Model-guided Geosteering for Horizontal Drilling

Horizontal drilling is an important development in the petroleum industry and it relies heavily on guiding the drill bit with the aid of sonic logging, i.e. geosteering. The quality of sonic imaging depends heavily on the effective suppression of borehole waves and enhancement of weak reflection signal. To this end, we propose an approach to image the near-borehole structure using acoustic logging data. We model the borehole wave propagation using log-derived velocities. The modeled borehole waves are removed from the raw data, leaving reflected signals for imaging interfaces. We tested this method with three sets of data. First we calculated synthetic waveforms for a horizontal well with an interface parallel to the borehole using the 3D finite difference method. The processing result with our method clearly shows the parallel reflecting interface. Next, we conducted an ultrasonic laboratory measurement in a borehole with a parallel Lucite-water boundary. In this case, the interface was also visible in the final image. Finally, we applied this method to a field dataset. In the field dataset, the acoustic logging data were continuously recorded along the well, which enabled us to reject the borehole modes in both common shot gather and common offset gather. The large amount of common offset gather data also allowed us to apply migration to the data. The migrated image of the near-borehole structure is in good agreement with available geological and petrophysical information of that field.