HHT算法改进及其在储层预测中的应用研究

发布时间：2018-01-24 15:23

本文关键词： 时频分析 HHT 地震信号处理瞬时属性储层预测　出处：《中国地质大学》2016年博士论文　论文类型：学位论文

【摘要】：随着勘探程度的不断提高,油气勘探的对象日趋复杂,常规的资料解释和储层预测技术难以适应勘探、开发的需要。通过地震瞬时属性(主要有瞬时频率、瞬时振幅、瞬时相位等)能够准确反映出储层的几何形态与物性特征,对储层进行有效地识别。然而,由于地震信号是一种非线性、非平稳的复杂信号,而在信号处理过程中所使用的传统方法大多是基于平稳、线性信号的处理方法,所以这些传统方法的效果较差,难以准确地反映地层信息。希尔伯特—黄变换(Hilbert-Huang Transform, HHT)是一种处理非线性非平稳信号的新方法。该方法所得结果可以更好地反映地震信号自身的特点,这使得其时频分辨能力大大提高。但是,常规HHT方法在储层预测领域的应用中存在有模态混叠、端点效应与瞬时频率的误差问题。这些问题会对所求地震瞬时属性产生严重的影响,导致瞬时属性剖面的畸变和错位等现象,给解释工作带来极大的干扰和“陷阱”。因此,本文针对常规HHT方法中出现的模态混叠、端点效应与瞬时频率误差的现象进行了全面的分析,对这些问题产生的原因、对地震瞬时属性的影响以及前人的研究成果展开了系统地研究,提出了基于AR模型预测的CEEMD方法,并与归一化Hilbert变换(NHT)方法相结合,形成了一套完整的改进后的HHT方法。通过模型正演与实例分析对该方法的有效性、准确性与分辨率进行了验证,证明了该方法能够有效地抑制模态混叠、端点效应与瞬时频率的误差对结果带来的影响,所求瞬时属性更加精确有效,其分辨率与可信度也大大提高。并通过结合测井资料对合成地震记录进行瞬时属性分析,研究了瞬时属性与储层几何特征、物性特征之间的关系,对利用改进后的HHT方法在储层中的应用进行了详细探索。利用改进后的HHT进行储层预测,本文得到了如下几点主要认识：(1)基于AR模型的CEEMD方法能够有效抑制常规EMD方法中存在的模态混叠与端点效应问题,NHT方法能够明显降低常规Hilbert变换中存在的瞬时频率误差问题,改进后的HHT方法具有更高的分辨率。(2)得出了地震瞬时属性与储层几何形态、物性特征之间的关系：地震信号瞬时振幅横向分辨率更高,能够对地层的形态与连续性进行准确描述。该属性也与砂体发育相对应,能够反映储层的厚度、孔隙度与岩性差异；瞬时频率能够对地层速度变化进行体现,可以反映储层的厚度与流体性质；瞬时相位可以对地层上下界面进行准确定位,反映出储层的横向连续性。(3)油气富集区往往具有强振幅、低频率与相位局部异常的特点。通过瞬时振幅与瞬时频率属性相结合,在纵向上能够很好地反映地层的孔隙度,泥质含量与地震波传播速度变化,区分砂岩层与泥岩层；利用低频率、强振幅的特点,可以对储层进行有效地识别；瞬时相位属性能够体现出地层物性的变化,瞬时相位在储层上界面出现极性迅速大幅的反转,说明该层段富含油气,能够有效反映储层的含气性特点；瞬时斜率能够对储层的边界、厚度、埋深等几何特征进行更为准确地定位。结合地震、测井与地质资料对工区进行多属性综合解释,能够更加准确有效地对储层进行识别与预测。本文在研究的过程中做出了如下的创新：(1)提出基于AR模型预测的CEEMD分解方法,并采用NHT对分解结果求取瞬时参数,形成了一套完整的改进后的HHT方法。减少了常规方法中因模态混叠、端点效应和瞬时参数误差对结果的影响。(2)发展了改进后的HHT方法在储层预测中的应用。通过对模型的正演与实例分析,得出了地震瞬时属性与储层几何特征和物性特征的关系。结合测井、地质资料进行综合分析,提高了对储层特征的响应,揭示出了更多的细节信息。综上所述,本文针对HHT方法在储层预测领域中的不足之处,对常规HHT方法进行了改进,并用改进后的方法提取地震信号瞬时属性；研究了这些参数与储层几何特征和物性特征之间的关系；最后利用地震瞬时属性结合地震、测井和地质资料,有效地将该方法应用于储层预测的工作中。
[Abstract]:With the continuous improvement of exploration targets for oil and gas exploration, more and more complex, it is difficult to adapt to the exploration of conventional data interpretation and reservoir prediction technology development needs. Through the seismic instantaneous attributes (including instantaneous frequency, instantaneous amplitude, instantaneous phase) can accurately reflect the geometry and reservoir properties, are effective recognition of the reservoir. However, due to the seismic signal is a complex nonlinear and non-stationary signal, while the traditional methods used in the signal processing is mostly based on the stable linear signal processing method, so the traditional method of effect is poor, it is difficult to reflect the formation information accurately. (Hilbert Huang Bianhuan Hilbert-Huang Transform, HHT) is a new method for processing nonlinear and non stationary signal. The result can better reflect the characteristics of the seismic signal itself, which makes the time-frequency Resolution is greatly improved. However, the conventional HHT method to predict the application in modal aliasing in reservoir, error of end effect and instantaneous frequency. These problems will have a serious impact on the seismic instantaneous attributes, resulting in distortion and dislocation phenomena such as instantaneous attribute profile, to explain the work of great the interference and "trap". Therefore, this article aims at the modal aliasing in the conventional HHT method, the error of end effects and instantaneous frequency phenomenon is analyzed. The causes of these problems, the effect of the seismic instantaneous attributes as well as the results of previous studies have been systematically studied, we propose a CEEMD method AR based on model prediction, and transform and normalized Hilbert (NHT) method are combined to form a complete set of the improved HHT method. The forward analysis of the effectiveness of the method and examples of model, The accuracy and resolution is verified that this method can effectively suppress the mode aliasing error, end effect and instantaneous frequency of the result, the instantaneous attributes is more accurate and effective, its resolution and reliability is greatly improved. And through the instantaneous attribute analysis of synthetic seismogram with well data measurement, research the geometric features of instantaneous attributes and reservoir, the relationship between physical characteristics, a detailed exploration of application using the improved HHT method in the reservoir. The reservoir prediction using improved HHT, this paper obtained the main conclusions are as follows: (1) CEEMD AR method can mode effective suppression of the conventional EMD method of aliasing and endpoint effect based on NHT method can significantly reduce the problem of instantaneous frequency error of conventional Hilbert transform, the HHT method is more improved High resolution. (2) the instantaneous seismic attributes and reservoir geometry, the relationship between the physical characteristics of the instantaneous seismic signal amplitude higher horizontal resolution, to the formation of the shape and continuity of the accurate description. The property and development of sand body should be relative, can reflect the thickness of the reservoir. The porosity and lithology difference; instantaneous frequency to the velocity variation of reflected, and can reflect the thickness of reservoir fluid properties; instantaneous phase of formation on the interface for accurate positioning, reflecting the lateral continuity of the reservoir. (3) oil and gas enrichment zone often has high amplitude, low frequency and phase of the local characteristics abnormal. By combining the instantaneous amplitude and instantaneous frequency characteristics, can well reflect the formation of porosity in the vertical, clay content and the propagation of seismic wave velocity changes, to distinguish between sand formation and clay layer; the use of low-frequency , strongamplitude, can effectively identify the reservoir properties; instantaneous phase can reflect the change of formation of the instantaneous phase in the reservoir on the interface of polar fast and dramatic reversal, indicating that the layers rich in oil and gas, which can effectively reflect the reservoir gas bearing characteristics; instantaneous slope to the reservoir boundary, thickness, depth and other geometric features more accurately. Combined with seismic, logging and geological data of the multiple attribute comprehensive interpretation of the work area, can be more accurate and efficient identification and prediction of the reservoir. This paper has the following innovations in the research process: (1) the decomposition method of AR model based on CEEMD, and using NHT to calculate the instantaneous parameters of the decomposition results, form a complete set of the improved HHT method. The decrease in the conventional method for modal aliasing, end effect and instantaneous parameter error on node The effects of (2). The application and development of improved HHT method in reservoir prediction. Through the example analysis and the forward model, the relation between instantaneous seismic attributes and reservoir geometry and physical characteristics. Combined with logging, geological data analysis, improve the response of the reservoir the characteristics, reveals more details. In summary, this paper HHT method in reservoir prediction deficiencies in the field of conventional HHT method was improved, and the extraction of seismic signal instantaneous attributes with the improved method; to study the relationship between these parameters and reservoir geometry and physical characteristics; by using seismic instantaneous attributes combined with seismic, logging and geological data, effectively applied to the reservoir prediction work.

【学位授予单位】：中国地质大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：P618.13

【相似文献】