GPS observed horizontal ground extension at the Hutubi (China) underground gas storage facility and its application to geomechanical modeling for induced seismicity

2019 
Abstract Induced earthquakes and ground displacements were seldom reported in relation to gas injection or extraction. The Hutubi underground gas storage (HUGS) facility is the largest one in China and is also a unique case with both earthquakes and ground displacements detected during multiple cycles of injection and extraction since 9 June 2013. Unlike previous studies with a primarily seismological focus, here, we conducted quantitative analysis on the geomechanics of seismicity induced by the HUGS through developing a hydrogeologic framework, which systematically integrated geodetic, geophysical and geological data. First, we measured horizontal ground extension and shortening on the order of cm in response to gas injection and extraction of the HUGS at depth using a local GPS network, which was not reported in other regions with induced seismicity. Second, we synthesized a variety of data, including seismic reflection profiles, a newly acquired local velocity model, rock physics measurements, well drilling and logging data, to build up a 2D geomechanical model for the HUGS. Third, based on fully-coupled poroelasticity, we proposed two methods to optimize the permeability of the upper aquifers as well as the reservoir porosity and permeability with constraints from well level, GPS and well pressure data. Numerical simulations using the calibrated 2D model revealed that the horizontal extension due to the reservoir dilation is larger than ground uplift. The observed seismicity on faults without hydraulic connections to the gas repository was probably induced by the poroelastic effect of reservoir dilation. Our study provided a prototype scheme for detecting and characterizing the geomechanical behavior of cyclic fluid injection and extraction in a deep reservoir, which would be applicable to other UGS facilities.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    40
    References
    18
    Citations
    NaN
    KQI
    []