| | | |

Case study

Bearing Fault in Large Motor

The following case study was taken from a regular survey of plant located at the site one of our customers.

Introduction

A faulty bearing in a 2800 kW motor was successfully detected using Enveloping filters.

Outline

As part of the initial survey at a gold mine in NSW, measurements were taken on a Semi-autogonous Mill. When the measurements were analysed, the spectrum from the non-drive end bearing showed the characteristic features of an inner race fault. This bearing was then monitored over the next six months before being replaced. When it was inspected, the inner race was found to have been damaged as had been predicted from the vibration analysis.

Method

A portable data collector/analyser and accelerometer were used to take the vibration measurements. The surface of the machine was prepared by grinding a small flat spot on the side of the bearing housing. The accelerometer was attached to the machine using a magnet, with a small amount of grease on the machine/magnet interface.

Measurements of velocity (0-500 Hz), acceleration (0-2000 Hz), acceleration enveloping (5 - 40 kHz) and Spectral Emitted Energy (SEE) spectrum (250 KHz centre frequency, demodulated to 0-1000 Hz) were made.

Results

At the first survey, it was noted that both the enveloping and SEE measurements showed spectra consistent with an inner race bearing fault (fig 1,2), while the fault was not at all visible in either velocity or acceleration spectra.

Figure 1: SEE Spectrum from Damaged Bearing

Figure 2: Spectrum from Acceleration Enveloping

Measurements were made monthly over the next four months, and the trend showed a rise in vibration level, particularly in the SEE spectrum (indicative of metal-to-metal contact). Prior to the next survey, some extra grease was applied to the bearing, which resulted in the SEE level being reduced, as can be seen in figure 3. However, one month later, the SEE level had risen again, as had the envelope levels, so the decision was taken to replace the bearing. Upon removal and inspection of the bearing, it was found to have inner race damage consistent with that seen in figure.

Figure 3: Trend of SEE

Figure 4: Damaged Bearing

Conclusion

Multi-parameter monitoring successfully detected this bearing fault, even though no fault was indicated by the velocity spectrum.

The customer stated that detecting the fault prior to failure justified the cost of the condition monitoring program for the next two years.

Top of page  |  Back to Case Studies