If your fan speed is 2200 but your hi vib (suspected resonance) is at 6600 There has to be a 3X force exciting the resonance. I don't see how 1X could do it. I would think that rather than concentrating on your 1X you should be looking for possible 3X forces. This may be over simple but could sheave alignment have changed with the sheave changes. I don't know if you can produce a waterfall plot with your gear but at least you should be able to do a coastdown while filtering around your 3X peak.

Rod,
I had a problem recently that mirrored your issues exactly. Found that the fan was starved for air (aerodynamic whip) and the predominant was 3x.
Check with your process folks and see if the delta p across the fan is within spec.

Ron, that's interesting. Got me to wondering about the aerodynamic limits of speeding up a fan. I would think that at some point the fan is simply going to stop working because it can't physically do what is being asked of it. Isn't this what "rotating stall" refers to?

Rod, maybe it's time to consult the fan mfg. about the limits of this fan. Do you happen to have a set of fan curves for this fan? Also, they have pressure transducers that work much as accels do that you could put on the fan discharge to examine the pressure pulsations/fluctuations. Oh man, sounds like it's time for an 8 or 16-channel data recorder!

One other thing, if there are any flow "obstructions" on the inlet to the fan (dampers, duct supports, etc) these will come more into play as you speed the fan up. What I call "voids" can develop at the inlet to the fan which will make the fan very noisy; this won't be a steady state condition though, so the noise/vibration will be broadband, possibly quasi-random.

That sounds just like what it's doing. I have blade pass going on both 3 & 4 bearings, and the fan makes alot of noise, almost like air is being beat up like a cavitating pump. I suspect some of that as well...Wish I had more experience in these problems. I can easily identify them on the spectrum, but solving the problem is harder because the company doesn't want to slow anything down or make the blade smaller, to eliminate the BPF...

quote:

Originally posted by rustythevibeguy:
Ron, that's interesting. Got me to wondering about the aerodynamic limits of speeding up a fan. I would think that at some point the fan is simply going to stop working because it can't physically do what is being asked of it. Isn't this what "rotating stall" refers to?

Rod, maybe it's time to consult the fan mfg. about the limits of this fan. Do you happen to have a set of fan curves for this fan? Also, they have pressure transducers that work much as accels do that you could put on the fan discharge to examine the pressure pulsations/fluctuations. Oh man, sounds like it's time for an 8 or 16-channel data recorder!

One other thing, if there are any flow "obstructions" on the inlet to the fan (dampers, duct supports, etc) these will come more into play as you speed the fan up. What I call "voids" can develop at the inlet to the fan which will make the fan very noisy; this won't be a steady state condition though, so the noise/vibration will be broadband, possibly quasi-random.

Have a customer with a variable speed I.D. fan and the VFD failed. While they wait on another VFD to arrive, they are trying to control air flow with a 2-blade shutoff damper on the suction side of the fan. Makes for extremely "noisy" operation, and it shows up as vibration as well.

I once had a maintenance superintendent tell me, "You can put about 12 lbs. of crap in a 10 lb. bag, if you pack it in real tight, but that's about it."

Hi Rod,

I have a few questions or maybe concerns from missing something.

quote:

From 2001 to 2004 it ran 1900cpm, then engineering wanted to speed it up to 2561. The vibrating started then but not as bad as now. 2005 we backed down to 2180cpm where we are now.

quote:

quote:
Originally posted by Ralph Stewart:
Have you got a Coastdown Plot of the Amplitude vs RPM.

Is it correct to assume (which one should never do) that it was running ok prior to the speed up and there was no 3x present?

Correct Ralph. It was running okay before the speed up, so we know what it is. Just don't know where to run it. I wish this did have a VFD on it. Sure would make it alot easier.

R. Bell
Baton Rouge, LA

Maybe I am missing something in all this discussion (I guess one thing was the fact that all this seemed to have started years ago) but if this thing started all this 3x and high vibration and destroying itself back in 2004 "The vibrating started then but not as bad as now." after a speed up and then, a year later, a slow down came into play and has ran for 3 years at this speed, is it now in a destroyed condition possibly due to being allowed to run for so many years having such high vibration and also now being so "old", why is it now thought to be "running at its critical" or ?even more questionable, being considered to be trying to move more air than it is designed to move?

So if this 3x has been there for 3 to 4 years and the concern for its condition is, what is going on now, (nearing its life's end due to a long time of high vibration) then maybe there is no hope for or reason for determinng what the critical speed is other than to extend its life a little longer. The fact that it might be running at its critical now could well be that, in its present condition of deteration, the critical frequency has changed from below to at or near at.

If it has been moving too much air for the design, seems like changing this now would not help or would it. If it is indeed now running at its critical, then changing the speed might lower the amplitudes and change the "too much air for the design" theory and extend the life. If "patch" work is needed, then do it and then determine the critical speed.

Like I said, maybe I am missing something in all this discussion.

OMOAICBTW

Well, I guess after we do the run up, test we'll know where the fan stands as far as the speed is concerned.

Soon as we do this I'll let ya'll know what it was, and what we ended up doing.

Thanks,

Heres the run up data on that fan...Our software does not show the overall, just the 1X..Not much help, unless you can tell something from the phase angles. Over all has been running about .9 to 1.20

Do not see much in the run up. If you have an increase in amplitude I do not see a change in the angle and when you had a change of angle (around 1380) amplitude stayed low. But around 1380 is where you show an angle of 387 and 434. My guess is 27 and 74 degrees.

But, as mentioned before the running speed does not seem to be the problem so it is not a critical speed you are worried about. However, you could have a resonance at the 3X. Did you get a waterfall?

If not a resonance then it is a strong possibility that you are trying to push more air then the fan or system is capable of. Look at the fan curve.

This data looks awfully odd.

A coastdown plot might show a better position than the runup, but plotting the smplitudes gives this which shows that it might be running just barely passed a critical. Phase doesn't say much but then theory never does hold 100% true in the real world.

What about a spectral waterfall plot with all these amplitudes and frequencies, have one? Just wondering when the 3x starts doing its thing.

Phase is somewhat steady ~ 60 to 90 Deg from 1400 - 2142 rpm, so it doesn't appear a critical is near. There are some small changes that could be structural, but with absolute (case) vibration it's difficult to get good phase readings with low vibration amplitudes.

Next time the fan is down, just do an impact test in the vertical direction to check fan critical. Tap the shaft lightly just enough to get a response.

As you guys mentioned this could very well be a fan performance issue due to the increased speed. 1X vibration is low compared to the overall vibration of 1.0 ips so it could be vane pass or low frequency pulsations(rumbling) from fan recirculation or stall. Fan recirculation or stall will cause the discharge ducts to vibrate and rumble.

Please post spectrum and waveform.

Wanted to give ya'll the final condition of that fan we've been working on. Sorry it took so long to get back, as I've been tied up with Hoffman blowers all week.

Okay, after welding a small crack in the base, and making sure all foundation bolts were tight, and tweaking that Dodge bearing a bit tighter, and finally CHANGING the sheave on the driven machine to slow it down about 150 rpm, the vibration levels dropped considerably.

Wasn't really anything loose on the fan, so we're pretty positive that it was a speed issue as suspected. I wanted to bring it on down another 100 or so RPM but they weren't too keen on that. Final running speed went from 2180 down to 2012.

I was very disappointed in the runup test because the only numbers it showed was the 1X, not the overall like we needed. Thats just how this SKF system is set-up. Didn't see a way to configure that part of it.

Just wanted to pass that along, and thanks for all your help and support once again!

Cheers,

Thanks Rod for the update.

quote:

I was very disappointed in the runup test because the only numbers it showed was the 1X, not the overall like we needed. Thats just how this SKF system is set-up. Didn't see a way to configure that part of it.

Not sure if there is a "legal" way of capturing a 3x on a runup with the SKF, but there might be a sneaky way to do it. The tach light counts the frequency it sees the reflector tape go by and figures the speed to be 1x from this, so by adding 3 pieces of tape to the shaft circumference, equally spaced, the system would sense this as 1x but is really 3x. Of course the real 1x would be lost but the 3x was what was actually wanted.
This is assuming that the SKF uses reflective tape and a tach light and is capable of counting that fast.

I don't know if anyone on the board has ever seen or used the Zonic WCA. It was a wonderful multi channel digital signal processor based on a MAC. Their software, RMA could be dedicated to saving the digitized time data from all channels at the real time of the analyzer (all channels was 5 kHz), but for real fast events, such as an internal combustion engine, it could be set up to order track and save only the orders desired to the hard drive, like so many tracking filters. Would have been ideal for this application.
Rockin' Rod, did you ever get your hands on the fan curve or confirm the fan was operating 'off the curve'?