Development of lithology-based static Young's modulus correlations from log data based on data clustering technique

• In this paper we developed new correlations for the static Young's modulus based on clustering.
• Clustering will group the reservoir to clusters that have the same facie (lithology) and this will help in the process of prediction.
• Perfect match was obtained using the new technique compared to the previous correlations.

The static Young's modulus is one of the most important geomechanical parameters that are used in the evaluation of the wellbore stability during drilling operations. The static Young's modulus is also important during the design of the hydraulic and acid fracturing operations for conventional and unconventional reservoirs. Static Young's modulus also is important in the evaluation of the in-situ stresses profiles and it can be used to evaluate the reservoir pressures. The existing correlations that determine the static Young's modulus either depend on the dynamic Young's modulus or compressional shear velocity. No previous studies considered the different log parameters to estimate the static Young's modulus. Some of the previous correlations were developed for specific type of lithology and did not consider the lithology variation within the single well.In this paper and for the first time we developed correlations for the static Young's modulus from the log data based on clustering technique. More than 300 measured static Young's modulus values were correlated to the log parameters such as shear transit time, compressional transit time, and bulk density. Using R-project statistical software the clustering was performed for the measured data along with the corresponding log data. Six clusters were identified based on the shear transit time because it has the highest relative importance to the measured static Young's modulus. Six correlations were developed for the six clusters that can be used to determine the static Young's modulus based on the log data. The developed correlations were tested on three cases from the field given the measured core data for each case.The developed correlations based on clustering predicted the static Young's modulus perfectly when compared to the measured one for three different cases with different lithology sets. Other correlations did not match the measured static Young's modulus and bad match was obtained in several cases . Log parameters such as shear transit time, compressional transit time, and bulk density showed high relative importance to be included in the correlations that predict the static Young's modulus. Data clustering is a good method to apply to obtain better match between the estimated and measured static Young's modulus.