Prediction of shear wave velocity in three sedimentary rocks in East Baghdad oilfield using multiple regression analysis

Abstract

Shear wave is a crucial parameter for assessing the wellbore stability, the stress response, and rock deformation. It is essential for constructing the mechanical earth model (MEM) for many applications related to reservoir geomechanics including wellbore stability, sand production, hydraulic fracturing, and fault reactivation. However, shear sonic data is often omitted during the well-logging measurements for cost and saving purposes. To overcome this challenge, recent research has been focused on determining shear wave velocity through the use of core plugs, empirical correlations, artificial intelligence techniques, and multiple regression to quantify and evaluate the mechanical properties of subsurface formations without performing direct measurements at the wellbore. The greatest difference between this study and the literature is to predict the shear wave velocities for three sedimentary rocks based on conventional well logs. This study has been conducted on datasets of two wells drilled in the East Baghdad oilfield, for which there is a lack of shear wave data. Two formations (Tanuma and Zubair formations) within the production section of this field were conducted to develop new models for determining the shear wave velocity using multiple regression analysis. These two formations primarily consist of three lithologies: limestone, sandstone, and shale. Before the model development, data analysis on the selected data was applied to figure the most influential parameter(s) in determining the shear wave velocity. The results of the developed models are then compared with the previous models in the literature. The results showed that the multiple regression analysis technique is a powerful technique in determining shear wave velocity with high-performance capacity. The correlation coefficient ( ) and the root mean square error (RMSE) were 0.84 and 0.092 for limestone, 0.84 and 0.0972 for sandstone, and 0.86 and 0.0796 for shale respectively. Furthermore, the performance of the developed models is well matched to the actual shear wave data rather than the Castagna correlations. The findings of this study are effective in determining shear wave velocity for future applications related to reservoir geomechanics without needing costly well-log or core measurements.

Keywords

• shear wave velocity; multiple regression analysis; compressional wave velocity; sandstone, limestone; and shale