Restricted (Penn State Only)
Li, Liwei
Graduate Program:
Energy and Mineral Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
December 05, 2017
Committee Members:
  • Russell T. Johns, Dissertation Advisor
  • Russell T. Johns, Committee Chair
  • Hamid Emami-Meybodi, Committee Member
  • Li Li, Committee Member
  • Xiaofeng Liu, Outside Member
  • Enhanced oil recovery
  • Carbon storage
  • Gravity-enhanced process
  • Multiphase flow and hysteresis
  • Well injectivity
Gas injection processes are the most widely used EOR processes. CO2 is the most popular solvent used to arrest declining oil production and also to aid in CO2 storage. Well planned and optimized CO2 flooding can be highly profitable and thus crucial to model accurately the CO2 injection mechanism in heterogeneous reservoir systems, where considerable amount of oil are not swept after water flooding. Estimates of oil recovery and EOR efficiencies are critical for optimal design of the process. The process recovery is usually divided into two terms, i.e., local displacement efficiency (ratio of produced oil over the amount of contacted oil) and sweep efficiency (volume of oil contacted over the amount of oil in place). Sweep and local displacement efficiencies are complex functions of fluid and petrophysical properties of the reservoir, geometry of the reservoir, well pattern and injection rates. This research examines the effect of different scaling parameters on sweep efficiency. Furthermore, a new model is proposed that includes hysteresis effects on multiphase flow in porous media. Finally, we develop a new framework to obtain the solutions during cyclic injections of fluids. Achieving sufficient sweep efficiency during gas flooding can be a challenge. A new recovery scheme based on first-contact-miscible flow is proposed in this dissertation, which achieves nearly greater than 80% for heterogeneous reservoirs using a gravity-assisted process. As a side benefit, the gravity-assisted processes also have the potential to store more carbon dioxide. A total of eleven scaling groups are determined based on inspectional analysis and found to model accurately storage and oil recovery. Latin hypercube sampling is used to generate the response surfaces based on the eleven scaling groups. The response surfaces can also be used as screening models for large databases of reservoirs to determine the most attractive CO2 EOR and storage candidates. Correlation-based relative permeability and capillary pressure are commonly used in commercial simulators to determine the recoveries. A new equation of state based approach is proposed. That is, relative permeability and capillary pressure can be calculated using both phase saturation and topology as state variables. Normalized Euler characteristics is defined to quantify the connectivity of fluid phases regardless of phase saturation and sample size. Capillary pressure is modelled as a state function of phase saturation, normalized Euler characteristics, and wettability index (defined by contact angle). Coupled tuning of both relative permeability and capillary pressure requires only a few coefficients. Furthermore, the new capillary pressure model is demonstrated to achieve satisfactory predictability using experimental data away from measurements. A general semi-analytical framework is developed to investigate the multiphase flow of CO2 and water within porous media with variable initial and injection conditions. Effects of relative permeability hysteresis and phase behavior are modelled by constructing an analytical Global Riemann solver. Furthermore, a robust front-tracking algorithm is implemented to solve wave interactions during CO2-alternating-water injections. Solutions from the semi-analytical approach are used to validate finite-difference numerical simulations. CO2 injection into a reservoir could possibly form discontinuous ganglia. The connectivity and amount of these ganglia determines the gas relative permeability. Both the conventional relative permeability and a new equation-of-state (EoS) relative permeability model are used to evaluate the residual trapping during semi-immiscible displacement. Dynamics of CO2 trapping is investigated semi-analytically and the results provide compelling evidence and motivation for using cyclic injection schemes to enhance trapping and for the long-term storage of CO2 in the subsurface. The simplicity of the semi-analytical solution yields an efficient and quick method to investigate the impact of uncertainty in the parameters on the CO2 WAG injectivities. CO2 residual trapping and solubility are shown to significantly affect the well injectivity. The semi-analytical approach may also provide a tool to downscale the saturations (or total mobilities) near the wellbore.