Detection and Location of Microseismic Events with Low Signal-to-Noise Ratios

Abstract

Abstract Microseismic monitoring of hydraulic fractures is the process of monitoring the small earthquakes induced by fluid injection during hydraulic fracture stimulation. The primary goal of microseismic monitoring is to map the source positions of microseismic events. However, in actual cases, location of the microseismic events could become problematic because a considerable part of the microseismic events show a low signal-to-noise ratio (SNR) and conventional event detection and location techniques tend to omit these events even though they may provide useful information for delineation of the hydraulic fractures. To address this problem, several improved methods are proposed in this study to increase the detectability and location accuracy of microseismic events with low signal-to-noise ratios. It has been well known that microseismic data processing generally composes of several critical steps, such as event detection, arrival picking, velocity modeling and source location. As for the event detection process, we develop a multi-channel semblance coefficient based method to identify the low SNR microseismic events. Our method utilizes a semblance coefficient to quantitatively determine the waveform similarity of windowed record segments after moveout correction, and uses it as a detector of the presence of a microseismic event. After identifying the events, a robust method is employed to pick their P- and S-wave arrival times. This method, referred to as SLPEA algorithm, is developed by integrating the differences in amplitude, polarization and statistic properties between seismic signal and ambient noise. Once the arrival times are gained, a simulated annealing (SA) based joint inversion algorithm is adopted to invert the velocity models and microseismic source positions. Innovations in this method include the use of a probability distribution function for determination of the source azimuth and a joint objective function for inversion of the velocity model and source position. The performance of the proposed methods is illustrated using field dataset recorded during an 11-stage hydraulic fracture treatment. 521 locatable microseismic events are detected from this dataset and nearly one third of them show a low signal-to-noise ratio. The source positions of these events are successfully gained through using the proposed methods, and analysis of the location results indicates that the low SNR microseismic events can reveal more details about the hydraulic fractures. Introduction Hydraulic fracture stimulation is a crucial tool for the development of unconventional oil and gas reservoirs, as it can create connected pathways within the low-permeability reservoirs, such as coalbed methane and shale gas. Over the last few years, microseismic monitoring (MSM) has rapidly grown as the most effective technology to monitor various reservoir processes, especially for mapping hydraulic fractures (House et al., 1996; Maxwell et al., 2000; Urbancic and Rutledge, 2000; Maxwell et al., 2002; Rutledge and Phillips, 2003; Maxwell, 2010). During a hydrofracture treatment, pore-pressure changes due to fluid injected into hydrocarbon reservoirs are usually accompanied by the occurrence of numerous small earthquakes, namely microseismic events (Warpinski et al., 2001; Shapiro and Dinske, 2009; Mayr et al., 2011). Monitoring of these microseismic events can be used to understand the fracture geometry, including the orientation and dimensions, and further used to optimize the late-stage treatment and design future stimulation programs (Fisher et al., 2002; Le Calvez et al., 2005; Wolhart et al., 2005). Other applications of the microseismic technology include monitoring of steam injection (Maxwell et al., 2007; Maxwell et al., 2008) and CO2 storage (Maxwell et al., 2004; Verdon et al., 2010).