The Use of Sensitivity Analysis for the Probabilistic-Based Seismic Assessment of Existing Buildings

Abstract eng:
Seismic assessment of existing buildings is treated by international codes and guidelines with semi-probabilistic approaches based on the use of a Confidence Factor (CF) directly related to a target Knowledge Level (KL) of the structure under assessment. Since for existing buildings, the knowledge reached after investigations and deepening could not be complete, the CF is addressed to solve the problem of residual uncertainties still remaining. It has been revealed by many authors the inadequacy of such approach, since it does not cover the probabilistic aspect of the problem in a robust way, and in some cases it is leading to unsafe results. Alternative procedures have been proposed by some authors or in some international and national guidelines. The main idea behind these approaches is to frame the seismic assessment of existing buildings by a fully probabilistic approach capable to take into consideration in an accurate way the propagation of uncertainties in the response of the structure. Although these methods can detect accurately the safety of the structure, they are still computationally demanding and difficult to be integrated in the engineering practice field as standard tool. In this paper, the sensitivity analysis is proposed as a tool to improve the seismic assessment under different points of view: to point out in an explicit way the influence each uncertain parameter has on the structural response; to support the set of an effective investigation plan in order to improve the knowledge of the structure; to compute on basis of a limited number of analyses the essential parameters to pass to a verification based on a probabilistic approach. To the latter aim, results from these analyses are used to determine the median Intensity Measure (IM LS ) and, with the help of the Surface Response technique, its dispersion β LS that define the fragility curve representing the capacity in the safety assessment. The proposed methodology is applied on two cases of study where the values of IM LS and dispersions β LS are calculated and compared with reference values obtained from nonlinear static analyses performed on a large number of models generated using Monte Carlo simulations. Results obtained from the procedure based on the sensitivity analysis show a good estimate of the IM LS and β LS parameters, providing, in terms of annual probabilities of occurrence, safety in almost all cases.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket