Sistemas urbanos de drenagem sustentável: uma abordagem entre BIM e aprendizado de máquina

• Ano de defesa: 2024
• Autor(a) principal: Sousa, Luciano Hamed Chaves Haidar
• Orientador(a): Studart, Ticiana Marinho de Carvalho
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Área do conhecimento CNPq: CNPQ::ENGENHARIAS::ENGENHARIA SANITARIA::RECURSOS HIDRICOS
• Link de acesso: http://repositorio.ufc.br/handle/riufc/79582

Sustainable Urban Drainage Systems (SUDS) are a non-conventional and necessary response to the challenges related to stormwater management in cities experiencing rapid urban growth. These systems aim to reduce the negative impacts of urbanization on the hydrological cycle, such as increased soil impermeability and urban flooding. To optimize decision-making regarding the solutions to be implemented, technologies such as Building Information Modeling (BIM) and Machine Learning (ML) have been increasingly explored in the development of drainage solutions. In the context of sustainable urban drainage, BIM provides an integrated platform for managing project data, modeling scenarios, and analyzing the performance of devices such as infiltration trenches, permeable pavements, and detention basins. Machine Learning, on the other hand, offers predictive and analytical tools that enhance the efficiency of planning and the performance of SUDS by analyzing hydrological data. This study aims to develop a workflow using BIM and Machine Learning, through Recurrent Neural Networks (RNN) and satellite-extracted data, for the design of sustainable urban drainage systems. To this end, Machine Learning techniques were applied using the Python programming language, and visual programming algorithms were also developed to assist in the design of SUDS devices. For validation, a practical application was carried out to compare the three implemented solutions (permeable pavement, infiltration trench, and detention basin). Based on the results, it was possible to identify the potentialities and areas for improvement in the integration between the BIM tool used and Machine Learning, contributing to paving new pathways for this integration.