Modelling a fluvial reservoir with multipoint statistics and principal components

Traditional reservoir modelling techniques use oversimplified two-point sta tistics to represent geological phenomena which are typically curvilinear a nd have other complex geometrical configurations. Use of multipoint statist ics has shown some improvement in recent years to reduce such limitations. This paper compares the performance of the use of conventional and multipoi nt data for estimating porosity from seismic attributes in a fluvial reserv oir using neural networks. According to the results of the study, the neura l network trained on multipoint data gave smaller error and higher correlat ion coefficient of porosity in a blind test. Further improvement is also ob tained by reducing the dimensionality of the input space using principal co mponents. This study shows a successful integration of neural networks and principal components for modelling multipoint data in practical reservoir s tudies. (C) 2001 Elsevier Science B.V. All rights reserved.