Lessons learned from Wireline Formation Testing with Single and Multiple Probes

Abstract

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractPreviously there were few RFT data available for the Wara formation of the Greater Burgan field. Therefore, during recent drilling, pressure measurements were made to investigate if vertical pressure equilibrium exists in the formation after a production history of more than fifty years. The erratic nature of measured pressure profiles indicates that vertical communication is poor in the reservoir.In each of the pressure measurements, a single probe was inserted into the mud filtrate zone, and then about 20 cc of liquid was collected. In addition to obtaining the pressure value, we also derived the permeability value by diagnosing the pressure derivative signatures. Multiple probes were also used to evaluate vertical permeability over an interval of 2.3 feet. This paper describes how we analyze the data and overcome the problems encountered. We demonstrate that 3D-reservoir simulation of the formation test can improve the quality of data interpretation. A summary of the analytical method development is also given.Due to the high precision, 0.01 psi, of the quartz gauge, the single-probe pressure derivative data quality was satisfactory for low permeability sands and the analytical method could derive both the horizontal and vertical permeability values successfully. However, for high permeability sands, there often existed a pressure hump immediately after shut-in and it resulted in unrecognizable signatures.A non-linear regression method was used to interpret multiple-probe test data. However, sometimes a simultaneous pressure match for all the probes was difficult to achieve by the analytical model. 3D finite-difference simulation was required to fit the data and characterize the geological heterogeneity. Single-probe pre-test results were used to provide a good initial permeability guess for the numerical model. Gamma ray and formation image loggings were proven valuable to determine the layering of the numerical model. Finally, core measurement and buildip test results were compared with formation test results.