Numerical simulation and interpretation of formation-tester measurements acquired in the presence of mud-filtrate invasion
MetadataShow full item record
Wireline formation testers (WFT) are widely used to measure fluid pressure, to perform downhole fluid analysis in real-time, and for estimating permeability through pressure transient testing. Formation testers can measure a range of fluid properties such as color, viscosity, density, composition, pH, optical refractive index, pressure, salinity, fractional flow, and gas-oil ratio (GOR). However, WFT measurements are influenced by the process of mud-filtrate invasion because overbalanced drilling promotes radial displacement of in-situ fluids by mud filtrate. Oil-base mud (OBM) is first-contact miscible with native oil and can lead to contaminated fluid samples, erroneous estimates of petrophysical properties, and changes of composition, viscosity, compressibility, GOR, and fluid density. The objective of this dissertation is three-fold: (1) to quantify the effect of OBMfiltrate invasion on WFT measurements, (2) to estimate in-situ petrophysical properties concomitantly from transient measurements of pressure, flow rate and GOR acquired with formation testers, and (3) to quantify petrophysical, geometrical, and fluid properties that can minimize the time of withdrawal of uncontaminated fluid samples. In order to quantify the effect OBM-filtrate invasion on WFT measurements, we develop a two-dimensional axial-symmetric compositional simulator and subsequently use a commercial adaptive-implicit compositional simulator, CMG-GEM1. History matching of three field data sets acquired with probe-type formation testers in light-oil formations accurately reproduces measurements of sandface pressure, observation-probe pressure, GOR, and flow rate. Further, we demonstrate that history matching enables the detection and diagnosis of adverse data-acquisition conditions such as plugging, noisy data, and presence of OBM-filtrate invasion. We introduce a dimensionless fluid contamination function that relates GOR to fluid-sample quality. Sensitivity analysis of simulated fluid-sample quality to petrophysical properties clearly indicates that sample quality improves in the presence of anisotropy and impermeable shale boundaries. A computationally efficient dual-grid inversion algorithm is developed and tested on both synthetic and field data sets to estimate in-situ petrophysical properties from WFT measurements. These tests confirm the reliability and accuracy of the inversion technique. Results indicate that permeability estimates can be biased by noisy measurements as well as by uncertainty in flow rate, relative permeability, radial invasion length, formation damage, and location of bed boundaries. To quantify petrophysical and geometrical factors that can optimize the time of withdrawal of uncontaminated fluid samples, we compare the performance of focused and conventional probe-type WFT in the presence of mud-filtrate invasion. Simulations indicate a significant reduction in fluid-cleanup time when using a focused probe. The specific amount of improvement depends on probe geometry, fluid composition, and petrophysical properties of the probed formation. Finally, we develop an inversion method to estimate Brooks-Corey parametric saturation-dependent functions jointly from transient measurements of fractional flow and probe pressure. Results show that estimating Brooks-Corey parameters can be nonunique if the a-priori information about fluid and petrophysical properties is uncertain. However, we show that focused fluid sampling consistently improves both the accuracy and reliability of the estimated relative permeability and capillary pressure parametric functions with respect to estimates obtained with conventional-probe measurements.