Diagnostic Techniques in Condition Monitoring

Abstract

The main purpose of permanent monitoring of rotating machinery is the early detection of fault conditions. This goal can be achieved by identifying abnormal changes in current data in respect with reference values defined on the basis of the data collected during a first monitoring period with the machine in normal state. Furthermore, acceptance regions can be evaluated for each typical operating condition of the plant in order to recognize abnormal changes in the machine vibrations due to fault conditions, deviating from normal changes caused by variations in the process parameters. This monitoring strategy allows to obtain useful information for diagnosing the causes of malfunction conditions.

Some different techniques used to define acceptance regions are shown in this paper. Furthermore, some results obtained by two experiences in monitoring large turbine-generator units are illustrated. The advantages of different diagnostic techniques used to define acceptance regions of monitoring data collected both in steady state and transient conditions are discussed.