Research And Application Of High-precision Processing Method For Potential Field Data

As the increasing of the exploration precision and efficiency of gravity and magnetic equipment, many traditional data processing methods can not satisfy the requirement of existent high resolution interpretation of potential field data because of the disadvantages of low precision and complicated computation. So improving and developing fast and high-precision methods have important significance. This paper present some improved algorithms for the potential-field transformation and separation and edge detection. These proposed methods have get some fine application effects.Transformation and separation are the most conventional processing and analysis methods of potential-field anomaly. The transformation accuracy and separation thoroughness directly affect the accuracy and resolution of some forward and inversion methods, such as Euler deconvolution and interface inversion. In this view, I present an approximate iterative operator method for derivative calculation and downward continuation, and an iterative method in wavenumber domain for potential-field separation which are based on the iterative compensation theory.Approximate operator iteration method is proposed to improve the stability and calculation accuracy of potential-field derivative calculation and downward continuation. For this method, the approximate operators of derivative calculation and downward continuation are derived based on the Taylor series, Laplace equation and transformational relationship among different order partial derivatives. By introducing these operators to the generalized iterative formula in wavenumber domain, a rapid iterative algorithm is developed for the derivative calculation and downward continuation. The filtering and convergence characteristics of this iterative method are analyzed for the purpose of estimating the optimal interval of number of iterations. Model tests of derivative calculation and downward continuation indicate that results of the approximate operator iteration method has stronger stability and higher precision than other iterative methods, and the correlation coefficient curve of successive iterative results can provide the optimal interval of number of iterations, to ensure the high computation efficiency and precision of the approximate operator iteration method.Based on the iterative method, an iterative filtering for potential-field separation is proposed by designing a regional field separation operator which amplitude is low. For this method, a part of the spectrum compositions of regional anomalies which are stranded in the spectrum of local anomalies can be extracted and returned to regional field. Such a scheme can make up for the loss of regional anomaly signal and achieve the purpose of improving the separation effect finally. Model test verifies that the extracted field by iterative filtering is more closed to true value than the conventional upward continuation method.Application of potential-field data to inverse the location of source boundary can provide gravity and magnetic evidences for tectonic interpretation. Existent edge detection methods are mostly associated with the gradient change of potential field, but these methods have the disadvantages that the detected edges are blurred and the results are susceptible to interference. In order to solve these problems, I propose three new edge detection methods including the cosine of pseudo total gradient calculating derivative with difference method, weighted small subdomain filtering technique, and improved Euler equation based on enhanced balanced filter. The three methods have their own advantages, and their calculation results can verify and make up for each other to improve the accuracy and resolution of edge detection.Edge detection technique based on cosine of pseudo total gradient calculating derivative with difference method is based on the directionality of boundary anomaly, regards the cosine of angle between vertical gradient and pseudo total gradient which horizontal gradient amplitude is compressed as the filter output. In this computational process, gradients are calculated using difference method to the mean of anomaly in the sliding window, and the maximum features at the boundary location have been tightened for two times to enhance the differences of anomalies between boundary and non-boundary, its maximum corresponds to the geological boundary. Model test results show that this proposed method is stable and can obtain fine result.Weight small subdomain filtering technique is proposed based on the gravity anomaly of vertical dike, numerical difference and small subdomain filtering. The filtering results indicate that the tightening gradient belt in the planar contour is the detection marker of the geological edge, and the positive/negative-anomaly trap is the analysis model for the detection of the residual density of the local geological body, which realize the weighted integration of the three small subdomain filtering in the space domain. Model test shows that the proposed method not only has a higher detection accuracy of the geological edges, but also has the ability to clearly reflect the relationship in the anomaly traps between the regional and local fields(the residual density relationship of the geological bodies).The conventional Euler equation has many dispersive solutions, and the solution accuracy is greatly affected by the structural index. To solve this problem, a new Euler equation is derived based on a new enhanced balanced filter. This equation can be used to calculate the plane position and depth of the edge, without prior information regarding the structural index and background field. Model test demonstrates that inversion results of location parameters using the proposed Euler method are more accurate than the conventional Euler method.Above three methods for edge detection, the first one can detect tectonic boundaries showing in gradient belt clearly. The second one not only make up for the imperfection on detecting tectonic boundaries which are shown in non-gradient belt, but also clearly reflect the relationships of anomaly amplitudes on both sides of boundaries. The last one can fill the blank of the first two while inversing the depth of geological boundary. These three methods being used together can better reflect the three-dimensional distribution characteristics of tectonic boundaries.At last, new methods proposed in this paper are applied to gravity and magnetic anomalies of Zhenlai zone in Songliao basin, and obtain some research results. According to the potential-field separation results, geological, physical properties of rocks and drilling data, I analyze the characteristics of the gravity and magnetic field, their residual anomalies, vertical derivatives and apparent physical properties of the middle and shallow stratums. According to the edge detection results, 11 main faulted structures have been detected. Analyzing the distribution characteristics of these faults and geological data, this zone is divided into three second-order tectonic units and nine secondary-order tectonic units. On the basis, I inverse the depth of Mesozoic bottom surface and forecast four sags as favorable zones for oil and gas exploration, using physical properties of rocks, drilling data and seismic sections as constraint conditions.