Arrival-time Picking and Slowness Estimate on Sonic Data

In this abstract, we introduce a novel technique that is suitable for picking compressional arrival-times. High depth-resolution slowness estimate are then derived from the obtained arrival-times. This technique is designed for sonic array data as follows.

First, a fast and robust model driven multi-waveform algorithm is applied to the data to guarantee a reliable detection of the onset arrivals. For this purpose, the Short Time Average/Long Time Average operator is applied to every waveform. The output of this operator is used to generate a set of candidate arrival-times. Finally, a Hough transform is applied to the set of arrival-time candidates available in a shot gather. Thanks to the robustness of the multichannel model driven approach, a detection of the onset of compressional arrivals is obtained, yielding a good estimate of both compressional arrival-times and slowness values.

The results of the model driven approach are then, used to guide the picking of the true arrival-times of the compressional waves. Finally, the obtained arrival-times are utilized to derive a high depth-resolution slowness log.

The introduced algorithm is fully automatic. We demonstrate the efficiency of our new technique on real sonic data acquired with both Wireline and Logging While Drilling sonic array tools.