Application of parallel neural networks in reservoir characterisation from well logs

A new class of neural networks for quantitative analysis of reservoir properties from well logs is demonstrated in several practical applications. The parallel neural network consists of a number of identical networks (experts) trained on identical or overlapping patterns. We demonstrate that the new artificial neoral network approach is a pragmatic and accurate alternative for converting well data to common reservoir parameters such as porosity, permeability, fluid saturation and for identification of lithofacies. Application to measurement white drilling is feasible.