Numerical Modeling of Multi-Fracture Horizontal Well for Uncertainty Analysis and History Matching: Case Studies from Oklahoma and Texas Shale Gas Wells

Abstract

AbstractAccurate modeling of multi-fracture horizontal shale gas wells is the holy grail of shale gas reservoirs. Modeling shale gas wells is difficult because of the lack of subsurface data and inability to model correct physics. Some of the biggest uncertainties in shale gas reservoirs are the length of fractures, existing natural fracture network, matrix permeability, adsorption, and size of stimulated rock volume (SRV). Numerous techniques are used to reduce these uncertainties, such as Diagnostic Fracture Injection Test (DFIT), micro-seismic monitoring and Special Core Analysis (SCAL).In this paper we present a new numerical approach to model multi-fracture shale gas wells. The problem is simplified in terms of modeling the natural fracture network intersected by the induced fractures by creating a SRV around the well with size and permeability of the SRV used as uncertainty parameters. Individual fracture stages are modeled explicitly by utilizing logarithmic local grid refinement (LGR).We apply this methodology to two field cases. First case corresponds to a multi-fracture shale gas well in Oklahoma. A detailed design of experiment study is also performed to identify main uncertainties and generate a probabilistic production forecast. Numerical simulation model, coupled with an in-house probabilistic tool, is used for this study. Second case corresponds to shale gas horizontal well in east Texas. Daily production and pressure data is history matched using a numerical model. History matching provided us with great insight and better understanding of different uncertainty parameters and how they affect early to late production of horizontal shale gas wells.