In the past, we developed an inversion scheme for the determination of anisotropy in a vicinity of a receiver situated in a borehole from the data obtained during a multi-azimuth multiple-source offset VSP experiment. In this contribution, we are using this inversion scheme to determine orientation of the axis of symmetry of a TI medium from P-wave polarization and slowness data. As before, as input data we use polarization vectors and vertical components of the slowness vectors of P waves. On synthetic data, we study accuracy and sensitivity of the results of the inversion to varying types of the waves used (direct, reflected or direct and reflected), varying levels of noise (separately for each of the considered waves, for polarization and slowness data, for each component of the polarization vector), varying configuration (number and orientation of profiles and their extent), etc. We show that accuracy of the determination of the orientation of the symmetry axis depends strongly on the accuracy of the determination of the elastic parameter A11, which characterizes horizontal propagation. Generally, we can get satisfactory estimate of the axis orientation if the level of noise in the slowness data is less than 5% and if the orientation of polarization vectors is determined with error less than 4°. If we have available well estimated parameter A11, then the errors of slowness and polarization data may be upto 20% and 15°, respectively. The estimate of the orientation is more sensitive to errors in the slowness vector than in polarization. The study indicates that it is better to use both direct and reflected waves in the inversion. The use of direct wave only leads to the loss of resolution and stability.
The image of the paper in GIF 120dpi (283 kB) is designed for an instant screen preview.
The paper is available in PDF (642 kB).