Modeling hydraulic fractures in finite difference simulators: Application to tight gas sands in Montana

Abstract

Hydraulic fracturing allows numerous, otherwise unproductive, low permeability gas formations to be produced. The interactions between the fractures and the heterogeneous reservoir rock, however, are quite complex, and often multiple fractures are created in a single well. These factors make it difficult to model production from hydraulically-fractured systems. For example, current analytical techniques assume a single rectangular shaped fracture in a single phase homogeneous reservoir. The current work demonstrates how to capture more complexity and model these systems with a finite difference simulator. It is proposed to treat the hydraulic fracture as a discrete object that is neither gridded nor included in the skin term of a traditional well model. It is modeled as an object that interacts both with the well and the gridded reservoir. From the commercial simulator standpoint, the fractures are modeled as “wells,” so no additional code or add-ons to the simulator are required. A “well” is simply connected nodes that are in pressure communication and interact with respective gridblocks. This is the same as the fracture being modeled. Nothing in the well model forces it to connect to the surface. The first part of this paper verifies that designating fractures with wells provides solutions that are equivalent to analytical models for simple cases. Once the technique was verified, it is used to model four hydraulically-fractured wells from a low permeability gas reservoir in Montana. The simulation models are matched to available production data by changing fracture lengths. Then the models are used to predict future response from the wells. The advantages of this technique are: (1) multiple fractures can be modeled in a single well and (2) heterogeneities within the fracture and reservoir can be accounted for. These advantages will lead to more realistic models and more accurate predictions; however, the most useful application of this technique may come in the fracture design stages. With this tool, various fracture geometries and scenarios can be tested quickly in the simulator, and the most economic scenarios selected and implemented. This will cause better fracture placement, and ultimately greater production.