Unconventional Resources
Understanding Basic Mechanisms in Natural Gas Production Using Reservoir-Scale Modeling Last Reviewed June 2017

FWP FE 406/408/409

Production of natural gas and other hydrocarbons from unconventional sources involves hydraulic fracturing and horizontal drilling, to establish connectivity and increase the permeability. The current recovery rates are only 10-30 percent with the production rapidly declining in the first couple of years. This inefficient extraction leads to drilling of multiple wells, which in turn increases the environmental footprint. The hypothesis is that there are several transport mechanisms that are responsible for hydrocarbon recovery, and a fundamental physics-based understanding of these mechanisms and the production curves will provide information on improving recovery efficiencies. The primary objective of this project is to develop a combined process-level and systems-level reservoir-scale modeling approach that will isolate the key process parameters and their influence on the production curve. This reservoir-scale modeling approach will be built on high-performance computing tools that have been developed at DOE national laboratories.

Los Alamos National Laboratory (LANL), Los Alamos, NM 87545

Unconventional hydrocarbon reservoirs (e.g., tight shale) have naturally existing fractures with very low matrix permeability (nanodarcy). Discrete fracture network (DFN) approach, where fractures are modeled as two dimensional planes in three dimensional planes, is known to be an effective approach in characterizing such reservoirs, provided the fracture stochastics are known. The process-level part of the approach is based on dfnWorks, which is a workflow built on the DFN approach. This workflow involves a DFN generator dfnGen, meshing toolkit LaGriT, flow simulator PFLOTRAN, and a particle-tracking toolkit dfnTrans. The challenge with this approach is characterizing fractures at smaller scale (damage) and other smaller scale processes such as matrix diffusion and desorption. Using the systems-level decision support toolkit (MADS) in the framework, the aim is to characterize these smaller scale phenomena using site data (geology, fractures) and production data from different sites (for instance, using data from collaborator Apache Corp and from Texas Railroad Commission). This approach can lead to multiple solutions for the process parameters when calibrating with the site production data. The results from the other two LANL projects (PIs: Xu and Carey) will enable further constraining some of these parameters (e.g., diffusion coefficient from Xu’s project). Additionally, using the combined process-level and systems-level framework, the researchers will be able to perform sensitivity analysis and identify how the production curves depend on the transport process parameters.

The tasks in place will help to identify and isolate the key process parameters, and analyze how the production curve depends on these parameters.

The first phase focused on capability development – implementing the small scale processes into the flow simulator PFLOTRAN, implementing multi-phase flow for flow-blocking analysis and implementing the systems-level decision support toolkit, MADS, into the overall framework with dfnWorks. The second phase is focused on using the combined process-level and systems-level framework to perform calibration on the various site datasets (from literature, Apache Corp, and Texas Railroad Commission), perform sensitivity analysis on the process parameters, identify the key parameters, and then summarize recommendations for improving production efficiency.

The accomplishments to date are as follows:

  • Project researchers have successfully implemented the various smaller-scale process mechanistic models in the flow simulator PFLOTRAN that is part of dfnWorks.
  • The project team has successfully added the systems-level decision support toolkit, MADS, to the overall workflow. This gives a powerful tool for performing production curve analysis in Phase II.
  • The team successfully implemented the methane-water equation of state in PFLOTRAN for performing multi-phase flow blocking analysis.
  • A new algorithm for performing particle-tracking using time-varying (transient) velocity fields on DFNs has been developed and implemented.
  • Researchers developed an analysis and visualization approach using the concept of a flow topology graph for characterization of flow in constrained networks.
  • The team performed simulation studies to evaluate the effect of aperture variability in fracture networks on hydrocarbon transport in these networks.
  • The team studied how the extent of damage influences flow using a continuum-based flow model.
  • For a simple fracture network, the team’s DFN simulations compare well with analytical solutions, giving confidence in the DFN predictions of flow.
  • Continuum approaches do not account for anisotropic permeability, whereas DFN models do.
    • Because fracture networks are the main flow pathways; flow permeability in a fractured reservoir is anisotropic.
    • Anisotropic permeability is difficult to represent in a continuum approach and current flow codes cannot work with anisotropic permeability.
  • Continuum approaches can inaccurately predict flow in the near term (e.g. during the peak production period), causing more diffuse production curves and longer transport times. However, the DFN models can produce more accurate predictions of the physics.

Current Status (June 2017)
Work under this project continues to examine large-scale fracture controls on hydrocarbon production in the Marcellus Shale. Fracture-network geometry/topology is underway and will extend to the impact of the fracture-network properties.

The effort to compare the Los Alamos Discrete Fracture Network with conventional approaches, and to identify key gaps in understanding the contribution of tributary zones and matrix processes has been completed with accomplishments listed above in numbers 8, 9, and 10.

Project Start: October 1, 2014
Project End: June 31, 2018

Project Cost Information:
Phase 1 – DOE Contribution: $233,000
Phase 2 – DOE Contribution: $233,000

Planned Total Funding:
DOE Contribution: $467,000

Contact Information:
NETL – Bruce Brown (Bruce.Brown@netl.doe.gov or 412-386-5534)
Los Alamos National Laboratory – Satish Karra (satkarra@lanl.gov)

Additional Information:

Mechanisms in Natural Gas Production Using Reservoir - Scale Modeling (Aug 2017)
Presented by Satish Karra, Los Alamos National Laboratory, 2017 Carbon Storage and Oil and Natural Gas Technologies Review Meeting, Pittsburgh, PA