Acoustic Emission Techniques for Condition Monitoring and Diagnosis of Rotating Machines

Detection of abnormalities and defects in rotating machines is of prime importance for
industry and equipment. It avoids the occurrence of unplanned down time of machines that
causes serious damage and costly downtime. Several condition monitoring techniques have
been applied to improve the plant reliability and reduce the downtime. Generally, bearing
represents the commonly used part in all rotating machines and equipment. Various rotating
machine problems are encountered during plant and equipment operation. These include
overload, unbalance, misalignment and over speed. Any of these problems can cause the
bearings to deteriorate. It is therefore necessary to detect and diagnose the bearing faults in
order to avoid serious problems which could lead to catastrophic machine failure.
This thesis explores the effectiveness of acoustic emission (AE) technique for incipient
detection of faults related to bearings. To understand the pattern of acoustic signals a series
of measurements are performed in an experimental test rig which is developed to examine
the bearings health condition. Initially a number of experiments are conducted for a
collection of healthy bearings. Each bearing is tested at various motor speeds and loads.
Consequently, significant amounts of acoustic data are recorded as reference.
Measurements are extended for poorly lubricated and seeded defect bearings which are
created by electrical discharge machine. Comparison of the results indicated that the
amplitudes of AE signals provide valuable information regarding bearings health condition.
On the other hand, fast Fourier transforms implemented for the acquired time domain
signals have significantly distinguished between the normal, onset, and defective bearings.
The study shows the importance of couplant grease in the transmission ability of the high
frequency acoustic signals.
The AE technique is implemented 111 an industrial plant to monitor and diagnose the
operating condition of the rotating machinery. A couple of AE measurements are conducted
on motors and pumps used in chemical process plant. The results obtained discriminate
between the machines working at normal conditions and that start to deteriorate.
Statistical methods are used to recapitulate the data obtained to demonstrate the
significance and valuable characteristics of acoustic signals. In this regards, the statistical
analyses are carried out for the time domain acoustic data acquired from the experimental
test rig. The results clearly differentiate between normal and defective bearings.