The multicomponent induction log provides both the resistivity measured transverse (vertical) to the bedding (Rv) and the resistivity measured longitudinally (horizontal) to the bedding (Rh). Typical multicomponent induction interpretations model these resistivities in terms of an anisotropic shale fraction, an isotropic sand fraction, shale anisotropic resistivity, and sand isotropic resistivity. There are more parameters than unknowns, but given the appropriate parameters from another source, it is possible to solve for the sand resistivity with at least three models:
given the shale fraction and shale horizontal resistivity, solve for the sand resistivity and the shale vertical resistivity;
given the shale horizontal and vertical resistivity, solve for the sand resistivity and the laminated shale fraction; and
given the shale fraction and the ratio of the shale vertical over horizontal resistivity, solve for the sand resistivity and the shale horizontal resistivity.
However, there are cases in which the sand can be macroscopically anisotropic; in these cases, the three classical models fail.
This paper extends the multicomponent induction interpretation model to account for the macroscopically anisotropic sands. The proposed three-component model, dependent upon an anisotropic shale fraction, an isotropic macroporous sand fraction, and an isotropic microporous sand fraction, is applied to generate some synthetic multicomponent induction resistivity data.
The research described in this paper indicates that it is possible to generate the trimodal synthetic multicomponent induction data (synthetic parallel and transverse resistivity) using the following parameters: laminated shale fraction, shale vertical, and horizontal resistivity; macroporous sand fraction and its associated porosity, water saturation, and water resistivity; and microporous sand fraction and its associated porosity, water saturation, and water resistivity. The two sand water saturations are modeled from capillary pressure curves. In the process, some uncertainty is applied to the generated synthetic vertical and horizontal resistivity data that is consistent with the existing tool uncertainty characterization database. The paper presents results for the four models, summarizing the interpretation accuracy (water saturation and net pay) as a function of parameter error and tool uncertainty. Finally, the paper includes a discussion of the results obtained when the four models were applied to some actual deepwater Gulf of Mexico (GOM) well data.
