An Improved Frequency-dependent AVO Inversion Method and Its Application for Fluid Detection

Recently, more and more attention has been paid on fluid-related dispersion, which carries much useful information for fluid identification. Though frequency-dependent AVO inversion is viewed as an effective resolution on this aspect, yet its reliability greatly depends on the accuracy of AVO approximation and Vp/Vs. For this reason, we present an improved algorithm by employing Aki & Richards approximation instead of the routinely used Smith & Gidlow approximation. With the help of pre-stack inversion, variant Vp/Vs is also taken into account instead of a constant value (usually 0.5). Feasibility and advantages of this method are first verified by tests on a 2D physical model, which contains 3 caves filled with different fluids. The tests show that a more reliable P-wave dispersion is resulted from the improved frequency-dependent AVO algorithm, which is more suitable for identifying the fluid types and suppressing the nonreservoir energy than traditional algorithm. In practical application, we apply this improved algorithm on dissolved-cave carbonate reservoirs, and the obtained results are in good agreements with drilling information, such as fluid type and practical production.