One should always calculate or know the force being applied and determine that it is safe to apply at the balance location.

One can look at balance weights that others have used (be careful of the big shop weights). If the rotor has provision for balance planes, one knows what size weights may be appropriate (one may be too light to fix the problem).

What weight do you use when there is more than the rotor shaking in something like a spring isolated fan? Is it really isolated? If so, it is running well above the resonance, and there is a lot of mass moving.

Any one that would put enough weight on a machine with 100 mils to bring it down to 9 mils in one shot, must have looked at where to place it. Otherwise, one would be risking 200 mils of vibration; at least this seems like a risk to me.

Today some utilities are trying to make money vs the old regulated days. Not all are willing to go to the OEM; they have to look for value for the money they spend rather than relying upon the rate commision to give them a profit. I bet that I have balanced 100+ turbines by shutting them down and adding a weight; by my count that cost the customer no runs, but if you like 1 run. All it has to be is good enough to meet the desired goals; the balance doesn't have to be perfect.

The "unbalance equation" that I offered is not a 'rule of thumb' at all. It simply represents that the amount of unbalance is equal to the mass of the rotating element multiplied by the distance the center of mass is displaced. For example, a rotor weighing 1000 lbs that is running with 10 mils peak-peak vibration has an imbalance of 1000 lbs x .005 inches = 5 in-lbs. Put another way, the center of mass is orbiting the geometric center at a radius of 0.005 inches.

Of course, this is the simplest case and ignores lots of dynamic factors which may be present in varying degrees, or not at all, depending on the system. But, it is still a good starting point.

In the orginal equation, A is the peak-peak displacement (expressed in 'mils'). The 2000 in the denominatior converts the peak-peak displacement to the "half-amplitude" (amount the rotor is displaced from the centerline) and allows you to write the amplitude as a whole number (10 mils vs. 0.010" for example).

There is no 'guesstimate' involved... this is simply the physics of a simple rotating system with an imbalance.

I didn't catch that relationship the first time through, but the basis for the rule makes sense now.

m*r = M*e for a single degree of freedom system far above resonance
(e is one half of pk/pk displacement.)

Anywhere far below resonance, that SDOF system would have a lower M*e and the rule would be conservative (underestimate the m*r associated with the initial unbalance).

As Rusty says, the dynamics of the real system may not reflect that simple SDOF, but at least we know where the rule comes form and can factor it into our thought process.

This message has been edited. Last edited by: electricpete,

quote:

Put another way, the center of mass is orbiting the geometric center at a radius of 0.005 inches

For this to work one needs the rotor (assumed round) to orbit about the mass center. Then with proximity probes at the center of mass one would measure this double amplitude (pk-pk value). If it orbits about the geometric center then one should get 0 vibration.

quote:

The "unbalance equation" that I offered is not a 'rule of thumb' at all.

Really! What connects this to an arbitrary rotor imbalance situation? I think this fits squarely under the definition of ‘rule of thumb.’

As ElPete points out the logic would fit for a rotor running above a resonance with some approximation to modal mass (whatever that really is). This used to be found in an old Bently Nevada publication. I remember pointing out to people when I was at Bently Nevada that this could produce very large balance weights when the rotor was near resonance. Also, well below resonance this can underestimate the imbalance, but it may be ok as a ‘rule of thumb’ if one understands its foundation.

For most rotors running sub-critical and perhaps with some amplification from a natural frequency the logic does not directly apply. With a good deal of amplification from the resonance the vibration increases, but the imbalance does not. The application using casing readings is also unclear as to the direct application.

I first used this for forced response modeling of simple rotor systems that ran above the critical. This works fairly well when using probes near the mass on a Bently Nevada rotor kit with one mass after you pass through the critical.

Rules of thumb are rules of thumb; use them with care.

Bill, you can call this what you will, but for 99% of the equipment that most of us deal with, it works very well as a starting point for trial weight estimation.

This is really Art Crawford's baby, so I had no reason to question it. It works. It's that simple. You can throw rocks if you want to, but it still works.

For the sake of argument, I have a large center-hung fan and I locate a vibration pickup horizontally at one of the bearing housings. I measure 10 mils peak-peak displacement. Assuming that I mount my sensor carefully with no rocking and no cable noise, then the 10 mil measurement closely approximates the physical movement of the vibrating mass. If I could mount a dial indicator rigidly to the stationary floor, and indicate off the bearing housing (horizontally), I would see 0.010" total travel of the indicator. (I've actually done this a number of times... it's pretty close). The zero point of this peak-peak oscillation would fall on the geometric center of the mass, assuming we had a round rotor, shaft, bearings, etc. If you left your indicator in place and stopped the fan, it would read "zero" when the fan stopped. When the fan stops, it 'settles' into position along its geometric centerline, regardless of the imbalance condition.

That explanation is not at all technical, but to me it seems intuitive. Does it not? Am I the only one who sees this?

And of course it doesn't apply if you are running up the resonance curve. I have had the simple equation give me ridiculously large trial weights when I was close to resonance, but a ridiculously large balance weight is also intuitively obvious (to me at least) and you wouldn't just blindly stick such a weight on your machine. Nor would you just blindly stick a trial weight "anywhere"... I always calculate the position as carefully as I can, and the trial weight based on what I know of the unit, but always very thoughtfully. 20% of the time, the calculated trial weight will not make sense, and I will do something different. But 80% of the time, the simple equation gets you close.

I think Mr. Foiles and I work in very different worlds... I'm pretty sure I would have a hard time working effectively in his world, and I'm certain he'd have problems working in mine.

This message has been edited. Last edited by: rustythevibeguy,

I fail to see what the geometric center of the mass has to do with this. Perhaps you are looking for the center of mass. This measurment does not know the center of mass, nor the geometric center. With your seismic sensors you do not measure down to DC (unless you have less typical transducers like the strain gage accels or the feed back type) - the average (center?) value of an AC signal is 0 (with no position type informaton other than dynamic).

quote:

When the fan stops, it 'settles' into position along its geometric centerline, regardless of the imbalance condition.

Are you saying the shafts have a bow, and you are trying to measure the bow?

The shaft/mass does not spin about the geometric center (except upon coincidental occassions). You have talked about coast down plots. Did you notice that the amplitude changes as well as the phase? If so, which amplitude do you pick for your formula (Shouldn't each amplitude be valid?), and why? Also, how does any of these different amplitudes and phases relate (directly) to the geometric center?

There is nothing intrinsically wrong with this rule of thumb, as long as one is aware of its assumptions. I would use it on rotors above the critical (perhaps spring mounted fans with an interesting guess/selection for mass).

Me; I weigh a little over 16 stones. Typically I'll use an accelerometer on the bearing's case to acquire a 1X signal for balancing but it produces a voltage at 1X and I could call it a double fortnight but why? And then in turn put a dial indicator on the shaft to measure 1/2 of the P-P fortnight or mil (why draw attention) - it's still an accelerometer producing a voltage signal at 1X that is not magically turn into mils P-P so why bother to go to a relative signal anyway? Yes, I balance in g's if using an accelerometer, but don't think it will tell me the movement of the rotor in mils.

Like Gary F I went through the IRD balancing seminars back in the '70's. And at that time in those classes under that particular situation, random placement was 'normal' or typical for the day's teaching and I think they were interested in imparting the mechanics of doing things and just simple how-to recognizing that basically all in the class were not experienced balance guys, so I don't fault them. Walt's description and Olov's photo was typical what was used for demo and hands-on experiments ocassionally sticking the puddy on the ceiling.

Bill's taking time to provide valuable knowledge and insight is an asset to this board and provides a learning curve to all that will listen and wants to further his/her knowledge with a deeper understanding. It can also keep you from making a huge blunder on a balance job.

Thanks Sam for the kind words. I've never gotten that stone thing, especially since a 'metric' country uses it - Must be one of those 'metric' units, then.

quote:

I think Mr. Foiles and I work in very different worlds

I’ve only crossed the bridge at Memphis, Tennessee to go into Arkansas (while attending a yearly meeting of the Vibration Institute – I think this was the year of the Gulf war, around 91. It was good to get out of Saudi for a trip.) to say that I had been there. I quickly turned around and went back to Tennessee. So, we may see things differently.

There may be similarities too. I believe we have flamingos here in Texas, but I haven’t seen many of the pink plastic type. I don’t know if Arkansas has the real birds.

In all seriousness most companies are in business to make money. Maintenance is done (or should be done) at a level to enhance making money, bottom line stuff. As such balancing should be as efficient as possible, and one need not balance to the finest possible level, just a suitable level. This level may include consideration for future degradations to the balance condition, hence a finer balance that required for the immediate present time. The bigger the more critical the machine the more money this means, usually; so, the relative effect of efficiency may differ from machine to machine. With fewer critical machines than non-critical ones the non-critical machines multiply efficiency by their volume.

Ever take a machine out of service in the oil and gas business? Do you want to see grown people cry?