A new joint inversion strategy using a priori petrophysical information as constraints
Abstract
Multiple geophysical data collected over the same area but based on fundamentally different physics usually contain complementary information about the subsurface, because different geophysical methods are sensitive to different physical properties and their variations on different scales. Joint inversion combines the complementary information by integrating all the geophysical data into a single inversion scheme. Thus, models resulting from joint inversion are more likely to represent the subsurface better than models derived from a single type of data. One important need in joint inversion is to find a reasonable relationship between different models so that they can constrain each other. This relationship could be an empirical relationship between different physical properties based on petrophysical measurements or a structural similarity measure between models. In this study, we consider the statistical petrophysical relationship between two physical properties as the way to link different models. There are two challenges concerning this approach. First, when multiple petrophysical relationships exist in the subsurface, it is usually difficult to specify correctly the spatial applicability of these petrophysical relationships, and thus, applying appropriate petrophysical information to appropriate regions in joint inversion remains a problem to be answered. Secondly, in reality we may have only partial information about the physical parameters under consideration. The incompleteness of a priori petrophysical information makes the task of utilizing petrophysical information in joint inversion even harder. In this study, we develop a new joint inversion algorithm that effectively builds a priori petrophysical information into inversion by means of guided fuzzy c-means (FCM) clustering technique. Synthetic examples show that this method can effectively deal with the two problems mentioned above, i.e. the non-uniqueness and incompleteness of a priori petrophysical information. We also present the inversion results by applying this joint inversion algorithm to SEG/EAGE salt model.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2012
- Bibcode:
- 2012AGUFMNS34A..05S
- Keywords:
-
- 3260 MATHEMATICAL GEOPHYSICS / Inverse theory