Integration of Horizontal Well Log Information in Fractured Reservoir Characterization

Abstract

Abstract Representation of reservoir heterogeneity for simulation models is one of the most challenging aspects of reservoir engineering. Naturally fractured reservoirs present an exceptionally difficult challenge in this respect. Vertical wells provide extremely limited information about the spatial distribution of fractures and fracture density as the hole size is small with respect to the characteristic spacing and distribution of natural fractures. Vertical wells fail to provide short scale (up to interwell distance) correlation information for geostatistical models in any type of reservoir. Outcrop representations of fracture distribution are generally inadequate and fail to correspond to in situ fracture characteristics. Formation evaluation in horizontal wells can provide both detailed information on fracture distribution and short-scale correlation information. This information can be integrated with geostatistical tools such as Sequential Indicator Simulation (SIS) models to provide much more realistic reservoir descriptions. These reservoir descriptions result in dramatically improved simulation matches. Further, this reservoir characterization technique can explain behavior that is far more difficult to explain with conventional and dual-porosity models. This paper presents the following: A description of how horizontal well formation evaluation is used to obtain the spatial distribution in a naturally fractured reservoir,examples of variogram estimation with the derived data for both traditional and indicator variograms,examples from horizontal well performance used to develop additional correlation information,SIS models derived from the conditioning data from the horizontal wells and the indicator variograms, andreservoir simulation results for horizontal well performance. This paper does not explain variograms, geostatistical models, correlation ranges, conditional simulation, etc.. However, the geostatistical software used was developed by the Stanford Center for Reservoir Forecasting1 and is available in Reference 1.