Abstract

Incremental Dynamic Analysis (IDA) has been applied lately in order to evaluate the maximum lateral strength of buildings under variable seismic intensity. This variation on intensity is achieved by means of the progressive scaling of the amplitudes of the dynamic signal imposed to the models of the buildings. This numerical procedure was used in order to assess the seismic performance, via the assessment of the behavior factor q (used in many seismic codes worldwide) of the recent version of the Ecuadorian Construction Code. Likewise, the used approach is based on the adaptation of the FEMA P-695 Methodology including issues like the seismic input selection and the definition of the collapse threshold. In order to perform the above mentioned assessment, a set of archetypes representatives of the design space have been defined. These archetypes were designed according three different procedures: acceleration-based and displacement-based procedures and an energy-based alternative procedure. Results depict that some of the adopted values for the behavior factor does not fulfill the requirements of the methodology used, so the design is not able to guarantee an adequate performance of the buildings when they withstand destructive earthquakes.

Abstract

Behavior factor of the structures is a coefficient that includes the inelastic performance of the structure and indicates the hidden resistance of the structure in the inelastic stage. In most seismic codes, this coefficient is merely dependent on the type of lateral resistance system and is introduced with a fixed number. However, there is a relationship between the behavior factor, ductility (performance level), structural geometric properties, and type of earthquake (near and far). In this paper, a new optimal correlation is attempted to predict the behavior factor (''q'') of EBF steel frames, under near-fault earthquakes, using ''Particle Swarm Optimization (PSO)'' and ''Simulated Annealing (SA)'' algorithms. For this purpose, a databank consists of 12960 data created. To establishing different geometrical properties of models, 3-,6-, 9-, 12-, 15, 20- stories steel EBF frames considered with 3 different types of link beam, 3 different types of column stiffness and 3 different types of brace slenderness. Using nonlinear time history under 20 near-fault earthquake, all models analyzed to reach 4 different performance level. 6769 data were used as training data of the ''Adaptive Neural Networks (ANNs).'' Moreover, to validate the algorithms, 2257 data (as test data) were used for calculating mean squared error (''MSE'') and correlation coefficient (''R'') between the predicted values of (''q'') and the real values. Results shows the high accuracy of proposed correlation, established by PSO algorithm.