Diagnosing Bearing Failures Using Ultrasound Spectrum Analysis

 

Ultrasound technology has become one of the essential tools in predictive maintenance, condition monitoring and reliability, due to its quick learning curve, ease of use and flexibility. Leak detection has been one of the most common applications for ultrasound, but we now see the technology more and more being used together with sound analysis software to diagnose failures on rotating equipment.

Ultrasound Spectrum Analysis

Ultrasound technology can be used in different applications such as bearings condition monitoring, leak detection, steam traps & valves inspection, and electrical inspections. In some cases, for example when trying to diagnose failures on mechanical assets, it may be necessary to use an instrument with sound recording capabilities. This allows the inspector to load the recording into a sound analysis software to more accurately diagnose the type of fault.

Diagnosing Bearing Failures

Mechanical inspections with ultrasound include diagnostics such as bearing faults, pump cavitation, and valves condition. When it comes to bearings, users usually monitor their condition by relying on what they hear through the headset or by trending decibel levels. This is a simple and effective method.

However, in some cases, maintenance professionals will need to dig deeper and record the sound from the asset for further analysis on the software. This practice is especially useful in two situations: inspecting slow speed bearings and pinpointing the failure’s root cause.

When it comes to inspecting slow speed bearings, in many cases there is not enough “noise” to trend the condition using decibel levels. In this case it’s necessary to look at the sound spectrum.

                                                       Sound Spectrum of a 1 RMP bearing

Here we can see the sound spectrum from a 1 rpm bearing on a furnace application. Note all of the anomalies that appear in the Time Wave Form from the “crackling and popping” sounds that were produced by bearing fatigue. This issue could only be properly diagnosed by using a sound spectrum analysis software.

We can also use this type of software to identify where the fault is, if there is an integrated bearing fault calculator. By entering in the speed (rpm) and the number of balls (bearings), an outer race, inner race, ball pass, and cage frequency will be calculated.

                                                                 Use of a bearing fault calculator.

In this case, the speed was 1708rpm and the number of balls was 8. The fault frequency calculated by the spectrum analysis software that was of interest was an outer race fault at 91Hz.

Bearing Sound Samples

By using sound recording and analysis inspectors can also learn how a good bearing/bad bearing sound, what happens when a bearing is being lubricated, etc. Here are some examples:

Good Bearing

FFT of a good bearing. Since there are no defects, the sound will be a smooth rushing sound. The spectral view will not show any harmonics or large peaks.

Bad Bearing

This is an FFT spectrum view of a bad bearing. As a bearing enters the failure stage there is a rise in the decibel level of 12- 16 dB over a baseline. This rise in amplitude is usually accompanied by a change in the sound quality. Note that on this bearing we can now observe fault frequency harmonics which we can use to confirm & analyze. The integrated bearing fault calculator can confirm inner/outer ring, ball pass or cage defects.

Faulty Slow Speed Bearing

This is a Time Series view of a bad slow speed bearing (less than 25 RPM). The heterodyned audio signal we can listen to gives us already a clear indication of bad condition. When analyzing slow speed bearings it may be difficult to get a good reading in FFT view. However, the defects are very apparent in the Time Series view.

Bearing Being Over Lubricated

This is a time series view of a bearing being over lubricated. If too much lubrication is added the amplitude will rise. If this occurs, immediately stop adding lubricant. You can observe and hear the drop in amplitude and the rise in amplitude when too much lubricant was added.

 


Gruetzner