Estimation of flushed zone and mudcake parameters using a new micro-resistivity pad device

摘要

An accurate estimation of the flushed zone and mudcake parameters is important information for formation evaluation. The desired parameters are the resistivity of the flushed zone, Rxo, the mudcake thickness, hmc, and the mudcake resistivity, Rmc. Ideally, the sensor that detects aforementioned parameters is only sensitive to the parameters of the near-borehole environment and is not influenced by deeper undisturbed formation. Pad-mounted devices generally provide these parameters at relatively high vertical resolution. As an example, such a device consists of a metal pad with a number of embedded buttons. The pad is mounted on an arm extending from the main tool body and presses against the borehole wall. The pad injects current into the formation and the amount of current from each button is measured. The size of the guard surrounding each button controls the depth of investigation. The information gathered can then be processed simultaneously to obtain the desired parameters. However, the response of the buttons to the formation parameters is nonlinear. The traditional processing method involves the use of correction and cross-plot charts, which map these measurements into the parameters of interest. These charts (or their electronic equivalents) are generated by forward modeling the tool response for a variety of postulated parameter values. Frequently, the curves on the charts cluster together, posing a challenge in resolving the parameters. A small error in measurements can result in large variations of the parameters. It is also difficult to account for the error in the measurements and uncertainties in the underlying model in this processing scheme. An alternative approach is to perform iterative inversion where the statistics of the underlying parameter variation and measurement errors can be taken into account. A number of methods have been developed using different optimization engines. The most popular ones based on Gauss-Newton, or Levenberg-Marquadt methods, require computation of a sensitivity matrix that makes processing timeconsuming and impractical for well site interpretation. The formulation discussed in this paper involves a special transformation that reduces the non-linearity and thus speeds up the inversion. The resulting algorithm is suitable for real-time application. The synthetic and real data log examples show that the micro-resistivity instrument gives excellent estimates of the parameters of interest while maintaining sufficient vertical resolution. Therefore, the results can be incorporated with other measurements for an integrated formation evaluation.