Preliminary studies on using Artificial Neural Networks to predict sedimentary facies of the Permo-Carboniferous glacigenic Al Khlata Formation, Oman

An automated facies interpretation technique was developed for the very heterogeneous Permo-Carboniferous glacigenic Al Khlata Formation of south Oman. The technique uses two neural network models based on core- and partially outcrop-calibrated sedimentary facies. The first model predicts a set of six wireline facies based on quantitative analysis of conventional wireline log data. The second can predict all predefined sedimentary facies and facies associations using qualitative information of the internal formation organization and the quantitative data of the first model. The highly robust output can directly be used as input for 3D modeling of reservoir architecture and ultimately in 4D dynamic simulations for optimized field development.