【作者】 张显文；

【导师】 韩立国；

【作者基本信息】 吉林大学 ， 地球探测与信息技术， 2010， 博士

【摘要】 储层预测是油气勘探开发的核心技术,论文以地震方法进行储层预测的相关技术展开,进行了储层地震综合预测技术、储层地震波传播特性分析与储层地球物理参数反演的研究。储层预测是油气勘探开发的核心技术,论文以地震方法进行储层预测的相关技术展开,进行了储层地震综合预测技术、储层地震波传播特性分析与储层地球物理参数反演的研究。地层品质因子Q作为地震波的动力学参数之一,在岩性和含气储层识别方面较速度参数更为敏感,本文分析总结了国内外常用的Q值提取算法,进一步提出一种改进的时频域Q值反演算法,利用该算法对实际数据进行Q值分析,并利用Q值进行反Q滤波的应用与油气储层的识别。地震信号谱分解技术已广泛应用于储层预测中,通过分析比较了目前常用的时频分析算法,进而提出地震信号谱分解的两步法MP快速算法,通过理论模型与Gabor变换、小波变化、S变化、Wigner分布的谱分解效果进行对比分析,利用峰值瞬时频率对薄层厚度进行反演,并通过实际地震数据进行了储层预测。考虑双相各向异性介质理论能更真实地描述实际地下储层,本文基于BISQ机制推导了三维双相各向异性介质的地震波频散方程,给出用于确定相速度和逆品质因子的波数方程及其表达式,研究了固体骨架各向异性、固流耦合效应各向异性、渗透率各向异性以及孔隙度和流体粘滞系数对衰减及频散的影响,并对频散和衰减的方位特性进行分析。进一步分析孔隙度和渗透率的变化对速度频散影响,利用遗传算法,通过不同方位的速度频散对孔隙度和渗透率进行了反演,并分析了速度频散的方位对反演精度的影响。针对反演精度对速度频散方位的依赖问题,提出利用多方位速度频散对孔隙度和渗透率进行联合反演的技术思路,并对反演精度的误差影响进行了分析。更多还原

【Abstract】 Reservoir prediction is the core of oil and gas exploration and exploitation, the techniques of reservoir seismic prediction are a comprehensive and intercross technology series, which include many methods. This paper starts off with the reservoir seismic prediction technique, strive to make further study on this aspect, and provide some technical support for the reservoir prediction, and then reducing the risk of the reservoir exploration and exploitation.This paper firstly elaborates the current studies on the methods of seismic attenuation and seismic spectrum decomposition to detect hydrocarbon, and further makes the detailed summarization on the research progress of two-phase anisotropy. As one of the seismic dynamic parameters, Q is more sensitive than velocity to the application of lithologic discrimination and hydrocarbon detection. Because of the large attenuation characteristic of hydrocarbon reservoir, we usually depict the gas reservoir with Q-profile. In the paper, the current Q estimation methods, including the spectral ratio algorithm, the centroid frequency shift algorithm, the analytic method and two time-frequency analysis algorithms are analyzed by theory model. Furthermore an improved Q inversion algorithm in time-frequency domain is proposed, the new method takes Q-V empirical formula as the initial solution of Q inversion, realizes Q analysis in time-frequency domain with Gabor transformation, and then processes interval Q inversion using interior-point algorithm. By analyzing the dependency of Q inversion accuracy on initial solution, I verify that the new Q inversion algorithm is robust and accurate, and does not depend on the initial solution. The new algorithm is applied to the field data to make Q analysis, and then using the Q value to the application of inverse Q filtering and hydrocarbon reservoir detection.Seismic spectral decomposition is a time-frequency analysis technology based on seismic attribute analysis. It is widely used to determine the formation thickness, reservoir description and the hydrocarbon detection. The thesis describes several the current popular time-frequency analysis techniques at home and abroad, including Short Time Fourier Transform (STFT), Wigner-Ville Distribution (WVD), Continuous Wavelet Transform (CWT), S Transform (ST), as well as Matching Pursuit (MP) algorithm. Comparing all of these algorithms by model analysis, we obtain the advantages and disadvantages:STFT is a older time-frequency analysis technology, it is constrainted by the type and length of the time window, the long time-window will affect the time resolution of high frequency components, vice versa will affect the frequency resolution of the low frequency components. Time-frequency analysis technique based on CWT, decomposes signal into different scales of time-frequency atoms, which can better adapt to the time-frequency characteristics of seismic signals. However, the wavelet family is built on the basis of that the scale and frequency is inversely proportional, and the expanding coefficients within the wavelet framework can not estimate the wave frequency components precisely. S-transform, as the extension of CWT, which takes Morlet wavelet as the mother wavelet, is based on a moving and changing scale localized Gaussian window, and provides the frequency related resolution while keeping a direct relationship with the Fourier spectrum. WVD is another method of time-frequency analysis, which has well time-frequency localization properties, but vulnerable to cross-term impact. Although smoothed WVD can reduce the interference caused by the cross-term, it has to be at the cost of decreasing time-frequency resolution.MP algorithm makes up insufficiency of the conventional time-frequency analysis methods, and it decomposes the seismic record into a series of wavelets. The time-frequency characteristics of the signal is matched by optional wavelets in the redundant dictionary, which can most adapt the structure and characteristics of signal. The paper makes use of the model analysis to compare the advantages and disadvantages of the above time-frequency analysis methods. Further, based on time-delay of the optimal wavelet, a two-stage MP algorithm of seismic signal spectrum decomposition is presented, by multi-points simultaneous searching-style to optimal wavelets, which improve the operation efficiency without losing algorithm accuracy. Based on theory model, this paper discusses the effects of MP spectrum decomposition algorithm, and makes comparison with the algorithms of Gabor Transform, CWT, ST and WVD algorithms. In comparison with the conventional thin-bed inversion algorithms, peak instantaneous frequency method is applied to wedge model, and the results indicate the algorithm is not sensitive to the reflection coefficient variation of reservoir. At last, the two-stage MP algorithm is applied to field-data, and results the different frequency slices, we observe the phenomenon of gas shadow, which can directly indicate the existence of the hydrocarbon.Considering the theory of the two-phase anisotropy can truly depict the real reservoir, on this relevant aspect studies, the paper introduces the theory of anisotropy and BISQ, the elastic wave equation of the two-phase anisotropy, which includes the wave equations based on micro-flow and solid/fluid coupling anisotropy. Further, base on BISQ mechanism, I derive the dispersion equation of 3D two-phase anisotropic medium, and give the wave-number equation and expression used to derive phase velocity and 1/Q. Then I investigate the effects of the solid skeleton anisotropy, the solid/fluid coupling anisotropy and permeability anisotropy on attenuation and dispersion, and also make analysis to azimuth characteristics of attenuation and dispersion. The results show that in the two-phase anisotropic media, the attenuation and dispersion anisotropy of seismic wave are common affected by solid frame anisotropy, solid/fluid coupling anisotropy and permeability anisotropy. On orthogonal fracture orientation, qP1 has the largest phase-velocity, and has the largest attenuation when the azimuth angle is 45°and 135°. As for the effect of solid/fluid coupling anisotropy:qPl has the largest attenuation and velocity dispersion in the vertical direction of the largest solid/fluid coupling density, while qSV appears in the direction of largest solid/fluid coupling density. Relatively, the effect of permeability anisotropy is more complex:in the range of low-frequencies, qP 1 has the largest attenuation and velocity dispersion in the direction of the largest permeability and qSV appears on the direction of the smallest permeability. However, in the range of high-frequencies, the rules of qPl and qSV is reverse. In addition, as for the effect of phase-velocity anisotropy, solid/fluid coupling anisotropy and permeability anisotropy is less than solid frame anisotropy. The curves of phase-velocity azimuthal anisotropy mainly embody the effect characteristics of solid frame anisotropy.As to the effect of porosity and viscosity on attenuation and dispersion, the velocity dispersion of P1-wave increases with the increasing of porosity or viscosity, and the velocity dispersion of S1-wave nearly do not change with the alternation of porosity or viscosity. Along with the increasing of the porosity,1/Q of the P1-wave and S1-wave with different frequencies increases gradually. However,1/Q of the P1-wave and S1-wave with different frequencies decreases with the increasing of viscosity. Relatively, the effect of porosity or viscosity on 1/Q of S1-wave is small. The above research results provide strong evidence for predicting the existence, distribution and pore structure of reservoir fluid, and make multi-azimuths reservoir geophysics parameters joint inversion feasible.As to the studies on reservoir, the porosity and permeability are the key parameters in hydrocarbon reservoir detection, and the reservoir parameters inversion plays the significant role on dynamic monitoring and optimal management in oil exploitation. Based on the dispersion equation of 3D two-phase anisotropic media derived from BISQ mechanism, I establish the relationship between porosity, permeability and phase velocity, attenuation. After the effects of porosity and permeability variation on velocity dispersion are analyzed, then the fitness function to make reservoir parameters inversion is constructed. The porosity and permeability inversion is studied by different azimuth velocity dispersion with genetic algorithm. and the azimuth characteristic of inversion accuracy is analyzed. The results show that the porosity and permeability inversion accuracy is influenced by velocity dispersion azimuth used in inversion, which can be elaborated in single-azimuth velocity dispersion to make porosity and permeability inversion by 45°incident-angle and 0°、45°、90°azimuth respectively. As to qPl velocity dispersion incident with 0°azimuth to make porosity and permeability inversion, along with genetic algorithm iteration, the porosity and x-permeability, z-permeability can fast converge their optimal solutions, and the object function also converge rapidly, however, the y-permeability diverges during the whole inversion process. Similarly, as to 90°azimuth incident, the porosity and y-permeability, z-permeability can fast converge their optimal solutions, however, the x-permeability diverges during the whole inversion process. For 45°azimuth incident, the object function also converge rapidly, and the porosity and x-permeability, y-permeability, z-permeability can fast converge their optimal solutions.Aim at the dependence of porosity and permeability inversion accuracy on velocity dispersion azimuth, I propose an improved inversion thought, which makes use of multi-azimuths velocity dispersion to make porosity and permeability inversion, by improving the fitness function, the inversion error is analyzed using the theory data. By multi-azimuths velocity dispersion to make porosity and permeability inversion using 45°incident-angle and 0°、45°、90°azimuth jointly, the inversion results show that the multi-azimuths velocity dispersion to make porosity and permeability joint inversion, the inversion accuracy compromises the every azimuth dominance. Along with inversion iteration, the object function fast converge and the porosity, the three directions permeability converge their optimal solution quickly. More over, the multi-azimuths’ inversion accuracy is higher than porosity and permeability inversion of the single 45°azimuth. As to using multi-azimuths velocity dispersion to make porosity and permeability inversion, considering the possible error in the process of the velocity dispersion estimation in field-data, I add noise to the theory data to depict the error. Then the effect of input dispersion error on inversion accuracy is analyzed, the results show that the slight dispersion error affects the inversion accuracy little, and as to the strong dispersion error, the inversion accuracy is also satisfied, under this case, the object function can fast converge, the porosity and permeability also converge their optimal solution fast. But the more strong dispersion error can change velocity dispersion curve manner, which make inversion failed.Joint inversion of porosity and permeability based on the multi-azimuth velocity dispersion, the inversion compromises the advantages of various azimuths. Considering the possible errors of velocity dispersion estimation in field-data processing, the inversion results also can achieve the better inversion precision. Therefore, the multi-azimuth reservoir parameters joint inversion in 3D two-phase cracked orthorhombic media based on BISQ mechanism has the good application prospect, and the algorithm provides a novel technique for the reservoir geophysical parameters inversion.更多还原