Leakage locating and sampling optimization of small-rate leakage on medium-and-low-pressure unground natural gas pipelines

Abstract

The leakage on underground natural gas pipelines leads to not only energy loss but also a significant threat to the nearby region. Because of atmospheric dilution, strong wind, and complex underground conditions, small-rate leakage is particularly challenging to locate accurately for reparation. In this study, a series of field experiments with real-gas are carried out. Base on experimental data, a steady dispersion model of underground gas leakage is established and calibrated. This model with inversion method can be used to guide the locating of leakage. To furtherly improve the locating accuracy, two optimized sampler placing strategies based on the Fisher information matrix (FIM) and the covariance matrix are compared to the currently-used method in both simulations and field experiments. with optimized sampling strategies, the locating errors can be reduced by 14–54% with the same number of sampling points. This study show that first-principle steady-distribution model can be used to locate the leakage base on underground concentration samples and achieve the required precision. The current sampling strategy can also be optimized to improve the accuracy. Overall, this work could improve the safety of gas supply and reduce the emission of greenhouse gas.