Effect of Pressure Sampling Methods on Pipeline Integrity Analysis

Abstract

A computing program has recently been developed to predict corrosion fatigue crack growth in pipeline steel in near-neutral pH environments. Supervisory control and data acquisition (SCADA) data are used as inputs for crack growth computation. The accuracy of crack growth prediction depends largely on whether the SCADA data have captured all the crack growth–contributing events of pressure fluctuations during the pipeline operation. In this work, statistical analyses of pressure fluctuations during oil and gas pipeline operation were performed to extract those pressure fluctuation parameters that affect corrosion fatigue crack growth in pipeline steel in near-neutral pH environments. High-resolution pressure data, which were recorded either with a very small sampling interval or at the time points whenever a measurable change in pressure was detected, were also modified to generate pressure spectra with different sampling intervals. The pressure points in the spectra could be the actual value of pressure at the time of recording (Method I) or an average value of all the pressure points within a given sampling interval (Method II). It was found that SCADA data generated by choosing large sampling intervals can miss both underload and minor load cycles and yield very conservative predictions, especially for oil pipelines. The SCADA data recorded by averaging pressure points within a given sampling interval (Method II) attenuate the amplitude of pressure fluctuations and yield more conservative predictions than those recorded by Method I. It is recommended that the data of pressure fluctuations during oil pipeline operation be recorded whenever a measurable pressure change has occurred in order to achieve more accurate predictions. In contrast, consistent predictions can be obtained when the pressure data of gas pipelines are recorded at a fixed sampling interval up to 20 min.