Search

Aimed at the disadvantages of Fuzzy C Means Clustering (FCM) algorithm which is sensitive to initial values and easily falls into local convergence, a new FCM algorithm named as CQPSO-FCM is proposed in this paper, which makes a combination of FCM algorithm and Chaotic Quantum Particle Swarm Optimization (CQPSO) algorithm. In order to solve the problems mentioned above and improve the ability of fuzzy classification, CQPSO algorithm is used to initialize the membership matrix, which significantly improves the global optimum search ability of the proposed algorithm. What’s more, this method could be used to predict accurately the fluid of complex fracture-cavity carbonate reservoir. Because the common fluid director of AVO inversion would not be available enough in fluid identification of carbonate reservoir, this method is then employed to establish a relation of fluid and reservoir between well logs and elastic properties of AVO inversion. By this accurate relation, the fluid enrichment and fluid type are both identified. The case study results not only prove a practical applicability of the method in predicting the fluid distribution in carbonate reservoir, but also provide a new way for fluid prediction in carbonate reservoir by taking full advantage of pre-stack seismic data.

A new fracture parameters inversion method based on azimuth anisotropic elastic impedance in HTI media is introduced in this paper. The azimuth anisotropic elastic impedance formula derived from Connolly’s elastic impedance theory and Ruger’s equation is deduced. The explicit relation of elastic impedance and fracture parameters (fracture density and direction) is built. The formula is further used to form an integral workflow of fracture parameters inversion based on azimuth anisotropic elastic impedance. Particularly, the EI inversion for each azimuth is constrained by the well data of anisotropic elastic impedance. The proposed method is successfully applied in both numerical data with different SNR and field data in the Tarim Basin. It is proved that the fracture parameters can be efficiently inverted, which shows an excellent stability in fracture density extraction even for low SNR data. This method provides a new idea and approach for the fractured reservoir prediction.

Seismic anisotropy is relevant to frequency. When fractures in reservoirs are saturated with different fluid types (oil, gas and brine), the variation rates of anisotropy versus seismic frequency (which is called anisotropy dispersion) are different. Thus, anisotropy dispersion can be used to distinguish different fluid types in fracture. In this paper, P-wave frequency-dependent azimuthal AVO inversion algorithm is proposed and applied to real seismic data. The studies focus on the analysis of P-wave anisotropic dispersion response for different fluid saturation. Continuous wavelet transform is employed for spectral decomposition, aiming at well time-frequency localization. Real data application indicates that the magnitude of P-wave anisotropy dispersion of brine is far stronger than that of oil and gas in fractures. Our study firmly proves that the method aimed at calculating the P-wave anisotropic dispersion for different fluids saturation is valid and reliable, and this brings us the confidence to discriminate fluid by the calculated anisotropic dispersion values in fractured reservoir.