时移地震油藏参数的差异性反演研究及应用

发布时间：2018-04-25 10:52

本文选题：时移差异性反演 + 岩石油藏参数　；参考：《中国地质大学》2016年博士论文

【摘要】：经过几十年的勘探,陆上的大型构造油气藏都已基本被发现并进行了开发。近年来有很多老油田以现有的油气田为基础,对老油田进行二次勘探,期望通过研究油气田不同时期的地震响应,直接表征储层油藏性质的变化。地震反演技术一直是地震响应与储层参数之间最有效的桥梁,将地震反演技术应用到时移地震领域可以直接解决地震响应与油藏性质变化之间的关系。目前时移地震反演多采用对时移前后的地震数据分别进行反演,通过两次反演结果的差异得出储层参数的变化信息。而时移差异性反演可以直接对时移数据差异进行反演并获得储层参数的变化信息,与常规方法相比有着明显的计算效率优势,并可以简化反演流程。本文从赵凹油田的时移数据出发,主要开展对时移差异性AVO反演方法和时移差异性全波形反演方法的研究。重点解决以下几个问题：1)实际工区中储层的油藏参数与弹性参数的关系不明确,且时移地震资料的一致性较差,无法提供较准确的时移差异数据。2)目前的时移差异性AVO反演算法直接进行储层油藏参数反演时,反演的稳定性较差,反演的多解问题严重。3)全波形反演是近年来反演领域中获得高精度速度模型的有效方法之一,但在时移差异性反演领域尚未得到运用。针对目前时移反演中存在的问题,本论文从岩芯的岩石物理实验出发,完善工区内储层岩石弹性参数随油藏参数变化的规律,利用数值模拟对时移地震的波场特征进行讨论,分析有效压力及含水饱和度变化与地震响应差异之间的关系,建立储层油藏参数与弹性参数之间的联系,对时移AVO差异性反演中关键参数的敏感度进行讨论,发展了时移差异性全波形反演算法,为实际数据的时移差异性反演提供了理论基础。本文取得的主要成果有：1)对赵凹油田的典型岩芯进行岩石物理实验,测试结果表明岩芯样品的孔隙度对储层的弹性参数变化规律影响很大,当储层岩石的孔隙度大于15%时,水驱情况下的地震可检测情况较为理想。2)通过正演模拟测试储层参数变化对地震响应的影响,储层厚度对波阻抗界面的地震响应影响较大,在薄互层模型中体现的尤为明显,当薄互层模型的储层等效厚度小于四分之一波长时,储层内部的波阻抗界面将无法识别；储层含水饱和度的变化会影响储层的弹性参数,因此也对波阻抗界面的地震响应产生影响；储层中的有效压力变化引起的地震响应差异的变化规律不呈线性,当有效压力小于20 MPa时振幅能量的相对变化较大3)通过对时移地震数据的分析,在保证两期数据相对一致性的原则下,以最大限度地减少或消除由于非油藏因素引起的地震资料的不一致性,保留和突出油藏变化引起的差异为中心任务,完成对非重复性采集时移地震数据的一致性处理,为后续的时移差异性反演研究提供数据基础。4)理论模型合成记录的时移AVO差异性反演试算结果表明,在无噪声条件下,采用阻尼最小二乘反演算法进行时移AVO差异性反演可以较为稳定地反演出地层纵波速度、横波速度和密度的时移相对变化以及储层含水饱和度与有效压力的变化。但该方法对仍受到Aki-Richards近似式的局限,地震数据的入射角范围会对反演结果的准确性产生影响,当地震数据的入射角范围在0-30°内时,反演结果的相对误差随着入射角度的增大而减小。同时,地震资料的信噪比也会影响时移AVO差异性反演结果的准确性,当地震资料的噪音强度超过7%会导致反演结果的相对误差变大,导致反演稳定性变差,且当地下构造的复杂程度越高,噪音引起的反演误差越大。5)对Marmousi模型正演模拟记录进行时移差异性全波形反演试算结果表明,该方法对前期资料采集、常规建模、反演参数设置等方面的要求较高。其中资料采集观测系统将直接影响反演结果的分辨率和准确性,同时排列长度较长的观测系统在反演迭代计算时收敛速度较慢；初始模型对反演结果的分辨率及精度将会产生严重的影响,且当反演初始模型的精度较低时,目标函数的迭代收敛速度明显降低；多尺度反演采用的频率组内的频率个数会直接影响反演的计算量,且增加反演的频率数目可以有效提高差异性波形反演结果的分辨率和准确性；当反演数据中的包含丰富的低频数据时,在重建异常波数域频谱有效影响时,使其影响范围变宽,反演数据中的低频数据能使反演结果具有更好的保真度。本论文的创新之处主要表现在以下两个方面：1)改善了时移AVO差异性反演的方法,建立反射系数差异与含水饱和度变化、有效压力变化之间的非线性关系,利用Gardner公式约束反演过程以提高反演的稳定性,并对反演参数的敏感度进行了量化分析。2)发展时移地震差异性全波形反演方法,设计并实现频率域多尺度时移差异性全波形反演的反演关键框架流程,探讨了时移差异性全波形反演参数与反演结果的内在联系,并实现了在复杂模型中的时移差异性反演试算。
[Abstract]:After several decades of exploration, the large tectonic oil and gas reservoirs on land have been basically discovered and developed. In recent years, many old oilfields are based on the existing oil and gas fields, and have carried out two exploration to the old oil fields. They are expected to directly characterize the reservoir properties by studying the seismic response of different periods of oil and gas fields. It has been the most effective bridge between seismic response and reservoir parameters. The application of seismic inversion technology to the domain of time-lapse seismic can directly solve the relationship between seismic response and the change of reservoir properties. At present, the seismic inversion of the time shift is mostly used to inverse the seismic data before and after the time shift, and the difference between the two inversion results is obtained. The time shift difference inversion can directly inverse the time shift data and obtain the change information of the reservoir parameters. Compared with the conventional method, it has obvious computational efficiency advantage and can simplify the inversion process. From the time shift of Zhao depression oil field, this paper mainly carries out the inversion of time shift difference AVO. The research on the method and the time shift difference full waveform inversion method mainly solves the following problems: 1) the relationship between the reservoir parameters and the elastic parameters of the reservoir is not clear, and the consistency of the time shift seismic data is poor, and the accurate time shift difference data.2 can not be provided. The present time shift difference AVO inversion algorithm is directly stored in the reservoir. When the reservoir parameters are retrieved, the stability of the inversion is poor, the inverse problem is serious.3). The full waveform inversion is one of the effective methods to obtain the high precision velocity model in the field of inversion in recent years. But it has not been used in the domain of time-lapse inversion. On the basis of the experiment, the variation of reservoir rock elastic parameters with reservoir parameters is perfected, the wave field characteristics of time-lapse earthquake are discussed by numerical simulation, the relationship between effective pressure and water saturation change and the difference of seismic response is analyzed, and the relation between reservoir parameters and elastic parameters is established, and the difference of time shift AVO is made. The sensitivity of the key parameters in the inversion is discussed, and the time shift difference full waveform inversion algorithm is developed to provide the theoretical basis for the time shift difference inversion of the actual data. The main achievements of this paper are as follows: 1) the rock physics experiments are carried out on the typical core of the Zhao depression oil field, and the test results show that the porosity of the core samples is to the reservoir. The variation law of elastic parameters has great influence. When the porosity of the reservoir rock is greater than 15%, the seismic detection in water flooding is more ideal.2). Through the forward modeling, the effect of reservoir parameters on the seismic response is tested. The reservoir thickness has a great effect on the seismic response of the wave impedance interface, especially in the thin interbed model. When the reservoir equivalent thickness of the thin interbed model is less than four minutes, the internal wave impedance interface in the reservoir can not be identified; the change of the reservoir water saturation will affect the elastic parameters of the reservoir, so it also affects the seismic response of the wave impedance interface; the variation of the seismic response caused by the effective pressure change in the reservoir is changed. The law is not linear. When the effective pressure is less than 20 MPa, the relative change of amplitude energy is 3). Through the analysis of the time shift seismic data, under the principle of guaranteeing the relative consistency of the two periods, it can reduce or eliminate the inconsistency of seismic data caused by non reservoir factors to the maximum extent, and retain and highlight the reservoir changes. The difference is the central task, complete the consistency of seismic data for non repetitive acquisition, and provide data base.4 for the subsequent time shift difference inversion research. The time shift AVO difference inversion test results of the theoretical model of the theoretical model show that the time shift AVO difference is carried out by the damping least square inversion algorithm under the condition of no noise. The inversion of the velocity of the formation, the relative change of the time shift of the velocity and density of the shear wave and the change of the reservoir water saturation and the effective pressure can be more stable, but the method is still limited by the Aki-Richards approximation. The incidence of the incident angle of the seismic data will affect the accuracy of the inversion result, when the seismic data is entered. When the angle range is within 0-30 degrees, the relative error of the inversion results decreases with the increase of the incidence angle. At the same time, the signal-to-noise ratio of seismic data also affects the accuracy of the result of the time shift AVO differential inversion. When the noise intensity of the seismic data exceeds 7%, the relative error of the inversion result becomes larger and the inversion stability becomes worse, and it is under the ground. The higher the complexity of the structure, the greater the error caused by the noise, the greater the inversion error.5). The test results of the time shift difference full waveform inversion of the Marmousi model forward simulation record show that the method has higher requirements for the previous data collection, the conventional modeling and the inversion parameter setting. The data acquisition and observation system will directly affect the inversion results. The resolution and accuracy of the observation system with long length are slower in the iterative calculation, and the initial model will have a serious influence on the resolution and accuracy of the inversion results. And the iterative convergence rate of the target function is obviously reduced when the initial model's precision is low, and the multi-scale inversion is used. The number of frequency within the frequency group will directly affect the calculation of the inversion, and the increase of the number of inversions can effectively improve the resolution and accuracy of the differential waveform inversion results. When the inversion data contains rich low frequency data, the influence range is widened and the inversion number is changed when the frequency spectrum of the anomalous wave number domain is rebuilt. The low frequency data in this paper can make the inversion result better. The innovation of this paper is mainly shown in the following two aspects: 1) improving the method of time shift AVO differential inversion, establishing the non linear relation between the difference of the reflection coefficient and the change of water saturation, the change of effective pressure, and using the Gardner formula to restrain the inversion process. In order to improve the stability of the inversion, and to quantify the sensitivity of the inversion parameters,.2) is used to develop the time-lapse seismic difference full waveform inversion method. The key frame flow of the multi scale time shift difference full waveform inversion is designed and realized, and the internal relation between the inverse parameters of the time shift difference and the inversion results is discussed. And the time lapse difference inversion calculation in complex model is realized.

【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】：P631.4;P618.13

【相似文献】