Electromagnetic resistivity pseudo-log as a new instrument for estimating reservoir properties beyond boreholes

Abstract

A new approach to the reservoir properties' prediction using so called “electromagnetic resistivity pseudo-log” built in the target location is suggested. It is based on artificial neural network continuation of resistivity well logs beyond boreholes using electromagnetic resistivity profiles determined from adjacent ground electromagnetic survey. As a result of this transformation, we obtain a resistivity profile which possesses high resolution of the borehole resistivity log in the reservoir scale. We demonstrate advantages of this approach by neural network modelling of the porosity prediction below boreholes and in the interwell space. To this end we use electrical logging data collected in two boreholes located in the Northern Tien Shan and results of magnetotelluric survey in their vicinity. It is shown that porosity prediction using the developed algorithm provides better results than its estimation using resistivity logs or electromagnetic resistivity alone. In particular, the mean relative error can be as low as ∼2% in the case of predicting at double borehole depth and around 8% in the interwell space.