Gas migration in soil from a distribution network leakage: an experimental and numerical study to assess time and spatial scales

摘要

Gas distribution networks are submitted to various aggressions and loads that may sometimes lead to involuntary leaks: external interference, corrosion, sealing issues, etc. Gas may therefore be released and may follow different paths below ground, through cracks, along pipelines or below the road surface. The objectives of this work are to quantify buried gas leak flow-rates and to determine the extension of the flammable gas volume, and the time needed to reach a steady state. In the framework of a GERG (Gas European Research Group) project with Gas Natural, National Grid, E.ON Technologies and GDF SUEZ as partners, realistic configurations that are representative of gas distribution networks in Europe were defined in terms of pipe diameters, burying depths, structure of urban near-surface soils, etc. Then, based on the previous findings, GDF SUEZ developed an experimental method to quantify buried gas leak flow-rates as well as time and spatial scales of gas migration in soil. A test rig of 9.5x8.5x2.7 m~3 filled with compacted sand was set up, and methane was injected at 1 m depth. Two leak diameters (1 mm and 5 mm), and several inlet pressures varying from 40 mbar_g to 15 bar_g were tested. Resulting flow-rates ranged between 0.2 L(n)/min and 24 L(n)/min. Gas concentrations were measured all around the sand volume, thanks to numerous vacuum probes. For all experimental cases, a steady state was reached, which means that flammable gas volume does not increase without bounds, due to the existence in this case of a balance between lateral spreading and buoyancy. The large amount of experimental data acquired during this campaign first allowed to validate a formula linking the gas leak flow-rate to known soil characteristics. This formula mainly depends on the inlet pressure, the leak diameter, the soil permeability and the Forchheimer coefficient. These two last parameters directly depend on the type of soil but are not well referenced in the literature. Thanks to the instrumentation developed within this project, it is now possible to estimate them by performing simple instrumented injections of compressed air below the ground surface. Then, numerical codes were compared to the experimental results to assess their ability to simulate gas migration in soil. TAGS, which is a 3D code developed by GDF SUEZ, OSAKA GAS and TOKYO GAS in the 90's and after only by GDF SUEZ, shows satisfactory results. It is able to properly predict the time constants of the phenomenon and it is conservative in the calculation of the flammable gas volume for the simple experimental set-ups explored until now. In parallel, other codes have been tested, as ANSYS CFX, with encouraging results. In order to qualify numerical simulation codes on more complex and more realistic experimental set-ups and to produce additional results about weather and soil influence, the test equipment was transferred to a test field in Germany. Here, another test series took place and data from a rural gas pipeline situation as well as from a scenario with service lines close to cellar walls were acquired.