重力及张量数据联合反演研究及在Vinton Dome地区的应用

发布时间：2018-01-23 05:58

本文关键词： 全张量重力梯度误差滤除正演快速算法深度加权联合反演交叉梯度联合反演　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着重力勘探技术的发展,特别是重力梯度测量的广泛应用,重力数据处理及解释技术也亟待提高。重力梯度测量是一种相对较新的技术,具有探测信息量大、测量精度高、抗干扰性强等优点,近年来在地球物理勘探中得到广泛重视。重力梯度数据重力位的二阶导,其高频信号正好与重力数据低频信号互补,所以重力与重力梯度数据联合反演具有重大意义。重力梯度的五个分量数据分别包含不同的地质体异常信息,联合重力梯度不同分量进行反演能够更好地反应地下异常体的形态分布。因此研究重力及重力张量数据联合反演可以达到信息互补,减少反演结果多解性的目的。首先本文建立模型,根据正演公式计算出全张量重力梯度数据,并根据实测噪声特征人为的加入了随机噪声即高斯白噪声,利用韦尔奇功率谱密度图对加噪后三个水平分量(Vxx1,Vxy1,Vyy1)进行分析,根据分析结果设计合适的Butterworth和Gaussian两种低通滤波器并对比滤波前后的效果。为了选择参与联合反演的物理量,本文对重力及张量数据进行对比分析。在频率域内推导出重力梯度张量之间的关系,对比分析发现:重力数据Vz在低频部分的能量高于重力梯度数据分量Vzz,而在高频部分的能量均低于重力梯度数据Vzz,说明重力数据Vz能比较好的反应深部信息,重力梯度Vzz能更好的反应浅部异常。用张量数据奇异值分析了各分量及组合分量间的相关性,证明在所有分量中Vzz包含信息最为丰富,Vxy信息量最少。联合反演大大增加了反演数据量,其中反演需要的各分量灵敏度矩阵也随之增加,这就需要更多地计算时间和储存空间。为了解决这个问题,本文在前人的研究基础上设计了新的快速计算灵敏度矩阵的方法,减少了内存需求,缩短了计算时间。本文重点研究了重力反演理论,并建立两个模型对重力及梯度数据进行三维反演,为了抵消“趋肤效应”对反演结果的影响,本文使用了两种深度加权函数,并选取效果更好的一种进行反演。为了使反演结果更光滑,本文在反演目标函数中加入了粗糙度因子。本文建立了两个异常体模型进行重力及张量数据的单独反演。从反演结果上看:在垂向上Vz和Vzz能够较好地反应地质体深部信息,但两者深部反演发散,出现拖尾现象。在横向上Vzz分辨率高于Vz,对于多个地质体的识别能力较高。全张量重力梯度分量Vxx和Vxz反演结果在X方向上收敛较好,对地质体南北走向具有很高的识别能力;Vyy和Vyz在Y方向上收敛较好,对地下异常体东西走向边界识别能力强;Vxy数据包含更多的地下异常体的角点信息。总体来看,重力及重力梯度分量的反演结果都能基本表现地质形态。多分量联合反演较之单分量反演结果有明显改观,主要表现在:反演密度更接近真实值,抗噪性增强,横向和纵向上的分辨率都有所提高,收敛性更好,反映地质体的边界更明显,呈现出良好的结果。本文还引入了交叉梯度函数进行重力梯度数据间的联合反演,并将该方法对文顿盐丘实测数据进行反演,反演结果证明了交叉梯度联合反演的优越性和实用性。
[Abstract]:With the development of gravity exploration technology, especially the wide application of gravity gradient measurement, gravity data processing and interpretation technology also needs to be improved. The gravity gradient measurement is a relatively new technology, can detect a large amount of information, high accuracy, the advantages of strong anti-interference, widely attention in geophysical exploration. Two order derivative of gravity gradient data of gravity potential, the high frequency signal is low frequency signal and gravity data is complementary, so the joint inversion of gravity and gravity gradient data is of great significance. The five components of the gravity gradient data contains geological anomaly information respectively, combined with different components of gravity gradient distribution of inversion can better response the underground abnormal body. So the study on joint inversion of gravity and gravity tensor data can achieve complementary information, reduce the inversion results of multiple solutions to the first. According to the model, forward the formula to calculate the gravity gradient tensor data, and according to the measured noise characteristics of human with random noise is Gauss white noise, using the Welch power spectral density map of noise after three horizontal components (Vxx1, Vxy1, Vyy1) are analyzed, according to the analysis results of two kinds of low effect pass filter and compare before and after filtering the appropriate Butterworth and Gaussian. In order to choose to participate in the joint inversion of physical quantity, based on gravity and tensor data comparative analysis. Derive the relationship between the gravity gradient tensor in the frequency domain, the results show that the gravity data Vz in the low-frequency part of the energy is higher than the gravity gradient data component Vzz. In the high-frequency energy was lower than the gravity gradient data of Vzz, shows that the deep gravity data Vz can better response, response of shallow gravity gradient Vzz can better use a exception. The amount of data to analyze the correlation between the singular value of each component and combination between components, that in all components of Vzz contains the most abundant information, the Vxy information is the least. The joint inversion has greatly increased the amount of data in which each component of inversion, the sensitivity matrix inversion needs also increases, which requires more computation time and storage space. In order to solve this problem, this paper designed a new method for the fast calculation of sensitivity matrix on the basis of previous studies, reduced memory requirements, shorten the computing time. This paper focuses on the theory of gravity inversion, and establish two models of 3D inversion of gravity and gradient data, in order to offset the "skin effect" on the inversion results, we use two kinds of depth weighting function, and select a better inversion effect. In order to make the inversion results more smooth, the inversion objective function in Join the roughness factor. This paper established two abnormal body model separately and inversion of gravity tensor data. From the point of view: the inversion results can better reflect the deep geological information to Vz and Vzz in the vertical, but the two deep inversion divergence, trailing phenomenon. In the horizontal resolution of Vzz is higher than Vz for the higher, multiple geological body recognition. Gravity gradient tensor components Vxx and Vxz inversion results in the X direction of convergence is good, has the very high recognition ability of geological body to North-South Vyy and Vyz; good convergence in the Y direction, the underground abnormal bodies west boundary recognition ability; corner point the Vxy data contains information of underground abnormal body more. Overall, inversion of gravity and gravity gradient component were basic geological form. Multi-component inversion than single component inversion results have significantly changed, mainly in: Play the density is more close to the real value, enhances the noise immunity, the horizontal and vertical resolution are improved, better convergence, which reflect the geological boundary is more obvious, showing good results. This paper also introduces the cross gradient function of joint inversion of gravity gradient data, and the method for inversion of measured data, salt domes, the inversion results show the superiority and practicability of cross gradient joint inversion.

【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P631.1

【相似文献】