The maintenance team probably won’t be asked to go out and correct the resonance problem. Instead you will treat the symptom of the resonance problem. You will experience a bearing failure. You will experience high vibration from unbalance. You may experience broken welds or bolts. And you may even experience product quality problems.

The list is endless. But what is the root cause? Resonance.

What Is Resonance?

The simplest way to describe resonance is the amplification of vibration. The vibration being amplified could originate from a number of sources: unbalance, misalignment, pump vane vibration – almost any source of vibration can cause resonance. And that vibration does not have to be a problem in itself. For example, resonance can cause slight unbalance to appear as if the rotor is seriously out of balance.

Every mechanical structure resonates at multiple frequencies if excited at its natural frequencies. Let’s take an example, a tuning fork. If we strike the tuning fork it will vibrate at a specific frequency.

A tuning fork designed for the musical note “A” will vibrate at 440 Hz. If you had a machine that only vibrated at 50 Hz and you touched the tuning fork to the machine, the tuning fork would vibrate at approximately the same amplitude as the machine. But if you had a machine that vibrated at 440 Hz, the tuning fork would vibrate at much higher amplitude, perhaps 20 times the amplitude of the machine.

440 Hz is the natural frequency of the tuning fork. When the tuning fork is attached to the machine its natural frequency is being excited which makes it resonate.

All rotating machinery, and their support structure, has these natural frequencies. In an ideal world the natural frequencies of the machine and its support structure would not be the same as the frequencies generated by the machine.

But that is not the case. You may think that it is unlikely that the machine vibration would coincide exactly with these natural frequencies. Unfortunately, they do not need to match exactly; they just need to be within approximately 20 percent.

Tuning fork mounted to motor and pump combination.

If our machine was vibrating at 420 Hz, and we touched the tuning fork to the machine it would amplify the vibration of the machine – not as much as if the machine was vibrating at 440 Hz but the vibration would still be amplified.

Therefore, it is very common for a machine to generate vibration that excites a natural frequency. It results in higher levels of vibration, which then potentially leads to the bearing failure, or the product quality problem, or the failure of a weld or fastener, or the mistaken belief that a rotor is seriously out of balance.

As mentioned, the vibration is being amplified – if the source of vibration is low in amplitude, the resultant amplified vibration will not be as high as if the source vibration was itself higher. The closer the frequency of the source vibration is to the actual natural frequency the amplification will again be higher. These two facts are the keys to dealing with resonance.

Resonance with a linear scale and with logarithmic amplitude scale.

Why do we have resonance problems? We have them because the vendors of your pumps and fans and other equipment, and the engineers who design the supporting structure, are not taking resonance into account. They simply design a structure with sufficient strength and it is commonly just a matter of luck whether or not resonance occurs.

It should be said that it is not normally the motor or the pump that is resonating, it is the base that the motor or pump is mounted on that is resonating. The same is true for fans, compressors and other equipment. (Yes, it is true that some machines do resonate, and rotors in turbines resonate – but that is not our focus in this article. The most common reason for equipment failure is resonance of the supporting structure.)

How Do You Know if You Have a Problem with Resonance?

There are basically two ways to go about answering this question. First, if you are experiencing a high level of equipment failure but you feel that your installation and maintenance practices are adequate, you might look for resonance as a root cause. This is especially true if you are experiencing failures of fasteners, a high level of rolling element bearing failures, difficulty balancing a rotor, or high amplitudes in a vibration spectrum that cannot be easily explained.

Second, there are a number of ways to perform tests on the equipment in order to detect the presence of resonance. This can range from the simple examination of the vibration spectrum through to quite detailed vibration testing. We will briefly summarise these tests – but please understand that there is a great deal more than can be said about each of these techniques; that is what vibration analysis training is for!

Examining Vibration Spectra. The simplest way to check for resonance is to look at the vibration spectra you have already collected. A peak in a spectrum would normally be quite narrow – if the peak is broad at the base it suggests that it may be amplified by a natural frequency.

The best way to check this hypothesis is to switch the amplitude scale to logarithmic. There you will see a hump in the noise floor. The hump relates to the 20 percent region of amplification mentioned earlier. The centre of the hump coincides with the natural frequency.

Simple Bump Tests. Have you ever felt like beating one of your machines with a piece of timber? Here is your chance.

If you strike the tuning fork it vibrates at its natural frequency; 440 Hz in the example above. The same is true for your machine and its supporting structure. If we strike the machine – taking due care not to damage the machine or injure yourself of course – we can excite its natural frequencies.

This can be done if the machine is running, but it is best to do it when the machine is not running. You would use your vibration analyser set to peak hold mode while capturing at least 50 averages. The spectrum will have peaks that coincide with the machine’s natural frequencies. Some vibration analysers have more sophisticated techniques for performing this test.

Impact Tests. The impact testing technique is more sophisticated than the simple bump test. In this case the machine is not running. We use a special hammer with a sensor attached that is connected to one channel of your two-channel analyser. A second sensor is attached to the machine.

The analyser is set up to collect a frequency response function and when the hammer strikes the machine or structure the analyser is able to compare the input signal from the hammer with the response signal from the machine.

This test provides much better information than the simple bump test but it requires a little more time and experience.

Run Up or Coast Down Test. In the previous two testing techniques we have excited the natural frequencies via an impact. Another way to excite the natural frequencies is from the machine’s vibration. As the machine runs up to speed the vibration from the turning of the shaft will sweep through the frequencies of interest (hopefully) and thus excite the natural frequencies. This is called a run up test.

Conversely, when the machine loses power and runs down to rest again the vibration from the machine will excite the natural frequencies. This can be called a run down test or a coast down test.

The test can be performed in a number of ways, but most commonly a once per revolution tachometer signal is fed into the analyser and the analyser monitors the vibration at the running speed. By observing how the vibration changes in amplitude, and how the phase changes, we can determine where the natural frequencies are located.

ODS Testing. ODS stands for Operating Deflection Shape. The ODS test enables us to visualize the vibration. We perform the ODS test while the machine is operating.

If we take vibration and phase readings at the frequency (or frequencies) that we believe are exciting natural frequencies, at locations all over the machine and structure, we can then compare the amplitude readings and phase readings to build a picture of exactly how the machine and structure vibrates. Does it bounce up and down? Does it sway from side to side? Does it twist or does it rock?

If we have this information we are in a strong position to make modifications to solve the problem. The normal output of the ODS test is an animation of the machine and structure – but you need special software.

Modal Analysis. A modal analysis test is a more sophisticated combination of the impact test and the ODS test. The machine is not operating during the modal test.

Instead of just impacting the structure in one location and measuring the response in one location, we either impact the structure in one location and repeat the test as we move the sensor to multiple locations around the machine or structure – or we do the opposite; we measure the response in one location and we impact the structure in multiple locations.

As you can imagine there are lots of details we could go into, but the most important information that we achieve by performing multiple impact tests is that we build a picture of how the entire structure vibrates, and we know exactly where the natural frequencies are located – unlike with the ODS test. We can therefore animate how the structure vibrates at each of its natural frequencies, which provides us with the essential information necessary to solve the problem.

How Do We Correct for Resonance

If the source vibration that is exciting the natural frequency is low in amplitude, the resultant amplified vibration will be lower in amplitude and it will do less damage. That is one way to solve a resonance problem. But in most cases we take a different path.

The natural frequencies of a structure are related to its design and are greatly affected by the mass and stiffness of the structure. If we made the tines of a tuning fork stiffer, the note made by the tuning fork would increase in frequency. And the same is true for the structure supporting our machine; which is normally what is resonating. If we know that we have a resonance problem we can consider either adding mass to reduce the natural frequency or increasing stiffness to increase the natural frequency. The aim is to change the natural frequency so that it is no longer excited by the machine.

This is a very sophisticated process and there are a number of issues that must be considered. In essence this is what we are trying to do – modify the structure so that the vibration generated by the machine is no longer amplified and therefore harming the machine or the structure or the product being manufactured – or generating noise that affects workers or homes located near the plant.

If you would like to know more we do have pre-recorded webinars on our Website, and resonance and natural frequencies are covered in the Category II, III and IV training courses in varying levels of detail.

Jason Tranter

Managing Director, Mobius Institute , Australia,

jason@mobiusinstitute.com