The Magnetic Gradient Tensor Research And Data Interpretation

In this thesis the author deduces magnetic gradient tensor forward formula of point dipole, sphere and cube model. The tensor of these model shows better properties than the total magnetic anomalies. The tensor is symmetrical about the center of the geological body or exist extremum in the center of the geological body. There will be abnormal at the boundary of the geological body. Tensor can reflect more direct and clear space geometric features of the target body. Different direction tensor reflect different spatial information of the geological body. Tensor is less affected by clino-magnetization.The author extend boundary recognition algorithm based on the total magnetic anomalies to boundary recognition algorithm based on the tensor. Through the model the author found some Conclusion about magnetic gradient tensor. Boundary recognition algorithm based on tensor is less affected by the oblique magnetization and is not sensitive to direction of magnetization. Through the model, the author compares each boundary recognition algorithm based on the tensor. At the end of the thesis,through processing the observed data, the author verify that only five elements are independent elements in the nine elements of tensor. And tensor recognition algorithms that do not through the pole processing can identified the magnetic body underground by clino-magnetization. Through comparing the results of model and the observed data, the author find that the anti-interference ability of tensor invariant I1 is stronger.Solution of the euler deconvolution is solution the equations that coefficient matrix make up by tensor. The author deduces the euler deconvolution formulas that do not contain structure index, to solve the problem that solution euler deconvolution need to know in advance the structure indexes. Through the results of boundary identification delineate the scope of geological body, reducing the computational. The author use abnormal body distance criterion to rule out the unqualified solution, solving the problem of understanding divergence. Through the model to verify improved eulerdeconvolution method, the author find that the area to solve is reduced, and accelerate the speed of operation. The inversion effect is significantly increased.