Modelo de machine learning aplicado na previsão do desempenho de fibras como reforço para concretos reforçados com fibra

Detalhes bibliográficos
Ano de defesa: 2023
Autor(a) principal: Silva, Victor Hugo Peres
Orientador(a): Não Informado pela instituição
Banca de defesa: Não Informado pela instituição
Tipo de documento: Dissertação
Tipo de acesso: Acesso aberto
Idioma: por
Instituição de defesa: Universidade Federal de Mato Grosso
Brasil
Instituto de Ciências Exatas e da Terra (ICET) – Araguaia
UFMT CUA - Araguaia
Programa de Pós-Graduação em Ciência de Materiais
Programa de Pós-Graduação: Não Informado pela instituição
Departamento: Não Informado pela instituição
País: Não Informado pela instituição
Palavras-chave em Português:
Link de acesso: http://ri.ufmt.br/handle/1/5607
Resumo: Concrete is currently the most used manufactured material in the world. Despite its high resistance to compression, this material has low tensile strength, and its durability may be reduced by the appearance of cracks caused by tensile stress present in the structures. By allowing water infiltration and allowing corrosion of the reinforcement of reinforced concrete structures, these cracks, which can also occur during the concrete drying process, make it necessary to carry out repairs to the concrete structures. In this scenario, the use of new technologies such as self-repairing concrete and crack control in concrete structures becomes relevant as they reduce the costs needed for structural repairs. Aiming to facilitate the study of new cracking control technologies in concrete structures, the present work aims to compose a machine learning (ML) model that can predict, based on a dataset provided to the model, the performance of a given fiber as a reinforcing material for fiber-reinforced concrete. To create the model, data collected from 18 (eighteen) published scientific works were used, composing a dataset with 13 (thirteen) different types of fibers. The computational model was written in Python language, using multiple linear regression (MLR) methods and Support Vector Regression (SVR) techniques to develop the prediction model. From the results obtained, it could be observed that the SVR model (R²=0.857 and RMSE=0.3710) performed better than the RLM model (R²=0.592 and RMSE=0.6267), therefore being more suitable for predicting the performance of fibers as a reinforcing material against cracking in fiber-reinforced concrete.