Reflection Imaging of Egs Reservoirs Using Microseismicity as a Source

The authors have been developing a new variation of passive reflection technique, where microseismic multiplets are used as a source (multiplet reflection imaging using microseismicity: multiplet RIM). Closely similar waveform and precisely determined hypocenter of the multiplet can be effectively used to image structures inside/around the seismically activated EGS reservoirs. In the multiplet RIM, artifacts and blurred reflection image due to limited number of sources and detectors is suppressed by identification of relative/absolute time of arrivals of reflected phase from the multiplets and new concept of inversion. The multiplet RIM has been applied to a data set collected at Soultz, France. The coherence index, which is a measure of arrival of reflected phase in the multiplet RIM, suggested that coherent reflected phase among the multiplets were successfully identified. The reflection image showed that there are possible reflectors at the bottom of shallow reservoir and surrounding zone. The index of reflectivity along a borehole at the site was consistent with results from FMI and PTS logging. INTRODUCTION Techniques which enable us to delineate structures inside granitic basement are of importance for development and monitoring of EGS reservoirs. However, in practice, conventional seismic techniques, which have been mainly developed in oil industries, have insufficient ability for this purpose in most of the cases. This is because of high reflection coefficient at the top of the basement and strong attenuation in overburden with a thickness of several kilometers. While, high temperature, pressure, and geothermal liquid and gas inside/around the reservoirs avoid deployment of downhole source or detector for VSP and crosshole measurement. Soma et al. (2002) have developed a reflection method, which uses natural or induced microseismic events (acoustic emission: AE) as a source (AE reflection method), and its capability for imaging structures inside and around the hydraulically stimulated reservoir at Soultz, France, has been successfully demonstrated. Reflected waves are identified by evaluation of linearity of three dimensional motion of the seismic wave (hodogram) at an observation point and diffraction stack migration is used to obtain reflection image in the AE reflection method. One of the significant advantages of the AE reflection method is that downhole triaxial monitoring of the microseismic signals at one station enables reflection imaging. However, ellipsoidal reflectors appear because of limited number of sources and detectors, and, therefore, techniques to image structures with higher reliability has been expected. We commonly observe groups of microseismic events which have highly close waveforms despite of their origin time and magnitude. Such events are referred to as “microseismic multiplet” and coherence-based relative phase picking and relative mapping technique of their hypocenters have been developed (Moriya et al., 1994). It has been accepted that the relative hypocentral location of the multiplets are more precisely determined because relative time of arrival among the events can be estimated with high accuracy in the frequency domain. The authors have been trying to incorporate characteristics of the microseismic multiplets in principles of the AE reflection method, because highly similar waveforms of the multiplet can be used like a “repeatable source” and precisely determined relative hypocenters would improve reliability of reflection image. In this paper, we describe principles behind the “multiplet reflection imaging using microseismicity (multiplet RIM)” and its application to a dataset collected at Soultz, France. PRINCIPLES OF THE MULTIPLET RIM The concepts of the original AE reflection method and the multiplet RIM are compared in Figure 1. In both methods, time of arrival and state of polarization of the direct arrival and the reflected phases are estimated by the analysis of 3D hodogram. Figure 1: concepts of the original AE reflection method and the multiplet RIM. Principles of the signal processing and migration in the multiplet RIM is schematically shown in Figure 2. Detail of the main procedures is described below. Figure 2: Principles of the multiplet RIM. (a) identification of reflected phase. (b) inversion. (a) Identification of reflected phase: It is assumed that the state of polarization of the 3D hodograms of reflected waves from one reflector has close similarity among microseismic multiplets. Meanwhile, there is a difference in the time of arrival of the reflected wave, because the distances from the hypocenter to the reflector are unique for each event. Spectra ) , , ( τ f t Sij (i: event identification, j: component identification) are estimated using a short-time Fourier transform for one pair of the multiplet set. We evaluate the similarity of the 3D hodogram in the time and frequency domains using a 3D timefrequency coherency (3D-TFC) function (Asanuma et al., 2001). Here, we calculate the 3D-TFC function shifting a window for one event by τ for every window time t . The 3D-TFC function used in the multiplet RIM, thus, has three parameters; time t in the reference event, frequency f , and shift τ relative to the reference event, and it is represented as,