This is the predominant technology employed in condition monitoring to determine the condition of various mechanical components and be able to make a prediction on service life expectancy and, hence, urgency of repairs. It is a non-invasive technique that collects vibration data from machines under normal operating conditions. Vibration readings are typically taken at bearing locations following accepted industry practices and conventions.  This raw waveform data is then transformed into spectral format and analyzed with proprietary software that allows analysis to take place.

Vibration Analysis provides the best indication of machine condition and the presence of faults as it can detect imbalance, misalignment, looseness, bearing and gear defects, lubrication problems, resonance, and cavitation, amongst others. Trending data over time allows wear severity and progression to be monitored, improving the predictive capabilities.

A high degree of skill and training is required to be able to interpret vibration data. As such, reports are tabled in non-technical terms and use a traffic light system to illustrate the severity of the fault detected. Green suggests monitoring is within allowable limits, amber means a fault/issue has been identified and requires some attention or further monitoring, whilst red indicates a severe fault that requires immediate attention to avoid equipment failure.

Limitations of this technology are largely related to the frequency and operating conditions under which the data is collected. Occasionally, faults can develop quickly and failures occur before the next scheduled survey. Identifying critical equipment and understanding the maintenance history will assist in developing an appropriate program.