Development of a Compositional Simulator for Liquid-Rich Shale Reservoirs

Open Access
Author:
Rajput, Vaibhav Hiralal
Graduate Program:
Energy and Mineral Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
May 18, 2016
Committee Members:
  • Turgay Ertekin, Dissertation Advisor
Keywords:
  • Compositional Simulation
  • Liquid-rich shale reservoirs
  • Liquid-phase adsorption
  • Multi-component diffusion
  • Type curves
  • Multi-mechanistic flow
Abstract:
Hydrocarbon production from shales has gained significant momentum in recent years with the advancement in hydraulic fracturing and horizontal drilling technologies, and production from shales (and unconventional sources in general) is beginning to garner greater share in US energy portfolio. However, storage and production mechanisms in these ultra-tight reservoirs is not well understood. It is widely believed that adsorption accounts for a significant portion of stored gas in shale gas reservoirs. However, whether this mechanism is important in liquid-rich systems is not well established. In addition to this, due to the matrix permeabilities existing in nano-darcy ranges, it is hard to establish physics of flow on Darcy’s law alone. In this work, we have developed a new thermodynamically consistent adsorption model that is made applicable to liquid-rich shale systems. Standalone calculations reveal that neglecting this storage mechanism could result in under-estimation of reserves by about 5-15%. The model is based on the ideal adsorbate solution theory (IAST), which has been successfully applied to coalbed methane and dry-gas shale systems earlier. Additionally, a new approach for multi-mechanistic flow formulation is applied in this study. Previously, multi-mechanistic studies include modeling diffusional flow based on the difference in concentration or molar density. However, this approach becomes handicapped when we have a single phase condition (gas/oil) in the matrix and the other single phase condition (oil/gas) in fractures, since it is not possible to consistently define concentration gradient across discontinuous phases. Such a condition is frequently expected to take place in shale systems, where pressure in fractures would be significantly different from that in the matrix, and therefore fractures may have two hydrocarbon phases, while matrix will still be in single phase condition. In our work, we have defined diffusive flux based on gradient in chemical potential, with the resulting equation being mathematically equivalent to the one defined based on concentration gradient. This approach is consistent across all the thermodynamic conditions (single and/or two phasic conditions). Finally, flow modeling in near-wellbore region is of utmost importance, especially in shale systems where early production phase is characterized by depletion through the hydraulically fractured region. It is established in literature that flow in near-wellbore region of horizontal well is of ellipsoidal nature. This is more emphasized when we consider that micro-seismic studies state that the fracturing process forms an ellipsoidal region. Thus, in order to model the flow pattern correctly, we have modeled the reservoir in ellipsoidal coordinates. A comparison of our model’s performance is made with analytical models presented for horizontal wells in homogenous regions. In addition, we also generated pressure-transient and pressure-derivative type curves using the ellipsoidal model. These type curves were validated using type-curve matching process, with satisfactory results. At the end, an in-depth sensitivity analysis was performed on certain important parameters and presented. Also, a case study is shown, using reservoir parameters from Utica and Marcellus shales. Sensitivity analysis is performed on drainage area and SRV volume, with some recommendations provided on economically-feasible drainage area per well. In summary, we have developed a three-phase, 3D, dual-porosity, dual-permeability compositional reservoir simulator in this study. The features presented above are incorporated in this model. Case studies illustrating the effect of important parameters in each of the above phenomenon are carried out and results are reported.