Optimization methods for definition of geological model

• Grant number: 21/07122-0
• Support Opportunities: Scholarships in Brazil - Doctorate
• Effective date (Start): September 01, 2021
• Effective date (End): January 31, 2024
• Field of knowledge: Physical Sciences and Mathematics - Geosciences - Geology
• Acordo de Cooperação: Equinor (former Statoil)
• Principal Investigator: Alexandre Campane Vidal
• Grantee: Juan Francisco Villacreses Morales
• Host Institution: Centro de Estudos de Energia e Petróleo (CEPETRO). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Host Company: Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Mecânica (FEM)
• Associated research grant: 17/15736-3 - Engineering Research Centre in Reservoir and Production Management, AP.PCPE

Abstract

In the use of all available information and the application geostatistical, seismic inversionmethods, and flow simulation data, a large number of auxiliary variables are produced. Theseauxiliary variables must be interpreted in conjunction with primary variables to obtain morerobust and reliable models.Seismic attributes are secondary seismic variables obtained exclusively from the originalseismic data by direct measurement or by experience/logical reasoning. As with seismicinversion, a considerable body of work has been published in this field of research during the lastdecades (Brown 2001). Depending on geological conditions, each attribute may providecomplementary information about the reservoir or its surrounding structures. However, asignificant level of redundancy can occur.Despite being recognized as an important tool for minimizing uncertainty of structural,lithological, and stratigraphic interpretation, seismic attributes require special attention, becausea bad choice of attributes may produce different results from those expected (Chopra and Marfurt,2007).For these reasons, an effective choice of seismic attributes to characterize reservoirs andthe combination of data with different scales are delicate tasks. In order to improve accuracy andtime of interpretation, we purpose the application of optimization methods to establish thesignature between seismic attributes and the chosen well logs, estimating petrophysicalparameters in the region between boreholes by neural network. Similar works, as described byDorrington and Link (2004), show advances of combining such techniques, which were able todetect channels and predict porosity in siliciclastic reservoirs.The goal of this project is to address challenges of reservoir characterization and predictionby applying mathematical and computational methods to the available reservoir dataset. Thisproject will produce a workflow and set of parameters that will contribute to the construction ofreliable and accurate geological models focused on uncertainty analysis in this, and comparable,reservoir system. (AU)