Stratigraphic oil trap detection using seismic attributes and neural network

Abstract

Seismic data interpretation mainly is classified to structural, stratigraphical and lithological interpretation. Reservoir characterization based on quantitative and qualitative seismic data evaluation is very useful and inevitable for reservoir studies, therefore integrating seismic data inversion, seismic attributes and neural network classification will be helpful to detect oil traps. Stratigraphical interpretation evaluates litho facies lateral changes of reservoirs based on high quality of 3D seismic data cube. Evaluation of various time slices related to desired targets using relevant seismic attributes will illuminate strata anomalies. First appropriate seismic attributes were used to detect probable stratigraphic oil trap which are illustrated by lineament. Instantaneous frequency, instantaneous phase and envelope attributes potentially be able to indicate structural and stratigraphical anomalies. Variance, Chaos, Semblance coherence attributes are completely susceptible to lateral facies changes and RMS amplitude validates channel shape in reservoir. In the following, stratigraphical facies were classified by neural network based on unsupervised data. To facies classification and analysis by neural network, the maximum correlation is relevant to using instantaneous frequency, variance and envelope, which illuminated channel trend. Finally to confirm stratigraphic oil trap in this reservoir, seismic inversion and porosity estimation were applied.