Research Of Velocity Analysis For Low Signal To Noise Ratio Seismic Data

With the exploration and development of oil and gas increasing, the seismic exploration is gradually turning to the piedmont zone, loess plateau area with complex surface conditions, large lateral change of velocity, complex deep geological condition, high steep structure, overthrust belt and salt and other complex structure. The seismic datas of this area have a low SNR, large trace time changes. The original shot gathers or common midpoint gathers have no obvious reflection events, velocity spectrum energy is not focused which brings great difficulties to the velocity analysis prosess. For the low SNR data, how to establish an initial velocity model accurately becomes a key problem to be solved.This paper firstly extracts the pre-stack gathers which can improve the SNR, then take a high precision velocity analysis based on this gathers.The traditional common-reflectionsurface stack(CRS) method considers the local feature of the reflector and the whole reflect energy of the first Fresnel zone, uses the set of high coverage to improve the signal-to-noise ratio of the data efficiently. But the output of this method is high quality stacked section which has no help for the subsequent velocity analysis process. So we improve the method in the first part of the thesis to obtain the high SNR pre-stack gathers.Common scattering point(CSP) gather based on equivalent offset concept accumulating all scatter energy may be from the same scatting point with higher coverage and larger offset range can improve the quality of pre-stack data. But this method has some problems: scatter energy nearby map to this scattering point in the extraction of one scattering point gather, it is difficult to judge whether an event belongs to this scattering point or not; in the construction of a CSP gathers, all of the original gathers used for mapping, the memory requirement is high, the calculation is large; if the mapping aperture is too large, the coverage of CSP will increase large; During mapping we use the root mean square velocity at zero offset operator, without considering the transverse velocity. Therefore, based on the viewpoint of common offset Kirchhoff diffraction migration,this paper takes the common offset vertex time into the double square root equation, obtaines a new parametric common scattering point gathers mapping formula, the method directly put the common offset stack amplitude on the common offset apex, able to control the stack gathers’ quantity, save storage space, improve the computation efficiency; and the mapping process less depends on the formation of velocity speed.We obtain high quality prestack gathers for velocity analysis, however, the velocity spectrum energy calculated by superposition and related discrimination criterion is not obvious, the velocity spectrum resolution is not high, affecting the accuracy of velocity analysis. This paper uses the correlation between seismic data, apply the principal component analysis(PCA) method to the velocity analysis. But the conventional PCA method is not sensitive for low signal-to-noise data, we use a weighted factor to enhance the sensitivity in time and velocity direction, improve the resolution of velocity spectrum which bring convenience for collection of velocity spectrum. After getting the velocity spectrum, how to carry out interactive pickup becomes another problem. In this paper,we use Su platform and iCluster software to realize the interactive pickup of velocity spectrum and the establishment of smooth RMS velocity field.