Aplicação de redes neurais na previsão do custo ótimo econômico e ambiental de vigas mistas aço-concreto
সংরক্ষণ করুন:
| প্রধান লেখক: | |
|---|---|
| প্রকাশনার তারিখ: | 2024 |
| বিন্যাস: | Master thesis |
| ভাষা: | por |
| সম্পদ: | Repositório Institucional da UPF |
| Download full: | https://repositorio.upf.br/handle/123456789/10176 |
সংক্ষিপ্ত: | Society, driven by technological advancements and sustainable principles, incessantly seeks solutions in civil engineering for more economical and efficient projects, ensuring structural integrity. An innovative approach is the integration of composite structures with artificial intelligence. The increasing use of steel and concrete composite beams, especially in the Brazilian context, stands out for structural efficiency and cost reduction compared to conventional systems. This efficiency stems from maximizing the advantages and distinct characteristics of each material used. This study proposes a prediction model based on artificial neural networks and grounded in machine learning to size and optimize steel and concrete composite beams, aiming for precise and effective results, reducing costs, and contributing to environmental preservation by minimizing carbon dioxide (CO2) emissions. This process is carried out using the R programming language and the H2O library, with an interface in RStudio. The considered beams are simply supported and subjected to distributed loads, following current national standards, notably NBR 8800 (ABNT, 2008). The prediction model is based on a database with pre-established dimensions drawn by the Monte Carlo method and validated through available literature databases, including results from laboratory tests. This approach highlights the applicability of neural networks in civil engineering. The optimization algorithm's primary function is to define the lowest cost and minimize CO2 emissions, ensuring structural safety and considering the material's ultimate limit states. Performance metrics, R² and RMSE, corroborate the effectiveness of the proposed model, signaling a significant advancement in the field of civil engineering. In the optimization process, good results were found, achieving a reduction of over 50% in total cost and CO2 emissions. The obtained results are promising, demonstrating the validation of the sizing process, safety verification, effectiveness of the prediction model, and the success of the optimization process. |