Archie equation is a historical mark of Formation Evaluation establishing a relationship among the physical properties and the petrophysical properties of reservoir rocks, which makes possible the identification and quantification of hydrocarbon in subsurface. Water saturation is the solution of Archie equation obtained from the measure of formation deep resistivity and porosity estimated. However, the solution of Archie equation is no trivial, in the dependence of previous knowledge of formation water resistivity and Archie exponents (cementation and saturation). This thesis introduces a set new intelligent algorithm to solve Archie equation. A modification of competitive neural network, nominated as bicompetitive neural network produces the log zonation. A new genetic algorithm with evolutionary strategy based in the mushrooms reproduction produces estimates for the matrix density, the matrix transit time and the matrix neutron porosity, which associated to a new rock model, produces realistic porosity estimates considering shale effects. A new model of competitive neural network, nominated as angular competitive neural network is able to accomplish the interpretation of Pickett plot, supplying the information about formation water resistivity and cementation exponent. All results of the methodology hereintroduced are presented using synthetic data and actual wireline logs and core analysis results.
|
