Research On Quasi-3D Inversion Method Of Airborne Time-Domain Electromagnetic Data

Airborne time domain electromagnetic method can be widely used in mineral resources exploration,petroleum and hydrogeological surveys,and environmental monitoring and other fields.For some areas where it is difficult to carry out ground geophysical prospecting,such as lakes and swamps,airborne time-domain electromagnetic exploration has great advantages.In order to accurately obtain the underground medium information,it is necessary to invert and interpret the airborne time-domain electromagnetic data.With the improvement of computer performance,high-dimensional inversion has become a current research hotspot,but because of the huge amount of airborne time-domain electromagnetic data and the great large calculation cost of high-dimensional inversion,and it has not been applied in actual engineering.Therefore,one-dimensional inversion method is still an important inversion imaging technique worthy of research in practical algorithms of underground interpretation imaging.Airborne time-domain electromagnetic data uses full-time continuous sampling.The data density is large and the distances between measuring points is small;In most geological environments,especially sedimentary geological environments,the physical property changes between adjacent measuring points are small;The data signal of airborne time-domain electromagnetic acquisition is weak,and it is easy to be affected by external electromagnetic noise.Therefore,it is difficult to obtain accurate inversion results by only inverting the collected data without constraining the inversion model.For those reasons,it is of great application value to use the spatial location information of the collected data to establish the spatial constraint relationship between adjacent measurement points and apply it to the model constraint of airborne time-domain electromagnetic inversion.This paper mainly studies the airborne time domain electromagnetic quasi-3D inversion method.Starting from the quasi-2D inversion method,the effects of vertical roughness constraint,lateral roughness constraint and prior information on the inversion are studied,then the quasi-2D inversion is extended to 3D space to study the quasi-3D spatial constraint inversion method.Specifically,this paper is divided into the following aspects:(1)Airborne time domain electromagnetic 1D and 3D forward modeling.For the1 D forward modeling,Hankel transform is used to calculate the frequency electromagnetic response of the center loop source in a uniform layered medium,and then using the G-S transform to calculate the 1D forward electromagnetic domain electromagnetic response.For 3D forward modeling,the mimic discretization method is used to discretize the Maxwell equations to obtain the formula for forward modeling.By calculating the forward response of the half-space and three-layer model,the onedimensional forward programs used in this paper are in good agreement with the calculation results of the free software EMMA,and the relative errors are less than 5%,which can be used for inversion method research.(2)Quasi-2D inversion method based on lateral and vertical roughness constrained.In ATEM inversion,one-dimensional inversion with lateral and vertical roughness constraints is used,and the inversion result is a two-dimensional profile,which is a quasi-2D inversion method.However,the quasi-2D inversion is simultaneous inversion of data of multiple measurement points,and the calculation time is affected by the number of measured points.Taking into account the problem of long calculation time of quasi-2D inversion,we propose an inversion scheme with lateral segmentedcombined constraints.First,observation points are segmented;Then the lateral and vertical roughness constraints are used simultaneously in each data segment,and the lateral priori information constraint is used between the segments;And finally inverse for each segment of data.This inversion scheme reduces the calculation time of quasitwo-dimensional inversion by segmentation,ensures high accuracy of the inversion results by using the lateral and vertical roughness constraints,eliminates the discontinuity between segments by using the lateral prior information,and can be effectively processed the actual airborne time-domain electromagnetic field exploration.By constructing the lateral constraint relationship between the measurement points on a single measurement line,and combined with the vertical roughness constraints between the model parameters within a single measurement point,the proposed inversion method not only has high computational accuracy,but also can accelerate the convergence rate of inversion.(3)Quasi-3D inversion method based on spatial constrained.Because there is no constraint relation between the measuring points on the adjacent measurement lines,the results of quasi-2D inversion have poor continuity in 3D space.In this paper,the spatial constraint relationship is established by Delaunay triangulation.In order to reduce the calculation time of inversion,a sub-regional calculation strategy is used,and at the same time,the modified prior information constraint and the secondary inversion are used to reduce the error generated by the partition calculation.Through the inversion results of simulation and experimental data analysis found that,using vertical roughness constraints,spatial constraints and the priori information constraints at the same time,the inversion results have good continuity in both the measured line direction and the depth direction,have a better clear and accurate geological boundary contour,consistent with the geological background data,and the airborne electromagnetic field data.Compared with the quasi-2D inversion results,it can be found that the quasi-3D inversion results are more accurate than those of quasi-2D inversion and more consistent with the geological background data.Of course,compared with quasi-2D inversion,the computation time of quasi-3D inversion method is longer.