Scientific Computing and Imaging


contents up previous next
Next: About this document ...

CS 523 Scientific Visualization
Spring Quarter 1997

Project Report

Dmitriy A. Pavlov
dept. Geology and Geophysics

Visualization of the results of fast inversion of geophysical time-domain sounding data

The goal of electromagnetic survey is to describe the conductivity in the subsurface of the earth and indicate the location of structures with anomalous conductivities with respect to surrounding medium. In time-domain observations the time derivative tex2html_wrap_inline21 of the vertical component of magnetic field is measured. In the frame of this project it was considered two types of fast inversion. 1. - inversion on the base of the thin sheet approach, which is called S-inversion. 2. - full inversion , based on solution of integral equations using some modeling computations and precomputations. Solution of the inverse problem has been made on the base of regularized gradient method in real-time regime.

In the frame of project it will be used some programming codes existing at Consortium of Electromagnetic Modeling and Inversion (dept. of Geology and Geophysics, U of U), and also new original codes. Initial codes were written on Fortran and C. It has been created programm complexes which runs numerical codes and Matlab scripts for visualization of results.

For presentation of this project it have been created special Matlab scripts, which just show results of computations

Figures 1 - 3 show results of modeling on the base of S-inversion for the model of horizontally stratyfied Earth. Figures 4-7 show model and graphical results of S-inversion for local conductive body in homogeneous half-space (approximately the same results we can obtain for the conductive body in the horizontally-layered Earth). Figure 8 corresponds to the 1D full inversion for the model of horizontally-layered Earth. Because the method of inversion also based on the model of the horizontally-layered Earth, here we received practically 100

Problems

Till this time I did not finish the part of the computational code which solves 3D inverse problem. This work in progress now. And will be finished in a few week.

Conclusion

Because this work on fast inversion now is just in the beginning we don't have a lot of interesting results. We just on the stage of modeling study. And from this point of view results of project look not bad.




contents up previous next
Next: About this document ...

Scientific Computing and Imaging
Sun Jun 15 19:35:24 MDT 1997