Predicting Petrophysical Parameters in a Fuzzy Environment
Jose Finol, Jing Xu · Studies in fuzziness and soft computing · 2002
SummaryThis paper concerns with the use of fuzzy models to predict petrophysical rock properties. Fuzzy models can approximate any continuous function with any degree of accuracy by covering its graph with a finite number of fuzzy patches. Every patch is identified with a fuzzy If-Then rule which describes a local input-output relationship between the variables of the model. A methodology to construct the fuzzy rules from a set of input-output data using fuzzy clustering algorithms is presented. The structure and parameters of the fuzzy model are extracted from the detected fuzzy partition. In order to demonstrate the effectiveness of the proposed fuzzy modeling method an application is presented using well log and core data from the Lake Maracaibo Basin.