Using artificial intelligence to predict permeability from petrographic data

Petrographic data collected during thin section analysis can be invaluable for understanding the factors that control permeability distribution. Relia ble prediction of permeability is important for reservoir characterization. The petrographic elements (mineralogy, porosity types, cements and clays, and pore morphology) interact with each other uniquely to generate a specif ic permeability distribution. It is difficult to quantify accurately this i nteraction and its consequent effect on permeability, emphasizing the non-l inear nature of the process. To capture these nonlinear interactions, neura l networks were used to predict permeability from petrographic data. The ne ural net was used as a multivariate correlative tool because of its ability to learn the non-linear relationships between multiple input and output va riables. The study was conducted on the upper Queen formation called the Sh attuck Member (Permian age). The Shattuck Member is composed of very fine-g rained arkosic sandstone. The core samples were available from the Sulimar Queen and South Lucky Lake fields located in Chaves County, New Mexico. Nineteen petrographic elements were collected for each permeability value u sing a combined minipermeameter-petrographic technique. In order to reduce noise and overfitting the permeability model, these petrographic elements w ere screened, and their control (ranking) with respect to permeability was determined using fuzzy logic. Since the fuzzy logic algorithm provides unbi ased ranking, it was used to reduce the dimensionality of the input variabl es. Based on the fuzzy logic ranking, only the most influential petrographi c elements were selected as inputs for permeability prediction. The neural net was trained and tested using data from Well 1-16 in the Suli mar Queen field. Relying on the ranking obtained from the fuzzy logic analy sis, the net was trained using the most influential three, five, and ten pe trographic elements. A fast algorithm (the scaled conjugate gradient method ) was used to optimize the network weight matrix. The net was then successf ully used to predict the permeability in the nearby South Lucky Lake field, also in the Shattuck Member. This study underscored various important aspe cts of using neural networks as nonlinear estimators. The neural network le arnt the complex relationships between petrographic control and permeabilit y. By predicting permeability in a remotely-located, yet geologically simil ar field, the generalizing capability of the neural network was also demons trated. In old fields, where conventional petrographic analysis was routine , this technique may be used to supplement core permeability estimates. (C) 2000 Elsevier Science Ltd. All rights reserved.