Scalable High-Fidelity Hydraulic Fracturing and Geo-Mechanics Simulator for Unconventionals

Abstract

Abstract Development optimization is the key to ensuring economic extraction of hydrocarbons in unconventionals. Every dollar saved per fracture grows by 3-4 orders of magnitude per pad. Typical questions of business interest include optimal number of perforations per fracture, fluid and proppant rates, cluster spacing, well spacing, influence of pre-existing fractures and depletion on infill wells/fractures etc. Evidently, these questions span broad length scales, wide variety of governing equations, and nonlinearities. This is compounded by the challenge of tight balance between maintaining the accuracy of numerical simulations and completing simulations fast enough for sensitivity analyses. In this work, we describe how we addressed all these challenges in building our proprietary hydraulic fracturing and geomechanics simulator. We solve the single phase porous flow equations along with Terzaghi’s effective stress based force balance equations representing the reservoir. The model includes fluid and proppant flow from well head to the specific stage through 1D compressible pipe flow. The well flow is coupled with reservoir through perforation orifice flow. Flow in fractures is considered to be the typical plane Poiseuille flow with Darcy leak-off into reservoir. We adopted scalable data structures and parallel algorithms that enable solving ~50 million degrees of freedom with relative ease. The pore-pressure is solved on a Finite Volume grid, and the deformation on Finite Element grid. All these governing equations are solved simultaneously in a fully coupled approach using an iterative solver. The fracture is assumed to be planar, defined by element interfaces and prepared a priori with deactivated duplicate nodes to avoid modifying distributed data structures during runtime. Multiple different models are implemented for well and perforation and can be chosen based on suitability. The perforations belonging to various clusters, stages and wells are activated along the timeline of completion clock, mimicking a frac-job. The article will focus primarily on the techniques that allowed us to build such a scalable fracturing simulator, followed by simulation results. Using this simulator we are able to perform stress update calculations due to depletion, multi-well and multi-stage hydraulic fracturing scenarios. The fracture interference across stages and wells is directly captured through the governing equations. We will also present strong scaling test results to show the scalability of the simulator to millions of degrees of freedom.