Mean refracted ray path method for reliable downhole seismic data interpretations

The downhole seismic method is one of the most widely used field seismic methods because it is cost-effective and simple to operate compared to other borehole methods. For the interpretation of the data, the direct method is generally used, but this method determines the VS profile roughly and requires an interpreter׳s subjective interpretation. To evaluate the VS profile in detail, the refracted ray path method is used. However, the VS profiles evaluated by these methods often show meaningless repetitive fluctuations with depth and it is because the estimated travel time data is somewhat inaccurate. In this paper, the mean refracted ray path method (MRM), which combines the advantages of both the direct method and the refracted ray path method, is proposed. It provides the VS profile more reliably and automatically. The travel time data is corrected based on the refracted ray path and the R2 value of the regression curve is employed for automation. To verify the proposed method, the synthetic travel time data were generated by forward modeling based on Snell׳s law with some amount of random error added. As the amount of random error increased, the meaningless repetitive fluctuations in the VS profile determined by the conventional methods also increased. On the other hand, the VS profiles determined by the MRM matched the model well and the superiority of the proposed method was thus noted. Finally, the proposed method was applied for the data reduction of several instances of field data. The determined VS profiles were compared with the drilling logs, the SPT-N values, and/or the CPT result, and the reliability and applicability of the MRM was thus verified. •We propose the mean refracted ray path method for downhole data interpretations.•This method has advantages of both direct method and refracted ray path method.•The corrected travel time data is based on the results of refracted ray path method.•The R2 value of the regression curve is employed for dividing layer boundaries.•The VS profiles determined by the proposed method matched well with the models.