IEEE Open Journal article of shaft failure of soft-starter fed electric submersible pumps 2

[electricpete]

Purely for info
This article appears in "IEEE Open Journal of Industry Applications"
I believe that journal is available to everyone.
Here's a link:

https://ieeexplore.ieee.org/document/8920120

=====================================
(2B)+(2B)' ?

 

[electricpete]

In fact, you can make any search you want at ieeexplore.ieee.org and then check "open access" on the left side to limit the results to freely available articles.

=====================================
(2B)+(2B)' ?

 

[Strong]

Do I understand the paper that the soft-starter is providing a ramp function, but does not eliminate the transient when breaker is closed? Do I correctly understand that the paper offered a good model, but no corrective action to prevent torsional induced shaft failure?

Walt

 

[EdStainless]

Having been in ESP business I had to read this.
1. What pump is this? 750hp with less than 500' of head? This is a mine dewatering pump and must be very large diameter. These impellers will have an order of magnitude higher inertia than a conventions 5" OD deephole pump. They are also missing about +10,000' of cable from the system, which changes the dynamics a lot.
2. What is with the low shaft strength? The UTS of the shaft should be about 900MPa min, or in some cases 1050MPa.
3. If this a large diameter high flow pump what gives with the very small shaft diameter? 30mm? in pumps this size we used more like 51mm or 57mm shaft.

Under-starting one of these always breaks things because you hang up in a resonant speed and shake it to death.
The most common issue though is starting and having it fault. Then restarting while fluid is still backflowing and everything is spinning backwards. That shock snaps a shaft every time. We used back EMF sense for interlock. On deep hole systems sometimes the re-start delay was 20 min.

This sounds like amateur hour as far as design of the system goes. I wonder if the motor/pump supplier told them that this combination wouldn't work and used it anyway.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy

 

[LionelHutz]

They use a stupidly low current limit which hangs the pump acceleration at a resonance speed and then predict the pump will self destruct. Well duh. You will have difficulty starting a motor with an open shaft using their 125% current limit.

Nothing is perfect. VFD's can do the same thing going through a resonance if you use too slow a ramp time. They can also introduce a steady state torque oscillation that can destroy things.

 

[jraef]

125% is WAY too low and usually leads to the motor just hanging there until a time limit is reached and it goes across the line, ready or not. That’s basically just a paper written on how NOT to do it...


" We are all here on earth to help others; what on earth the others are here for I don't know." -- W. H. Auden

 

[waross]

In the early days when forbidden frequencies were being discovered. I saw two 60 HP centrifugal pumps destroyed by reflected pressure waves.
It has to do with the time between each impeller blade passing the pump discharge and the time for the pressure wave to travel to the first 90 degree bend and be reflected back.
When the times coincided, the pressure spikes built up and bad things happened.
The process engineers were using variable pump speed to control the discharge flow.
The multiplying pressure spikes blew out a piece of the pump housing about 5 inches in diameter.
They switched over to the number two pump while waiting for a replacement pump and destroyed the second pump the same way.

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

 

[EdStainless]

Our drive designers had tables of the blade pass frequencies for various pumps and the shaft stiffness factors. They could estimate critical frequencies for various combinations based on the test information. Then these frequencies would be blocked out. We had one popular pump that had a critical at 41Hz, so 39-43Hz was blocked, going up and down it would simply skip over that range.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy

 

[waross]

Hi Ed. I understand that as well as the blade pass frequencies, the distance to the first bend or fitting that may cause reflections is also critical.

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

 

[electricpete]

That sounds like an acoustic natural frequency which represents a standing pressure/velocity wave in the fluid. One case where it can occur is when the length of the straight-run of pipe is certain multiples or fractions of the wavelength of sound in the fluid at the frequency of interest (lambda = speed / frequency) where speed of sound in the fluid depends primarily on the material properties of the fluid (density and bulk modulus)

There can also be mechanical natural frequencies that most of us are more familiar with.

When excitation coincides with natural frequency, resonance occurs. For acoustic natural frequency the excitation can be blade pass. For mechanical it can be blade pass frequency or running speed or more.

In my 20 years working on power plant motors (sitting next to the vibration engineer and aware of most of what he does), we've seen a lot of mechanical resonances but haven't been unlucky enough to see an acoustic resonance (that we know of).

=====================================
(2B)+(2B)' ?

 

[EdStainless]

In most ESP systems there are no bends or fittings, just thousands of feet of straight pipe.
We worried about the interaction of the blade pass with the shaft stiffness. We measured frequencies for each shaft size, and then knew knew the blade pass values for each pump design. It created a matrix of no-go speeds.
It is funny that many oil wells are not straight, not even ones that they intend to go horizontal. When an ESP is installed in a slight bind it will vibrate less, but too much of a bind and it will cause rapid wear.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy

 

[waross]

That sounds like an acoustic natural frequency which represents a standing pressure/velocity wave in the fluid. One case where it can occur is when the length of the straight-run of pipe is certain multiples or fractions of the wavelength of sound in the fluid at the frequency of interest (lambda = speed / frequency) where speed of sound in the fluid depends primarily on the material properties of the fluid (density and bulk modulus)

You said it much better than I could Pete.
That was the conclusion when the pump housings were blown out.

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