| The array induction logging tool (AIT) is different from the conventional induction tools with signal focusing. It is composed of several different spacing three-coil arrays, operates three frequencies, and measures many logging curves which have different investigation depth and discernibility through modern logging technology. So the new multi-array induction log can provide more abundant formation information than Dual-laterolog, exactly recognize the invasion characteristic of formation, and make qualitative analysis to the bed of oil and gas. However, in the thinner formation, the measured resistivity is often far from the true value of the bed, and can't reflect the invasion character accurately due to the influence of the borehole, mud environment, invasion zone and shoulder beds, and so on. Therefore, an appropriate numerical process and effective inversion method are introduced to elimate the effects of various factors in order to obtain the true parameters of formation, this is very important in the logging explanation. In this paper, I have studied from the following aspects.In Chapter one, it systematically summarizes the development process and signification of the multi-array induction log, at the same time, the main research contents and measurement are given in the paper.In Chapter two, it establishes the fast forward simulation theory of AIT based on the model matching method and character of the multi-array induction logging tool in the horizontally layered inhomogenous medium. Using this measurement, we can quantificationally investigate the unique feature of multi-array induction logging response in a variety of complex formation, and analyze the invaded zone resistivity, the original formation resistivity, the influence of the bed thickness, invasion radius and transmitter frequencies, etc. The forward simulation results show that the responses are different because of the different spacing coil arrays. Even though the frequency reaches 105.0khz, there is still error between the apparent resistivities and the true value of formation resistivities, Simultaneously, the skin effect of the apparent conductivities on subarrays with the long coil spacings exits. The apparent resistivity of every subarray is on the increase with the original formation resistivity increasing. In the high invasion formation, the extent of apparent resistivity increasing is larger, however, in the low invasion formation, the apparent resistivity slowly grows. The larger the invaded zone resistivity and the coil spacing are, the smaller the range of apparent resistivity is, especially, the induction logging response of the eighth subarray has almost no changes. The influence of the mud resistivities on subarrays with the short coil spacings is larger than the long coil spacings. With the invasion radius increasing, the apparent conducticities of all subarrays decrease, Furthermore, the change of the bed depth also affects the logging responses. The apparent conductivities are larger along with the bed thickness increasing, but the logging curve of the subarray with the long coil spacing about 1.8m is anisomerous.In Chapter three, the stratum model parameters are usually made up of the original resistivity of formation (Rt), the resistivity in flushed zone (Rxo), the invasion radius (rxo) and the interface depth, etc. in the horizontally inhomogenous layered medium. Hence, the paper constructs a multiparameter inversion algorithm about the model parameters which are composed of the virgin formation resistivity, invaded zone resistivity, invasion radius and bed boundary depth in each bed according to the multi-array induction logging datum. First, applying the connection between the formation conductivity and model vector, we could deduce the relation both the relative perturbation of the formation conductivity and small comparative change of the model vectors through differential calculus disposal method. Secondly, using the semi-analytic forward simulating technique and theory of Born approximation calculation, we establish the fast inversion arithmetic of Frechet derivative matrix about the multi-array induction log. Then, the formation parameters are continuously revised through the normalization process and dependent singular values decomposition technic. Ultimately, the paper surveys the inversion results of the theoretical model under all kinds of factors including the bed thickness, the invaded depth and the original resistivity of formation, and so on. Numerical simulation results are best appropriate to the multi-array induction logging data.In theory, the fast iterative inversion method can availably and simultaneously rebuild the original resistivity of formation, resistivity in the floushed zone, invasion radius and stratum boundary depth from the logging datum. In the inversion process, taking full advantage of the difference both investigation depth and vertical resolution of coil system, it mostly divides the iterative inversion course into two parts. For one thing, the single frequency multiparameter inversion has been used to the coil-system subarrays with proper surveying depth and vertical resolution. Whereafter, multifrequency multiparameter iterative inversion algorithm is applied to the same theoretical model and initial model parameters, determing all formation model parameters, then the inversion results realize to fufill optimum suit with the real model. Finally, synthetic data results show that the invasion character around borehole is exactly reflected by their relative sizes. Although the multifrequency inversion measurement can't completely return to the real value in the thinner bed, the error between inversion results and the true model parameters is greatly reduced in comparison with single frequency multiparameter inversion results. This method enhances the inversion stability and increases the reliability of inversion results, so that the better estimation of all formation parameters can be obtained by the inversion datas. |
PDF Full Text Request