Carbonate Lithofacies Prediction Using Neural Network and Geostatistical 3-D Modeling of Oolite Shoals, St. Louis Limestone, Southwest Kansas
Lianshuang Qi1, Timothy R. Carr1, and Robert H. Goldstein2
1 Kansas Geological Survey, University of Kansas, Lawrence, KS
2 University of Kansas, Lawrence, KS
In the Hugoton Embayment of Kansas, reservoir units in the St. Louis Limestone consist of relatively thin (<4m) spatially scattered, highly heterogeneous oolitic grainstone deposits. Quantifying the distribution of such oolitic deposits is challenging, but essential for improving understanding of sedimentologic processes and developing efficient reservoir management strategies.
A single hidden-layer neural network was used to train and establish a non-linear relationship between lithofacies assignments from detailed core descriptions and selected log curves. Neural network models were optimized and used to predict lithofacies on the entire dataset of the 2,023 half-foot intervals from the 10 cored wells. Predicted lithofacies compared to actual lithofacies displays absolute accuracies of 70.37 to 90.82%. Established quantitative relationships between digital well logs and core description data were applied in a probabilistic sense to predict lithofacies in 90 uncored wells across the Big Bow and Sand Arroyo Creek fields.
Predicted lithofacies were integrated with well data to build 3D geostatistical models and stochastic simulations of oolitic reservoirs. The models provide insight into the distribution of facies, the external and internal geometry, and the sedimentologic processes that generated the reservoir units. The depositional pattern and connectivity of modeled oolitic complexes suggest an accumulation of oolitic deposits during pulses of relative sea level rise followed by deepening near the top of the St. Louis. Neural networks and geostatistical 3D modeling can be applicable to other complex carbonate or siliclastic reservoirs in which facies geometry and distribution are the key factors controlling heterogeneity and distribution of rock properties.