Intermittent vibration on large electric motor 1

[podobing]

Hello,
Underground coal mines use large electric motors to operate their ventilation fans. The number of fans required depends on the size and depth of the coal mine. One fan, in particular is a 2000 HP, 4kV, 890 rpm motor that is started directly across the line. The motor has two vibration sensors on it, one on the drive end (DE), and one on the ODE. The fan is a 10 foot diameter axi-vane fan. Both fan bearings are monitored with three axis accelerometers mounted to the fan pillar block bearings. The fan also turns at 890 rpm. This fan operates 24/7/365. If the fan is stopped, and the ventilation in the coal mine is disrupted, the mine must be evacuated. All sensors are monitored at a remote site 24/7/365.

I have not witnessed this personally, but from what I am told, occasionally, for no apparent reason, the fan motor "growls" for 3-4 seconds. Originally, when the motor growled,the motor vibration sensors tripped off and de-energized the motor, resulting in a mine-wide evacuation. The sensitivity on the motor vibration sensors has since been decreased, and now the motor growls, but the fan continues to operate. The acclerometers on the fan bearings do not pick up anything unusual. I was told that the "growl" sounds like a large transformer being energized.

Currently, we have a power system analyzer connected to the fan motor starter. The power system consists of a 138:69 kV substation, 4 miles of 556.5 ACSR line and a 5 MVA 69:4.16 kV transformer. The line is lightly loaded with other mining loads at this time. Since we installed the power system analyzer, we have had two vibration trips and no events being recorded on the power analyzer.

Any ideas?

Regards,
Podobing

 

[ScottyUK]

What are you using as a trigger for the power analyser? The local distribution / subtransmission system is large relative to the load, so voltage disturbances caused by the motor are likely to be fairly small, other than some sag during the motor start as measured at the 4kV busbar.

 

[waross]

You may have the voltage going either low or high on one phase feeding the motor. Will your power analyser monitor min. and max voltages per phase?

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

 

[podobing]

The monitor triggers at 10 percent (sag or swell) on any one phase. It also monitors frequency, and I believe that the tolerance is plus or minus 0.2 Hertz.

 

[itsmoked]

How far is the starter from the fan?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

Other than during starting I doubt you'd ever trigger with those settings. Can you use an external digital trigger signal such as the vibration trip command?

Most modern protection relays constantly update the internal event recorder to hold records corresponding to a period of a few hundred milliseconds or a few seconds, discarding the old data as new data comes in to displace it, and in the event of a trigger event you see the pre-trigger and post trigger data. If your analyser can't accept an external trigger then maybe you can configure a protection relay to do the job for you.

Have you considered that process conditions such as specific airflow rate could be triggering a resonance in the fan itself?

 

[jraef]

A 10% sag or swell on one phase represents a very significant imbalance actually, over 3%. Negative sequence currents from phase imbalance become significant at levels over 2%. NSC causes negative torque in the rotor that "fights" the normal torque, the result is a added heat PU of Amps in the motor, but as it gets worse becomes a noticeable growling sound because the NSC is essentially pulsing and can interact with natural mechanical harmonics in the load. You might want to set your trigger for a lower level of phase voltage drop. The specs allowing for +-10% are based on symmetrical voltages, imbalance is much worse.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[ScottyUK]

jraef,

I read 'sag' and 'swell' as voltage events, not current. Maybe the OP can clarify whether he has configured a voltage trigger or current?

 

[jraef]

No, I read it as voltage too. But a small voltage imbalance results in a much much larger current imbalance, I think a 2% E = 10% I, something like that. We typically monitor for voltage imbalance because it's simpler to look for than current imbalance and most of the time, the current imbalance is caused BY a voltage imbalance (or a circuit problem, which becomes otherwise evident as well). So most specs will refer to the voltage imbalance and a 2% limit, even though technically the damage is caused by the current imbalance it produces.

I've been heavily promoting the use of Solid State Overload Relays for motors lately because they typically monitor for CURRENT imbalance rather than only voltage, since they are monitoring current anyway. It's technically more precise in terms of what might be doing harm to your motor, because it WILL also pick up those other causes, often before the collateral damage takes place.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[waross]

As a possible cause, are there any nearby large motors with autotransformer starters?
The auto transformer starters and the old manual "compensator starters" used an open delta auto transformer configuration.
The phase currents drawn during starting is not balanced. This is generally not a problem but on a "soft" supply system a large otor starting on an auto transformer may cause a voltage dip on one phase that may cause issues with other nearby motors.
Try setting the monitor to trigger a recording, not a trip, on phase voltage dip of 1.5% or 2%.
When your motor is loaded it may be drawing 1.8 or 1.9 MVA. A similar motor on the same 5 MVA transformer starting on an auto-transformer may pull the common phase down enough to cause growling.
3 to 4 second sounds about right for the heavy starting current of another motor.

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

 

[ScottyUK]

Hi jraef,

I see what you were getting at now. Thanks. From the description I think the supply is going to be fairly stiff, so unless there are some other large undisclosed loads fed from this 4kV busbar I doubt an induction motor in an operable state will cause a significant voltage disturbance. Perhaps the OP can clarify that too.

 

[podobing]

Guys,
Thanks for all of the responses and interest in this problem.

To address some of the questions that were asked:

Itsmoked: The motor starter is about 25 feet from the motor. It is an across the line starter, and it is the only load on the 4 kV bus.

Waross: Other loads operated by this customer are all across the line start with some smaller (200 HP) VFD loads. All of these loads are at least 1.5 miles away. The nearest other customer is a large natural gas compressor station. They tap off of the 138 kV line, so they would be at least 4 miles (and 3 -4 transformers) away. I do not know what loads they have or how their motors are started.

ScottyUK: This analyzer does not have the external trigger capability, but I like the idea. If nothing shows, we will rent a recorder that does have that capability. The motor protection relay that is being used is a MultiLin 369. The work done by a ventilation fan in an underground coal mine is damn near constant. There are no adjustments that can be made to blade pitch or dampers while the fan is in operation. The load on the fan changes with air density, but that change is very slow and gradual.

Jraef: I am now triggering on voltage dips and swells. As you recommend, I will tighten the voltage imbalance to 3 percent when I get to the site again.

General Information: The three phase short circuit capability on the 138 kV bus is 3500 MVA, and the short circuit capability on the 69 kV bus at the fan site is 270 MVA

Take care, and thanks again.
Dave

 

[ScottyUK]

Have a closer look at your 369 - I am fairly sure that relay has an event recorder built in, although it possibly doesn't trap a long enough time period (a few cycles rather than a few seconds).

 

[3DDave]

Any chance the vibration sensors have auxiliary outputs that can drive an event recorder?

In planning this out, have you made a list of items that won't require replacing/repairing the motor and those that will?

Maybe it's audio frequency distortion from the natural gas compressor or other motor.
Maybe there's a resonance that it moves through.
Maybe the resonance is because of a progressing fatigue crack in a fan blade.
Maybe it's one of the bearings.
Maybe it's winding coming loose.

 

[BrianE22]

Can you sweep it through its speed range just to check if there's a natural frequency that's being exited?

 

[electricpete]

Are there by any chance spectra available from the times when the anomaly occurs?
Failing that, spectra from normal operation?

I can appreciate that it is very logical to focus narrowly on the power system based on very intermittent nature described. If that is not panning out, consider also broadening your review to look for other possibilities… send a vib guy to inspect the machine, check housing temperatures, draw bearing oil sample, measure all three phase currents if easily accessible, and most importantly to collect lots of vib data if your remote vibration signal does not include anything other than a magnitude.

Some other questions for general curiosity (feel free to ignore). Is it horizontal or vertical motor? What type coupling? What type of bearings on both machines?


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

 

[keyepitts]

How often does the "growling" occur? In addition to what Electricpete suggests, I would look at the inlet air plenum conditions; maybe when the wind blows from a certain direction an occasional inlet resonance is excited.

 

[waross]

Can you post the complete model number of the MultiLin 369? That will allow us to see what options you have available to monitor line conditions.
Where are the revenue meters located and who owns the meters? Your firm or the supply utility?
It is possible to program the electronic KWHr meters to store a lot of data. eg. max and min voltages and currents per phase for 15 minute windows.
I have received floppies with Max/min current, voltage, kW, KVA, KVAR, demand, for 15 mimute intervals for a couple of months from a friendly utility supervisor.
I was having a problem with an Automatic Transfer Switch. The recordings showed that the utility was having issues with a voltage regulator on one phase. I was able to solve the issues by setting the set-points lower on the ATS.
I had another issue where a contactor based ATS was dropping out for a couple of cycles and pulling back in.
I spoke to an engineer at the utility who described an seasonal issue we were going through. During the dry season, the sea winds would deposit a layer of sea salt on the insulators near the sea shore. As we progressed into the rainy season the insulators on the distribution lines near the shore were prone to flashovers. This may happen a couple of times a week. I was told that when a flashover occurred all the meters in the substation would swing madly and the whole city may experience a severe voltage drop for a few cycles.
It may be worthwhile to visit the local utility and see if they are aware of any recurring system upsets.
Dropping in to the office is much more productive than a phone call.


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

 

[Skogsgurra]

Yes, Bill. Talking to the guys is far more productive than a phone call. Seen that a few times. After years of phone conversations, me and neighbors visited the utility office. Next year, we had a new 11 kV line. Eight miles. And zero power outages. We had scores of them before.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[podobing]

Guys,
I have been back to the site, and have some new information, but nothing concrete yet.
The motor, shaft, and bearings are all horizontal. The couplers used are grid-pack couplers. The system RPM is 890, and all bearings are grease lubricated.

The power system analyzer was changed from recording voltage sag-swell to voltage imbalance. There were no events recorded by the power system analyzer while recording sag-swell information, but the motor vibration occurred 6 times during that week. The problem with tightening the sag-swell tolerance to the 1 - 2 percent range is that those sags are normal during normal mining equipment. I was looking for something extreme.

The analyzer has been monitoring voltage imbalance on the 4160 volt bus for about 1 week. When I left the site, the "normal" voltage imbalance is 0.3% measured on the 4160 volt bus. I plan to visit the site again tomorrow (Tuesday) and see if have been any imbalance events logged, the magnitude of those events, and if there is any correlation between these events and the motor vibrations.

All bearings (fan and motor) have had vibration analysis readings taken by people that I consider "very good", and there have been no defect frequencies found. The way that I think of a bearing is that if there is a bearing problem, it does not come and go, it just gets worse. That being said, once we can clear the electrical system of problems, we will probably get a recording vibration analyzer.

Waross - I should be able to get the specific information on the Multi-Lin 369. I know that I can use it to alarm on current imbalance. Those parameters are now set at:

Alarm (current) imbalance = 5 %; Alarm (current) imbalance delay = 10 seconds.

Trip (current) imbalance = 10 %; Trip (current) imbalance delay = 15 seconds.

If nothing else, we should be able to tighten the Multi-Lin alarm parameters to 2 % imbalance with a 3 second delay to correspond to these intermittent vibrations without shutting down the fan.

Regards,
Dave