The facies zoning produces the identification and localization of several layers crossed by a borehole and may be considered as the main geological product from core analysis and can be used to geologic modeling and reservoir characterization. The term facies joins a group of macro and microscopic characteristics described on rock samples collected in outcrops or retreated from cores obtained during the well drilling that can be related with petrophysical properties of a layer. In reason of the reduced number of cored wells, the operational difficulties for collect representatives cores and, in some cases, the impossibility of coring operations, as in horizontal wells, we introduced a methodology based on intelligent algorithms, here represented by a new model of competitive neural network, the minimum angular neural network, which is able to produce the recognition of angular patterns presented in porosity log readings represented in the L-K plot. Thus, it may integrate the geological core analysis to log interpretation to produce the facies zoning for non cored boreholes. This method is evaluated with synthetic data and actual well logs from Namorado's Field, Campos' Basin, Brazil.
|
