Abstract: The existence of underground shallow gas will pose significant hazard and risk to the investigation, construction, and subsequent operation of subway projects. In this study, the probabilistic method based on geostatistics and reliability analysis is used to analyze the resistivity piezocone penetration test data of the Tangxinline station site of Hangzhou Metro Line 7. Regression analysis is performed to remove the trend term of resistivity data, and the variogram function is used to describe the anisotropy of the resistivity residuals in the vertical and horizontal directions. Kriging interpolation is used to obtain the optimal linear unbiased estimation of non-sampling points. By considering the resistivity background values of different soil types, three-dimensional probability maps of the subsurface shallow gas are obtained using the first-order reliability method. The results show that the resistivity residual has significant anisotropy, and the autocorrelation distances in vertical and horizontal directions are 5.1 m and 55.6 m, respectively. In the vertical direction, the shallow gas is distributed in the vicinity of 25 m and the depth interval of 30−35 m, and in the horizontal direction, it is mainly distributed around RCPTU1 point. The results of this paper can provide basis for shallow gas monitoring plan, exhaust hole layout and other treatment measures.

Key words: shallow gas, resistivity piezocone penetration test, geostatistics, reliability-based analysis