Bibliographic citations
Guillen, L., Correa, W. (2023). Evaluación de modelos de Machine Learning para determinar la respuesta sísmica no lineal de pórticos de CA en el Perú [Trabajo de Suficiencia Profesional, Universidad Peruana de Ciencias Aplicadas (UPC)]. http://hdl.handle.net/10757/669958
Guillen, L., Correa, W. Evaluación de modelos de Machine Learning para determinar la respuesta sísmica no lineal de pórticos de CA en el Perú [Trabajo de Suficiencia Profesional]. PE: Universidad Peruana de Ciencias Aplicadas (UPC); 2023. http://hdl.handle.net/10757/669958
@misc{renati/409316,
title = "Evaluación de modelos de Machine Learning para determinar la respuesta sísmica no lineal de pórticos de CA en el Perú",
author = "Correa Castillo, Wilfredo Bryan",
publisher = "Universidad Peruana de Ciencias Aplicadas (UPC)",
year = "2023"
}
History has shown the occurrence of disastrous seismic events in our country, this added to the informal construction has increased the need to implement innovative methods to evaluate the structural nonlinear seismic response before construction and after the occurrence of an earthquake, in order to propose structural strengthening measures and make effective and efficient decisions. In this sense, the present research focused on the evaluation of three Machine Learning models with the purpose of determining the best prediction model to determine the behavior of the structural nonlinear seismic response of reinforced concrete frames in Peru. The models analyzed were the Artificial Neural Networks (ANNs), Regression Trees and Gaussian Progressive Regression (GPR) model, for which a database of buildings of 3, 4, 5 and 6 stories aligned to NTE E.030 was made. Seismic Resistant Design, the results obtained from the regression metrics revealed that the Artificial Neural Networks model had a better performance with an R2 of 0.94, it also stood out for its accuracy and ability to effectively explain the variability of the data, and a lower mean absolute error in the predictions with a value of MAE 6.77. Finally, a graphical user interface was performed with the ANNs model, in order to compare it with traditional evaluation softwares such as ETABS and OpenSees, the results indicate that the ANNs based model represents a better option for the evaluation of structural nonlinear response, in terms of efficiency and ability to perform the analysis in less time.
This item is licensed under a Creative Commons License