 12!

[URL unfurl="true"]http://www.factorymation.com/Products/FM50_230V/FM50-203-N4FS.html[/url]

Do, however make sure the direct sun never hits it and do your best to minimize the ambient. I'd consider putting a tank in the booth with it so the cold ground water "air conditions" the space a little, but that's probably a bit over the top.


As for complexity. These really are pretty simple. You can set it up to S-start gently. Set the parameters up right to protect the pump, (now three phase). Also forbid reverse operation. Don't forget you need to make sure of the operational direction since the motor, being three phase, is happy to go either direction though the pump won't be happy about it and can be destroyed in seconds. With a VFD you can turn it very slowly in front of you to see. You can laminate an instruction check list that includes the correct parameter settings in the fine print.

You want to minimize the wire length between the drive and the motor. If it's more than a about 100ft you start adding an output filter.

Caribbean huh? Consider if some lightning protection is in order. Are you going to have automatic controls? I'm guessing not if you run all this with a generator. A common problem with lightning prone areas(and I'm not sure if Haiti has a lighting issue) is when a water tank is far up on some hill and takes near/direct strikes, with the well pump becoming the grounding point thru all the hardware between the two.


Keith Cress
kcress -

http://www.flaminsystems.com

 

[sjohns4]

Cable length is a problem: The location where we may be drilling the well is at the bottom of the hill, the location where we need the water is on the top of a hill at least one ridge over.

Total cable length is approaching 3,000 LF. With single phase DOL starting I was thinking of a buck-boost transformer to overcome the voltage drop.

We have to keep everything that's not buried located where the water is needed, otherwise it will be stolen.

Got anything else up your sleeve?

 

[itsmoked]

IeeeEEEeeye.

OK.

DOL would choke on this several ways also.

Let's discuss this.

How are you planning to run it all control wise, walk us thru it.

Is the well-head secure?
Pump set depth?
How is the pump start/stop envisioned to happen?



Keith Cress
kcress -

http://www.flaminsystems.com

 

[sjohns4]

IeeeEEEeeye exactly!
I figured our in house electrical engineer would trip an overload or go into melt down trying to figure this one out! Even a water/wastewater guy (me) knows we're committing too many sins to sleep well at night here! I'm committing a few hydraulically; but in a situation like this it needs to work, reliably, but not necessarily by the book, sooo...


The well head site wont be secure. I'm planning on a locking well cap, possibly buried just below grade. If we were in the US we'd just drill at the campus site deep enough to get into the water table. There are enough people around nothing should disappear. When I was there you see evidence where A LOT of stuff has disappeared. Given the visible geology I have no confidence we'll hit water within reach of Hatian drilling equipment there.

Pump depth should be fairly shallow ~20-30', but we're pushing another 130' or so up the mountain. We're drilling not far from a spring fed river in granular soil. I'm planning on drilling up on the first low bluff out of the defined flood way.

The storage tank will be within sight of the end use point. The pump & generator would be manually started when someone walks up the hill and sees the tank is low. It'd be stopped after it starts overflowing + an adequate irrigation time. There's a garden below the tank so the overflow will be ditched to provide irrigation.

Controls: about as simple as I can make it. Pray you bought decent gas, start the generator, hit the pump switch, pay attention to volts, current, etc., run until it overflows, turn off the pump, kill the generator. Pray to give thanks it worked and pray more it works again next time.

 

[itsmoked]

After some thought (walked the dog for an hour) I know what I'd probably do to crack this nut.

I'm not exactly clear on what your one way cable distance is, 3,000ft? or 1,500ft? Running 3HP at the end of either of those is a bit daunting. The problem is the dynamic load requirements. If we assume a DOL you need cable that will support a 3HP motor start, which is going to be large and painfully expensive. Trying to use a single boost transformer does not work because if it is adjusted to get the needed starting voltage down to the pump then the running voltage will be way too high. One alternative is to use two transformers and step the voltage way up and back down so the wire size isn't absurd. 30A (not even your 600%) wire 3,000ft long needs to be 350mcm or the copper in each conductor has to be about 1/2" in diameter. Six dollars a foot. $36,000 If you're talking 1,500ft it's a little better.

Ideally you would put the generator and a VFD within feet of the well head and just have them turned on remotely as this would save big bucks on wire and complexity. You could easily go wireless for the on/off signal(s). Maybe use DOL and the bigger generator too.

The next option is leave the generator "up the hill" and put the VFD at the well head. The advantage here is that the VFD monumentally reduces the starting current to below 150% instead of the 600+% of DOL. This means you can use much smaller wire on the run. Since you are supplying 1ph power you also use one less wire though they all need to be a little bigger to squeeze the same 3HP down them. With the VFD at the well you can set it to startup using only a couple percent more than FLA, keeping the dynamic voltage drop to that required to keep the VFD ramping happily on 6A capacity wire.

The above plan requires a box/vault/pillbox/cinderblock structure with steel door and appropriate locking systems to keep the contents from crawling away. While not currently the plan, perhaps it should be given some serious thought as the expense of a small structure would pale compared to the copper costs. This would allow some other niceties like protecting the well head and a safe place for pressure gauge and perhaps a water meter. It greatly mitigates the lightning issues almost completely.

If you included the generator in the structure the noise would be away from the farm - possibly a plus. In another thread this same sort of situation was done in the Philippines. The guy used a small light that could be seen for a mile to see if the generator was running. In this case you'd wirelessly start the generator then when the light was visible for, say, 60 seconds you'd wirelessly command the pump to start. Reverse the procedure for shutdown with perhaps a 3 minute cool down for the generator. I'd modify this a bit to have several colored lights (LEDs) which would show how much fuel was remaining. Green plenty, Yellow some, Red imminent empty stop pumping to allow cool down - better when the Red is on the VFD loses it's "run" signal dropping the generator load completely.

Running the VFD 1,500 to 3,000ft might be ok. As the cable between a VFD and motor gets longer the waveform on the cables can, at a "golden length" or maybe it should be called the "putrid length" will cause standing waves that result in high voltage spikes whacking the motor. Hence my suggestion to put it at the well-head somehow. However, Gunnar in here has stated that after the cables get longer than the putrid length (which is between about 200 and 800ft the problem goes away again, because the high voltage spikes are "dulled" by the large capacitance presented by super long cables. So, perhaps a tank mounted VFD is still in the cards. Maybe he'll chime in.

Otherwise, I don't really have a good solution for running a 240V 1ph pump down more than 1,500ft of wire run. A bunch of calculations for stepping up the voltage to 600V then back down to 240 at the well head, which still requires a structure again, for a DOL setup is lousy too in my estimate.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[sjohns4]

I'd say the building will probably be a necessary evil.

Rather than size the wire for 125% FLA, I was thinking I'd use a current somewhere between FLA and actual Amps at the pump operating point to keep conductor size manageable / reasonable. I'd scratched out a scenario with a 20% boost and #10 or #12 wire that looked good on paper, but not considering the witchcraft VFD's practice or ability to support the DOL inrush...

If we would keep the generator at the campus and have the VFD near the well head, would you see a problem with a pair of buck & boost transformers at the generator and line side of the drive to further reduce the conductor size?

Might need to look into a 480V generator...don't recall ever seeing a small portable though...

 

[itsmoked]

Ah, you don't want to go near a 480V generator on an novice run and maintained system. Further all 480V generators are assumed to be driving large loads and you won't probably find one less than... what, 50kW?

A pair of transformers at both ends would work and I'm not meaning buck boost but actual step up/ step down. That jumps the voltage up to 480 or 600V and hence the current drops in half or to a third. The higher voltage is bottled up between the two transformers presenting less user hazard.

I just looked back and am having a bit of a problem with pump current. If it is 17A FLA (1ph) why would one expect only 5.7A running? Considerably less than 1/3 of the FLA??? That large dynamic exacerbates the whole long wire run issue. That huge difference is suspect to me. Most motors draw about 1/2 their FLA even unloaded.

The 125% thing is trashed by starting current voltage-drop on overly long runs. You would still need some crazy big wire to pull off DOL. I suspect 125% might work for a well-head VFD but it would be better to have more, like 200% because the VFD sucks power in thru a rectifier front end. That means in little gulps at the sinewave peaks meaning the current is drawn in the peaks only. That means while the average current may be, say, 10A, the actual current is drawn in 15A pulses 120 times a second. As you can imagine, the 15A pulses will be subject to 15A related voltage drops.

I woke up this morning thinking this:
Reducing the pump size would reduce all the hassles considerably.
Any chance you could use maybe a 1hp pump and allow it to run longer? That way you could run just a few solar panels and the entire gasoline nightmare would go away. Pumping could occur for probably 7 or 8 hours a day (this is sunny Haiti after all).. Panels are pretty cheap these days. They're about 200W so you'd need 5 or 6. No fuel, no, "We forgot to check the oil". No noise. No, "Oh my gawd, where did you pour the oil you changed??!!", or "You've never changed the oil, ever?!?", or, "The generator won't start!". "Why is it knocking now?"

Ultimately your system's use and function is hostage to the generator. The pump will run years, the VFD will run for years, the wire run will last for years (if not stolen), the generator will run for maybe a year or so before it needs technical service. It can be damaged by mis-operation in minutes. There is also the ever present safety issue of gasoline and the extreme hazard of someone hot-fueling the generator. Two news articles in the last two weeks of people critically burned while attempting gasoline hot-fueling. With solar, no moving parts, little maintenance - clean them - once in a while. Harder to steal as they are not easily throw in a car trunk or directly applicable like a generator.

If the solar angle could work we can go into more detail there.


Keith Cress
kcress -

http://www.flaminsystems.com

 

[sjohns4]

I pulled the current draw from the pump manufacturer's cut sheet. I'm inclined to clamp the fluke on some of the local installations where, as typical, the motor should not be running anywhere near full load. May get back to you on that one (all would be 3-phase though).

I thought, according to NEC, even continuous motor loads you'd size the cable for 125% but to handle starting inrush you could and should upsize the overcurrent protection by up to 250% for a thermal magnetic breaker?

I'd quickly dismissed solar due to the amount of panels, battery, and inverter size. Maybe worth re-considering. I suppose first I need to see what I can do with the pump. HP is a direct function of flow and head, I could probably knock down the flow by a third and be OK, assuming the head remains constant since the pipe diameter will likely decrease, we may be in business. If the generator isn't running, the pumping time may be irrelevant too. You're right about the generator too. Gas is typically bought in a 12 ounce beer/soft drink bottle...who knows what's really in there. I've been told it's hit and miss.

 

[waross]

Hi Keith;
I suspect the 5.7 Amps running condition will at high head conditions. On start-up the pump will draw the initial 610 Amp inrush. If this creates too much voltage drop, the pump may never get up to speed. Once the pump has filled the discharge system and developed full head, the current may drop to 5.7 Amps. A check valve at the surface at the well location to keep the long discharge pipe full may help the motor if the pipe can be kept full.
For 5% drop at 240 Volts you want #1/0 AWG cable.
At 610 Amps starting current, the Amps will drop as the voltage drops until they meet at some low voltage and current lower than 610 Amps.
You won't get near 610 Amps but you won't get near 230 Volts either.
Try for 600 Volts. Or seriously look into Keith's solar solution.
BTW Diesel sets start at about 15 to 18 KVA, single phase. Above about 25 KVA at 1800 RPM you get a reconnected and de-rated three phase set.
For 3600 RPM sets single phase starts around 25 KVA and the jump to three phase is around 35 KVA. Probably lots of exceptions.
35 KVA single phase will be a derated 50 KVA set. Pretty close to your estimate Keith.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[itsmoked]

Thanks for the chime-in Bill!

sjohns4; The 125% is for normal stuff and is the minimum not for crazy long runs where the issue is the starting current. Up sizing the protection doesn't avoid voltage drop.

A longer running system allows for a smaller pump, smaller cables, small pipe diameters, and stresses the well less. Normally it would incur longer generator time as the negative trade-off.

I am just toying about here but a back-of-the-napkin solar scheme I came up with is as follows.
Note: No batteries, No inverter.

Rudimentary Recipe: (probably a few details to work out still)
1HP 3ph PUMP
1HP 230V VFD NEMA 4X version $290
Cable Run: Two only + and return DC rated to 324VDC (Use standard 600V rated insulation) You sink ground rods at both ends.
Solar panels 24V/60W (Ebay) 27"x26" 15 lbs (14) @ $119/unit = $1,666

The panels nominally add up to 324VDC when put in series. This is the DC link voltage that is delivered to the VFD's energy storage capacitor by the rectifier front-end normally supplied with 240VAC.

You deliver the power down the hill to the wellhead as 324VDC using wire that will allow only about a 5% drop at the motor's normal running load. The non-pulsating DC will allow the least loss because there are no rectifier peaks involved. The drive is setup to very softly start the pump keeping to about 110% of load current. The drive's braking resistor connection is plumbed to a large external brake resistor which is used to prevent the solar panels from over-volting the DC bus.

External Brake resistor. $28

Approximate total cost excluding pump and cable: $1,984

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

Comments on generator cool-down:
The prime reason for cool down is to allow the turbo charger to spool down. At full load the turbo is turning at maximum speed and is at maximum temperature. If you stop a generator when it has been running at full load, The oil circulation stops and the heat transferred from the turbo casing to the bearings is no longer removed by the oil. The turbo has been at maximum speed and will coast coast for quite awhile. With no flow of cooling oil through the bearings the bearings may be overheated. The last of the very hot oil may drain out of the bearings and the combination of very little, very hot oil may result in severe damage to the turbo bearings.
The issue is not as severe for a turbo aspirated generator on normal motor service as the load is often one third or less than the set rating. Cool down is still recommended.
Cool down is not important for a normally aspirated set but is still a good idea.
When either a turbo aspirated engine or a naturally aspirated engine is stopped abruptly from full load, There is a heat soak down and the coolant temperature rises. This may increase the pressure in the cooling system and stress the radiator and the radiator hoses. Failure of hoses or other radiator components is most likely to occur when a heavily loaded engine is stopped without a cool down period.
I suggest playing with the voltage drop and wire size numbers based on a voltage of 480 Volts or 600 Volts, and the VFD located close to the well.
I had a similar situation in Central America. There was a couple of miles between the water supply and the mill site.
We had a large storage tank at the mill site. When the water was getting low, some one would go to the pump location. We were pumping from a spring fed pool and did not need a submersible pump. The pump was directly driven by a diesel engine. The oil. water and fuel would be checked and then the pump would be started. The pump would then be left running on a timer. The engine was naturally aspirated and not fully loaded. We did not have a cool down time. Keep it as simple as possible.
The biggest problem was the fuel supply and the mud hole about 1/4 mile from the pump house. The boys got the Quad stuck and kept trying to drive it out until mud was sucked into the carburetor. The final solution was to walk in from the mud hole when the pump needed to be started and to hire a local farmer to deliver a couple of barrels of diesel fuel by ox cart every few months.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[sjohns4]

The generator was the primary reason I was trying to keep pump run times low as I figured the motor was the biggest wear item not to mention the locals may get tired of days filled with engine noise.

It makes sense the inrush current could be problematic for extremely long cable runs. Off the top of my head the longest one I recall putting on paper was probably 300'-400' starting a 480V, 15 HP well pump across the line. I cant remember if I had to upsize the cable due to voltage drop in that instance or not, otherwise I'm sure I'd have sized the conductors to 125% FLA? I don't recall the EE telling me I did anything stupid on his review though, and it's been pumping for 3 years now...

I'm liking the solar idea more and more. We've got a small 12V solar powered system there now powering an egg incubator. Since it has to maintain temperature 24/7 there are several panels connected in parallel through a charge controller and batteries to provide the load at night. That's the extent of my solar & DC knowledge.

I didn't know we could connect panels in series - I assumed at some point the voltage potential to ground would get too high for the panels?
So we series ~14 panels to get the nominal 324VDC. We lay two conductors from the panel array to the VFD at the well head. Can we bring DC directly into the VFD, or do we need to somehow bypass the front end of the drive? In this scenario we simply pump during the day. I suppose we could configure the drive to cut out on low voltage with an auto fault clear and restart? The incubator panels would drop out with minimal shading...I could see where periodic sky obstructions could cause multiple low/no voltage conditions throughout the day.

Well pumps typically have an integral check valve. You typically install one, or two, just outside of the well casing for pump removal. In short the operating current should be fairly constant once the system is filled, and if necessary we can choke off a valve to simulate a full pipe during filling.

 

[waross]

I worked on a small solar system in Central America. When I inherited the system there were several solar arrays and a small wind charger. The cables down from the wind charger acted as an antenna and if there was a lightning strike within a few miles, the induced voltage in the down leads would reverse bias the SCRs and then battery current would follow through and destroy the invertor. The solution was to dismantle the wind charger. (It did not supply much power in any event.)
The point is:
Be aware of and avoid possible lightning induced voltages in your leads from the solar arrays.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[itsmoked]

Bill re: cool down. I understand the turbo cool-down need but what about 5~10kW commodity generators? Any point?



sjohns4;

The generator was the primary reason I was trying to keep pump run times low as I figured the motor was the biggest wear item not to mention the locals may get tired of days filled with engine noise.

I totally agree. Very valid points.

Your pipe length and cable lay distance (one way / one conductor) is still a complete mystery to me. You mentioned 3,000 feet. Can you cough up a clarification?

Honestly from what I'm imagining, solar seems the only realistic solution in a bad-gas, poor community, realm. Otherwise sooner than, no, much, much, sooner than later there will be project derailing problems that are 'urgent' since now crops and/or a community's dependance have be established.

Pretty much the egg incubator is no different than the well. Just substitute a generator for the egg solar panel and consider the exponential pain-in-the-buttness created.

You seem to have an excellent grasp of the issues with regard to a solar solution.

--------------------------------------------------
Issues to be investigated/solved/confirmed:

The drive I linked you is specifically a single phase input 240VAC drive, hooking DC to its input should present no issues as the rectifiers will just conduct continuously.

The 1HP drive does not include an external brake resistor connection. That could still not be a problem or it could require using the 2HP version which does. Using the brake function to protect the system from high voltage buss situations during times when the pump is not needed/operating and the panels try to rise to their open-circuit voltage needs to be confirmed. Presumably 800W might need to be dissipated, so an external scheme might be needed. Alternatively, maybe unneeded energy could be put to other uses like medical refrigeration.

Protection is needed for the DC +300V. Possibly fuses, maybe a molded breaker.

The drive manual sez "Do not start or stop the AC drive using the main circuit power." I'm not sure why it cares but an alternative 'power is failing - command stop' circuit might be needed.

There might be a need for some simple logic to prevent too frequent power up/down cycles in the event of a weird cloud day.

Over voltage or under voltage faults should be settable to be ignored and not need human "clear fault" intervention.

Pump starting and stopping methods should be examined. A low voltage control cable? Wireless? Power cycling?

A capable lock-outable disconnect at the panels is required.

Perhaps additional capacitor buffering would be very useful at the drive.

Sunrise and sunset limits/control are likely needed to prevent boundary condition issues.

The panels-in-series causing panel to panel voltage issue needs further checking. Most homes now have all panels in series resulting in hundreds of volts in the strings so this isn't a problem usually. It might require a different model than the very first eBay one I came to.

The obvious panel mounting system. 15 lbs x 14 = 210 lbs. Isn't onerous. Wind loading would, of course, need to be considered.

--------------------------------------------------

These issues are all readily solvable via various techniques. It would certainly be worth the effort to set it all up including the spool(s) of wire, panels, VFD, pump, and a valve to throttle for pipe-run/head simulation. Then run the system closed loop back to the tank for a couple of days to work out any issues so there are no surprises in a remote spot. Everything could then be labeled for easy re-assembly on the site. If you want to crate it up and send it to me I'd be happy to do it for you.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

Hi Keith. I thought that I was agreeing with you on your estimate. " Further all 480V generators are assumed to be driving large loads and you won't probably find one less than... what, 50kW?"
With the distances involved I shudder at the thought of running power lines.
You seem to be well on the way to solving this challenge with solar. I hope you can build a system for the OP.
I like your approach. BTW Consider wasting some water after the storage vessel is full. Use a run time shut down with a generous setting. Itmay make the controls simpler rather than trying to run a shutdown circuit between the well and the storage facility.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[waross]

Nice work on this Keith. Probably lps time.

Bill
--------------------
"Why not the best?"
Jimmy Carter