Aprendizado de máquinas em dados geoestatísticos de perfis geofísicos para identificação de camadas rochosas com fluidos

Abstract

The oil industry requires advanced tools to process large volumes of data and identify complex patterns. Machine learning stands out in this context, offering efficient solutions for this demand. This work presents the development and application of a machine-learning-based algorithm to identify fluids in rock layers, using geostatistical data and geophysical logs (gamma ray, spontaneous potential, caliper, and porosity). The algorithm was able to recognize patterns and efficiently classify the types of fluids present, contributing to safer and more efficient exploration of petroleum resources. In addition to its relevance to the industry, the project also made an important contribution to the teaching of Petroleum Engineering and Geology by providing a modern tool for analyzing and interpreting reservoir data.