Estimation of anisotropy parameters for shales based on an improved rock physics model, part 2: case study

Part 1 of this paper presented an improved shale rock physics model to enable the prediction of anisotropy parameters from both vertical and horizontal well logs. The predicted elastic constants were demonstrated using the published laboratory measurements of a Greenhorn shale in paper 1, and are more accurate than the estimations in the existing literature. In this paper, this model is applied to the well log data of an Upper Triassic shale formation to predict the VTI anisotropy parameters, which are usually difficult to measure directly in the borehole. The effective elastic constants are calculated for solid clay, aligned clay-fluid-kerogen, a rotated clay-fluid-kerogen mixture and shale step by step using different effective medium theories. The input to this workflow includes the volume fraction of minerals, kerogen and two different pore spaces. Two parameters (the lamination index and pore aspect ratio) need to be inverted simultaneously by fitting the vertical or horizontal logs. An estimation of the anisotropy parameters from the vertical well logs uses a least square inversion in terms of C-33 and C-44. The result is demonstrated by calibration with the seismic amplitude versus angle (AVA) response. Correlations are found between the anisotropy parameters (epsilon and delta) and rock properties (pore aspect ratio, lamination index, clay content and total porosity). In the horizontal well, the anisotropy parameters are predicted by minimizing the objective function in terms of C-11 and C-44. The overestimated qP-wave velocity of clay-rich shales in the horizontal well is anisotropy-corrected and thus provides a more appropriate V-p-V-s relation. The impact of strong VTI anisotropy on Poisson's ratio is also overcome by the anisotropy-correction, thus improving the brittleness characterization of shale reservoirs.