this is for holding javascript data
Lorenzo Perozzi edited The_inversion_results_shows_a__.tex
about 9 years ago
Commit id: a6044b04359464a3be39f44377160c2f8db8ab23
deletions | additions
diff --git a/The_inversion_results_shows_a__.tex b/The_inversion_results_shows_a__.tex
index 5533840..42dc2f0 100644
--- a/The_inversion_results_shows_a__.tex
+++ b/The_inversion_results_shows_a__.tex
...
The inversion results shows a SSIM index of about 0.6 for $V_p$, $V_s$ and $\rho$. Compared to the random SGS realizations set, the inversion results are significantly improved in term of similarity with the reference models. The $\phi$ field shows a SSIM index close to 1 for both SGS realization and inversion results. This
means that both are very similar to the reference porosity field. This is quite normal as the $\phi$ distribution shows a low variance for each layers
(see (refer to table \ref{tab:well-log}). If we focus at the reservoir level (1050m to 1300m) we observe a
substantial clear improvement on the similarity for the inversion results (SSIM=0.89) over the SGS realization result (SSIM=0.82) as shown in figure \ref{fig:SSIM_res}. In effect, as we perform the
CO$_2$ CO_2 flow simulation within this unit, we increase the number of parameters to optimize (i.e., the optimization is done for both static and dynamic data), resulting in an increased similarity
with at the
respectively reservoir level
compared to the rest of the
reference model.
\\