基于交叉梯度约束的重力、磁法和电阻率法二维联合反演研究

发布时间：2018-05-12 19:01

本文选题：剩余密度 + 磁化率　；参考：《中国地质大学(北京)》2015年硕士论文

【摘要】：在实际工区中同时开展重力、磁法和电阻率测深法工作是目前很普遍的做法,在推断地下信息时,不同地球物理方法之间可以相互补充,来得到更加可靠的地质-地球物理模型。考虑到单一方法各自的不足和局限性,以及反演本身固有的多解性,不同方法单一反演得到的地质模型往往差异较大。我们认为,当不同物性具有结构分布上的一致性时,如本文的剩余密度、磁化率和电阻率具有相同的物性结构分布时,把由不同物性梯度算子构成的交叉梯度引入到目标函数中,进行联合反演,理论上可以得到一个更加可靠的结果,该算法使反演沿着剩余密度、磁化率和电阻率结构分布一致的方向进行下去。文中采用了解析法来进行重磁的正演计算,以及有限单元法来进行电阻率法的正演计算。重力、磁法和电阻率法反演采用的都是正则化光滑约束反演,同时,因为需要对剩余密度和磁化率进行物性正约束,在正则化光滑约束的基础上,引入了对数障碍项,来避免反演迭代中剩余密度和磁化率出现负值。文中通过多个理论模型,包括完全一致模型和不完全一致模型,对该算法进行了测试,模型算例表明,在物性数值和模型形态上,交叉梯度联合反演都比单一方法反演得到的结果更加接近真实理论模型,从而证明了交叉梯度联合反演的正确性和有效性。
[Abstract]:It is a common practice to carry out gravity, magnetic and resistivity sounding methods simultaneously in actual working areas. In the process of extrapolating underground information, different geophysical methods can complement each other. To get a more reliable geological-geophysical model. Considering the shortcomings and limitations of the single method and the inherent multiplicity of the inversion itself, the geological models obtained by the single inversion with different methods are often quite different. We think that when different physical properties have the same structure distribution, for example, the residual density, magnetic susceptibility and resistivity have the same physical structure distribution. The cross gradient composed of different physical gradient operators is introduced into the objective function for joint inversion, and a more reliable result can be obtained theoretically. The algorithm makes the inversion along the residual density. The direction of the distribution of magnetic susceptibility and resistivity structure is consistent. The forward calculation of gravity and magnetic field and the forward calculation of resistivity method by finite element method are carried out by analytical method. The inversion of gravity, magnetic method and resistivity method are all regularized smooth constraint inversion. At the same time, the logarithmic barrier term is introduced on the basis of regularization smooth constraint because of the physical positive constraint of residual density and magnetic susceptibility. To avoid negative values of residual density and magnetic susceptibility in inversion iteration. In this paper, the algorithm is tested by several theoretical models, including completely consistent model and incompletely consistent model. The model examples show that, in terms of physical properties and model shape, The results of cross gradient joint inversion are closer to the real theoretical model than that of the single method, which proves the correctness and validity of the cross gradient joint inversion.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P631

【参考文献】