Response Analysis And Inversion Of Multicomponent Induction Logging

As the development of oil exploration has gradually entered the post, more and more oil and gas exist in thin inter bed of sand and shale. The characteristics of oil and gas reservoir is low horizontal resistivity and high vertical resistivity. The conventional induction logging system is only due to the direction of the axis and can get horizontal resistivity, and cannot obtain vertical resistivity. It is unable to accurately assess the thin inter bed of sand and shale, extremely easy to cause the oil and gas underestimation and leakage.In order to solve the challenges encountered in oil and gas exploration, multi component induction logging(MCIL) is born at the right moment. It has three orthogonal coils direction and can obtain the formation of horizontal and vertical resistivity at the same time. Compared with the conventional induction logging, it can provide multiple logging curve, which can reflect the comprehensive physical properties of strata. On the other hand, in order to get more information of the reservoir, we need to combine multiple curves to the simultaneous inversion of multiple parameters. This also puts forward higher requirements for our data processing and explanation.Induction logging includes forward modeling and inversion processing. For forward algorithm, this paper studies the multi-dimensional approaches to solve the problem of electromagnetic field in inhomogeneous medium. For inversion algorithm, we studied several methods based on the physics and mathematics process to solve the problem of electromagnetic inversion.According to the conventional induction logging and multicomponent induction logging, this paper compared their similarities and differences from the coil system structure、working principle、measurement and inversion of the data. Then, starting from the most basic homogeneous medium, this paper deduced the induction voltage and extracted the formula of the apparent conductivity. According to one dimensional layered medium model, this paper deduced one dimensional Green function in layered media for the forward calculation. We studied the conventional algorithm for solving Sommerfeld integral. What is more,we proposed and optimized filtering method based on Fast Hankel Transform(FHT). For the response of multicomponent induction logging, this paper investigated the influence of different physical properties of the strata from many aspects.According to one dimensional layered medium model, this paper analyzed the principle and process of inversion of layered medium for inversion processing. We studied the Gauss Newton method and Broyden method based on parameter optimization and joined the linear search and nonlinear transform in the inversion process. In this paper, the forward algorithm was applied to the inversion, which reflected the influence of different forward algorithms on the efficiency of the inversion and the calculation results. Finally, we combined several numerical examples to study the influence of different models for the efficiency and accuracy of Gauss Newton inversion, and the Broyden method was successfully applied to the inversion of the numerical simulation, and achieved good results.