Feasibility Study of Improved Reservoir Mapping Using Joint Inversion of Extra-Deep Azimuthal Resistivity and Deep Shear Wave Imaging

Abstract

Downhole deep electromagnetic and acoustic measurements, i.e. Deep Shear Wave Imaging (DSWI) and Extra-Deep Azimuthal Resistivity Measurements (EDARM), have tens of meters detection range, enabling them to map geological structures at a reservoir scale. However, several problems occur when imaging the formation distributions using DSWI and EDARM individually, e.g. non-unique solutions due to the inherent azimuth ambiguity of DSWI and expensive time cost of EDARM data processing caused by the limited prior information.

We propose a joint inversion scheme to improve the reservoir mapping capability by using DSWI and EDARM datasets. First, the acoustic impedance image of bed boundaries around the borehole are obtained using DSWI dataset associated with pre-stack Kirchhoff migration. These resulting boundaries serve as prior information for EDARM data processing. Accordingly, a series of possible formation models are established. Finally, Using Bayesian inversion with Markov Chain Monte Carlo (MCMC) sampling, the true formation resistivity distribution is recovered rapidly. Combined with the determined resistivity bed boundaries, the DSWI 180° azimuthal uncertainty is eliminated simultaneously.

Numerical results performed over synthetic examples show that the joint inversion method of EDARM and DSWI is feasible. It not only improves the speed and accuracy of resistivity imaging, but also eliminates the inherent azimuthal uncertainty in DSWI dataset. More specifically, taking a five-layered formation model for example, inversion of EDARM without the prior boundaries information requires more than 10² seconds per logging point. Based on multiple boundary distances to borehole extracted from DSWI, 16 possible inversion models are generated, leading to the increased complexity of resistivity imaging from EDARM. In each possible model, the invert parameters are reduced from 12 to 6. Subsequently, the samples of MCMC are reduced from million to several thousands. At least an order of magnitude improvement in speed over that previously offered using individual EDARM data is observed. From the above analysis, the resistivity and acoustic impedance images of the bed boundaries show their respective limitations. More importantly, we show how each dataset contains complementary important information of these boundaries. While the images from individual DSWI and EDARM datasets may bias the reality, our joint inversion scheme leads to much improved geological boundaries.

Application of the joint-inversion scheme to the synthetic data supports the feasibility of mapping the reservoir using EDARM and DSWI. The attraction of the joint scheme is that different geophysical measurements are sensitive to different properties of the sub-surface, so through joint inversion we significantly eliminate the null or blurred bed boundaries and achieve a true geological model.

AAPG Datapages/Search and Discovery Article #90350 © 2019 AAPG Annual Convention and Exhibition, San Antonio, Texas, May 19-22, 2019