A new approach for the prediction of petrophysical rock parameters based on
a rule-based fuzzy model is presented. The rule-based fuzzy model correspo
nds to the Takagi-Sugeno-Kang method of fuzzy reasoning proposed by Sugeno
and his co-authors. This fuzzy model is defined by a set of fuzzy implicati
ons with linear consequent parts, each of which establishes a local linear
input-output relationship between the variables of the model. In this appro
ach, a fuzzy clustering algorithm is combined with the least-square approxi
mation method to identify the structure and parameters of the fuzzy model f
rom sets of numerical data. To verify the effectiveness of the proposed fuz
zy modeling method, two examples are developed using core and electrical lo
g data from three oil wells in Ceuta Field, Lake Maracaibo Basin. The numer
ical results of the fuzzy modelling method are compared with the results of
a conventional linens regression model. It is shown that the fuzzy modelin
g approach is not only more accurate than the conventional regression appro
ach but also provides some qualitative information about the underlying com
plexities of the porous system. (C) 2001 Elsevier Science B.V. All rights r
eserved.