Machine learning in the determination of soft phyllite strength parameters

• Autor(a) principal: Pinto, Lívia Aparecida Gonçalves
• Data de Publicação: 2023
• Outros Autores: Silva, José Margarida da
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Research, Society and Development
• Texto Completo: https://rsdjournal.org/index.php/rsd/article/view/39693
• topic: Soft rocks; Phyllites; Strength parameters; Machine learning; Artificial neural networks; Linear regression.; Rocas blandas; Filitas; Parámetros de resistência; Aprendizaje automático; Redes neuronales artificiales; Regresión lineal.; Rochas brandas; Filitos; Parâmetros de resistência; Aprendizagem de máquina; Redes neurais artificiais; Regressão linear.

Geological-geotechnical aspects must be taken into account cince geological exploration studies in mining projects. Some accidents that occurred in mines in recent decades could be avoided if the geological and geotechnical conditions of the rock mass were understood. There is still great difficulty in classifying some types of rocks, especially rocks considered soft, through known geomechanical classification systems. The vast majority of existing classifications were developed based on hard rock characteristics. For rock masses with little strength, it is necessary to adapt current systems or develop new classification systems, which are specific for practically continuous soft rock masses. The objective of this article is to propose the use of Machine Learning techniques to predict geotechnical parameters of soft rocks, especially phyllite. Were used historical data from the results of geotechnical laboratory stress measures campaign of rocks from mines in the Iron Quadrangle are used, which, through optimized interaction, and with the aid of Artificial Intelligence techniques, such as the Artificial Neural Network and Linear Regression, are capable of generating results of interest for stability analysis and geotechnical modeling. Of the techniques used, results showed that the Linear Regression method was satisfactory in determining the strength parameters of soft phyllites and with good prospects for expansion and use for other parameters, as well as other types of rocks.