Tao Zhang¹, Roy E. Plotnick¹
(1) University of Illinois at Chicago, Chicago, IL

Abstract: Predicting biofacies distributions using a sequence stratigraphic model and artificial intelligence methods

Artificial intelligence techniques, particularly fuzzy logic and neural networks, have been integrated with the SEDPAK sequence stratigraphic model. This integrated approach predicts the distribution of biofacies within the context of sequence stratigraphy. This makes it possible to test and calibrate existing models for sequence stratigraphy in particular basins using biofacies data.

This approach is illustrated using the intensively studied Neogene stratigraphy of the Baltimore Canyon Trough, offshore New Jersey. A Sedpak simulation of this basin was used to produce a sequence framework and to generate the physical environmental inputs for the biofacies model. Fuzzy logic was used to define biofacies categories based on a set of biofacies rules. Prediction of these biofacies is then made by a neural network that has the ability to learn and generalize the biofacies rules. The prediction results were compared with data from industrial wells COST B2 and B3. Prediction and observation show good agreement. Artificial intelligence methods have great potential as means to test and calibrate sequence stratigraphy models using paleontological data.

AAPG Search and Discovery Article #90914©2000 AAPG Annual Convention, New Orleans, Louisiana