Elaboração de algoritmo para aplicação de aprendizado de máquina a fim de reconhecer armadilhas geológicas do tipo Domo Salino em seções sísmicas

• Autor(a) principal: Amaral, Robertom Guedes do
• Data de Publicação: 2021
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UFAM
• Texto Completo: http://riu.ufam.edu.br/handle/prefix/5903
• topic: Machine Learning; Big Data; Processamento de dados sísmicos; Armadilhas geológicas; Domos salinos; ENGENHARIAS; Saline dome; Método sísmico de reflexão; Método geofísico; Scikit-Learn

The 4.0 Industry advances permits the creation of powerful technologies and computational tools capable of working in the most diverse areas of science, among these tools are artificial intelligence and Machine Learning. Through programming and creation of algorithms is possible to use the Machine Learning, which is a technique that the previously developed algorithms “teach” a computer to improve its “abilities” in a given field of study for the most diverse purposes, constantly and autonomously improving its own behavior and performance similarly to humans. Among the geophysical methods, stands out, in petroleum industry, the seismic reflection method, this method utilizes the propagation of waves in subsurface for data acquisition related to local geology. The amount data acquired is extremely numerous and to this quantity of data the name Big Data is given. After all the process of acquisition, it is necessary to accomplish the processing of these data and this phase can take weeks or even months, sometimes. This process is humanly massive, even with the support of some software. After the acquisition phase, there is a stage of data interpretation. The methodology is based in the use of programming language Python in application of Machine Learning, specifically aims at the elaboration and creation of algorithm capable of processing data from seismic in search of geological traps of type saline dome. The training took place through the SVC module of the Scikit-Learn library, which proved to be quite efficient within the proposal. The use of Machine Learning is very promising in the field of Geophysics applied to Oil and Gas Engineering with regard to the identification of saline domes in seismic sections, which can be extended to other uses.