Traction Motor Brush Wear Problem

[mtwentyc]

Hi all, thanks for the forum. Have searched by keyword with no results - maybe someone has seen this problem before. We have a 1000HP GE752 AUT, series wound in an offshore drilling environment. The motor is mounted vertically on a top drive drilling unit driven by a Ross Hill SCR system. I'm mechanical so am struggling a little bit, thanks in advance. Here's the issue;

Have had extreme problems with brush wear, needing to replace brushes after 16hrs run time at approx 80% load. The weird thing is that only the lower brushes, closest to the armature are wearing out. The upper two appear to have no wear at all. It's as if only the lower brush in the group of three on each of the four brush holders is carrying an current which is hard to explain. We cannot explain the reason for this selectivity. Vibration analysis has been done with no adverse results. SCR wave forms look OK, have had brushes analysed by Morgan, right grade, no contaminents etc etc. One similar case history has been found on a shunt wound mud pump drive where an AC ripple was nominated as the cause but inconclusive.
Has anyone ever seen such wear before? If so what was the cause/fix cause.
Cheers

 

[edison123]

Are the spring pressures same at the top and bottom arms ?

Try interchanging the brush holders along with the springs.

 

[Skogsgurra]

It sounds like a typical vibration problem. The brush holder arms are attached to a ring which sits on the end bell? Right?

Imagine what can happen at the end of these arms when running near or at their resonance frequency. Large movement close to rotor and much less close to end bell. I have had this on large vertical pumps and a simple fix was to attach supporting beams (isolating!) between the brush holder arms' extreme ends. Took the problem away instantly.

Did you actually do vibration analysis on the brush holders?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[polvalrot]

Is it possible to measure the current in each brush with a DC/AC mapere clamp? there is a possibility that not all brushes are needed to deliver the current under these loading conditions. This is a common problem on underloaded Ac and DC rotors.
If brushes are deligned e.g. for 10Amps / cm2 but youonly need 3Amps, then the systme protects itself, by trying to develop more conductive routes, through the brushes to the slip rings. Thus wearing a lot the brushes and quite often grooving the slip rings.
In this case you consult the manufacturer, in order to use a different brush grade or less bruses.
Have you changed the brushes quality (or suspect change of grade/quality)?
Also have you checked the insulation resistance in the slip rings (leakage)?

Best,
chronis

 

[mtwentyc]

Thaks all;
Spring pressures all Ok - hace replaced the drive three times with a new unit each time due to flashover risk in open hole, problem so bad eventually comm was worn on the section closest to the armature windings. Problem does not follow motors, it stays with the TDS. Also, machine in same config has drilled extensively with no issues before. Problem only recently occurred, accelarations measured at drive end and NDE and wuhin spec. Also, we can't check brush holder vibration as the unit operates in a hazardous ares so running with covers off impossible. Really have discounted vibration as an issue, was looking for an electrical root cause...

 

[Skogsgurra]

Still can't think of anything but brush holder vibrations. All other possibilities seems to have been ruled out.

Does the commutator look anything like this (yes, old machine)

Vibrations in the brush holder cannot be measured from outside.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[mtwentyc]

No - it's a standard hi-torque 752, has a spiral grooved comm, four axially arranged brusholders with three brushes per holder. Is vertically mounted and very typical, thousands in use in similar drilling applications around the globe. Vibration seems so unlikely to me, no evidence of any nor of any mechanical cause for vibration. Also, if vibration was the root cause I would not expect to see such a consistant result with the lower brush only wearing down as the upper brushes do not seem to be carrying current (hence lower is overloaded and in thermal runaway). My hunch os it's a control (SCR) system issue but I can't support this case with anything concrete.

Heres a link to a similar issue;

http://www.westengineer.com/newsletter/nl4expand.htm

 

[Skogsgurra]

OK. That brush holder doesn't vibrate - I am convinced there.

I can also understand (to some extent) what happened. The ripple is an AC component on top of the DC. The brush arrangement shown means that the three parallel current paths from input via the three brushes and the commutator segments to the armature winding have differing impedance - brushes closest to armature risers have lowest impedance and that makes the AC ripple go the easiest way; through brushes closest to armature. AC ripple is tough on the commutator, so it is natural to see arcing under these brushes.

But, you say that you do not have ripple? And problem doesn't follow motor... Hmm... Sounds like a ripple problem then. Record armature current with scope or memory recorder and see if ripple is a coming-and-going thing.

I would luvv going there to see what happens, but I avoid off-shore rigs. Never had the rescue training. And do not intend to have it either...

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[waross]

Try pulling the problem brushes and running on the others.
respectfully

 

[mtwentyc]

Thanks again all - Can't pull brushes and test as are currently between drilling campaigns, in fact cannot simulate any load at all now, so have to fix before the next well begins.......

So how do we find this ripple if it is the cause? We did take snapshots of SCR output wave forms (on a low end Hoki ocilliscope I think), SCR vendor reviewed pictures and said it looked OK but I imagine it would be hard to be sure. Is there a definitive test we can do and do we need DC load?

 

[Skogsgurra]

The low end Hioki is probably OK. Is it an 8804 or similar? If so, very good. Use a DC current clamp adapter and record at least 20 ms worth of armature current. It should be available if your SCRs are not also in Ex area. Use Memory mode - not RMS!

Check the six current "hills" for even height. If difference is below ten percent of mean DC value, then you do not have any serious problem. If more than 20 %, then watch out! Values in-between may be worth investigating more in-depth.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[mtwentyc]

Thanks Gunnar, appreciate it.
Anyone else seen this before?

 

[petronila]

Dear mtwentyc,

Some DC Motors manufacturers use designs with Series Parallel interpoles connections.If this is your case,Use a Clamp-on DC Ammeter and take readings of each parallel path current,we are looking if is circulating by one path the 75% load current, if this is happenning, for the other path you will have a very light current and you will have a problem with the low current density of this light loadded brushes. To solve the problem is necessary to install equalizing jumpers to connect the brush holders of like polarity.(Contact a repair shop )

Regards

Petronila

 

[markcraig]

Easy way to determine if it is the SCR is to Re assign it to one of the mud pumps or draw works motors and assign a known good SCR to the top drive. If the problem persists then it has to be a top drive problem.

 

[mtwentyc]

Thanks all,
assignments on the SCR have been changed and tested (this one will drive a mud pump as well), no signs of similar brush wear on the pump (but they are a different style of dc motor), this supports the waveforms on SCR outputs, which show OK.

Present thinking is that we have a field cabling issue - there is a VFD that steps 415VAC50Hz up to 60Hz for some ancillaries on the machine, we have had trouble with it and it also draws more current than it should. Suggestion is that we either have a termination problem in one of the common j-boxes feeding AC & DC to the machine, or a service loop (umbilical) with damaged cores, which may be feeding AC from the VFD onto the DC side.

Does this sound logical?

 

[Skogsgurra]

I don't know. But I have several DC machines in a steel mill that have worked flawlessly for about fifteen years. Now, after a couple of 100+ kW VFDs have been installed in same area, they (DC machines) develop bearing problems.

I will have a look at those machines next week. It certainly doesn't sound logical that the mere installation of these VFDs can influence bearings on DC machines. But I shall have a go on it and I really hope that I can find out the whys and hows of it. Stand by.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[edison123]

Gunnar

That is interesting. One of my valuable clients, who makes automobile tyres, has plenty of dc motors. They are planning to go in for a 1500 HP VFD motor (remember my thread about the noise levels in VFD motors) for the first time for a rubber mill.

I will pass along this caveat to them.

*Why make the same mistake twice when there so many new ones available*

 

[Skogsgurra]

Back from the steel works. Yes, they have a problem coming from the fan installed a couple of years ago. The supply voltage is 500 V coming from a 2 MVA DY transformer with isolated star point.

It turns out that PWM components from VFD is present on all phases. Nothing phase-phase but about 400 V pk-pk 2.5 kHz common mode voltage. I will start a new thread on that - it might be of some interest.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[mtwentyc]

Skogs,
thanks for your input, I'll look for the new thread.
We're still trying to find teh fault, will post when have some new info
Cheers

 

[eddykurent]

I agree with the other threads that spring pressure & vibration could be issues. If you have replaced all the brush holders for new, this removes them from the equation.

Has the neutral plane been interfered with? Where the brush holders positions marked before replacement, or can they only be installed in a fixed position.

Has a commutator run out check been performed. Using a clock gauge (DTI) check the deflection over approx 10 commutator bars. Shouldnt be more than a couple of th"diff. Check with manufacturers. If its larger you could have some wear issues. As this machine is vertically mounted you will have a thrust bearing at the DE, and the upper bearing could be worn or damaged if a braking device is installed near by.

You mentioned brushes were lasting 16 hours at high load. How are you bedding your new brushes in? I would expect to sacrifice some brush life to gain a good commutator filming, although i would expect them to wear relatively evenly.

What is the Hi torq commutator like. Rough, grooved, burned, any bars discolored over a single revoloution of the machine.

Has has the DC bus been checked for AC, either by a scope or a meter? Possible snubber network failure allowing AC through the bridge. Has the current limiting, or stall, circuits been altered or checked to protect the motor operating in severe overload conditions.

I hope some of the above may assist you.
Let us know the outcome.
Eddy............