Integration of conventional well logs and core samples to predict porosity of tight reservoir: a case study from Ordos Basin

Abstract

For reservoirs of oil wells with no cored data, predicting porosity from wireline logs and core samples is an effective approach. Integration of conventional well logs and core samples to predict porosity with high accuracy is particularly challenging owing to complex logging responses of tight sandstone. Therefore, a novel prediction workflow based on a linear interpolation algorithm (LIA) is described to estimate porosity from well logs in the present study. Based on core reposition, porosity correction under overburden pressure, core–log data matching and calculation of shale content, two multi-regression formulae to estimate porosity values are obtained using the nearest neighbour algorithm and LIA respectively. The formulae are applied to the tight sandstone in the Chang 9 member of the Yanchang Formation in the NE Wuqi Oilfield, Ordos Basin. The comparison results indicate that the porosity predicted from the formula obtained by the LIA is in better agreement with the measured porosity, showing a better prediction effect. The application example demonstrates that the LIA formula is applicable for core porosity prediction in the study region. This approach can also be applied for porosity prediction in other oil regions that have similar geological background.