PREDICTION OF RESERVOIR PERMEABILITY FROM WIRE LOGS DATA USING ARTIFICIAL NEURAL NETWORKS

This paper presents a methodology to predict reservoir permeability from well logs data by using an artificial intelligence technique namely artificial neural network. A multilayered perceptron trained by backpropagation algorithm was used to build the predictive model. The performance of the net results was measured by correlation coefficient. The implemented artificial neural network model is demonstrated by applying it to Mishrif limestone reservoir at Nasyria oil field, south of Iraq. The results show that artificial neural network was capable of reproducing permeability (horizontal and vertical) with very high accuracy, so that the calculated correlation coefficients for horizontal and vertical permeability were 0.85 and 0.90, respectively. The results could be generalized to other field after examining new data, and a regional study might be possible to study reservoir properties in south of Iraq with cheap and very fast constructed models.