A transformation technique for imaging offset VSP

Abstract

The objective of this research was to develop a transformation procedure to image offset VSP. The transformed VSP is readily interpretable and easier to tie with surface seismic. A model-dependent transformation approach applicable to a general 2-D subsurface model has been developed. The transformation works by first determining a velocity-depth model consistent with the VSP data. Then using the model and raytracing we determine and relocate each primary event at its correct reflection point. The transformation generates both time and depth sections. The tune section generated from the depth model using the concept of image ray tracing. Since the transformation is basically an application of a moveout correction to the VSP traces), its effect is to distort the original waveform on the VSP trace and consequently smear the reflection event in both space and time. This smearing of the events degrades the lateral resolution of the transformation and therefore must be minimized by using optimum aquisition and processing parameters. An important step in the transformation procedure is the stacking of the moveout corrected traces at constant com m on reflection point (CRP) intervals. In this research the criteria for the determination of the optimum bin size is defined. The minimum bin size must be such that with in the zone of subsurface coverage there is at least one reflection point in each bin on each reflector. The maximum bin size must be such that the distortion across the bin is less than some acceptable value which depends on the objectives of the survey. The above criteria are particulary advantageous when stacking multi-offset VSP data. One of the main objectives of this research is to investigate the class of 2-D models for which the transformation is applicable. Time transformation is applicable to all classes of 2-D model for which the VSP time T is single-valued in the two-way timer and also the emergent image ray offset E is not multi-valued in r. Depth transformation is applicable to all classes of 2-D model for which the VSP time T is single-valued in the reflection depth Z and also the reflection offset X is not multivalued in Z. Since the transformation is model-dependent, it correctly transforms only those events whose dip, depth and velocity are included or interpolated from the model. All other events though not imaged perfectly, will be positioned in a manner that facilitates interpretation. The transformation was successfully tested on both synthetic and real VSP data.