Fracture aperture calculation of electrical imaging logging in oil-based mud based on numerical modeling and machine learning

Abstract

One of the primary tasks of electrical imaging logging in oil-based mud (OBM-EIL) is fracture evaluation. At present, there is no research on the fracture aperture calculation of OBM-EIL similar to that of electrical imaging logging in water-based mud (WBM-EIL). First, numerical modeling is employed to analyze the fracture responses of OBM-EIL under the influences of multi-parameters such as the background formation resistivity, fracture resistivity, fracture aperture, fracture dip angle etc. Then, a strategy for determining the fracture aperture is proposed, which includes establishing multi-parameter response databases, adopting a stepping strategy, and developing calculation models of fracture apertures based on back propagation neural network (BPNN). Finally, several tests are conducted to assess the effectiveness of fracture aperture calculation models. The results indicate that the fracture apertures in the range of 0.01 to 100 mm can be accurately calculated and the determination coefficient can reach 97.1%. The fracture resistivity and the fracture dip angle are essential parameters for the fracture aperture calculation. The lack of background formation resistivity and gap distance has slight effect on the fracture aperture calculation. These findings will greatly enhance the researchers' confidence in the fracture aperture calculation using the proposed method, even when some of these parameters might be acquired through inversion or machine learning methods.