Vol. 60, n.1, March 2019
A Meta attribute for reservoir permeability classification using well logs and 3D seismic data with probabilistic neural network
M. Hosseini, M.Alì Riahi and R. Mohebian
Received: 10 September 2018; accepted: 13 November 2018
Permeability is one of the most important reservoir properties for its crucial contribution to the productivity of the reservoir. This study presents a new approach to characterise reservoir parameters, specifically the permeability. The presented method consists of three major steps. Firstly, several seismic attributes such as acoustic impedance, geometric attributes, and instantaneous attributes are computed. In the second step, seismic attributes, which have good correlation with permeability logs, are selected. Finally, a Meta attribute, which extracts the permeability classes, is created using the probabilistic neural network. The results of this study indicate that the combination of several seismic attributes, each exploring a different feature of the seismic data, can be effectively used for determination of reservoir parameters, especially permeability. The current method is applied on the Asmai reservoir of an Iranian oil field located SW of Iran. A comparison of the results from this study to well-test data recorded in the field indicate that the generated Meta attribute can qualitatively predict the permeability values obtained from the well-testing.
Download PDF complete
back to table of contents