重力梯度张量数据的3D聚焦反演方法研究

发布时间：2018-10-30 06:43

【摘要】：重力引力位在各个方向的二阶导数称为重力梯度张量,可以通过重力梯度测量仪器测量获得重力梯度数据,与常规重力勘探仪器测量的重力位垂向一次导数数据相比,重力梯度测量的优势是其测量数据对区域地质构造中的局部密度异常更加敏感,而且包含可以联合解释的多个分量,用梯度数据进行反演,能够更好地描述场源的空间形态及位置。本文详细分析了国内外重力梯度张量数据3D聚焦反演的发展现状,在此基础上,提出了基于双重加权的重力梯度张量数据的3D聚焦反演方法。针对观测点远远小于反演参数而产生的欠定问题,在经典Tikhonov正则化理论框架下,引入最小支撑泛函对反演模型进行约束以避免反问题解的不稳定性;针对重力梯度数据聚焦反演中存在的趋肤效应,在模型目标函数中引入双重加权—深度加权函数和加权系数矩阵;针对正演时几何格架计算量和存储量大的问题,给出了快速正演算法。通过理论模型,对重力梯度张量各分量进行了单独以及全张量联合聚焦反演,验证了基于双重加权的聚焦反演方法的有效性,同时也证明了重力梯度全张量反演效果比单分量的反演效果要好。反演时还将聚焦反演和传统的最平滑约束反演结果进行对比,结果表明,聚焦反演的异常体更集中,边界更清晰且异常体密度值也与理论模型更吻合。将该反演方法运用到涩北一号气田区的实际数据中,反演结果较好地反映出气田位置。
[Abstract]:The second derivative of gravity potential in all directions is called gravity gradient Zhang Liang. Gravity gradient data can be obtained by gravity gradient measurement instrument, compared with the gravity potential vertical first derivative data measured by conventional gravity exploration instruments. The advantage of gravity gradient measurement is that the measured data are more sensitive to the local density anomalies in the regional geological structure, and contain multiple components that can be interpreted jointly, which can be inversed by gradient data. It can better describe the spatial form and position of the field source. In this paper, the present situation of 3D focusing inversion of gravity gradient Zhang Liang data at home and abroad is analyzed in detail. On this basis, a method of 3D focusing inversion based on double weighted gravity gradient Zhang Liang data is proposed. In view of the problem that the observation point is far less than the inversion parameter, under the framework of classical Tikhonov regularization theory, the minimum support functional is introduced to constrain the inversion model to avoid the instability of the inverse problem solution. Aiming at the skin effect in focusing inversion of gravity gradient data, a double weighted depth weighting function and a weighting coefficient matrix are introduced into the objective function of the model. A fast forward algorithm is presented to solve the problem of large computation and storage of geometric lattice in forward modeling. Based on the theoretical model, the combined focusing inversion of each component of gravity gradient Zhang Liang is carried out, and the validity of the focusing inversion method based on double weighting is verified. It is also proved that the inversion effect of gravity gradient full Zhang Liang is better than that of single component inversion. The inversion results show that the focus inversion is more concentrated, the boundary is clearer and the abnormal volume density is more consistent with the theoretical model. The inversion method is applied to the actual data of Shibei No. 1 gas field area, and the inversion results reflect the location of the gas field well.
【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

【相似文献】