RVAT starting question for large well motor 3

[bimr]

I have a new installation of a water well pump rated at 450 HP, 2300 V. We are having problems getting the pump going. The pump starter (RVAT) fails on pump acceleration and shuts down. The shutdown internal by the controller is set at 5 seconds.

Utility power is 12,000 Volts. The utility then has a 480 V KVA transformer to power the entire site. In the building, a 500 KVA step up transformer is used to step up the power to 2400 V for the well pump. This in turn is connected to an RVAT.

I also have a 750 KVA 480 V. generator that is not completely installed, but will be finished shortly.

Several questions:

1. Why won’t the pump start?
2. Are the transformers (s) too small?
3. Is the utility transformer too small?
4. The electrical engineer stated that the system gear that is used is 480V was because of the high cost for 2300 V gear. Is that as valid statement?
5. Is the pump starting time too short?

Thanks

 

[burnt2x]

As Keith and dpc said, you need to secure needed data (from vendor, mfr or local experts in your area). Your weak electric system compounded to the problem of your long motor acceleration time.
Just make sure you don't exceed the number of starts allowed so that you don't waste your pricey equipment.
To compute for the acceleration time of your pump, you need these data:
1) total rotational inertia of your pump and motor,
2) motor full-load torque,
3)motor breakdown torque,
4)motor locked-rotor torque, together with
5)rated horsepower of pump and rpm.

             total rotational inertia x pump rpm
Time (sec) = ----------------------------------------
             308 X (ave. mot. torque-ave.load torque)

where:
                   Rated pump HP X 5252
Ave. Load Torque = --------------------
                    pump rpm X 3
                    (FL torque + BDwn torque)/2 + BDwn torque + LR torque
Average motor torque =--------------------------------------------
                                           3

 

[bimr]

The motor manufacturer says that the motor should start in 0.8 secs at 80% voltage.

The current also does not drop off significantly until the motor reaches about 90% of synchronous speed.

 

[dpc]

Right, but you may not have anywhere near 80% voltage AT THE MOTOR. You have about 15% voltage drop across one transformer then the autotransformer is taking another 20% off that. The key is the voltage at the motor terminals.

At 80% voltage, you have 64% torque, at 70% you have 49% torque and at 60% voltage you have 36% torque. Lower torque means longer acceleration time.

 

[jraef]

The motor manufacturer says that the motor should start in 0.8 secs at 80% voltage. ...

That is only the motor, not the motor plus the load. Remember who is saying that: the motor manufacturer has no idea what the load is, their statements are always about the motor alone (unless you contracted with them to determine the starting time with all available load data).



"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[rockman7892]

"The motor manufacturer says that the motor should start in 0.8 secs at 80% voltage.

The current also does not drop off significantly until the motor reaches about 90% of synchronous speed."

Wow 0.8 sec seems really quick for a motor to get up to speed. Are you sure you didn't mean 8sec. As dpc said this may only be for the motor by its self and the time will most likely be longer once you factor in the load inertia. If the motor supplier does not have this information then you can request a speed vs torque curve from the equipment manufacturer for your load and ask the motor manufacturer to factor that into his acceleration time calculation.

If the motor does indeed only take .8sec to get up to full speed then I suspect that with the load factored in the acceleration time is more than 5sec because the current is not dropping. This also has to do with even more of a voltage drop as dpc is suggesting.

Once you know the total acceleration time with all factors such as motor inertia, load inertia, and volage at the motor terminals during the start, you can plot this acceleration time and find out where the motor reaches 90% speed. If the 90% speed point is over 5sec then you will need to increase the timer on your relay our it will never see the current drop take place between 85-90%

 

[bimr]

The well pump manufacturer says that the pump will be up to full speed in 0.8 sec. This is because the well pump is a skinny column pump and therefore low inertial mass.

 

[dpc]

You should check with the motor and pump suppliers to see if they are OK with making a transition to "full" voltage before the current comes down, recognizing that the actual voltage at the motor will be much less than 100% of nominal.

That may at least allow you to verify that the motor can reach full speed and that there is not some issue with the motor and/or pump.

 

[bimr]

I received that information from the well pump (and motor) manufacturer. The well pump manufacturer actually prefers 100% line voltage for starting. The well pump manufacturer stated that the well pump (and motor) does not need the RVAT. The RVAT is something that the utilities require because they expect a 6:1 d ratio of starting amp to running amp.. This particular well pump has a 4:1 ratio of starting amp to running amp.

So the utility is requiring an RVAT in order to get the 4:1 ratio. But, the well pump will already provides the 4:1 ration

The well pump manufacturer is saying that the well pump will start with 80% of the rated nameplate voltage of 2300V. I am thinking that the well pump will now start if the RVAT is reprogrammed to transition to line voltage in 1-2 seconds.

 

[dpc]

Seems to be where things are headed. Basically, you don't really need the RVAT starter so the faster the transition the better.

This is another reason why utility "rules" regarding RV starters are so pointless.

 

[itsmoked]

bimr; I curious. What is the approximate physical size of this down hole pump/motor?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[burnt2x]

That large voltage drop made your acceleration time long enough for the multilin to trip! I tried finding a similar data of your submersible pump and my calculations yielded 1.5 - 2 seconds run-up (if the voltage at motor terminals are within specs).

 

[rockman7892]

It sounds like since the motor will accelerate completely in a couple of seconds you will be able to transfer within 1-2 sec and still have a reduced starting current.

One of the things that I struggled with in setting up the transfer time for my motor was when to make the switch from reduced (80%) to full voltage. From everything I researched the most efficient time to switch was when the motor was motor was around 90% speed or else any earlier would be like a RVAT was not even used and there would't be a decreased starting current. I'm not exactly sure why this is, but I believe it has something to do with the current drop-off.

One of the things that complicates things is the longer time that 80% voltage takes to reach 90% speed and how this time compares to the rotor damage curve. For my particular motor the acceleration time was about 10s. With 80% voltage this acceleration time would increase and if I waited to switch at 90% speed I would bump into the rotor damage curve due to this long acceleation time and the associated current. This kept causing an overload on my multilin so I had to switch the transfer time to a sooner time in order to avoid the damage curve. Moving this transfer time sooner does not produce the complete reduced current that you would expect because of switching less than 90% as mentioned.

To complicate things it appeard that I had a motor that was giving a higher LRC than was on the datasheet. The datasheet showed a LRC of 256A and even with a reduced voltage start of 80% I am seeing 312A. Current is increasing rather than decreasing with reduced voltage and I am talking with the motor manufacturer trying to determine what is going on.

 

[bimr]

The Contractor was able to start the well pump using the onsite emergency generator. This is a 1000 kVA at .8 PF generator that is capable of starting a 4,200kVA motor.

Since then, the well pump will now start using utility power. Prior to the emergency generator start, the well pump would not start on utility power.

Nobody seems to be able to explain this.

I do have another question. Should the incoming voltage be expected to drop when starting a well pump motor like this?

 

[dpc]

Yes.

 

[bimr]

OK, this site has a nominal 480 Volt service from the utility. The voltage is typically around 495V to 500V.

What would you expect the voltage to drop to? Is 10% too much of a drop?

At what point should someone go back to the utility and ask for a bigger service?

 

[dpc]

Unless the utility is getting complaints from other customers, it just depends on whether or not the voltage drop is causing problems, and how often the motor is being started.

A momentary 10% dip should not be a major concern for other motors, provided their running voltage was at nameplate voltage or above prior to the motor start.

One benchmark: NEMA requires that motor starters stay pulled in down to 85% of rated voltage. Below that, there is a risk of contactors dropping out or chattering, which is generally a big drag for everyone involved.

 

[NeonJohn]

hello guys. I just joined this site so that I could comment on this thread. I'm a retired nuclear utility engineer who for the last 15 years or so has moonlighted in my friend's electric motor repair shop and served as his field troubleshooting consultant.

I'm cringing at this thread. The central problem that no one has pointed out yet is that there is no central technical authority controlling this startup. This person should be an engineer who knows motors, pumps, utility practices and electronic controls. Trying to rely on a series of manufacturer's tech support reps' advice is a good way to burn down this installation.

- Bimr, I should point out that if this is a standard TEFC surface mounted motor driving the pump through a long shaft, the value of that motor is in the $20,000 range. If it's a submersible pump, double that amount. The starter, transformers and other apparatus are valued at least that much again. You're taking great risks with some very expensive hardware and with almost no hard data to work with.

- I don't understand the electrical architecture, first stepping the voltage down and then back up. Your system engineer's statement that a 480 volt transformer was less expensive than a 2300 volt transformer of the same kVA rating is certainly odd and contrary to my experience. Here, within a wide range, price is based on kVA regardless of the voltages involved.

- 500kVA transformers are borderline too small. If the motor is 92% efficient, that's 365kW. Figure a 0.9 pf and that works out to 405kVA. That gives you very little headroom for inrush or overload. They're probably workable but everything else is going to have to be just right. The utility transformer probably has the overload rating to handle the situation but the dry one may not unless it is also utility-service rated.

- You have essentially no test data to analyze nor to make decisions from. At the minimum, I would instrument the motor for voltage, amperage, wattage and RPM feeding into a data acquisition system. Ideally, I'd also instrument the output of the transformer feeding the motor upstream of the autotransformer, the utility transformer secondary voltage and amperage and utility primary voltage. You can probably tap the utility's revenue metering potential transformer for the primary voltage. I'd have thermal sensors on the autotransformer and the motor windings if the motor doesn't already have a sensor built in.

With as much magnetic hardware as is involved between the utility and the motor, you probably already have "reduced voltage starting" without the auto-transformer. No way to know without instrumentation. This sort of instrumentation is fairly inexpensive to rent by the month, especially when compared to damaging the motor or other apparatus.

- I do not believe the manufacturer's claim that the motor can accelerate in less than a second, even without an attached load. I've never seen an unloaded motor of that size spin up that fast in our shop.

- Has anyone done the inertial calculations on this setup yet? Especially if this is a surface mounted motor with a long shaft drive, there is a LOT of inertia involved. Depending on at what interval the shaft is supported, the start-up torque could cause it to wind up, bow and whip, adding a LOT of inertial loading. If that is happening it would be evident in the RPM and kW traces on the data acq system.

- Is this a totally new design or has one similar already been started up and operated in the area? The question I'm interested in here is whether there may be something wrong with this particular setup or whether it's an erroneous design.

- Bimr, I see that your contractor "got it running" by hitting with with a bigger power source (reminds me of a certain utility client - "hit 'er again and let's see what blows next".) That it now starts (for the time being) on utility power is easy to explain. The initial run bedded-in the bearings and seals and significantly reduced both the starting and running power required. If a 90 deg gearbox is involved then it required significant bedding-in power.

This is almost always the case for large machinery. In fact, our shop has a variety of large motors available for lease so that a larger than design motor can be used to do the initial bedding-in of the rotating machinery. Our most frequent customers are rock crusher and cement kiln operators.

- Even though the pump is running now, that does not mean that it is out of trouble. It is at its optimum point in its life. sliding surfaces like seals are bedded in, ball bearings are clean and new, the pump impeller is free of crud build-up and cavitation wear. Power demand will likely rise as normal wear and tear happens.

- Even though it's running now, I'd still want to instrument it and find out how much headroom I had to accommodate system deterioration. I predict very little. The cost of renting the instrumentation and a consultant to operate it is minor compared to what even one rewind of that motor would cost. Having presented that data to management NOW may save your job when something does happen in the future.

- I'd want to know the service factor on that dry transformer. If it is 2 or above you're probably OK as far as long-term loading goes. If it is close to 1 (1.1, 1.2, etc) then while it'll probably handle the load now, it'll run hot and will have a shortened life.

Though it's probably too late in the project to change things now, were I designing this system, I'd have around a 750 kVA 2100 volt transformer feeding the motor control panel, with a smaller 2100 to 480 or whatever feeding other smaller loads.

One last comment. As someone else mentioned, pay REAL close attention to the maximum start frequency specification of that motor. Motors that size typically are limited to 3 or 4 starts an hour.

I've seen WAY too many motors come into our shop where that specification was not respected. The process demanded more starts and stops (a clutch would have probably been appropriate) than the motor could withstand. The interior looks like it has been in a burn-out oven. Everything is burned to a char. Somethings there is molten aluminum laying around, slung from the rotor as it melted.

We generally charge half the retail price of a motor for a rewind when one is possible. Rotor melts send the remainder of the motor to the salvage yard, as do very bad in-slot arcs.

If this motor has a thermal sensor (most all motors that size do), make SURE that it is connected to the proper protective electronics. If your management doesn't want to do that then at least CYA by putting the advice in writing.


John DeArmond

 

[bimr]

John,

Appreciate your interest. You have many valid points.

1. This project was designed by a subcontracted Electrical Engineer. The Electrical Engineer left the firm between design and construction phases and there is little support available at the old firm. It is difficult to find a replacement engineer in a short time. You are correct in saying that a project engineer would have been useful.

2. The EE says that the RVAT is necessary to start the 2300V well pump using the emergency generator. I don't know if that is true. I also asked why he did not get a 2300V generator and his stated that all of the other equipment such as the transfer switch would be much more expensive and require special workers to repair. The special workers are required because of the higher voltage. He also said that the 2300V equipment would require larger footprints such as 12' in front of the cabinets. I don't know if this is true either.

3. This is a direct drive submersible pump with 15 stages. The pump manufacturer has actual test data showing the pump starts in less than 0.8 seconds.

4. This a new installation. This is a municipal well and will typically run 6-8 hours per day, with only one start per day.

5. The utility guy sized the incoming transformer at 500kVA. I don’t know for sure what the quality of the transformer is. I asked for a shop drawing and he did not supply it.

6. When the initial attempt to start the pump failed with utility power, the utility guy stated that if the pump would start with the emergency generator, he would change the transformer from a 500 kVA to a 750kVA.

7. The well pump is running at 2300V and 114 amps. That is basically the design conditions and is also pumping out the design water flow at the design water pressure.

8, This is a 17” dia. 450 HP 4-pole NEMA D motor.

 

[JIMGEN]

I agree that the secondary transformer may be too small.
But I would like to see the motor data that states a 450hp motor will accelerate to full rpm in .8sec.

We build machines with motors from 300hp to 800hp , 480volt motors. All are soft started. We currently set the soft starts at 30 seconds with current limitation of 300% of inrush. The soft start can be set up for voltage ramp, current limit or pump mode. The ramping time can be as long as 240 seconds. We set the ramp at 30 seconds for voltage ramp or current limit. The last machine on current limit of 300% of inrush I thought the unit got to full 1800 rpm in about 10-12 seconds. Our machines have much more inertia than a turbine pump. It is true that low voltage machinery is much lower priced than medium voltage. For that reason we will use two 480v motors on a machine to supply 800hp.
However I am familiar with turbine pumps up to 200hp with water columns of up to 300ft. Of course, there is no check valve since the turbine is immersed below the static water level. But as I recall the motor typically took about 4-5 seconds to get to full speed. (across the line starting) And it was to full speed before the water arrived at the surface of the well.

It is real hard for me to believe the .8 seconds of start time.

 

[waross]

The emergency generator started the pump and now the utility supply can start it.
From that I would draw two conclusions.
1> The emergency generator setup supplied a higher voltage to the motor terminals than the utility supply for some reason.
2> Something was sticking the rotor and it was not turning or turning very slowly. That obstruction has now been cleared by the higher voltage supplied by the emergency generator.

The utility is assuming 6:1 starting current. Your motor draws 4:1 and the 500 KVA transformer will reduce that as will your service transformer. The impedance of the motor leads and feeders will also introduce a voltage drop and reduce the starting current. I would present the data to the utility and ask for permission to start DOL. I would also consider setting the transition time below one second.

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