Fracture-cave system in compact carbonate reservoir is a complex, inhomogeneous and nonlinear system. Detection of the fracture-cave system is a worldwide difficult problem. Fractures and caves are main reservoir space, and key criteria for reservoir evaluation, therefore, it is significant to detect them and provide guidance for petroleum exploration and exploitation. The thesis begins with feasibility analysis on detection of fractures and caves, aimed at detection method and technology as well as integrated reservoir evaluation, and by way of analysis on geological characteristics, numerical simulation, ultrasonic experiment on physical model, theoretical forward modeling, the thesis discusses seismic response characteristics of fractures and caves in reservoir. Based on above study in the thesis attempted to detect fractures and caves by multiple methods to 3D seismic data, and then discussed reservoir classification and evaluation.According to fracture theory put forward by Hudson, the elastic parameters of EDA medium were calculated, and then the formula of the phase velocity of P-wave was determined by Christoffel equation. Numerical simulation based on above parameters indicates that velocity of P-wave has gentle anisotropy in fractured media at same time, the change of reflection coefficient is three times as much as the change of velocity resulted from the same condition of fracture density variation. Results of ultrasonic experiment on physical model testify that with the existence of fractures and caves, the value of seismic P-wave velocity and dynamic parameters decrease, and variance ratio of dynamic parameters are high two to three quantitative grade than that of velocity. The change of amplitude and frequency is bigger than that of velocity by theoretical forward modeling and wave-field characteristic analysis of fractures-caves contained on top part and inner part of buried bill. Three methods all discover that dynamic characteristic is more sensitive to fracture than kinematic characteristic. Forward modeling also indicates the wave-field characteristic of the buried hill surface is different to the characteristic of the buried hill internal phase, so amplitude is more sensitive to fracture in inner part of buried hill than that at the top part of buried hill, the center of fracture-cave zone is consistent with center of seismic response abnormity zone, combination of multiple fracture-caves generate a string of beads abnormal seismic reflection or chaotic seismic reflection characteristic, number and length of a string of beads are approximate with number of fracture-cave combination. That is to say, it is possible and feasible to detect fracture-cave system by way of seismic means in carbonate reservoir.On the basis of above studying, we have built the new seismic detection method system of fracture-cave contained in compact carbonate reservoir. It includes wavelet frequency-divided coherence to detect fractures, the four methods to detect combination of multiple fracture-caves: synthesizing parameter method based on dynamics, waveform analysis based on kinematics, the second order directional filter and GHT method based on edge detection. Cumulative energy difference method to detect karst caves.Data volume with different channel (or frequency band) obtained from wavelet frequency-divided is processed by coherence analysis, and high frequency coherence data is processed by wavelet decomposition and fusion. The result data clearly displaies the plane developed feature of minor fault and fracture.Under instruction of theoretical forward modeling, multiple dynamic parameters of seismic wave are fused to directly demonstrate plane distribution of fracture-cave system. Forward modeling also testify that waveform analysis based on kinematic parameters is effective to detect fractures and caves, therefore waveform analysis method is adopted. Using the second order directional filter method and GHT method based on edge detection, they are sensitive and powerful to identify fracture-cave abnormity zone, and detection effect is satisfied. Data fusion based on principal component analysis is that the methods mentioned above are coupled to an integrated parameter, and it can exactly describe plane developed feature of fracture-cave system.Karst cave is very important signification during reservoir evaluation. By cumulative energy difference method brought forward in this thesis, karst caves in Yijianfang formation are detected, results show that karst caves are mainly developed at the upper part of reservoir, and the plane distribution is zoning, and distribution of karst cave is mainly controlled by (paleo-) structure and fault. Detection result by this method is well consistent with drilling data, and is an innovative method.Combined with fracture detection by using wavelet coherence, integrated detection of fracture-cave system, and result of karst cave detection, the carbonate reservoir in Yijianfang formation is classified into 4 grades, and plane distribution is divided into 5 districts. Combined with the geological production including structure and fault, to evaluate 5 districts separately. At last, it is pointed out that these are the best prospecting targets of fracture-cave developed belts, area with well developed fracture and comparatively scattered karst cave cluster. |