The project goal is to develop a “new generation” hydraulic fracturing model that will, for the first time, provide an operator with the ability to model the simultaneous propagation of non-planar hydraulic fractures from multiple perforation clusters and create a realistic picture of the stimulated rock volume (SRV) around horizontal wells. The model will be used to simulate the performance of different fracturing fluids and fracture designs to maximize the effectiveness of the SRV to increase well productivity, improve ultimate recovery estimates, and reduce overall horizontal well costs.

University of Texas at Austin, Austin, TX, 78712
University of Texas–San Antonio, San Antonio, TX, 78249

According to a 2011 U.S. Energy Information Administration report, almost all incremental oil and natural gas production in the lower 48 states will come from unconventional resources—shales, low permeability sands, and heavy oil. Production of virtually all the oil and gas from unconventional reservoirs will rely on the application of multiple fractures in horizontal wells. The development of oil- and gas-bearing shale plays largely depends on the cost of drilling and fracturing horizontal wells. Rapid decline rates require that new wells be drilled to maintain existing production rates. A reduction in the cost and environmental footprint of drilling and fracturing will lead to a significant expansion of oil- and gas-bearing shale development. This project aims to develop better methods for substantially reducing these costs while maximizing oil and gas production from the shale reservoirs.

Virtually all current approaches to hydraulic fracture modeling rely on finite difference, finite element, or boundary element methods to solve a fracture formulation. These methods usually use linear-elastic fracture mechanics to determine crack lengths based on the internal pressure driving the fracture open. The discontinuous nature of the cracks causes problems with methods that rely on computing derivatives across domains containing discontinuities and severely limits the applicability of these methods to only the simplest geometries (usually single, planar fractures). 

The project’s primary objective is to develop a “new generation” hydraulic fracturing model, based on a peridynamics formulation, that models multiple, non-planar, competing fractures in heterogeneous shales for better well design, improved hydraulic fracturing, enhanced production, reduced drilling and completion costs, and improved shale oil and shale gas production economics. Peridynamics is a recently developed continuum mechanics theory that allows for autonomous fracture propagation. It has been demonstrated and applied to other geo-mechanical and structural problems where material failure was pervasive (e.g., cement structures). Peridynamics allows three-dimensional modeling of arbitrarily complex fracture geometries and the growth of competing and interacting fractures in naturally fractured media. No current models capable of modeling the propagation and performance of simultaneous multiple fractures in horizontal wells exist.

Reducing the high cost associated with drilling and completing long laterals with a large number of hydraulic fractures requires a better understanding of the geometry of these fractures and SRV around the wellbore. Significant questions remain about optimum spacing among horizontal laterals, fracture stages, and perforation clusters. The inability to conduct such an analysis stems from the difficulty in determining, with any degree of certainty, the fracture geometry created through multiple clusters and multiple stages along each lateral and between laterals.

The impact of this work is expected to be widespread and applicable to all oil- and gas-bearing shale resources in North America. Both the proposed modeling and the fracturing recommendations are expected to have an immediate and long-term impact and benefit.

The ability to realistically model hydraulic fracture propagation will provide a starting point for a better understanding of how fracture design affects the stimulated rock volume and well performance. It is anticipated that the new hydraulic fracturing model will lead to recommendations and guidelines regarding cluster spacing, stage spacing, stage sequencing, and fracture design in long horizontal wells for a given set of reservoir conditions. These recommendations should result in significant performance improvements and cost savings, thereby allowing more wells to be drilled and completed for the same annual budget. Increased reservoir drainage due to improved fracturing will result in more economic and longer producing wells, potentially resulting in a 5 to 10 percent increase in the recovery of oil and gas from these unconventional plays and a reduction in well costs of up to 25 percent. The models will be particularly useful for oil-bearing shales that are more likely to have natural fractures and more complex fracture patterns.

• The project team has completed the testing of the parallelized peridynamics code. New cases were run to test the code and to solve problems of fracture propagation that have never been modeled before. Heterogeneities at different length scales were used to study fracture propagation in heterogeneous media. The heterogeneities at different length scales can be captured in this manner.
• The parallel code was tested for many different cases. Changes to the code were made to accommodate special features that are needed to predict the behavior of layered media.
• Cases were run to show the impact of the mechanical and flow properties of layered media on the vertical propagation of fractures.
• Cases were also run to show the interaction of hydraulic fractures with natural fractures in both 2-D and in 3-D.
• One of the PhD students completed a daft of his PhD dissertation and will be defending it next month.
• To speed up the run times (particularly for 3-D problems) the peridynamics code was parallelized. This required a significant rewriting and reorganizing of the modules.
• The parallelized code was tested and was found to be quite scalable on multiple cores on a multi cluster machine. In the future this code will be modified and tested on a multi-core Windows machine.
• A paper titled, “A Fully Coupled Porous Flow and Geomechanics Model for Fluid Driven Cracks: A Peridynamics Approach”, was published in the journal Computational Mechanics. This is the first time that a peridynamics model has been built for hydraulically driven cracks.
• The newly developed peridynamics code was tested by running it for different heterogeneous porous media problems in both 2-D and 3-D. It was shown that this resulted in complex fracture geometries with turning and branching of fractures (effects that cannot be captured in other models).
• The peridynamics code was tested and run for some realistic 3-D problems. The results were satisfactory but the computation times were very large.
• Results of the simulations were presented to the 34 member companies of the Fracturing and Sand Control JIP. There were 103 industry representatives in attendance and another 107 remote viewers connected to the live web feed (with at least one person at leach location). This provided wide visibility for the project and its main results.
• A paper on the peridynamics fracturing model developed during this award has been submitted and accepted for the 2015 SPE Hydraulic Fracturing Technology Conference. This is the first time the peridynamics fracturing model has been applied to practical hydraulic fracturing problems and will be presented to an oil and gas audience.
• Dr. Sharma gave a keynote address at the SPE Hydraulic Fracturing Conference. He highlighted some recent results from the projects.
• UT Austin will be presenting a paper at the Advanced Computational Workshop in San Diego in July, 2015.
• The poroelastic peridynamics model developed during phase I of this award has been successfully expanded to include the possibility of arbitrary crack propagation. This now allows the model to propagate multiple fractures and include natural fractures in the domain.
• The poroelastic code was made parallel to increase computation speed, which will enable it to be run on massively parallel, multi-core machines. It is expected that this enhancement will reduce run times by an order of magnitude.
• Developed an efficient algorithm and C++ code for the parallel coupling of peridynamics formulation of porous flow and fluid pressure-driven poroelastic response of the reservoir.
• Project team worked to include the nonlocal poroelasticity model into Peridigm, the open-source DOE computational peridynamics code.
• Dr. John Foster organized a workshop in Tampa, FL, entitled “Workshop on Meshfree Methods for Large-Scale Computational Science and Engineering”. Nonlocal poroelastic formulation was presented.
• Dr. Foster presented papers on nonlocal poroelastic formulation at IUTAM/USACM Workshop on multi-scale modeling (Northwestern University) and USNCTAM (Michigan State University) during the summer of 2014.
• Developed the coupling strategy for the recently developed peridynamics porous-flow formulation and existing peridynamics solid mechanics formation. The coupled peridynamics model has been validated for a purely elastic deformation with a two-dimensional consolidation problem and for two-dimensional bi-wind planar fracture propagation. Results were presented at the Hydraulic Fracturing and Sand Control JIP meeting on April 22, 2014.
• The project team developed a novel, generalized, non-local, state-based peridynamic formulation for anisotropic transient fluid flow in an arbitrarily heterogeneous and fractured porous medium.
• Dr. John Foster organized a workshop in San Antonio, TX entitled “Workshop on Non-local Damage and Failure: Peridynamics and other Non-local Models".
• Dr. Kariyar presented a paper entitled “A Peridynamic Formulation of Coupled Mechanics Fluid Flow Problem".

The project ended on September 30, 2016. The final report is available below under "Additional Information".  

$1,038,087

$554,367

NETL – William Fincham (william.fincham@netl.doe.gov or 304-285-4268)
University of Texas at Austin – Mukul M. Sharma (msharma@mail.utexas.edu or 512-471-3257)

Final Scientific Report [PDF] (December 2016)