Streamline-based inversion of formation properties from formation-tester measurements acquired in high-angle and horizontal wells

• Petrophysical properties from a near-wellbore streamline-based inversion method.
• Integration of well logs, pressure, and contamination measurements.
• 90% reduction in CPU computation time achieved with the new method.
• Applying the new method successfully in two field examples.

A new method is developed based on streamlines to determine the reservoir properties from formation-tester measurements. To do so, a previously developed finite-difference reservoir model is coupled with a streamline method to simulate the near-wellbore dynamic measurements in the presence of invasion and arbitrary fluid distributions. The streamline method is specifically developed to overcome technical challenges in modeling deviated wells in heterogeneous reservoirs efficiently.In this study, synthetic reservoir models are constructed based on measurements acquired in offshore well A with 15o deviation angle. The streamline-based method is then implemented to simulate packer-type formation-tester measurements and to appraise permeability of multi-layer reservoirs. In addition, transient measurements from a focused-sampling probe-type formation tester is modeled with streamline method to estimate relative permeability and anisotropy in offshore well B. Inversion results indicate that the accuracy of estimated formation properties is higher for formations with larger mobility because more streamlines trace flow into probes from large-mobility layers. For offshore well A, in the presence of 5% zero-mean Gaussian additive noise, the uncertainty of permeability varies from 6% to 38% for layers with high and low mobilities, respectively. The uncertainty increases to 8% and 41% for high and low mobility cases, respectively, when 5% skewed-Gaussian noise contaminates the measurements.The coupled finite-difference and streamline-based inversion method (FDSM) is then compared to a previously validated finite-difference reservoir model (FDM). The coupled FDSM is 8 times faster than FDM on an average when estimating properties of heterogeneous reservoirs using inversion on formation-tester measurements acquired in highly deviated wells. Computational advantage of FDSM is due to application of one-dimensional solutions of fluid saturations and concentrations along streamlines instead of three-dimensional FDM numerical calculations. Despite the efficiency, history matching of measurements indicates up to 6% difference between FDSM and FDM in results.In high-angle wells, mud-filtrate invasion causes a non-symmetric distribution of invading fluid around the wellbore perimeter. Moreover, presence of large permeability–porosity contrast among layers increases complexity of numerical computations and uncertainty. This study proposes that the streamline-based inversion method is an excellent candidate to overcome these difficulties efficiently.