Compressibility predictions using digital thin-section images of rocks

We use numerical simulations on petrographically characterized thin-section images to predict threedimensional elastic compressibility of sandstones. We predict two key statistics of compressibility curves measured under uniaxial boundary conditions in a laboratory - i) minimum compressibility at 1500 psi or 10 MPa depletion stress, and ii) maximum compressibility. A new Digital Rock workflow was developed for predicting compressibility based on the simulation of stress field using a segmented two-dimensional thin-section image. We also propose linear and non-linear relationships of log base 10 (compressibility) with in-situ porosity that can be used for compressibility prediction in the absence of laboratory measurements or two-dimensional images. Based on the results of application of the proposed relationships on samples from different fields with laboratory measurements, we conclude that the best prediction for minimum compressibility is obtained using the Digital Rock workflow and the best prediction for maximum compressibility is obtained using the proposed non-linear relationship using in-situ porosity. The range of compressibility values given by the difference between maximum and minimum compressibility predicted using the proposed methods can be used in making better informed economic decisions in field development planning.