The Analysis Of Some Unsolved Induction Interpretation Problems Using Computer Modeling

Modern computers have allowed the development of increasingly sophisticated models of resistivity logging tool response to actual logging geometries. Significant progress has also been made recently in the inversion of logging data to obtain actual formation parameters. Until we achieve the goal of completely automatic real-time 3-D inversion of logs, such sophisticated forward models can be very helpful in analyzing unsolved interpretation problems. Several of these problems will be examined in detail in this paper. A particularly difficult situation to analyze occurs when features of interest are much smaller than the resolution of the tool. This is the case for induction response in highly laminated reservoirs containing thin layers of oil bearing sand and non-reservoir shale. In these reservoirs, we would like to determine the resistivity of each constituent of the laminated zone and the relative amounts of sand and shale present. Modeling ILD response to sequences of several dozen sand-shale laminations ranging from 1" to 1" in thickness, we verify that the tool response is predictable when expressed in terms of conductivity instead of resistivity; it is simply a linear average of the volume sand times the sand conductivity plus the volume shale times the shale conductivity. Using core data or a micro-resistivity log such as a micro-electrical scanner image to determine the relative amounts of sand and shale present in laminated zones, this relationship can be applied to solve for the unknown conductivity of the sand laminations and thus estimate water saturation and producibility. In conjunction with modeling laminations, we also analyze how severe dip must be to significantly effect ILD response in the limiting case of three thin beds. Another area where modeling is helping to solve an induction interpretation problem is in the analysis of erratic oscillations that occur on some dual induction North Sea logs in high resistivity formations logged using oil base muds. In this case, a finite element model is used to simulate several different sets of conditions that log analysts have suggested as possible causes of these oscillations. Some of these suggestions are, oil base mud filled cracks, laminations, or layers containing magnetic materials. By elimination, it is shown that the only way these erratic logs can be reproduced is by modeling a conductive anomaly near the borehole wall. The origin of this conductive anomaly is still open to discussion. Some suggestions have been, breakdown of the mud emulsion at high pressures and temperatures, or the build-up of a conductive mudcake from the saline water in the mud emulsion. More information is needed about borehole conditions at the time these logs were run before a definitive conclusion can be reached.