A COMPARATIVE STUDY OF TECHNIQUES USED TO GENERATE SYNTHETIC SONIC LOGS FROM AVAILABLE WELL LOG DATA

Abstract

Sonic Logging is an indispensable tool for characterization of subsurface lithology and quantifying other properties like permeability, cement bond quality, secondary porosity etc. Apart from using velocities in shales for predicting the mechanical properties of the well and hence, to foresee borehole collapse (Horsrud 2001), sonic log data could be used to estimate overpressure in shales (Lubandazio et al. 2006, Dutta et al. 2002, Bowers & Katsube 2002, Hermandud et al. 1998). But quite often sonic log data is either absent or highly erroneous. Road noise and signal attenuation may greatly decrease the quality of the sonic log data resulting in spiking and cycle skips. This may in turn lead to errors in interpretations. In order to counter this problem, sonic logs are synthetically generated using other available well log data. Data from logging techniques like Micro-Spherically focused Log, Neutron log; Density log etc. can be used to obtain transit times by applying suitable theoretical and empirical relations. In this study, Synthetic Sonic Log (SSL) for 6 different Wells in the PY-1 Field (off the coast of Pondicherry, India) has been generated using various techniques. The Archie and Faust equation have been used to get transit times from resistivity log data. Porosity value evaluated from Density Log has been inverted using theoretical and empirical relations like Wyllie's, Raymer-Hunt, and Raymer-Hunt-Gardener to get transit times. The RMS value of error percentage was found to be least for methods which accounted for the presence of shale in the rocks.