The Research On The 3D Automatic Inversion Of Gravity And Magnetic Anomaly

Automatic inversion method is the hot spot in exploration industry currently. With the exploration of more and more difficult, more and more high precision measurement data now, gravity and magnetic exploration proposed higher demand. Today's modern exploration operations, the three-dimensional inversion growing role, has become an indispensable means of a technique. This paper researched the automatic inversion algorithm, evaluated the characters of the various constrains and supported the utility of the automatic inversion.First, we studied the infinite rectangular cross-section of gravity forward formula for the level of objects, determined the subdivision strategy of the automatic inversion of gravity profiles, researched the basic formula inversion, calculated inversion results and made the evaluation. Initially, inverse calculation used the generalized inverse method. Its computational speed is slow and consumes large memory. In the optimization algorithm we found that the conjugate gradient method has better performance and results like the generalized inverse method's, so in subsequent research we using the conjugate gradient method instead of the generalized inverse method.Second, for the unconstrained inversion problems, the use of a depth-weighted overcame the automatic inversion of the skin effect exists; the constrained by physical boundaries, controlled the range of the inversion results. Then we got a primary inversion result.The mainstream of current research of automatic inversion is to make a different form of inversion constraints. The primary inversion results can be further improved by these constraints. The smoothness constraint inversion can get the most smooth result, it is the most widely used application of the inversion method. The compression constrained inversion can get the minimum volume results, this method's application has certain limitations. The axial constrained inversion have good resolution for axis of the distribution of field sources. This paper realized these three constrained inversion methods and analyzed these methods'advantages and disadvantages by model test, and then integrated axial constraint inversion and smooth inversion done multiple constrained inversion model testing to improve the inversion resolution.Finally, the study suggests that the applicability of smoothness constrained inversion is the most extensive and easiest ways to achieve constrained inversion methods. This article applies to this constraint on the surface of the terrain data with three-dimensional inversion tests , and achieved a certain understanding.