GAS CHIMNEY DELINEATION AND PROSPECT IDENTIFICATION FROM SEISMIC DATA USING ARTIFICIAL NEURAL NETWORKS IN AN AREA OF NORTH SEA, NETHERLANDS

Abstract

In this thesis, the study on the gas chimney delineation and prospect identification using Artificial Neural Networks is described. The study was mainly concerned with the identification of prospects by the use of Artificial Neural Networks as well as the proposal of well locations. To interpret and analyze seismic and well data, version 6.2.0 of the OpendTect software was used. The available data of the F3 block extends from the Zechstein unit of the Paleozoic era to the Upper North Sea unit of the Cenozoic era. In the interpreted horizons, three different horizons have been chosen for the identification of hydrocarbon presence. Three differently derived outputs have been generated for the interpretation, which is attribute attribute-derived, amplitude-derived, and spectral decomposition output. Chimney cube have been generated by the use of Artificial Neural Networks whereas the amplitude derived result was calculated by RMS amplitude and similarity. Lastly, the RGBA blending was used to calculate and generate the spectral decomposition result. Finally, all the outputs were evaluated and proposed three different vertical wells to drill at the site.