This paper presents a method for detecting deterioration of turbomachinery in pipeline service. Many pipeline compressors have data logging systems, but do not fully exploit the resulting data. This paper describes a method to predict how expected values for turbomachinery performance parameters vary with conditions, and provides options for assessing the severity of deviations between measured and expected values. Parameters considered include: centrifugal compressor discharge pressure, centrifugal compressor thermal efficiency, fuel flow, turbocharger speed, axial compressor discharge pressure, and air filter pressure drop. For most parameters, a method of predicting expected values is: first, to map performance over a range of operating conditions, then develop a regression model. Normalizing (dividing each measured value by its expected value) and comparing the result to unity represents one way to monitor for deterioration. Comparing the observed deviation against the statistical variance provides an alternative method to assess the significance of the deviation.

This content is only available via PDF.
You do not currently have access to this content.