Método híbrido para a inversão de dados geoelétricos baseado na identificação de padrões

Abstract

This work proposes to create a database, containing synthetic data, of geoelectric surveys of vertical electric soundings, VESs, with Schlumberger arrangement, in several situations of stratified environments, in order to allow a ranking with the VES to be interpreted. The database is normalized so as to have invariant scales in the openings and in the apparent resistivity values, which enables the selection of the best result through the shape of the curve apparent resistivity $\times$ openings, in a process of pattern identification, regardless of the survey values. The results are obtained by choosing the VESs that present the lowest residuals, which is evaluated by the square of the Euclidean distance, between the VES data to be interpreted and the database. Thus, the results obtained can be used both in the preliminary interpretation of VESs and also as a starting point in usual inversion methods, in order to obtain more accurate results, the ranking of the best solutions is useful in the evaluation of possible ambiguities present in geoelectric methods. In this work we used the genetic algorithm with inversion method in the second step of the process. The methodology proved to be efficient, with low computational cost and easy implementation. Database;Geoelectrical Section;Geoelectrical Data