I am trying to understand the difference in the spreading rate of supercritical-CO2 plume for a stacked reservoir. The hypothetical reservoir has two permeable horizons, separated by a non-permeable horizon. I am trying to account for which of the petrophysical properties (porosity, permeability, entry pressure, relative permeability) is controlling the radial plume spreading rate the most and for that I am trying to calculate the dissimilar behaviour of the plume spreading rate curves (which is a time-distance plot). Can anyone suggest which statistical measure should I follow to account for the dissimilarity between the two curves?
Greetings and Regards
No, unfortunately, I just entered this site about oil articles and I don't have any information.
Thanks
Of course! To compare the curves, you might consider using:
1. Statistical Tests: Employ tests like t-test to see if there are significant differences between the spreading rates at various times.
2. Dynamic Time Warping (DTW): This helps measure similarity between curves, accounting for variations in timing or speed.
These approaches can highlight where and how the curves differ.
Thanks to both of you Nestor Ramos
and Abdirasak Mohamed Siad for your valuable feedbacks on this question. I have performed both the K-S test and Mann Whitney-U test on my sample as I wanted to see the difference between the two datasets. I would also like to test the DTW approcah on my dataset as well. In addition, I have also performed Kruskal-Wallis test to see the effects of different petrophysical properties on the spreading pattern. For example, if the first layer has a higher spreading rate than the second, I assigned it as a dummy variable 1. If it is the opposite, then I assigned it as 2. And then I performed the Kruskal-Wallis test to see whether there are any difference between these two groups in terms of the petrophysical parameters. It performed well and good on my sample dataset.
I am unsure whether your work space is heterogeneous or not and this does impact substantially the nature (and difficulty) of the measure. If heterogeneous (realistic) I would 1) start by focusing upon the saturation front, 2) get interested in Hausdorff distances 3) get interested in Mahalanobis distance.
Nestor Ramos
Sure, I will send you a copy of the manuscript once it is published. Thanks for all your valuable insights.
Pierre Bergey yes my workspace is heterogenous and I would like to take into account for the Mahalanobis distance. Thank you for your valuable suggestion.
- Badges
- Science method
- Similar topics
- Teaching