Abstract
Induction logs run in wells drilled with oil-based mud frequently show evidence of filtrate invasion. Computing the diameter of invasion from a traditional tornado chart requires input from a shallow reading resistivity device such as a microlog or laterolog, but these are not available in non-conductive boreholes. A new technique using induction measurements alone circumvents the problem, while being both simple and robust.
The technique has also been applied to data from wells drilled with conductive muds. In these cases, a significant part of the total signal can be due to the borehole environment, and this must be backed-out in order to derive the formation conductivity. This introduces a degree of uncertainty in the computed curves that increases as the mud resistivity decreases. We quantify this by looking at the relative magnitudes of signals due to formation and borehole contributions.
The approach has particular application in low conductivity formations where signals are of low amplitude. Conventional wisdom, largely inherited from first generation tools, is that readings greater than one or two hundred ohm-m are unreliable. We demonstrate that, subject to appropriate processing, logs can be trusted to at least 1000 ohm-m. This has helped us characterize an oil reservoir with fresh formation water.
The technique has application where the environmental envelopes of laterolog and induction measurements overlap, and where current practice is usually to select one and reject the other. A better understanding of induction tool performance in highly resistive formations helps explain apparent discrepancies between the measurements in terms of real formation properties such as anisotropy or deviations from step invasion profile.