Story of Surprise failures of cooler fans

What Happened?

These cooler fans (three in number), which are operated at different speed through variable frequency drive are critical for operation. In case of a breakdown of any of these cooler fans, kiln production of the cement plant comes to a stop.

In this particular case two of the fans suffered surprise failures of the fan non-driven end bearing. Such bearing failures happened thrice. Though the bearings were being regularly monitored by vibration analysis the faults could not be detected in time for the maintenance department to take timely action so as to avoid failures and consequent plant stoppage. This was a matter of concern for the management as to why condition monitoring could not provide any early warning of the bearing failures.


To accurately ascertain the root cause of bearing (non-drive end bearing) failure of Cooler Fan.

Cooler Fan of a Cement Plant Fig 1
Another view of the same cooler fan Fig 2

Pictures of the failed bearing of Fan Non Driven End (FNDE)

Outer Race Fig 3
Outer Race (another view) — Fig 4
Inner Race, Fig 5
Inner Race, Fig 6
The Rollers Fig 7
Broken cage Fig 8
Detail of the fractured cage, Fig 9


To conclusively arrive at the root cause of failure of the said bearing we have based our analysis on the dynamics of the system. Dynamics of the system was observed through the collated vibration data of the fan in terms of overall vibration readings in velocity and relevant vibration signatures in terms of displacement, velocity and acceleration. In total, three sets of readings were taken on three different dates, under different speed conditions, which are as follows:

1) 24.09.2016 with operating frequency of 49.56 Hz

2) 25.09.2016 with operating frequency of 51.21 Hz

3) 26.09.2016 with operating frequency of 49.56 Hz

Basic Observations:

1) At 49.56 Hz speed 1 times running rpm (revolutions per minute) peak (fundamental frequency of 49.56 Hz) was observed at both bearings of the fan in velocity mode and high frequency peaks were observed in acceleration spectrum primarily in axial direction. .

2) At 51.21 Hz speed, non-synchronous peaks at 5.24 times running speed (frequency of 268 Hz) and its harmonics were observed in both velocity and acceleration spectrums and the high frequency peaks of high energy content were observed in the acceleration spectrums in all three directions — axial, vertical and horizontal.

3) However, the non synchronous peaks and high frequency, high energy content peaks simply disappear when the system is run at 49.56 Hz.

Root cause:

The root cause of rapid failure of the non-driven end bearing is Resonance.

Phenomena progression:

Speed goes up to 51.2 hz -> resonant frequency excited-> creates axial movement -> force -> distortion of cage and balls -> failure

Deeper explanation:

1. It can now be confirmed with certainty that the root cause of failure of the bearing is due to resonance which appears when the system is run at 51.21 Hz.

2. Since the resonance phenomenon acts as an alternating stress the failure happens as a low cycle fatigue. It means that the failure happens over a short duration of time after the system hits resonance.

3. Since the primary force is in the axial direction, the movement of the self aligning bearing would be more pronounced in the axial direction in the non-locating bearing which incidentally happens to be the non-driven side bearing. Hence the non-driven end bearing suffers more than the driven end bearing. Therefore, the failure was observed in the non-driven end bearing. Further, owing to excessive axial movement of a self aligning bearing under the impact of resonance the balls hit and rub with the cage and damages it in no time. This explains the extensive damage of the cage and the balls as observed at site, all of which happened over a very short interval of time.

4. Moreover, with the present bearing arrangement (floating bearing arrangement) it is important that one ring of each bearing should be able to move on or in its seat, preferably the outer ring in the housing. So, when resonance strikes; the bearing arrangement suffers both in terms of instability and varying loads. This aggravates the problem and hastens failure when subjected to resonant conditions.

5. In this context, it is not possible to monitor the health of the bearing with the explicit objective of giving a substantial period of time to respond to an observed incipient failure unless the system is modified to keep resonant zone well beyond the operating range. Even if the bearing health is monitored at the operating speed of 51.21 Hz it would provide the maintenance department too little time to proactively respond to a measured deterioration since deterioration, in this particular case, is subjected to low cycle fatigue, which causes rapid deterioration to failure.

Possible Solution:

1. Change the bearing arrangement so as to make it non-resonant over the entire speed range. Pay special attention to the non-driven end bearing in terms of allowance for movement in the axial direction and stiffness to keep the bearing resonant frequency well below the disturbing resonant zone or operating zone. Several arrangements are possible keeping the function of the bearing in mind. Bearing manufacturer may be consulted in this regard. This offers the best solution with long term benefits. To aid right selection, the following link may be referred to under section “Floating bearing arrangement.”

Some Vibration Data:

© Dibyendu De