Distribution of the Vp/Vs Ratio within the Basel 1 Geothermal Reservoir from Microseismic Data
Dr. Jörn Kummerow
We estimate the spatial distribution of the Vp/Vs ratio using microseismic events. High-precision P and S arrival time differences must be calculated for pairs of microseismic events. Then, Vp/Vs ratios are determined from the slope of distributions of differential P- versus S-travel times. If done for all elementary spatial cells, this will provide a 3-D Vp/Vs ratio image. Application of this approach to the Basel case study shows a good agreement with the sonic logs.
Microseismic reflection imaging of stimulated reservoirs and fracture zones
Dr. Anton Reshetnikov
In this work we present an approach which uses waveforms from induced seismicity to build a detailed high resolution image of a stimulated reservoir. The idea of Microseismic Reflection Imaging (MRI) is to treat a located microseismic event as an active seismic source and to apply migration techniques adapted from reflection seismics. To image microseismic reflection data we perform a number of preprocessing steps and use a directional migration approach. To obtain polarization of reflected waves, we consider auto- and cross-variances of a seismogram within a time window. We also apply a hodogram linearity threshold as a criterion to exclude parts of seismograms with unreliable polarity estimates. The MRI approach is applied to the data from the stimulation of EGS at Basel, where microseismic waveforms were recorded at 6 shallow (about 200 m) and the deep (2.5 km) receivers. We show separate images of PP and SS reflections for different receivers. Finally, we provide a joint interpretation of obtained images and event locations. We discuss how microseismic reflection imaging can complement to surface seismic imaging, what the reflectors in microseismic reflection image can tell us about the stimulated reservoir and how these images can contribute to the reservoir characterization.
Multiplet Based Extraction of Geological Structures
A sound analysis of microseismic waveforms was a subject of the following contribution. In recent studies we introduced two new measures, CM and TM, to analyze waveform similarity of events. These measures are based upon cross correlation of seismic traces and inter-event S-P travel time differences. Application of these measures to the Carthage Cotton Valley dataset reveal clusters of seismic events representing single natural fractures. We investigate pairs of seismic events characterized by higher relative S-P travel time differences and high waveform similarities. These pairs can be formed by events which do not belong to the same crack but to the same type of a geological regime. Assuming a system of parallel cracks within a geological formation and a similar stress distribution, high waveform similarity between events occurring at different cracks can be expected due to almost identical source radiation patterns. Multiplets formed by those events have the potential identify the existence of a system of similar cracks.
Reflections at a Thin Fluid Layer Representing a Hydraulic Fracture
Frequently we observe significant microseismic reflections indicating the presence of strong reflectors. The main goal of our work on this topic is to find reflection coefficients for a wave reflected at thin fluid layer modeling hydraulic fractures. We consider a thin layer of an ideal fluid which is embedded into an elastic and isotropic solid. We present full analytical solutions for the reflections of an incident P-wave, the P-P and P-S reflection coefficients, as well as for an incident S-wave, the S-S and S-P reflection coefficients. We find that for parameters in the range of microseismic monitoring, which means a layer thickness of h = 0.001−0.01m and frequencies of f = 50−400Hz, the reflection coefficients depends on the Poisson’s ratio. Furthermore the reflections of an incident S-wave are significantly high. This possibly impacts our images obtained from microseismic data.
The Influence of Anisotropy on the Location of Microseismic Events
We consider possible implications of not taking into account the velocity anisotropy effect while locating microseismic events. We compare two algorithms of computing P- and S-wave travel times in anisotropic media. The first one is based on weak anisotropy approximation and computationally relatively "cheap" while the second one is based on the solution of Christoffel equations and provides better accuracy but is less stable. The two methods give almost identical results for P- and SH-waves, but have a large difference for SV-wave. We show that not taking into account anisotropy may lead to significant location errors even in the case of weak anisotropy.
We are now developing a new algorithm for the location of microseismic events in anisotropic media. The influence of anisotropy is taking into account in an additional step after the conventional location procedure. The perturbations of velocity anisotropy and the resulting location misfits are estimated simultaneously in this step.
Microseismic events span a source area from a few meters to a few hundred meters in length. The understanding of the parameters that control processes at this scale is less developed than e.g. for earthquakes of larger size or acoustic emissions in the laboratory. Questions of direction, rupture length, magnitude, and dependencies on the surrounding conditions are of high interest for example in hazard assessment.
We investigate source characteristics of microseismic events using multiple approaches that were borrowed from global seismology. For example the use of back projection or empirical greens functions can be adapted to be applicable at reservoir scale. This then provides insights to rupture location, direction, dimension, or velocity. Together with fault plane solutions and cluster analysis e.g. activated faults can be identified and the understanding of the meaning of microseismic events can be driven forward at a fundamental level.