Construção de modelos estratigráficos de alta resolução a partir da aplicação de técnicas de análise de sinal: um estudo de caso na Sequência Balbuena IV da Formação Yacoraite, Bacia de Salta, Argentina

Abstract

One of the most valuable applications of high-resolution sequence stratigraphy is reservoir zonation and characterization. Establishing a hierarchical chronostratigraphic framework comprising sequences at multiple scales is crucial to identify main reservoir heterogeneities that compartmentalize the fluid flow in the subsurface. In this context, this dissertation proposes a workflow that integrates different signal analysis techniques to identify sequences of multiple hierarchies from a series of stratigraphic data. The Continuous Wavelet Transform (CWT), Detrend Error Log (DTEL), and Integrated Detrend Error Log (INDTEL) techniques are utilized to process gamma-ray data obtained from outcrops of the Yacoraite Formation, which are conventionally employed as analogs for reservoir characterization. Our results suggest that CWT fits better with higher frequency cycles, while INDTEL shows a good fit with medium frequency cycles. In addition to these techniques, Hidden-Markov Models were also applied, predicting T-R cycles as hidden states from the facies transition matrix and possible occurrence of missed beats. This approach brings parameters that can guide the construction of stratigraphic models through automated processes, streamlining some steps that are routine in the day-to-day of stratigraphic interpretation, and also reducing the subjectivity that exists in geological interpretation through the mathematical pragmatism of the methods being applied.