GeoLEAP, a geostatistical software for building heterogeneous petroleum reservoir models
Geostatistics, variography analysis, kriging, conditional simulation.
Geostatistics is an important tool in reservoir modeling. It promotes interdisciplinary, improving the relationship among geologists, petroleum engineers and statisticians, and creating a better interpretation of geologic reality. Its main objective is to improve the predictions by building of models that represent reservoir heterogeneity. The two main components of Geostatistics are the estimation and simulation. Estimation is used to infer properties at unsampled locations using (known) sample data. The most common method is kriging, a linear regression unbiased technique that provides the best estimation at each non-sampled location. Kriging tends to smooth details of spatial variations of properties, overestimating or underestimating at some values. Conditional simulation techniques were initially developed to correct the smoothing effect, the main conditional simulation techniques are the Sequential Gaussian Simulation and Sequential Direct Simulation. Nowadays, geostatistical software encompasses a wide range of products in terms of price, operating systems, user-friendliness, functionalities, graphical and visualization capabilities. In the present study, it was built to geostatistical software with an academic purpose and open-code that allows users to modify and improve the algorithms. Named GeoLEAP, the software allows 2D and 3D analysis with a user-friendly interface. The GeoLEAP is capable of variography and model-building using estimation and simulation. Two set of data were studied to validate the software, one two-dimensional and the other three-dimensional. In both cases, it was done the variography analysis. The results showed that the techniques implemented in GeoLEAP are capable of building anisotropic reservoir models that present global and local precision.