基于正则化下延方法的视密度全空间反演

发布时间：2019-04-02 18:32

【摘要】：重力反演就是通过已知重力场的分布,使用某种手段推算出地下的密度分布,它是地球物理勘探的重要手段之一。这种方法的基本思想就是对地下待反演地质体进行网格剖分,通过不停的反演迭代直至达到精度要求,来计算出所有剖分网格单元的视密度。但是,这种计算方法的主要问题就是反演多解性严重:第一、观测重力异常是地下的空间体积效应;第二、采集的数据是离散的而不是连续的。如何减少多解性,提高反演成像精度,成为解决重力反演的关键性问题。针对多解性问题,采用的是有效地提取更多有利的先验信息,增加反演数据体的数量来提高视密度反演成像的精度。如何有效地增加反演数据体的数据,成为本文的基本出发点,因此采用基于深度变化的正则化下延方法,得到不同深度的重力场,即三维重力数据体,然后利用三维空间数据体即不同深度的位场下延数据体进行反演,该反演方法的数据量远远大于观测面的数据量,然后采用黄金分割算法,并合理地加入约束进一步减少反演的多解性问题,从而得到接近于实际地质情况的密度分布和几何形态,提高成像精度。进行模型试算研究该方法的可行性,采用不同的地质模型,通过对不同下延场的选择,比较视密度成像精度,得出最合理的下延场选取规则。在反演过程中,也要对其他不同的参数进行比较,合理加入约束,得到最佳的反演结果,改进反演算法,提高反演精度。通过模型试算总结出:该方法确实提高了视密度成像在纵向上的分辨率,对目标地质体的位置反演的更加精确,对不同深度的地质的成像效果明显提高了,大大提高了反演精度。应用本研究方法对北三台地区重力异常进行了视密度反演,并根据视密度的三维展布进行了火成岩解释,反演结果与实际钻井钻井资料比较吻合。证明了该方法的可靠性与实用性。
[Abstract]:Gravity inversion is to calculate the density distribution of underground by some means by known distribution of gravity field. It is one of the important means of geophysical exploration. The basic idea of this method is to mesh the underground geological bodies to be retrieved and to calculate the apparent density of all grid elements by continuous inversion iteration until the accuracy requirements are met. However, the main problem of this method is that the inversion of multiple solutions is serious: first, the observed gravity anomaly is the spatial volume effect of the underground; second, the collected data are discrete rather than continuous. How to reduce the multiplicity and improve the precision of inversion becomes the key problem of gravity inversion. In order to solve the multi-solution problem, more favorable prior information is effectively extracted and the number of inversion data is increased to improve the precision of apparent density inversion. How to effectively increase the data of inversion data volume is the basic starting point of this paper, so the regularization descent method based on depth change is used to obtain the gravity field of different depth, that is, the three-dimensional gravity data volume. Then the three-dimensional spatial data volume, that is, the data volume of the potential field with different depth, is used for inversion. The amount of data of the inversion method is much larger than that of the observation surface, and then the golden segmentation algorithm is used. Furthermore, the multi-solution problem of inversion is further reduced by adding constraints reasonably, so that the density distribution and geometric shape close to the actual geological conditions can be obtained, and the imaging accuracy can be improved. The feasibility of this method is studied by means of model trial calculation. Different geological models are used to compare the precision of apparent density imaging with the selection of different drop-down fields, and the most reasonable rule for selecting the drop-off fields is obtained. In the course of inversion, it is also necessary to compare other parameters, add constraints reasonably, obtain the best inversion results, improve the inversion algorithm, and improve the inversion accuracy. Through the model test calculation, it is concluded that this method improves the resolution of the apparent density imaging in longitudinal direction, and the position inversion of the target geological body is more accurate, and the imaging effect of this method on the geology of different depths is obviously improved. The inversion precision is greatly improved. The apparent density inversion of gravity anomaly in Beisantai area is carried out by using this method, and igneous rock interpretation is carried out according to the three-dimensional distribution of apparent density. The inversion results are in good agreement with the actual drilling data. The reliability and practicability of the method are proved.
【学位授予单位】：中国石油大学(华东)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P631.1

【参考文献】