Loading...
Publication
Análise comparativa acerca do potencial de utilização de redes neuronais artificias para estimar danos sísmicos em edifícios tradicionais de alvenaria
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.62 MB | Adobe PDF |
Advisor(s)
Abstract(s)
O presente artigo tem como objetivo discutir estratégias alternativas para estimar danos sísmicos em edifícios tradicionais de alvenaria através da análise comparativa de resultados obtidos recorrendo a duas abordagens distintas: um método simplificado de avaliação da vulnerabilidade sísmica utilizado em inúmeros estudos de avaliação do risco sísmico à escala urbana, e uma abordagem inovadora baseada no uso de redes neuronais artificiais. Com recurso a um vasto conjunto de informação dano pós-evento recolhidos no rescaldo do terramoto que atingiu o arquipélago dos Açores no dia 9 de Julho de 1998, considerou-se uma amostra, composta por 90 edifícios da ilha do Faial, Açores. Esta amostra foi utilizada como input para a obtenção de resultados analíticos, recorrendo às duas abordagens supracitadas e para diferentes intensidades macrossísmicas, IEMS-98, resultados esses que foram então discutidos e comparados tendo em conta o seu nível de aproximação aos danos reais observados.
The present paper aims to discuss alternative strategies to estimate earthquake damage inflicted to traditional masonry buildings through a comparative analysis of the results obtained resorting to two different approaches: a seismic vulnerability index scoring method widely used in the past in numerous large-scale earthquake risk assessment studies, and an innovative approach based on the use of Artificial Neural Networks. Resorting to post-earthquake damage data collected in the aftermath of the magnitude VII earthquake that struck the Azores archipelago (in Portugal) on July 9, 1998, a sample of 90 buildings of Faial Island was considered. was used to generate real damage data for a set of traditional masonry buildings located in the island of Faial. This sample was used as input data to derive the analytical results obtained resorting to the referred approaches for different macroseismic intensities, IEMS-98. These results were then discussed and compared to post-earthquake damage data.
The present paper aims to discuss alternative strategies to estimate earthquake damage inflicted to traditional masonry buildings through a comparative analysis of the results obtained resorting to two different approaches: a seismic vulnerability index scoring method widely used in the past in numerous large-scale earthquake risk assessment studies, and an innovative approach based on the use of Artificial Neural Networks. Resorting to post-earthquake damage data collected in the aftermath of the magnitude VII earthquake that struck the Azores archipelago (in Portugal) on July 9, 1998, a sample of 90 buildings of Faial Island was considered. was used to generate real damage data for a set of traditional masonry buildings located in the island of Faial. This sample was used as input data to derive the analytical results obtained resorting to the referred approaches for different macroseismic intensities, IEMS-98. These results were then discussed and compared to post-earthquake damage data.
Description
Keywords
Vulnerabilidade sísmica Estimativa de danos Edifícios de alvenaria de pedra Índice de Vulnerabilidade Redes neuronais artificiais
Citation
FERREIRA, T.M.; ESTÊVÃO, JM.C.; MAIO, R.; ROMEU, V. (2019) – Análise comparativa acerca do potencial de utilização de redes neuronais artificias para estimar danos sísmicos em edifícios tradicionais de alvenaria. Em 11ª Congresso Nacional de Sismologia e Engenharia Sísmica (Sísmica 2019). p. 1-12.