PREDICTION OF RESERVOIR PROPERTIES FOR BLIND WELL USING NEURAL NETWORK AND SEISMIC KNOWLEDGE

摘要

Well Drilling costs a lot without knowing porosity distribution. Geoscientists use the seismic waves to overcome this problem and reduce the exploration risk. The current paper proposes a system to predict porosity of well from other wells already drilled incorporating with seismic data. This proposed workflow aims to estimate porosity values from three-dimensional seismic data and wells records from F3-block North Sea data. We used porosity interpretations from two wells (F2-1 and F3-2) and three-dimensional seismic attributes for neural network training. for assessing the result of porosity prediction, we used data from another well (F3-4) as a blind well. Correlation in the three stages of training, validation, and testing are discussed. Test results indicate the superiority of the proposed Neural Network to predict porosity compared to other techniques in current use. By implementing Neural Network to predict porosity in blind well it is found that correlation R=0.98.