TABLE OF CONTENTS Broadly speaking, the processing task is that part of the monitoring activity where data, which has been collected and suitably formatted, is operated upon or otherwise transformed so that a diagnosis of plant condition can readily be made. As we have mentioned in the previous chapter, it is here that the most significant scope for automation exists. Indeed, to perform many of the data processing functions now commonly used in condition monitoring, considerable computational power is required. Processing may be done on-or off-line and this choice will predominantly depend upon whether the monitoring system is one that operates on a continuous basis or not. It also depends on the understanding regarding how quickly the monitored faults can develop. Usually, condition monitoring in a continuously operating plant or those where faults that can develop quickly, should allow the obtained data to be processed on-line. However the distinction cannot be clearly drawn. Modern data acquisition units now permit data to be obtained from a running plant and analysed at a different location without any significant time delay.
In the past, dedicated signal analysers were applied to analogue voltage signals from the transduction process to obtain the spectra that could indicate the occurrence and level of signal components at particular frequencies. Their use has now almost been completely replaced by general computer algorithms acting on data acquired in a digital format.
Although commercial computer software packages are available for digital signal processing applications and can be used for condition monitoring of...
