INTRODUCTION
Detection and characterization of permeable fractures are of great interest in oil exploration because they are related to the transportation and storage of hydrocarbon and, therefore, give the direct information for the optimization of production. We conducted modeling studies to utilize Borehole Acoustic Reflection Survey (BARS) data for the detection of fractures. Schoenberg’s slip boundary condition (1980) is used to model the fractures. Significant differences in recorded signals from a fracture and bed boundary are predicted. We applied the modeling technique to interpret the field data acquired in China and migration of fractures was tried.
Detection and characterization of permeable fractures are one of the primary roles of downhole logging. The resistivity tool which images cracks on the borehole wall is widely used to detect fractures. The sonic tool can be also used to find Stoneley wave reflections and transmissions which imply the existence of fractures (Hornby et al., 1989). As the acquired information by these measurements is limited to the vicinity of the borehole wall, the fracture extent cannot be estimated. In order to obtain the fracture extent, measurements using waves which penetrate the formation with deeper depth of investigation are necessary. Although the event signals from fractures are expected to be recorded in sonic waveforms, Borehole Acoustic Reflection Survey (BARS) (Hornby, 1989; Esmersoy et al., 1998) is rarely acquired for the detection of fractures because the acoustic and elastic responses from a fracture are not well understood. If BARS can be available for the fracture detection, fracture extent can be estimated. The detection of fractures from BARS measurements was discussed by Yamamoto et al. (1999) focusing on the P-to-S, S-to-P transmitted and P-to-P reflected waves. We investigated appropriate event signals to be used for fracture detection using a ray based modeling technique. The boundary conditions across a fracture play fundamental roles in simulating a fracture’s response. The boundary condition expressing the fracture was first proposed by Schoenberg (1980) and many models have been proposed (e.g., Nakagawa and Schoenberg, 2007). In this study, we used the model proposed by Schoenberg (eq. 28, 1980) for the boundary condition because our main interest is to find open fractures which will allow perfect slip between two interfaces and an appropriate compliance is difficult to be imposed. In order to consider various types of event signals, radiation from the source and response of the receiver in fluid filled borehole are considered following the method by Schoenberg (1986). We focused on the waveforms in the common offset gather (COG) which is a correction of waveforms for a particular receiver station of the multi array sonic tool. Fig. 1 shows arrival times of event signals from a fracture displayed by the COG gather. As shown in Fig. 1, various event signals may be recorded in data. Therefore it is important to know the signals having dominant amplitudes for signal processing and migration. In this paper we only discuss events of an isolated fracture and P-SV response because of the limitation of space.
