ESTIMATION OF DAMAGE LEVEL AT URBAN SCALE FROM SIMPLE PROXIES ACCOUNTING FOR SOIL AND BUILDING DYNAMIC PROPERTIESReport as inadecuate




ESTIMATION OF DAMAGE LEVEL AT URBAN SCALE FROM SIMPLE PROXIES ACCOUNTING FOR SOIL AND BUILDING DYNAMIC PROPERTIES - Download this document for free, or read online. Document in PDF available to download.

1 ISTerre - Institut des Sciences de la Terre

Abstract : It has been observed repeatedly in the post-earthquake investigations that buildings having frequency similar to soil frequency coming from H-V for example exhibit significantly greater damage due to the double resonator concept Caracas 1967, Mexico 1985, L-Aquila 2009. However this observation is generally not taken directly into account neither in present-day seismic regulations small scale, nor in large-scale seismic risk analysis. We considered a theoretical analysis to study the effect of frequency coincidence between soil and building. As a first step, 887 natural soil profiles with linear behavior are associated to a set of single degree of freedom elastoplastic oscillators. The results obtained are used to quantify the damage increment related to the soil-building frequency coincidence and depending on different parameters such as the loading level characterized by the peak ground acceleration PGA, the soil profile impedance contrast, soil frequency and the building ductility, fundamental frequency. This statistical work is based on Artificial Neural Network ANN approach that does not require any prior knowledge, confirming that the main parameter controlling the damage increase is the ratio structure frequency to soil frequency fstruct-fsoil, with a synaptic weight exceeding 58% when PGA represents 27.05%, the impedance contrast 10.44% and ductility 4.24%. The leading parameter, i.e. the fstruct-fsoil ratio, controls also the damage increment when considering various ductility classes with a synaptic weight percentage of 45%; the parameter that follows is the PGA.

Keywords : spectral coincidence frequency damage vulnerability Neural Network building





Author: C Salameh - P.-Y. Bard - B Guillier - C Cornou -

Source: https://hal.archives-ouvertes.fr/



DOWNLOAD PDF




Related documents