The Separations Of P And S Waves Based On Curvelet Transform

With the depth development of oilfield, the difficulty of development gradually increased, the exploration showing more and more limitations of the surface seismic technology. Vertical seismic profiling technology placed the detectors in the borehole in most cases, the seismic waves signal has a more significant kinematics and dynamic characteristics; compared with the surface seismic technology, the VSP technology has abundant wave field information. The research of this paper is:the separation of the P wave, S wave in the VSP wave field. The traditional separation method of P and S wave, such as F-K filtering, Ï„-p transformation, etc. Although the P and S waves can be separated, but the effect of wave field separation is not very ideal in many practical situations.We understanding and master the basic theory and methods of the Curvelet transform based on the research of existing separation technology for the P and S wave, the Curvelet transform has unique performance, research to apply the method to the possibility and effectiveness separation of the P and S wave. In1999, Candes and Donoho proposed a new method of multiscale geometric analysis based on Ridgelet transform is the Curvelet transform. It except the scale parameters, location parameters, also have the direction parameters, which make Curvelet have a good directional characteristic, for the P wave and S wave data the slope is not the same size, after the Curvelet transform have the different coefficient characteristics, can make an effective separation of P wave and S wave data. The main content of this article is discussion its application in the separation process of P and S wave based on the study of Curvelet transform.Firstly, the introduction of the knowledge of the vertical seismic profile, a brief introduction to the proposed and development course of the VSP, after a simple understand of the common wave field separation method. And then introduces the Curvelet transform, derivation of its implementation process simply, focus on the implementation process of the discrete Curvelet transform and the nature of the transform. Next Curvelet transform is introduced into the field of wave field separation to lay the foundation.Do the early verification of the feasibility of the separation of P and S wave using the Curvelet transform by the Matlab platform based on the study Curvelet transform. First of all, the sharp pulse signal having a different angle, respectively Curvelet transform, in order to get its best suitable angle and the efficiency of Curvelet transform under different angle, analysis the distribution characteristics in the Curvelet transform domain of the spike pulse signal under different angles, looking for theoretical support for the separation of the P and S wave; and proposed separation program of the P and S waves according to the best Angle compensation.Next, replace analog pulse signal by Ricker wave, which more close to the actual situation and Curvelet transform in the best angles, analyzes the transformation matrix, contrast the similarities and differences of Curvelet transform with the spike pulse signal at the same angle. After have a vertical level signal, respectively Curvelet transform to compare the advantages and disadvantages of two kinds of processing mode. Draw the vertical flattening coefficient matrix distribution is more suited to the signal decomposition and reconstruction.Finally, added controllable noise to the experiments made on the step, and the separation of the P and S wave on this basis, achieved good separation effect.In this paper, learning the relevant theories of wave field separation and Curvelet transform in the early, after studied a series of signal by the Matlab platform, and discuss the feasibility and reliability for the separation of the P and S waves by the Curvelet transform, come to the following conclusions:1. For the separation of the P and S wave, need to use the vertical correction data to the information in the early, to obtain a better separation effect.2. When data angle less than15degrees, the Curvelet transform will not be able to carry out the separation of the P and S-wave data.3. The Curvelet transform can achieve better separation of P and S wave under the strong background noise.4. For the sharp noise and index noise, the suppression effect is better than the random noise, the gaussian is the worst.