深部储层叠前地震反演方法研究

发布时间：2018-03-12 15:29

本文选题：深部储层　切入点：小角度反演　出处：《中国石油大学(华东)》2015年博士论文　论文类型：学位论文

【摘要】：随着勘探程度的深入,浅层的资源探明度逐渐增加,具有商业开采价值的油气藏越来越少。因此,勘探的目标逐渐向中深层转移,深部储层的预测是地球物理油气勘探领域的研究热点之一。基于地震振幅信息的反演方法是储层预测的常用方法,可靠的地震反演结果可以提高勘探成功率、降低投入成本,对油气藏勘探与储层评价具有重要意义。现阶段针对深部储层的地震反演方法较少,常规方法在深部储层勘探中的应用效果不理想。本文针对深部储层的资料特点,开展了叠前地震反演方法的研究。深部储层埋深大、地震数据偏移距(入射角)有限,而常规的叠前反演方法通常需要大偏移距(入射角)地震数据,因此不利于常规方法的应用。本文从Zoeppritz方程的近似方程出发,借鉴Smith和Gidlow两项方程理论推导了适用于小偏移距(入射角)的小角度近似方程。在此基础上,利用弹性参数之间的转换关系,进一步推导出包含弹性参数的小角度近似方程。基于贝叶斯反演理论,建立了小角度弹性阻抗反演方法。模型试算与实际应用结果表明,本方法可以在缺少大角度地震数据的情况下提供精确的反演结果。在反演的过程中,角度的误差经常被忽略,由此引起的问题也没有引起足够的重视。在叠前AVO/AVA反演的过程中,反射系数近似方程中的角度是入射角和透射角的平均角,而用于反演的叠前地震道集通常是基于入射角抽取的,这种角度的误差会给反演结果带来误差。深部储层的地震数据信噪比低、保真度差,误差问题更加严重。本文发展了角度优化的近似方程,解决了角度误差带来的反演问题。在此基础上,建立了叠前AVO反演方法,并与小角度反演方法相结合,进一步推导了角度优化的小角度近似方程,建立了相应的概率化反演方法。模型试算与实际应用结果表明,本方法可以提高反演结果的精确度,为储层预测和储层流体识别奠定良好的数据基础。深部地层受到埋深、地层压力、温度等因素的影响,储层的物性与浅层不同,使用常规的流体识别方法容易产生识别的假象。本文基于孔隙介质理论研究了常规流体指示因子的敏感性,提出使用孔隙流体模量作为深部储层的流体指示因子,并推导了包含孔隙流体模量的近似方程。以四元柯西分布为先验分布,并假设似然函数服从四元高斯分布,建立了基于贝叶斯反演理论的概率化反演方法。模型试算与实际应用效果表明,使用孔隙流体模量作为流体指示因子可以很好地识别孔隙流体性质。
[Abstract]:With the deepening of exploration degree, shallow resources proven degree increased gradually, with commercial exploitation value of oil and gas reservoir is less and less. Therefore, the exploration aim gradually to the deep, deep reservoir prediction is one of the hot research of geophysical exploration of oil and gas fields. The seismic inversion method based on amplitude information is a common method for reservoir prediction, seismic inversion results can improve the success rate of exploration, reduce costs, has important significance for oil and gas exploration and reservoir evaluation. Less seismic inversion method at present in the deep reservoir, the application effect of conventional method in deep reservoir exploration is not ideal the data of this paper. According to the characteristics of deep reservoir, carried out the research of prestack seismic inversion method. The deep reservoir buried depth, seismic data offset (incident angle), while the conventional prestack inversion methods usually require large offset ( Angle of incidence) seismic data, it is not conducive to the application of the conventional method. The approximate equation of Zoeppritz equation, using Smith and Gidlow two equation is deduced for small offset (incident angle) of the small angle approximation equation. On this basis, the relationship between the conversion elasticity parameters, deduced out the small angle contains elastic parameter equation. Based on Bayesian inversion theory, set up a small angle elastic impedance inversion method. The results of model test and practical application show that this method can provide accurate inversion result in the lack of wide angle seismic data situation. In the process of inversion, the angle error is often ignored, thus the problem is not caused enough attention. In the process of prestack AVO/AVA inversion, reflection coefficient approximate equation of the angle is the average angle of incidence angle and transmission angle, and for anti Do the pre stack seismic data is usually selected based on the incident angle, the angle error will bring the error. The inversion results of deep reservoir seismic data with low signal-to-noise ratio, poor fidelity, error problem is more serious. The approximate equation of angle optimization, solving the inverse problem of angle error brought. On this basis, the establishment of the prestack AVO inversion method, and combined inversion and small angle method, small angle optimization is further derived approximate equation, established a probabilistic inversion method. The corresponding results of model and practical application show that this method can improve the accuracy of the inversion results, lay good foundation for reservoir prediction and reservoir fluid identification. The deep stratum by depth, formation pressure, temperature and other factors, the reservoir and the shallow layer, using the fluid identification method of conventional prone recognition The illusion. In this paper, based on the theory of porous medium on the sensitivity of the conventional fluid indicator, put forward the use of pore fluid fluid modulus as the deep reservoir indicator, and the approximate equation of pore fluid containing modulus is derived. The distribution of the prior distribution to four yuan Cauchy, and assuming that the likelihood function subject to four yuan Gauss distribution, established probabilistic inversion method based on Bayesian inversion theory. That model calculation and practical application effect, the use of pore fluid modulus as fluid factors can well identify the pore fluid properties.

【学位授予单位】：中国石油大学(华东)
【学位级别】：博士
【学位授予年份】：2015
【分类号】：P631.4

【相似文献】