Submersible Pump Oddities

[jraef]

I have some experience with submersible water pumps, but not as much as I would like, especially in light of this recent experience. Maybe someone with some more experience can offer some suggestions as to whether my observations / conclusions were off base of not.

 

Application is a 10" well pump, 100' deep with 20' of water, so 80 of head (pumping out onto the ground at this point). It is a Byron Jackson submersible pump, 150HP 460V 60Hz, 195FLA. Maximum locked rotor time is 10 seconds cold, 5 seconds hot.

 

Control system is a Robicon 200HP VFD with a soft starter as the backup / bypass controller in case the VFD is being serviced or faulted. A Multilin 269 relay is protecting the motor regardless of being run by the VFD or the Soft Starter. Multilin was set with a Class 5 OL curve (later tested with a Class 10 to see if we could do it).

 

This pump replaced a previous 200HP pump of different manufacture that failed after 4 years. The old pump had run fine on the VFD, but when they went to test it using the soft starter, they failed to realize the phase rotation may not be the same as the VFD. It wasn't, and the nut on the bottom of the pump unscewed, dropping the pump in the well and severing the cable causing a short circuit. The old pump was apparently recovered, but damaged beyond repair. Enter this new pump. The customer decided that they didn't need as much flow at the top of the well as they had with the 200HP pump, hence the reduction in size for the replacement.

 

At startup with the VFD, the pump would not acccelerate even with the VFD in constant torque (linear) V/Hz pattern. The Robicon VFD has a way of detecting actual rotation, and by boosing the current limit to maximum or the VFD (remember it was a 200HP VFD but now a 150HP pump), they were able to make the shaft turn about 30% speed with 450% FLA, but it would not accelerate any further. Most people at this startup (except me and the Robicon rep) were convinced that the VFD was faulty, so we tried it with the soft start. I gave it only 1 second of ramp time and no current limit (600%) to simulate an Across-the-line start. The motor immediately went to 370% FLA and stayed there, but there was no flow out of the pump. It tripped out on Overload in just under 9 seconds. The Multilin history showed the current as being exactly the same in each phase prior to tripping.

 

The pump rep swears that he had "spun the motor by hand" prior to putting it down the hole, thereby eliminating the possibility of a mechanical jam. We checked the motor leads and windings with a megger and they were fine, exactly the same in all phases. Upon investigating the motor design, we were told by the manufacturer that the Locked Rotor Amps of this motor was 723A, coincidentally 370% of FLA! So the motor was drawing LRA when across-the-line, yet there was no flow.

 

My conclusion was that since we were developing LRA with being across-the-line, and the VFD was forcing the motor to output above LRT before it showed any motor rotation, there is something mechanical in that motor/pump assembly that is binding as soon as any appreciable torque is applied.

 

The VFD rep is convinced that the new motor is unable to develop enough torque under any electrical circumstances to lift the water column. He is using the 370% LRA as proof of this point, because it is such an unusually low number. I disagree, because failure to lift the column would not prevent the shaft from rotating, it would just sit there at full speed with no flow, not at locked rotor. It could go to locked rotor with full open head, unrestricted flow, but not with no flow.

 

The pump rep insists that there is nothing wrong with his pump / motor, because he could "spin it by hand". I doubt this is reliable, because I think that the only time he would have had direct access to a shaft to be able to spin it by hand would have been prior to assembly, so something could have gone wrong after that point. The thing that bothers me about this is that the Robicon VFD did indicate that with extreme overtorque, it did rotate slowly. If it were jammed, I doubt it would do that.

 

Was I on the right track? If so, what could make it spin by hand, yet lock up with torque applied? Anyone ever run into anything similar?

 

And what about that 370% LRA? Isn't that oddly low or is it common place forthis industry? The pump mfgr said it was because it was a long skinny motor design to accomodate going into that 10" well casing. I fail to see how that would increase impedance to that extent.

[marke]

Hi Jeff

 

It is very late here and I have an early start in the morning, but I will make a few comments on the way to bed.

 

 

• The Locked Rotor Current for submersibles is always less than conventional motors because of the long thin rotor. The resistance of the rotor bars are significantly higher.

• The Locked Rotor Torque of a submersible pump motor is generally higher than a conventional motor because of the higher rotor resistance.

• Typcially, the start current on submersible pumps is in the order of 250% as opposed to 350% common for conventional pump motors.

• The higher the head that the pump is starting into, the lower the start torque required. To minimise the start torque (and current) close the valve at the top of the pipe unless it is a multistage positive dis[placement type.

  From your description, it sounds as though there is a definite problem down the hole. Possibly some sand or grit in the impellor that is causing it to bind. At Locked rotor current, I would expect the motor to develop a healthy 200% plus torque and it should really go!!

   

  Your description of the performance using both the soft start and the drive are indicative of something down the hole causing the pump to require a very high torque to spin slowly. Are you sure that the impellor has not been wound partially off the shaft and binding on the bottom of the housing? The other possibility is that it was not assemble correctly and there is a spacer missing, or in the wrong place causing mechanical binding.

   

  From the sound of thing, I believe that you have to pull the pump up!!

   

  Off to bed, I will look at it again in the morning.

   

  Best regards,

[jraef]

Thanks Marke,

As of today, they are pulling the pump based upon my report (gulp!), and word is that this will cost them close to $5,000! There is apparently only one contractor nearby with the right equipment, and he is busy this time of year. Pity. If I ever hear the outcome I'll post it.

[marke]

Been there, done that,

Usualy find either something in the pump that is caught in the impellor, or misassembled. Seen them spin OK by hand, but the moment the motor is energised, the rotor binds on the stator. Not centralized etc.

 

Good luck, hope they do not send you the bill!!

 

Best regards,

[jraef]

Well, they managed to pull the pump today after all, and I am vindicated! There are 2 x 2" IPT service openings at the crown coupling from the motor to the pump shaft. These openings provide access for making the final connection of the coupling set screws, and are to be sealed afterwards with 2" threaded pipe plugs. Apparently, after the pump technician did his spin test of the motor by hand, they made up the coupling and installed the plugs. But they must have gone out and purchased the plugs at Home Depot or from an electrical supplier, so they were not tapered threads. Consequently they were tightened into the holes so far that they were jammed against the coupling. When the VFD was giving excess tourque it was making the coupling scrape along slowly against these plugs, but when we went accross-the-line with the soft startrer, they just dug in deep.

 

Someone has some explaining to do... but not me

[marke]

Well, I was sure that it had to be mechanical. Just as well the pipe plug did not wander in while the pump was at full speed!! - could have been even more damaging.

 

Have a good day,