位场数据高精度处理方法的研究与应用

发布时间：2018-07-09 10:18

本文选题：位场转换 + 位场分离　；参考：《吉林大学》2016年博士论文

【摘要】：随着重磁勘探仪器探测精度与效率的提高,一些传统的位场数据处理方法因计算精度低、分析条件复杂等缺陷而难以满足当今位场勘探精细解释的需求。鉴于此,改进和开发高精度、高效率的位场数据处理方法具有重要意义。本文在位场数据转换与分离以及边界检测的反演工作方面提出了一些改进措施,取得了较好的效果。位场异常的转换与分离是最常规的异常处理解析方法,异常转换的精度以及异常分离的是否彻底直接影响一些正反演方法(如欧拉反褶积、界面异常反演等)的准确性和分辨率,本文以迭代补偿原理为基础,提出了用于导数换算和向下延拓的近似算子迭代法以及位场分离的波数域迭代法。近似算子迭代法是针对位场导数换算和向下延拓过程中存在的不稳定性问题提出的。该方法利用位场的泰勒级数、Laplace方程和不同阶次导数的转换关系,推导了导数换算和向下延拓的近似算子,将上述近似算子置入波数域迭代中,实现了位场导数换算和向下延拓的稳定计算。在分析近似算子迭代法滤波特性和收敛性的过程中,笔者发现可以利用邻次迭代结果的相关度确定最佳的迭代次数区间。含噪声模型试验表明,近似算子迭代法相比于其他方法具有较强的计算稳定性和较高的计算精度;相关系数与迭代次数的关系曲线可以快速提供最佳的迭代次数,保证了近似算子迭代法较高的运算效率和精度。同样以迭代补偿思想为前提,本文设计了一种幅值较低的区域场分离算子,给出了位场分离的波数域迭代法。该法能够将局部异常频谱中属于区域场的成分剥离出来,返还给区域异常,进而弥补区域异常的信号损失,最终达到改善异常分离效果的目的。模型试验表明,相比于常规向上延拓的异常分离方法,本文方法提取的目标场与真实值更加接近。利用位场资料反演地质体的边界位置可以为构造解释提供重、磁场划分依据。现有的边界检测方法大都与位场异常的梯度变化相关,因此这些方法常常受高频干扰影响,致使识别出的边界信息存在着模糊、缺失或虚假等缺点。针对于此,本文提出了基于差分求导的伪总梯度余弦值法、加权小子域滤波和结合增强型均衡滤波器的欧拉反褶积法三种边界检测技术。这三种方法有各自的特点,它们的处理结果可以相互弥补、验证,从而提高边界识别的精度和分辨率。基于差分求导的伪总梯度余弦值法是考虑了边界点异常的方向信息后,采用子域均值和差分方式计算梯度,以垂向梯度与伪总梯度(水平梯度的幅值被压缩)矢量夹角的余弦值作为滤波输出,结果中的极大值对应边界位置。该方法在计算过程中对边界位置的异常特征进行了两次紧缩,突出了边界与非边界异常的差异性。模型试验结果表明该法具有较好的计算稳定性和边界位置识别效果。加权小子域滤波技术是以直立岩脉重力异常、数值差分及小子域滤波为基础提出的,滤波结果以平面等值线中被紧缩的梯级带作为地质体边界的识别标志,以正、负圈闭异常反映局部地质体剩余密度的分析模式,在空间域内实现了对三方向小子域滤波的加权整合。模型试验表明,加权小子域滤波不但对地质体边界有较高的检测精度,而且能够清晰地反映出区域与局部场的圈闭异常关系(地质体的剩余密度关系)。针对常规Euler方程分散解较多,解的精度受构造指数影响较大的问题,笔者以一种新的增强型均衡滤波器为基础,推导了新的Euler方程。新方程不需要预先给定构造指数和背景场就可以计算出边界的平面位置和深度。模型试验表明,相比于常规Euler方程,新方程对边界的平面位置和深度的反演精度更高。以上边界检测方法中,基于差分求导的伪总梯度余弦值法能够清晰的检测出梯级带特征的构造边界;加权小子域滤波不仅弥补了前者检测非梯级带特征边界能力的不足,更能够反映出边界两侧地质体的异常强弱关系;结合增强型均衡滤波器的欧拉法反褶积法可以很好的填补前两者无法反演边界深度信息的空白。这三种方法配套使用可以更好的反映构造边界的三维分布特征。将以上方法应用到松辽盆地镇赉地区重磁场数据处理中,取得了一些成果。依据位场转换与分离方法的处理结果,结合收集的地质、物性、钻井资料,分析了研究区重、磁场、剩余异常、垂向导数及中、浅部地层的视物性分布特征;根据位场边界检测方法的处理结果,在研究区内划分出11条主要断裂,并将研究区划分为3个二级构造单元和9个次级构造单元;在此基础上,以岩石物性、钻井资料和地震剖面为约束条件,反演了研究区中、新生界底界面深度,预测了4处凹陷为油气勘探的有利远景区。
[Abstract]:With the improvement of the detection precision and efficiency of the gravity and magnetic exploration instruments, some traditional bit field data processing methods are difficult to meet the requirements of the fine interpretation of the potential field exploration because of the low calculation precision and complicated analysis conditions. In view of this, the improvement and development of high precision and high efficiency bit field data processing method is of great significance. Some improvement measures are proposed in the field of data conversion and separation and the inversion of boundary detection. The transformation and separation of the bit field anomaly is the most conventional method of abnormity processing, the accuracy of the anomaly conversion and the complete direct reaction of the anomaly separation (Euler deconvolution, interface) The accuracy and resolution of the anomalous inversion, based on the iterative compensation principle, is based on the approximate operator iteration method for derivative and downward continuation and the number domain iteration method for the position field separation. The approximate operator iteration method is proposed for the instability problems in the conversion and downward continuation of the potential field derivative. The approximate operator of the derivative conversion and downward continuation is derived by using the Taylor series of the potential field, the Laplace equation and the derivative of different order derivatives. The approximate operator is placed in the wave number domain and the stable calculation of the conversion of the derivative and the downward continuation of the potential field is realized. The filtering characteristics and the convergence of the approximate operator iteration method are analyzed. In the process, the author finds that the best iterative interval can be determined by the correlation degree of the adjacent iteration results. The noise model test shows that the approximate operator iteration method has stronger computational stability and higher calculation precision than the other methods, and the relation curve of the correlation coefficient and the number of iterations can provide the best iterative times. As a result of the iterative compensation idea, a kind of regional field separation operator with low amplitude is designed, and the iterative method of wave number domain for the separation of bit field is given. This method can remove the component of local anomaly spectrum and return to the regional anomaly. In order to make up the signal loss of regional anomalies and finally achieve the purpose of improving the abnormal separation effect, the model test shows that the target field extracted by this method is closer to the true value compared with the abnormal separation method of conventional upward extension. The inversion of the boundary position of geological body by using the bit field data can provide a heavy load for the structural interpretation. Most of the existing boundary detection methods are related to the gradient change of the anomaly. Therefore, these methods are often affected by high frequency interference, resulting in the shortcomings of blurred, missing or false boundary information identified. Euler deconvolution (eulla deconvolution) of an enhanced balanced filter (Euler) three boundary detection techniques. These three methods have their own characteristics. Their processing results can be made up to each other to improve the accuracy and resolution of the boundary recognition. The pseudo total gradient cosine value method based on the differential derivative is used to consider the direction information of the boundary point anomaly. The gradient and the cosine values of the vertical gradient and the pseudo total gradient (the amplitude of the horizontal gradient are compressed) are used as the filtering output. The maximum value of the results corresponds to the boundary position. In the calculation process, the anomalous characteristics of the boundary position are two times tightened and the boundary and non boundary anomalies are highlighted. The model test results show that the method has good calculation stability and boundary location recognition effect. The weighted subdomain filtering technique is based on the gravity anomaly of the erect vein, the numerical difference and the small subdomain filtering, and the filtering results are taken as the identification mark of the geological body boundary with the tight ladder zone in the plane contour line. The weighted integration of the three direction subdomain filtering is realized in the spatial domain by the positive and negative traps that reflect the residual density of the local geological bodies. The model test shows that the weighted subdomain filtering not only has a high detection precision on the boundary of the geological body, but also can clearly reflect the anomaly relationship between the region and the local field. In view of the residual density relation of the geological body. In view of the large dispersion solution of the conventional Euler equation, the accuracy of the solution is greatly influenced by the structural exponent. Based on a new enhanced equilibrium filter, the author derives the new Euler equation. The new equation can calculate the plane position of the boundary without the pre given structure index and the background field. The model test shows that the new equation has a higher inversion accuracy to the plane position and depth of the boundary than the conventional Euler equation. In the above boundary detection method, the pseudo total gradient cosine value method based on the difference derivation can clearly detect the structural boundary of the cascade characteristics, and the weighted subdomain filtering not only makes up for the former to detect the non ladder. The inadequacy of the characteristic boundary capacity of the band is more able to reflect the anomalous strong and weak relationship between the geological bodies on both sides of the boundary, and the Eulerian deconvolution method combined with the enhanced equilibrium filter can well fill the blank of the inversing boundary depth information of the former two. The three methods can be used to better reflect the three-dimensional distribution of the tectonic boundary. The above method has been applied to the data processing of heavy magnetic field in the Zhenlai area of Songliao Basin, and some results have been obtained. According to the results of the processing of the position field conversion and separation method, combined with the geological, physical, and drilling data collected, the distribution characteristics of the visual property of the study area, the magnetic field, the residual anomaly, the vertical derivative and the shallow strata are analyzed. The results of the field boundary detection method are divided into 11 main faults in the study area, and the study area is divided into 3 two stage structural units and 9 secondary tectonic units. On this basis, the rock physical property, drilling data and seismic profiles are restricted, and the bottom interface depth of the Cenozoic is retrieved in the study area, and 4 depression is predicted to be oil. A favorable scenic spot for gas exploration.
【学位授予单位】：吉林大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：P631

【相似文献】