During drilling of a well, the formation is exposed to mud filtrate invasion. The invasion displaces oil in the vicinity of the wellbore, much like a small water flooding experiment in the case of immiscible mud filtrate and formation fluid. Pressure transient or flow test of a wireline formation tester (WFT) commonly provides reservoir properties under the assumption of single-phase flow. However, a WFT sampling operation in a multiphase flow environment gives an opportunity for determining related properties in an inversion workflow by utilizing recorded bottom-hole pressure and water-cut data. In this paper, we present a novel methodology to estimate multiphase flow properties with the help of numerical simulation and optimization.
The numerical simulation model for mud filtrate invasion and cleanup consists of a proper definition of reservoir properties as well as WFT tool geometry, including size and shape of flow inlets, along with tool storage and fluid segregation effects. The model is embedded in an optimization workflow and relative permeability curves, damage skin and depth of mud filtrate invasion are then estimated by minimizing a misfit function between measured and modeled pressures and water-cuts. The relative permeability curves are parameterized using industry accepted models. The optimization workflow uses a distribution function of response parameters where the entire parameter range is included in the numerical runs, thus ensuring that a global optimum is found. Initial parameter estimates for the optimization process are determined from open hole logs, such as resistivity, and from pressure transient analyses.
The methodology developed in this paper is validated by application to a synthetic dataset with a known solution, and it is subsequently demonstrated on actual field data from a sampling job in an oil-water transition zone. The results of this paper demonstrate that it is possible to reliably estimate multiphase flow properties with defined confidence intervals from WFT sampling data. The key contributions of this study are to show the capability of estimating a variety of multiphase flow properties from routine WFT cleanup data and to establish an automated approach, including a novel inversion methodology, to reduce the turnaround time.