Exploring anisotropy in rock fluid flow and elastic behavior

Rasromani, Ebrahim Khalil
Journal Title
Journal ISSN
Volume Title

The difficulty in modeling hydrocarbon reservoirs stems from their inherent anisotropic and heterogeneous nature. Due to their simplicity, isotropic fluid flow and geomechanical models are often used in industry. In comparison to the isotropic models, anisotropic models require more data to be collected for initialization and higher computational power for processing. The increasing complexity of reservoirs today and demand for higher accuracy production forecasting and better optimization of production and drilling development programs makes the case for the use of these more complex models. This thesis first presents a scheme for efficiently calculating the anisotropic permeability tensor to better represent the initial reservoir flow behavior. Various methods for modeling the change in permeability throughout the reservoir development are then explored. As hydrocarbons are produced and reservoir pressure depletes, the change in stress state induces a change in the reservoir permeability. The error associated with the assumption that permeability change is the same in all directions, often used in industry, is assessed by comparing an isotropic permeability change model to the strain-induced anisotropic permeability change model developed by Wong (2003) through a case study of a production well in Tor formation of Valhall field. The final part of the thesis assesses the difference between the resulting fracture pressure predicted by the Kirsch equations and the Amadei solution to stress around a wellbore to demonstrate the importance of the incorporation of anisotropic elastic rock properties in geomechanical modeling of shale formations (Kirsch, 1898; Amadei, 1983). This is done through a case study of a horizontal well in the Lower Barnett Shale in Fort Worth Basin. Our results suggest that anisotropic reservoir behavior can be incorporated into reservoir models without a significant increase in computational power required and in some cases can significantly improve the prediction of the current and future reservoir state.