67 2. Well seismic surveying After wave separation, if the source and receiver are not located on the same line perpendicular to the layers, there is a difference in the processing sequence leading to a seismic image that is optimum for a geological interpretation. This is the most general situation that applies to the following cases, offset VSP, VSP in a deviated well, seismic walkaway and well-to-well seismic surveys. The processing sequence includes: • Deconvolution of upgoing waves. The deconvolution operator is unique. • It is extracted from traces at the bottom of the well and enables the removal of source signal effects. • Normal moveout correction and conversion into two-way time of deconvolved upgoing waves. The purpose of this correction is to compensate for the obliquity of the raypaths induced by the source offset. The aim is to take the acquisition geometry into account. Knowledge of the velocity model is necessary to perform this correction. • Migration: the method most commonly used with VSP is the one proposed by Wyatt and Wyatt (1982). The VSP seismic section obtained after migration is directly comparable to a surface reflection seismic section. The migrated VSP section has a lateral range of investigation of a few tens to a few hundreds of meters. The example shown here concerns data recorded in a highly deviated well on the Wytch Farm Field on behalf of BP-Amoco and partners. Well data were acquired in the F18 deviated well (which reached a maximum deviation of 88.5°) with a vibrator source located at a distance of 1,865 m (Jerry’s Point (JP)) with respect to the wellhead. Recordings were carried out with a CSI-type 3-component well geophone (Schlumberger’s Combinable Seismic Imager Tool). The well geophone was equipped with sensors with a natural frequency of 10 Hz. Acquisition filters were a 2 Hz low-cut filter with a 6 dB/oct slope, and a 330 Hz high-cut filter with a 30 dB/oct slope. The source signal was emitted within the 10 to 80 Hz bandpass range. The duration of the frequency sweep was 16 seconds. The velocity model used to process seismic data was created using the information provided by surface seismic and velocity curves from all wells in the vicinity of the F18 well. The velocity model was refined by inversion of first arrival time picks, minimizing the difference between measured times and the times calculated by the inversion algorithm. The difference between the calculated and measured times did not exceed 3 ms. Figure 2.13 shows the velocity model, the well trajectory, the different positions of the well geophone and the location of source points. For each source point, ray tracing shows the path followed by the downgoing wave.
RkJQdWJsaXNoZXIy NjA3NzQ=