A laboratory experiment has been conducted on a sandstone sample where a small hole has been drilled through the middle of the sample. The intention is to simulate borehole breakout effects and to observe nondouble couple acoustic emission events. We then determine full moment tensors for acoustic emission data from the triaxial laboratory experiment. The full moment tensor inversion uses the first motion polarity and amplitude information as input parameters. The observed data from the laboratory experiment are then compared with synthetic data from a Spectral Element Method simulation using hexahedron elements, which are discretized using highdegree Lagrange interpolants. We find good correlation between X-ray images of the sample and the determined locations, as well as similarities between the synthetic and the observed acoustic emission waveforms.
The interpretation of microearthquakes is an important means to improve the understanding of stress field changes in and around hydrocarbon reservoirs. The primary goal is an accurate location of the microseismic events, which has proven valuable during hydraulic fracturing and long-term reservoir monitoring, especially towards the end of a reservoir’s lifetime. For enhanced geothermal system the location of microearthquakes is a vital tool to estimate the extend of the hydraulically stimulated volume, and is commonly used to determine the target for secondary wells to establish fluid circulation. In addition to the location, source parameters provide insight on the magnitude of the event, stress drop, rupture size, slip, and corner frequency of the seismic phases and these are valuable to characterize each single event, and to investigate the statistics of the bulk of events (e.g. Gibbowicz et al., 1991; Oye et al. 2005). Such estimates, however, can only be determined reliably with sufficient coverage of receivers and knowledge of the velocity model. To determine a full moment tensor for microseismic data recorded in the field often fails due to insufficient instrumentation or due to only little high-quality data. In this paper we investigate acoustic emission data from a laboratory experiment conducted at NGI, where 12 receivers have been positioned around a sandstone sample. Due to the favorable source receiver geometry we conduct a full moment tensor analysis for selected events that have been recorded with sufficient signal-to-noise ratio on all receivers. We then compare the real data with synthetic data from a Spectral Element Method (SEM) simulation. The final goal is to upscale the SEM results and thereby be able to compare acoustic emission data with observed microseismic data from hydrocarbon reservoirs. Figure 1 shows the experimental setup (Aker et al., 2008) with the rock sample mounted inside a tri-axial pressure cell controlling the axial stress, the radial confining pressure and the pore pressure of the fluid inside the rock (Figure 1). The acoustic emissions are monitored by a set of 12 piezoelectric receivers (pinducers) that are positioned around the sample. The resonance frequency of the pinducers is about 1 MHz and the waveforms are recorded at 10 MHz sampling rate.
