Some prolific basins have large deposit thicknesses of different saline types that are often treated as a homogeneous layer. However, for drilling purpose it is extremely important to predict the different salt types and their behaviors due to their particularities and direct influence in the development of well design. To characterize the different saline types, in this paper we propose an inversion technique, called genetic inversion, which works with artificial neural net training, applying well information to find a global minimum. The results presented here (Impedance P and S, sonic P and S, density, Poisson and lithofacies) were tested in different oil fields from the Brazilian east margin basins, being very satisfactory due to the good correlations between the generated logs and the real well logs. They can also be used as secondary variables for geostatistical property distributions, allowing better results in the distribution between well and a more robust modeling.
The salt has a peculiar behavior, different from the other sedimentary rocks, that is the fluency, (Poiate et al., 2006). Its fluency impacts directly on the productive time of drilling, therefore the need to get closer to reality as possible on the knowledge of the heterogeneity of the different types of salt from the well location area.