The Use of Hydraulic Flow Units and Neural Networks to Improve Rock Types Estimation and Reservoir Models: A Case Study

One of the largest challenges in reservoir modelling is to obtain representative Reservoir Rock Types for permeability and saturation modelling. The present case study provides an example from a complex siliciclastic reservoir, where an integrated Fishbone workflow was applied. Rock typing concepts can be used to establish the relationship between petrophysical data from standard well logs and core data, through the utilization of Hydraulic Flow Units and Neural Networks. Flow Zone Indicator, Rock Quality Index and Hydraulic Flow Units were calculated using mathematical equations through a dedicated software. The application of Principal Components Analysis allowed the identification of the logs to be used in Neural Networks, thus enabling the propagation of the cored-based reservoir properties into the non-cored wells. Finally, a set of logs consisting of porosity, permeability and reservoir rock types were used in the reservoir modelling.