Extracting near-borehole P and S reflections from array sonic logging data

There has been widespread interest in acoustic reflection logging recently due to its higher resolution compared to seismic exploration and larger lateral extension compared to logging techniques. Acoustic reflection logging is particularly useful for delineating near-borehole subtle structures and fractured reservoirs. The primary data processing of acoustic reflection logging is to extract the reflection signals from the full waveform records which are overwhelmingly dominated by the borehole direct modes. We present in this paper a method combining a frequency–wavenumber (f–k) filter with a modified parametric prediction to separate the direct and reflection waves. Afterwards, we apply dip stacking to enhance and to further separate P–P and S–S reflection signals, and suppress the residual direct waves and other noises. Processing examples obtained with synthetic and real data show the capabilities of the proposed workflow and algorithms.