Study On Joint Inversion Algorithm Of Formation Resistivity By Dual Laterolog And Dual Induction Data

Reservoir fluid, hydrocarbon saturation, permeability and other geological parameters could be computed by resistivity logging data. Currently, resistivity always measured by dual laterolog and dual induction in oil field, the error must be in the actual logging data because the influencess of borehole、mud invasion、surrounding rock and so on. Therefore, inversion must be done by logging data. There are some well that dual laterolg and dual induction data are in the same well, joint inversion that multiple solutions inversion could be decreased is used in the paper.Dual laterolog response is computed by Finite Element Method and dual induction response is computed by Gianzero geometrical factor. And logging response when borehole、mud invasion and surrounding rock changed are computed in the paper.The improved differential evolution algorithm that could solve the problem of high dependence on the initial model and local convergence is used to compute real formation resistivity. Compared with genetic algorithm and marquardt algorithm in the same model, the applicability of improved differential evolution is analysised in layered medium model on joint inversion. The result indicatd that differential evolution algorithm better than other alrithom because of higher accuracy. And improved proposals could be used in real data inversion because the error is less than.Usually, resistivity is changed with invaded radius in invaded zone, but three-parameter inversion could made bigger error in resistivity inversion because the invaded zone resistivity is non-changed. Four-parameter inversion model considered invaded zone resistivty range is used in this paper. The four-paramaters joint inversion is studied by improved differential evolution added to the noise of 0%,5%,10% and 20% on the logging response. Result show that error of joint inversion has been decreased by improved differential evolution algorithm. error is little increased with the noise changed, but inversion data could be received. Therefore, the stability and noise immunity are proved on improved differential evolution algorithmThe method of activity function method is used to divide reservoir on the resistivity logging and Gamma curve. When the Gamma is higher than the set value, this formation would not be inverted because it is mudstone. Then the improved DE algorithm is used to joint inversion four parameters by dual logging and induction data in many actual well data, the result show that water saturation computed by inversion resistivity is more close to the test result, it also proved that the DE algorithm could be used in working, and could provide more accurate formation paremeters in logging interpretation.