Thinly laminated sand-shale formations represent a difficultrnchallenge for petrophysical evaluation. The conductive shalernlaminations have a profound effect on the traditional inductionrnlog, causing it to read the shale laminations significantly lowerrnthan the high-resistivity sand layers. This misreading leads tornpessimistic computations of water saturation and estimationsrnof reserve. The newly developed triaxial induction loggingrntool provides much more information than the conventionalrninduction logging measurement. It is not only sensitive to therneffective horizontal resistivity, which is dominated by thernconductive shale layers, but it is also sensitive to the effectivernvertical resistivity, which is determined by the conductivernshale and the resistive sand, at any dip angle.rnA key challenge of developing the triaxial induction arrayrnis the large borehole effect on coplanar couplings when therntool is eccentered in the transverse direction in water-basernmud boreholes. To reduce the borehole effect to a manageablernlevel, a new design with multiple electrodes has beenrnimplemented. Tank experiments and numerical modelingrnresults show the borehole effect is reduced dramatically byrnusing this design.rnA parametric inversion algorithm to simultaneouslyrndetermine the horizontal resistivity, vertical resistivity,rnformation dip, and azimuthal angle and bed boundary positionrnfrom the triaxial induction logging data will be presented. Therninversion problem is solved by employing a weighted,rnconstrained, and regularized Gauss-Newton minimizationrnscheme. To archive the practical application of the inversionrnalgorithm, a fast Jacobian matrix computation is implementedrnto improve the efficiency. Furthermore, a multiplicativernregularization technique is used to automatically determine thernregularization coefficient. Synthetic examples will bernpresented to indicate the robustness of the algorithm.A quantitative interpretation method of water saturation inrnlaminated shale-sand formation will be presented. Shalernanisotropy is taken into account for building the interpretationrnmodel. Two field examples demonstrate the laminated shalesandrnanalysis improves the estimate of hydrocarbons in place.
展开▼
