#TITLE_ALTERNATIVE#
Inversion modeling is very inhelpful predicting subsurface physical parameters of one-dimensional magnetotelluric method (resistivity and thickness of the layer). We can use non-linear inversion method with linear approach with an assumption that the relationship between model parameters and measure...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Subjects: | |
| Online Access: | https://digilib.itb.ac.id/gdl/view/20318 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Inversion modeling is very inhelpful predicting subsurface physical parameters of one-dimensional magnetotelluric method (resistivity and thickness of the layer). We can use non-linear inversion method with linear approach with an assumption that the relationship between model parameters and measured data of magnetotelluric is almost linear. In this case, Jacobian's Matrix, which contains partial derivative of forward modeling function with respect to model parameters, and misfit are used to modify an initial model iteratively. The main weakness of linear approach inversion method is the result of inversion parameters depends on initial model and can converge to non-optimum solution. To solve this problem, we add damping factor to the eigenvalue, which is obtained by decompose the Jacobian's Matrix with SVD (Singular Value Decomposition). This technique reduced the dependence on initial model, and the inversion process became more stable. The outcomes of synthetic data's inversion are the resistivity and thickness of layer which correspond to the synthetic model parameter with low data misfit (10 %). |
|---|