In a recent offshore Adriatic well, logging while drilling (LWD) data was acquired, and its petrophysical evaluation confirmed the expected gas presence was emphasized by a good resistivity contrast with water-bearing intervals. Nevertheless, during a five-day wireline acquisition including fluid identification (FI) no hydrocarbons were confirmed. Challenging borehole conditions requiring pipe-conveyed logging did not allow a wireline resistivity to be included. Consequently, to investigate the results the LWD resistivity was run again after completing the wireline program. A comprehensive data evaluation was then conducted to understand the reservoir conditions.
The evaluation focused on the differences between the time-lapse LWD resistivity measurements. A preliminary review of the raw resistivity curves suggested conductive and resistive invasion profiles were associated with the reservoir. To evaluate the data further, a quantitative evaluation of the resistivity data was performed to generate an array of resolution-matched curves using inversion techniques. This resulted in four, fixed depth of investigation (DOI) resistivity curves with improved resolution, diameter of invasion calculation, and a reduction of dielectric and resistivity anisotropy effects. The resolution refinement was also confirmed by comparison with wireline image data.
The drilling data inversion indicated zero or negligible invasion, as expected. The diameter of invasion, calculated from the time-lapse data in numerous intervals, was also negligible. However, the 60-inch DOI resistivities of the drilling and time lapse runs considerably diverged. Our explanation for this scenario is that invasion is considerably deeper than 60 inches. This would also explain the absence of hydrocarbons during the fluid identification.
An alternative approach to estimate the diameter of invasion was then implemented using resistivity forward modelling. Four radial zones were considered: wellbore, mud cake, invaded zone and formation at initial condition. The resolution-matched resistivity curves were used as an input in the simulation, and were modelled for varying invasion diameter. A high-confidence correlation coefficient of 0.99 was achieved between the synthetic and measured responses, indicating that the invasion may be as deep as 3.5 meters into the formation.
The next step was simulating the formation testing using parameters from the tests together with formation properties obtained from the forward modelling. This effort resulted in an estimate of the required cleanup time to achieve hydrocarbon samples in this environment. Study also comprises approximation of the invasion radius as a matter of real-time decisions support. Consequent sampling optimization allows for rig time savings and successful execution of the formation testing program.
