Study Of Processing Methods For Low S/N Ratio Seismic Data From Complex Areas

In recent years, with the advance of exploration technology and the continuous rising level of exploration, the easily found structural traps with simple geological conditions are becoming less and less and the complex oil and gas reservoirs in the complex areas are becoming more and more explored. These complex exploration objects present new challenges to the oil and gas geophysical technology with seismic exploration as the main method. Especially in the area of low S/N ration seismic data processing there are still many bottleneck techniques hard to be solved. Firstly, the drastically changed surface relief and the complicated surface velocities caused by the effects of complex surface geological conditions, together with the complexity of surface heterogeneity, lead to the original seismic data acquired with low S/N ratio and serious interference and in the mean time, the surface waves, the refraction waves and especially the comparatively evolutional scattering noises present many difficulties for raising the S/N ratio. Secondly, because of the effects of complex underground seismic and geological conditions, the data structures from complex areas are extremely complicated with developed fractures, complicated wave fields and very changed lateral velocities. Furthermore, the igneous rocks are comparatively developed and widely distributed. Due to the presence of igneous rocks, on the one hand, the igneous rocks have very strong energy and the strong reflection energy has very good isolation for the underlying effective reflection. On the other hand, the velocity is high and has very big wave impedance difference with surrounding rocks. Thus, the strong multiple waves are produced under the reflection interface. The existence of these factors makes the already complex wave fields much complicated and the accurate imaging of structures difficult.Based on the actual S/N ratio data from complex areas in Fangzheng Fault of Sanjiang and Yishu Graben of Peripheral Basin of Daqing Oilfield, the study in this paper aims at the methods for suppressing the pre-stack noises, especially the strong scattering and multiple-wave interference to improve the S/N ratio of actual data and ensure the accurate imaging of complex structures. Through method argumentation of principles, theoretical model design, software programming, as well as testing of actual data and final practical application of actual data and based on the results of previous researches, this paper puts forward some new innovative technical methods and understanding and has obtained the following five technological achievements suitable for low S/N ratio seismic data processing in complex areas.①Pre-stack scattering noise attenuation technique based on the spatial conversion of signal sub-space decomposition.②Multiple-wave removal technique with self-adaptive parabolic Radon transformation in the frequency domain.③Compensation technique for seismic signal transmission loss based on the wave theory.④High-density and point by point dual spectral velocity analysis technique.⑤Pre-stack time migration processing technique based on the anisotropy of surface relief. After making the systematic summary and conclusion as well as conducting the actual data processing, the technological process framework suitable for low S/N ratio data processing in complex areas has been obtained, which provides effective methods, ideas for reference and technological process for low S/N ratio data processing in complex areas. The work done mainly focuses on the following areas.1. By the analysis of formation mechanism and forward modeling for scattering noises, the characteristics and formation reasons of scattering noises have been obtained. The display forms of scattering noises in original records are generally linear. However, when the scattering anomaly is buried very deep, the similar hyperbolic characteristic, which has relatively strong energy and wide frequency distribution, will be displayed. It is very difficult to be removed by current conventional methods. As to the characteristics of scattering noises, this paper introduces the pre-stack scattering noise attenuation method based on the spatial conversion of signal sub-space decomposition. The thinking of this method is: to use the characteristic that the waveform of scattering noises has a certain similarity in the time-space domain for simulating and predicting the scattering noises and getting the estimation of scattering wave field and then remove the scattering wave filed from the seismic records and realize the attenuation of seismic scattering noises. This paper conducts detailed analysis and argumentation of this method on basic principles, theoretical model design and actual data processing and finally puts forward the pre-stack scattering noise attenuation technique based on the space conversion of the signal sub-space decomposition. With the help of corresponding developed software, this technique has been loaded to CGG processing system on PC clusters and can be mass-produced. It is a new method for pre-stack scattering noise suppression.2. Based on the review of previous Radon transformation technique for the removal of multiple waves, this paper also focuses on the characteristics and existing problems of Radon transformation in the time domain. Through the analysis of basic formulas and physical meaning of Radon transformation as well as multiple tests for actual multiple-wave separation with the theoretical model data, it is concluded that although PRT in the time domain has a clear physical and intuitive meaning, the unavoidable shortcomings of Radon transformation in the time domain still exist when conducting actual programming. After having found the drawbacks of Radon transformation in the time domain, this paper presents a new method that uses the parabolic Radon transformation to remove multiple waves in the frequency domain. Through different implementations of Radon transformation method in the frequency domain, inverse transformation least-square algorithm and forward transformation least-square algorithm, as well as comparison of results obtained by these two methods, it is concluded that the forward transformation least-square algorithm can effectively focus the energy and be convenient to distinguish and remove multiple waves and also prevent the presence of alias in the q-ωdomain. In the meanwhile, in order to further enhance the effects of multiple-wave removal and improve the accuracy of algorithm, this paper also focuses on discussing the application factors such as aliasing problems in the Radon domain, solution of double coefficient matrix equation, selection of white noise coefficient, removal of multiple waves in the Radon domain and self-adaptive filtering, etc. and makes a conclusion with a certain guiding significance. And then, after making a summary and conclusion, this paper puts forward the concrete realization process and procedures with the self-adaptive parabolic Radon transformation to remove multiple waves in the frequency domain. And with the characteristics of ultra low S/N ratio data from Suibin Depression in Sanjiang Basin in consideration, the Radon transformation method in the frequency domain was used to process the actual data and the results are very good.3. This paper also emphasizes the necessity of compensation for transmission loss, which is one factor that cannot be ignored for achieving high-fidelity amplitude compensation. Especially when the objective layer is covered by a large number of strong reflection interfaces or there is the presence of volcanic rocks with strong energy, the energy of effective reflection objective layer is very weak because of the severe transmission loss, which decreases the interpretation accuracy of final results. This paper points out the existing problems on the current conventional compensation method for transmission loss by analyzing the previous research results. The previous studies focused on the transmission loss of seismic waves in the random medium and only have theoretical significance. But the actual medium in sedimentation process is not at random and has cyclic characteristic. Thus, the transmission coefficient calculation in the actual non-random medium must be studied. In order to solve this problem, this paper puts forward the compensation technical method for seismic signal transmission loss based on the wave theory. Firstly, use the geometric theory and wave theory to calculate respectively the transmission coefficient for the thick layer model and thin layer model, describe the difference of physical nature between the two methods and then point out that the wave theory should be used to calculate the transmission coefficient of actual thin-layered medium. Secondly, establish the transmission compensation method based on the wave theory, use the modeling technique to realize the quantitative analysis of transmission loss mechanism, use the wave theory to calculate the compensation factor of transmission loss, point out the processing procedures for realizing post-stack transmission compensation with the velocity spectrum and seismic traces as the input data and then achieve the true restoration of seismic wave energy. Based on a large number of researches on theoretical methods and after making the summary and conclusion, this paper gives the concrete principles and technological process for post-stack transmission compensation. This compensation method was used to process the actual data from Suibin Depression in Sanjiang Basin and the result is very remarkable. The continuity on same phase axis in many parts of transmission-compensated sections is improved after the amplitude of reflection wave being restored and the basement as well as the form of deep strata is clear. Therefore, the post-stack transmission compensation is useful.4. This part is about the conventional velocity analysis method for the data from complex areas based on the time-offset hyperbolic curve of seismic reflection wave travel. Because of the approximation of hyperbolic hypothesis, the conventional velocity analysis cannot meet the requirement for establishing subsurface accurate velocity model with the data from complex areas. This paper puts forward the high-density and point by point dual spectral velocity analysis technique on the basis of previous research results. By introducing principles of high-density and point by point dual spectral velocity analysis, dynamic correction parameters to affect all the offsets, intelligent sorting of the pairν,η, geological filtering of high-density time differential correction parameter field (ν,η), unrelated dynamic correction parameters (dtn,τ0), interpolation and filtering ofνandηfields, this paper presents the theoretical formula of high-order non-hyperbolic time difference and puts forward the processing procedures used for high-density and point by point velocity analysis. This method uses the bi-directional scanning of dual-time parameters to obtain accurate velocity parameters. The actual processing of data acquired from Fangzheng Fault in Daqing Peripheral Basin demonstrates that this method can effectively solve the problem that the large-offset non-hyperbolic time difference correction gather couldn't be leveled. Meanwhile, the point by point velocity analysis and calculation effectively solve the problem of lateral change of velocity and thus the accurate velocity fields of underground complex structures can be obtained.5. In order to ensure the accuracy of migration imaging in complex areas, this paper analyzes the main problems faced by migration imaging in complex areas and puts forward the pre-stack time migration processing technique based on the anisotropy of surface relief. On the basis of imaging principles of surface common-reflection point, this paper uses the equivalent offset method for the data from the terrain with surface relief to eliminate the effect of surface relief on the structural morphologies and then formulates in detail the anisotropic pre-stack time migration processing technique. With the introduction of important concepts of anisotropy, phase velocity and ray velocity, NMO velocity in anisotropic medium, three dimensional description of NMO velocity in the anisotropic medium and principles of anisotropic pre-stack time migration method, this paper focuses on the anisotropic pre-stack time migration velocity analysis, high-accuracy calculation of velocity field, calculation of anisotropic parameters in VTI medium and the implementation process of anisotropic pre-stack time migration, and finally gets high-accuracy velocity modeling method and technical process for three dimensional anisotropic pre-stack time migration,ηfield method and process for obtaining anisotropic parameters in VTI medium with surface seismic data, three dimensional anisotropic pre-stack time migration processing method and procedures. It is proved by actual example that the surface–based anisotropic time migration processing technique can eliminate the effects of anisotropy on seismic wave travel path and velocity as well as the effects of surface relief on seismic data imaging and thus improves the accuracy of migration imaging. At the same time, because of the elimination of every anisotropic effect, it also solves the problem that the geophone offset on the CRP gather is over-corrected and in the meanwhile increases the effective multiplicity used for pre-stack inversion gather. And thus, the S/N ratio of migration result can be improved and it is also favorable for AVO analysis and pre-stack elastic impedance inversion.At the end of this paper, the five techniques used for processing seismic data from complex areas, obtained from this research, are systematically summarized and concluded and a set of technological processes, suitable for processing low S/N ratio data from complex surfaces and complex structures, is given. And then, the actual data of 100 km2 from Fangzheng Fault in Daqing Peripheral Basin was processed with the above techniques. By comparing the overall effect of the final results, the new processed results with the previous 2D and 3D results, analyzing quantitatively the S/N ratios of new and old data and comparing the new interpreted results with the old one, it is demonstrated that not only the S/N ration but also the resolution of data are improved significantly, especially the structural morphologies have very big changes. After comparing with the exploration well data, the new-processed results match perfectly with the well data, which proves that the processing methods in this paper are advanced and practical. This paper also provides a set of new techniques for processing the low S/N ratio data from complex areas, which can be very reliable technical support for accurate structure interpretation, inversion prediction of reservoirs and for speeding up the seismic exploration of complex surfaces and complex structures.