Seismic monitoring of the STIMTEC hydraulic stimulation experiment in anisotropic metamorphic gneiss

Abstract

In 2018 and 2019, we performed STIMulation tests with characterising periodic pumping tests and high-resolution seismic monitoring for improving prognosis models and real-time monitoring TEChnologies for the creation of hydraulic conduits in crystalline rocks (STIMTEC). The STIMTEC underground research laboratory is located at 130ĝm depth in the Reiche Zeche mine in Freiberg, Germany. The experiment was designed to investigate the rock damage resulting from hydraulic stimulation and to link seismic activity and enhancement of hydraulic properties in strongly foliated metamorphic gneiss. We present results from active and passive seismic monitoring prior to and during hydraulic stimulations. We characterise the structural anisotropy and heterogeneity of the reservoir rocks at the STIMTEC site and the induced high-frequency (>1ĝkHz) acoustic emission (AE) activity, associated with brittle deformation at the centimetre-to-decimetre scale. We derived the best velocity model per recording station from over 300 active ultrasonic transmission measurements for high-accuracy AE event location. The average P-wave anisotropy is 12ĝ%, in agreement with values derived from laboratory tests on core material. We use a 16-station seismic monitoring network comprising AE sensors, accelerometers, one broadband sensor and one AE hydrophone. All instrumentation was removable, providing us with the flexibility to use existing boreholes for multiple purposes. This approach also allowed for optimising the (near)-real-time passive monitoring system during the experiment. To locate AE events, we tested the effect of different velocity models and inferred their location accuracy. Based on the known active ultrasonic transmission measurement points, we obtained an average relocation error of 0.26±0.06ĝm where the AE events occurred using a transverse isotropic velocity model per station. The uncertainty resulting from using a simplified velocity model increased to 0.5-2.6ĝm, depending on whether anisotropy was considered or not. Structural heterogeneity overprints anisotropy of the host rock and has a significant influence on velocity and attenuation, with up to 4ĝ% and up to 50ĝ% decrease on velocity and wave amplitude, respectively. Significant variations in seismic responses to stimulation were observed ranging from abundant AE events (several thousand per stimulated interval) to no activity with breakdown pressure values ranging between 6.4 and 15.6ĝMPa. Low-frequency seismic signals with varying amplitudes were observed for all stimulated intervals that are more correlated with the injection flow rate rather than the pressure curve. We discuss the observations from STIMTEC in context of similar experiments performed in underground research facilities to highlight the effect of small-scale rock, stress and structural heterogeneity and/or anisotropy observed at the decametre scale. The reservoir complexity at this scale supports our conclusion that field-scale experiments benefit from high-sensitivity, wide-bandwidth instrumentation and flexible monitoring approaches to adapt to unexpected challenges during all stages of the experiment.