Research on pinhole accidental gas release in pipelines: Statistical modeling, real gas CFD simulation, and validation

Abstract

The successful risk mitigation of Natural Gas (NG) leakage to reduce its environmental and economic impact depends chiefly on timely detection leaks and predicting the amount of gas release. In the present study, we investigated pinhole leaks and predicted the gas release rate for various leak to pipe diameter ratios and operating pressures ranging from 2 bar to 110 bar. We first set up a laboratory-scale experiment. The generalized likelihood ratio (GLR) is used as an advanced statistical hypothesis testing technique to detect any shifts in the mean and variance of the process measurements in real-time. These results improved the understanding of several leak detection systems and contributed to a reduction in false alarms in these systems. The presence of a leak was flagged almost immediately after it occurred, indicating the speed and efficiency of statistical techniques in detecting microleaks. The present computational fluid dynamics (CFD) study provides details on the entire leak flow field that are not possible to obtain with experimental methods. CFD is a critical tool in process safety management, helping to identify, analyze, and mitigate potential risks in natural gas pipeline. The CFD study proposes new correlations to predict the nominal leak volume flow rate by investigating the influence of leak size, pipe diameter, and pipe pressure for wide ranges of pressures. The correlations were derived from three-dimensional transient detachable eddy simulation model in a commercial CFD code (ANSYS Fluent R3). The correlations enhance the conventional models by incorporating the gas compressibility effect for high-pressure conditions. The percentage of error between the results of the CFD and delivered correlation fluctuated between 4% and − 5%, demonstrating the high accuracy of the new correlation.