LI, HAIYAN, HUI-CHUAN CHEN, and ERNEST A. MANCINI, The University of Alabama, Tuscaloosa, AL
Abstract: Classification of Porosity and Permeability Category with Regression Trees
Permeability and porosity are two of the most important petrophysical parameters in reservoir characterization. In a carbonate reservoir, exact porosity/permeability values are often too difficult to be obtained from the low cost well-log data. Nevertheless, an accurate knowledge of porosity/permeability category (high, moderate or low) is useful for field development and production strategies. This paper presents a classification method capable of determining the porosity/permeability category by ft from well-log data.
This paper utilizes regression trees which incorporate regression method, tree structure, and expert knowledge to determine data category. A regression tree can be constructed based on a training well. This well contains the independent variables obtained from well logs and the dependent variables (in terms of desired categories) obtained from core data. The constructed tree is then used to explore the predictive structure of the given data and the interactions between various log signatures (seven well-log measurements are used). Finally a classification method is used to derive the corresponding category of porosity/permeability.
This paper used the data from seven wells in North Choctaw Ridge Smackover Formation, located in Choctaw County, AL. The proposed method results in a 86.7% accuracy for porosity and 82.8% accuracy for permeability. Using the same data, a back propagation neural network approach results in a 73.2% accuracy and 62.2% accuracy for porosity and permeability, respectively. The proposed classification method seems to be more powerful than the popular back propagation neural network in the classification of porosity and permeability category.
AAPG Search and Discovery Article #90924©1999 GCAGS Annual Meeting Lafayette, Louisiana