Study On Gravity And Magnetic High-efficiency Inversion And Joint Inversion Method Based On Adaptive Unstructured Meshing

Gravity and magnetic inversions can provide basis methods for structural analysis such as the distribution of mineral resources.During gravity and magnetic inversions,the underground space is usually divided into structural prism meshes.However,for undulating terrain areas and irregular geological bodies,the structural meshes are difficult to fit the characteristics of complex surface and irregular geological bodies.So,the inversion accuracy will be reduced.Unstructured tetrahedral mesh can flexibly fit the undulating terrain and irregular geological bodies due to the irregularity of its boundary,and can improve the inversion accuracy.We compare the inversion results of structured meshing and unstructured meshing,and the unstructured mesh inversion method has a higher vertical resolution which is much better suited for dip-slab models and undulating terrain than structural mesh.Therefore,in order to obtain the high accuracy density or magnetic susceptibility distribution characteristics of the geological body under undulating terrain,it is value to develop inversion method based on unstructured meshes.The refined inversion will trap into computational efficiency,we propose an adaptive unstructured mesh subdivision method to improve the inversion efficiency.The edge of the geological body is divided by edge detection method,and fine meshing is used inside the edge,and sparse meshing is used outside the edge.This inversion strategy can reduce the number of meshes.The edge detection method of gravity data selects the tilt angle method,which still has a better identification effect for deep and small-scale geological bodies.Because edge detection results will be affected by magnetization direction which resulting in a large deviation between the identified edge and the real range of the geological body.Therefore,we extract the edge of the geological body by analytical signal amplitude tilt angle method which is weakly sensitive to the magnetization direction.Through the synthetic model tests,the adaptive unstructured subdivision inversion can not lose the accuracy compared with the fine unstructured grid inversion in the whole region,and can effectively reduce the time consumption.The proposed method is applied to the measured gravity and magnetic anomaly data for density and magnetic susceptibility inversion research in Shandong Province.We can obtain the local three-dimensional density and magnetic susceptibility distribution.According to well logging data,we delineate three high density and magnetic susceptibility geological bodies which is consistent with the high value area of gravity and magnetic anomalies,and verify the accuracy of the proposed method.Compared with the single physical property inversion method,joint inversion method of multiple geophysical data can reduce the multi-solution problem.We still use the adaptive unstructured meshing method to obtain the results by cross gradient joint inversion method of gravity and magnetic data.Due to the different results of edge detection of gravity and magnetic data,the location and number of mesh points are inconsistent,it is impossible to carry out the results by cross gradient joint inversion method directly.Therefore,the gravity and magnetic data are interpolated to the same grid nodes to obtain the single inversion results under the same mesh by the least square method.Considering that the unstructured grids are not regularly arranged in the x,y,and z directions,the gradient values cannot be directly calculated.Therefore,we use the gradient calculation method based on the first-order finite difference function to calculate the physical property gradient changes.The model tests show that the nonconsistent adaptive unstructured mesh joint inversion can obtain higher resolution results when the gravity and magnetic data are homologous or non-homologous.After adding random noise to gravity and magnetic anomalies,the edge detection results and joint inversion results still maintain high accuracy and efficiency.This proof that the proposed method has good noise resistance and practical application value.Finally,the cross-gradient joint inversion method of gravity and magnetic with non-uniform adaptive unstructured mesh is applied to an iron ore area in Shandong Province to obtain the three-dimensional density and magnetic susceptibility distribution,which provides more reliable results for ore delineation in the study area.We propose gravity and magnetic property inversions and joint inversion methods based on adaptive unstructured mesh,and prove that the proposed methods can maintain the accuracy and effectively inversion results in the case of irregular topography and geological body shape.It provides a high-efficiency method for gravity and magnetic property inversions and joint inversion of unstructured meshes.