Practical Solutions to Machinery and Maintenance Vibration Problems
Chapter 2, Mechanical Resonance
Section 9, Detuning Resonant Part versus Reducing Vibration at the Source
When a part is resonant, such as a pipe or beam, it is usually assumed that detuning the resonant part will result in decreased vibration, not only on the resonant part itself but also at the machine's bearings. Although this would be true in most situations, it is not true in all. There are many situations whereby the vibration is drastically reduced on the resonant part, and yet there is almost no reduction at the machine's bearings. Why? The following explanation may not apply in all such instances, but does apply for most.
Vibration originating from its source has not only phase, frequency and amplitude but also has "vibration energy" or intensity. If the vibration energy level is small to medium, the vibration will travel along the machine's structure, pipes and support system only so far before the energy is dissipated by damping. If this low to medium level of vibration energy travels to a nearby fairly flexible section that is resonant to the vibration frequency, the part will resonate. When it does, the vibration amplitude on that part will be increased considerably. The increased vibration is then "telegraphed" back to the rotating machine itself. Therefore, the machine's bearings also have their vibrations increased. For such machines with relatively flexible resonant sections, detuning the resonant section will considerably reduce the vibrations at the machine, its bearings and the resonant section. However, there are situations where this will not work.
For this example, assume the same relatively low to medium level of vibration energy originates from the rotating machine (such as from unbalance or shaft misalignment). But now also assume that there are no relatively flexible resonant sections. Instead, the sections with natural frequencies that correspond to the source vibration are considerably more rigid (such as a large "I" beam under the concrete floor that supports the machine. Or it may be a support pedestal or a relatively rigid, large diameter pipe.) Although these sections have the correct natural frequencies to respond to the source vibration, the source vibration would not have enough vibration energy to overcome the flexural resistance and internal damping. Therefore, such relatively rigid sections would not resonate and, therefore, would not be vibrating excessively nor would they cause any increased vibrations at the bearings. Now assume the same relatively rigid sections have their natural frequencies in the same frequency range as the source vibration. This time, assume that the source vibration energy level is very large, such as due to large unbalance or large amounts of shaft misalignment. With the vibration energy level high enough to overcome the flexural rigidity and internal damping, the rigid part does resonate, thereby drastically increasing the amplitudes on the resonant section and at the bearings.
The resonant part will have all the usual symptoms of resonance (curling mode shapes, nodes, antinodes, phase change, etc.), but when the resonant part's natural frequency is altered, for example by making it more rigid, the previously resonant part no longer resonates. Therefore, its amplitude decreases and its resonance symptoms disappear. However, the high vibration amplitudes at the machine and its bearings remain approximately the same as before the part is detuned. (In some situations it actually gets worse.) Why?
If the source vibration's energy level is high enough to resonate a relatively rigid section, the source's energy remains high after the rigid section is detuned. After detuning, the high energy vibration simply travels further along the structure, pipes, beams and so on until it finally reaches another resonant section. Detuning the next section does not improve the situation either, as the source vibration energy is still very large. In such situations, the vibration must be removed at the source by balancing, aligning and so on.
A case history will reinforce this point. A very large reciprocating compressor recorded 12 mils displacement at a frequency of 2 x rpm of its crankshaft. The compressor was approximately 9 feet (3 meters) high and 21 feet (7 meters) long. The gas output, thick wall pipe diameter was about 18 inches (½ meter).
The pipe's first natural frequency was resonated by the frequency at 2 x rpm. The pipe was stiffened considerably with large angle irons. The amplitude on the pipe decreased from about 40 mils to about 10 mils. The vibration's phase also changed about 80°, indicating that the pipe was resonant before bracing. However, further along the pipe, past several turns and elbows, another section started to resonate (although it did not resonate before the first section was stiffened). Large angle irons were used to detune the second resonant section. Its amplitude also decreased, but the vibration at the compressor bearings remained as originally found. It was then noticed that the structural supports for a large "gas bottle" (connected to the pipe) now resonated. The supports were detuned, but the amplitudes at the bearings remained the same.
After giving up trying to decrease vibration by detuning resonant parts, the writer learned that the compressor was the first of its kind and that the designer left out an incredulous 2,000 pounds of reciprocating part counterweights! Obviously, this vibration had to be corrected at its source. But this example didn't seem to reveal any new principles as it appeared to be only the result of an exceptionally bad unbalance at its source. Certainly, it seemed, source vibrations of much less magnitude would not produce the same results! But this assumption was wrong. On the next three vibration problems, various relatively rigid pipes and beams were being resonated. Detuning produced the same results as for the compressor (the original vibrations remained the same). In other situations with more flexible resonant pipes and beams, detuning did lower the vibrations at the bearings as well as on the resonant sections. A pattern finally emerged.
As there are many variables to how much vibration energy it will take to resonate a section, a very specific rule cannot be given; however, a rough rule can be used. The writer suggests that for pipes of about 8 inches (200 millimeters) diameter or less, detuning the resonant part would most probably be effective in considerably lowering vibration amplitudes at the bearings. For larger diameters, the effectiveness would be more questionable. The same applies for structural members, such as beams, pedestals, floors and so on. If their thicknesses or depths are less than 8 inches or so, detune the resonant part; if a little larger, first go after the source. For diameters or depths of beam 10 inches (250 millimeters) on up, it almost always requires removing vibration at the source rather than detuning.
However, these numerical guidelines are not ideal for all the varied conditions found in relatively rigid machine structures, such as floors, bases, columns, beams, large diameter pipes and so on. For situations where certainty is important, a homemade vibration shaker can be used to prove whether detuning the relatively rigid part will work or whether the source vibration must be eliminated. (See section "Using Shaker to Determine if Fault is Due to Weak Structure or Vibration Source.")
The information above does not apply for rotors that are in resonance.
Large rotors such as papermachine rolls resonate considerably more easily
to their own rpm frequencies than a pipe of equally large diameter.
For example, for a pipe diameter of 2 feet (2/3 meters), it is best
to go after the source. But, for the same diameter resonant papermachine
roll, it is best to detune the roll. If not practical, the roll will
still require that its source for the vibration be removed (see sections
on "Resonant Whirl" and "Removing Whip.").
This textbook contains only part of the information in our Practical Vibration Analysis seminar.