An Improved Workflow for Evaluating the Performance of Hydraulically Fractured Wells in Unconventional Reservoirs through Production History Matching using Unstructured Grid-Based Reservoir Simulation

Abstract

Abstract Evaluating unconventional reservoirs presents several challenges because of a lack of modeling tools that can properly capture inherent heterogeneities and variable reservoir attributes. This paper discusses the development of a workflow for modeling complex fracture networks in hydraulically fractured horizontal wells and then subsequently validating them through the use of production-history matching with an unstructured grid-based simulation. In reservoirs with low stress anisotropy, stimulation generally creates fracture networks and induces fractures that can have varying orientation. During reservoir simulation, these fractures have typically been depicted as planar, orthogonal, bi-wing fractures for the simplicity of gridding, even though it is known that these planar models do not adequately describe the overall complexity of the induced fractures. These complex fracture network models can only be truly represented in a simulation model through the use of unstructured grids. In developing the workflow, a complex fracture network modeling tool that takes into account microseismic or image log data, as well as pressure and treatment data, was used to create and calibrate complex fracture networks. These were incorporated into an unstructured grid-based reservoir simulation model, which also included pressure-volume-temperature (PVT), rock and fluid, and completions data, as well as attributes from an earth model for the area. Nodal analysis software was used to generate the bottomhole pressure (BHP) from wellhead pressure (WHP) and production rate data for history matching. Results from simulation models with hydraulic fractures having similar orientation and dimensions in the structured and unstructured grids showed a very good match and gave confidence in the use of unstructured gridding to help ensure reservoir simulation. Several complex fracture designs with the unstructured grids were used during history matching. Fracture properties, such as propped fracture half-length, fracture conductivity, number and spacing of existing natural fractures, and natural fracture conductivity, were varied to determine the most representative models for the fractured reservoir based on production rates. Results showed that natural fractures were present in the stimulated reservoir volume. However, the closest matches were achieved with secondary/natural fractures having very low conductivity. The evidence suggested that the natural fractures in the reservoir were not being adequately stimulated and that a combination of finer proppants, fluid diversion, or other completion or treatment design changes might have resulted in increased production through better connection with the natural fractures. Complex fracture modeling and history-matching validation with unstructured grid-based reservoir simulators is a relatively new process, and this paper demonstrates its potential for optimizing fracture design and treatments by correlating a given treatment to the representative fractured reservoir model. Initial work performed with this workflow provided information that has enabled significant design changes, with encouraging production results. Continued work in this area of technology is now being performed to help understand reservoir, fracture, and fluid interaction to enhance drilling and completion practices based on specific reservoir conditions.