Unconventional resource plays are not typically homogeneous; therefore the reservoirs might not be contiguous. Geostatistical based earth modeling can create hundreds of reservoir property realizations. The challenge is to select a few optimal models from these realizations for further analysis. Traditionally, candidates are selected from ranked distributions representing quantiles (e.g., P10, P50, and P90) of the reservoir volumetrics, which assumes contiguous reservoir volumes. However, such methods lack any physical geometrical information.

In this paper, a two-way clustering method is used to take into account spatial property information. First, a three-dimensional (3-D) realization is transformed into a one-dimensional (1-D) array, so that each element of the array maps to one single grid cell that carries the spatial location information. Next, a matrix is created from the multiple realizations. Each column represents a single realization, and each row represents a single 3-D grid cell. Then, clustering algorithm is applied to both columns (R-mode) and rows (Q-mode). R-mode shows the similarity of different realizations, whereas Q-mode shows the clusters of different cells. Finally, connected geobodies are extracted from the clusters.

This novel method was applied to a west Texas Permian basin reservoir. Domain experts can first select realization candidates from different R-mode clusters and compare the results. Then, the location and connectivity of cells from Q-mode clustering can be mapped in addition to evaluating their statistical ranking. During Q-mode cluster analysis, Euclidean distance carries the spatial property location of the cell in a distance calculation. Statistical analysis of the geobodies suggests the optimal realization, which is not necessarily the P50. Through such quantitative evaluations of all the realizations, geologists and reservoir engineers can select optimistic to pessimistic realizations to aid economic assessment of the reservoir.

You can access this article if you purchase or spend a download.