Structural damage evaluation: theory and applications to earthquake engineering

Abstract

The further development of performance-based earthquake engineering (PBEE) is on the current agenda of the earthquake engineering community. A part of assessing the seismic performance of civil engineering structures involves estimation of seismic damage. The conventional approach to damage estimation is based on fragility functions that relate some chosen parameters of structural response to incurred damage. Therefore, damage prediction is based exclusively on the knowledge of the chosen structural response parameters, meaning that damage analysis is uncoupled from the structural analysis. The structural response parameters selected for use in damage analysis are usually referred to as engineering demand parameters (EDP). In the present study, it is shown that for structural damage estimation, the uncoupled damage analysis has deficiencies that lead to less accurate damage prediction. These shortcomings originate from two sources: first, dependence of practically all EDPs on structural damage and second, inexact damage description. To overcome these deficiencies, another approach to structural damage estimation is proposed. The proposed approach, besides using an EDP, uses all information available from structural analysis that is relevant to the damage to be assessed, implying that damage analysis is coupled with structural analysis. It is shown that utilization of this additional information provides more accurate damage prediction. The difference between the two approaches is studied by comparison of results of damage estimation performed for a 2-D structural model of a reinforced-concrete frame. The results show that difference between uncoupled and coupled damage analysis estimates could be significant and that it depends on specific characteristics of the chosen structural model and the damage model in a complex way, preventing the possibility of estimating this error in a general form that is applicable to all practically possible cases. Damage estimation is performed for various damage models that include both single and multiple damage states. Since the final goal of seismic performance evaluation is estimation of decision variables such as repair cost, downtime, etc., the two approaches to damage estimation are also compared in terms of repair cost that is calculated for the reinforced-concrete frame. A case where structural damage prediction is based on observation of EDP alone, without a structural model available, is also studied. It is shown that incorporating site-specific information can significantly change the damage estimates and, therefore, may be worth doing.