The Research On Fractured Reservoir Prediction Using P-wave Seismic Methods

The production of fractured reservoir account for more than half of thetotal oil and gas production of the world, and it is estimated that in the 21stcentury the fractured reservoir is the most important target for more reserves.But the fractured reservoir is different from the porous reservoir in that theformer has lower porosity, higher imhomogeneousness, and more complexity.So the fractured reservoir predicting becomes one of research hotspot. Theseismic data contain abundant information about the fracture, and seismicmethod is one of the key approaches to detect fracture. From the 50's of lastcentury to now, there have been three kinds of ways to predict fracture. One isbased on the cause analysis of the fracture, and it is an indirect method. Theother is using seismic shear wave(S-wave), and another is using seismic presswave (P-wave). S-wave has its advantage in detecting fracture, but the high costand the data processing difficult hamper this method's development. Therealistic method is using P-wave, but this method is at initial stage now and soneed systemic research. Most of these researches have been done on thevariation of P-wave amplitude and velocity with the offset and the anisotropicproperties of amplitude and velocity. But so far, the results of using P-wave todetect fracture are not very satisfactory. There are causes from acquisition andprocessing, but the effects of the formation structure and the inhomogenousproperties of overlay are also important. So there are still has difficulty in usingP-wave to detect and predict fracture.In order to simulate the real fracture as closely as possible, my researchstarts from forward modeling. I studied the effects of the gradient, the density,the aperture of the fracture on the P-wave field, and the P-Wave response of thecombination of different fracture and the random fracture. And then, I got thefollowing conclusion:(1) The aperture modeling shows that the smaller theaperture is, the weaker is the anisotropy;(2) It is thought that the bigger thedensity of fracture is, the more serious is the anisotropy , and vice versa. Butthis case not accords to the results of physical simulation. This indicates that thedensity of the fracture is not the key factor and there are many factors that affectthe anisotropy of the fractured media;(3) The combination of shallow and deepfractured formation modeling shows that the fractures existing in the overlayformation have great effect on the P-wave field. They will not only causesanisotropy response of P-Wave on the under interface between deephomogeneous formation, but also will they affect the P-wave response of thedeep fractured layer. So it is necessary to remove the P-Wave anisotropyresponse of the overlay formation to predict the deep fractured reservoir. (4)Thefractured random media modeling shows the amplitude and the travel-time ofthe P-Wave are depend on the development extent of the fracture. Where thereis high development extent of the fracture, there are large variation of theamplitude and travel-time of the P-Wave. This suggests that P-Wave can beused to detect random distributed fracture. All these experiments results providegood principle to use P-Wave to predict fractured reservoir.In order to inverse fracture quantitatively, assuming the offset is fixed, Istudied the amplitude of P-Wave varies with the angle between the direction ofthe survey line and the azimuth of the fracture and developed a relationship.Based on the relationship, I presented two approaches to predict fracturedreservoir. The first is a precise method that based on 2D multi-azimuth. Thesecond is a least square method that based on 3D multi-azimuth. In additional, Istudied the P-Wave AVO response, and presented a method that combined theanalysis of P-Wave velocity and P-Wave AVO.After studied the qualitative relationship between the attributions ofP-Wave, such as travel-time, amplitude, and frequency, and the fracture, Ideveloped a coherence analysis approach that based on eigenvalue ofcovariance matrix of seismic data, and a method that based on analogy principle,which integrates all of attributions information.Finally, I presented a working flow that using P-Wave azimuthalanisotropy to detect fracture. The working flow includes: forming the macro-binand pre-stack azimuthal gather, coherent noise attenuation, residual staticcorrection, stacking of the gather with fixed offset, azimuthal amplitude fitting,azimuthal AVO fitting, obtaining variation ratio of parameters, forming theattribution data in the fracture direction and vertical to fracture direction. Thisworking flow provides high quality data for fracture predicting using P-Wave.As an example, I used these methods to the predicting of fractured deepcompact sand reservoir of XuJiaHe in the west of the SiChuan. The seismicdata compose two blocks, all of them covered the main structure. After thespecial processing, I successfully characterized the reservoir using the methodsthat mentioned in the paper, which include: the azimuthal anisotropy analysis ofP-Wave, coherence analysis, amplitude, travel-time, and frequency analysis,and comprehensive analysis of attributions information. And then providedsuggests for two new well.The main achievements of this paper are:(1) According to the complexity of the real fractured formation, simulatedthe P-Wave in media that comprise tilted fracture, different fracture density,different fracture aperture and random fracture respectively, and find thevariation law of the P-Wave response to the gradient, density, aperture of thefracture. These provide reliable founfation for using P-Wave to predict theazimuth, density and scale of the fracture.(2) Developed a multi-trace coherence algorithm which based oncorrelation. This algorithm can adjust and compute the location and thedirection of the adjacent trace along the fracture. Presented a coherencealgorithm which based on the eigenvalue of covariance matrix of seismic data.This algorithm is nearly affected by the selection of the trace, and can suppressnoise effectively, at the same time can improve the resolution ratio of the data.(3) Developed a comprehensive analysis technique that based on analogyprinciple. This technique improves the reliability of the predicting resultsthrough threshold calculation, classification of the data according to thethreshold, and stacking the classified image.(4) Using 3D P-Wave data and the fracture predicting method that thispaper presented, I predicted the fractured deep compact sand reservoir ofXuJiahe, which located at west of the SiChuan province and provided thepromised region.