The Study Of Multi-gather Migration Velocity Modeling

There is a higher requirement on the precision of seismic data imaging with the increasing difficulty of the seismic exploration and more complication of the reservoir. In order to have a better service for the E&D of oil field, Pre-SDM imaging has become the focus. Nowadays, several Pre-SDM methods have been applied in seismic data processing successfully, and also get high-quality seismic sections. But in Pre-SDM, imaging quality directly depends on the precision of the velocity field used in migration. When the underground velocity-depth model is correct, we can get good seismic imaging result, or there will appear errors in imaging depth and amplitude fidelity, polarity and magnitude of errors depend on the precision of velocity-depth model. Thinking about the importance of Pre-SDM to target layer imaging and the sensitivity of it to velocity-depth model, seismic workers have tumed their attention to migration velocity modeling in recent years. There are many migration velocity modeling methods. According to the theories on which they base, it can divide into ray-tracing migration velocity modeling, migration velocity modeling of wave equation and comprehensive migration velocity modeling. Whet/based on gathers, it can divide into CMP, CIP, CRP, CRS and CFP gather. Wave equation method can divide into two methods: DFA and RCA. But these methods mentioned above were restricted by underground geological structure and the complication of velocity field, such as reflector dip, the complexity of overburden layer, offset and lateral variation of velocity field. In order to fit this constrain condition better to realize properly imaging of complex structure, we firstly discussed the quantitative relationship between velocity error and depth error, depth, traveltime error, velocity and offset in migration velocity modeling theoretically, and velocity-modified criteria. Then applied it to a variety of models (including Marmousi model), and reasonably selected ray-tracing method, tomographic method or Pre-SDM method used for migration velocity modeling. At the same time, we should select a kind of migration velocity modeling method according to the result and efficiency. Finally, we put forward four kinds of migration velocity modeling: One is CMP gather method based on ray-tracing, another is CIP gather method based on residual curvature analysis and Pre-SDM of wave equation, the third one is CFP gather method, the theory on which it depends is equal-traveltime principle and forward-modeling of ray-tracing or wave function, the last one is CRS gather method based on the theory of optimization of three parameters. Sometimes we can also adapt migration velocity modeling with multi-level and multi-gather, the applications to the models have showed the effectiveness and adaptability of the methods.This paper has a large range in content. It includes Pre-SDM, forward-modeling ,CFP migration and CRS stack. In PSDM, we discussed Kirchhoff integral method, SSF, FFD, FXFD and GS methods which involves PS method---the same as SSF, extended local Born approximation fourier (single reference slowness and multi reference slowness) and extended local Rytov approximation fourier methods. All of these methods were applied to Marmousi model. We derived the high-order FD equation of forward-modeling which fitted for strong lateral variation in velocity and simulated Marmousi model and Shengli model with zero offset. CFP migration is a new Pre-SDM method based on equal-traveltime principle. We did some theoretical model with CFP method. CRS stack is a seismic imaging method which is velocity-free. We derived all the theoretical formulation, and applied it to many models including Marmousi model and also processed a real data of Shengli area. These methods mentioned above are the base of migration velocity modeling. Taking computing efficiency, precision and stability into consideration, we used Pre-SDM of SSF and FFD, forward-modeling of high-order FD, CFP migration and CRS stack method. The key problem we discussed is migration velocity modeling. In migration velocity modeling, we used CMP, CIP(CDP or CRP), CFP and CRS gather. In CMP gather method, we used ray theory and maximum energy or maximum coefficient of correlation as the criteria to modify migration velocity, derived the analyzing formulation of the reflection travel-time of each trace in a variety of gathers including CMP gather using ray method with constant velocity or variable velocity, and put forward a CMP gather migration velocity modeling method based on velocity scanning. Then we applied it to the complex model. In CIP gather method, we derived the fundamental relationship between imaging depth error and migration velocity error, then made a quantitative description on the residual curvature appeared in post-migration CIP gather. To simple model, we gave the linear interpolation method to find updated velocity using the imaging error in depth of each trace in CIP gather. To complex model, we put forward single parameter and multi-parameter method to get derivative function using error in imaging depth of each trace in CIP gather, and following single parameter disturbing and multi parameter disturbing simultaneously to calculate modified velocity using generalized linear iterative inversion. We also applied it to many models including Marmousi model. The result showed good adaptability and effectiveness of this method. In CFP gather method, we put forward the method of constraint parameter inversion to get modified velocity based on least-square principle using equal-traveltime principle and DTS analysis. Then we applied this method to a model and the result is satisfactory. In CRS gather method, we derived the quantitative relationship between optimal three parameter and migration velocity field, and then put forward ray-tracing method to get migration velocity field using optimal three parameter in CRS stack. The applications proved its effectiveness and adaptability.According to the results we got above through four kinds of migration velocity modeling, the gathers and the theories on which they based, we can come to the conclusions as followings. To CMP gather method, it has a high efficiency but low precision which depends on the complexity of the structure. In a sense, the more complex the structure is, the lower the precision. There also exist the problems of accumulation error in velocity and non-uniqueness in solution. It fits for velocity modeling in geological body which is not very complex in structure. Ray-tracing method with variable velocity can improve greatly the adaptability of migration velocity modeling of CMP gather to complex media. To CIP gather method, its efficiency is low but has a high precision and stability. This method fits for velocity modeling in complex geological body. To CFP gather method, it has a high precision and stability, and fits for velocity modeling in complex geological body when based on wave equation modeling; when based on ray-tracing modeling it has a high efficiency, but the precision and stability is a little lower, and it fits for velocity modeling in general geological body. To CRS gather method, the precision of velocity analysis depends on the precision of optimization of three parameters. It fits for velocity modeling in complex geological body.In order to improve the efficiency, precision and stability of migration velocity modeling, we can also use the method of multi-level and multi-gather. For instance, we use CMP gather method in first level and CIP gather method in the second level, etc.. What we should pay more attention to is that CFP and CRS gather migration velocity modeling provide us with two kinds of new migration velocity analysis methods. With their continuous development and perfection, these methods will play more important role in migration velocity modeling and come into wide use in the field.