Aplicação de técnicas de inteligência artificial para avaliação de parâmetros de solo com foco em projeto geotécnico de poço

Abstract

Currently, the oil industry plays a crucial role in global energy supply; however, its exploration faces complex challenges. This is particularly evident in well construction in offshore environments and the high demand for pre-well studies in extreme locations. Precise knowledge of soil parameters is imperative for the safe and efficient construction of hydrocarbon wells. However, in complex environments, conducting soil studies at the point of interest is often unfeasible. Therefore, there is a need for alternatives capable of reducing associated costs throughout the construction process, mitigating human errors, and predicting characterizations without the need for in situ studies. Thus, this work focuses on the application of artificial intelligence techniques and the use of Kriging to evaluate soil tests and extrapolate data from geometrically correlated points. The application of these techniques allows for an estimation of parameters of interest in nearby points, as well as assessing the influence of distance in this process. In this context, this study aims to automate the application of Artificial Intelligence models, specifically using the Support Vector Machine (SVM) technique on public Cone Penetration Test (CPTu) data. Implementation will be carried out in the Python programming language, and the results obtained will be compared with traditional regression techniques, expanding the analysis in the context of geotechnical studies.