Report - PEER - University of California, Berkeley

in predicting damage as well as repair method given an EDP. The steps in thedevelopment process are discussed in the following sections.4.1 Damage versus EDPExperimental data characterizing the progression of damage for the test specimenswere used to generate data sets linking the thirteen damage states with the threeprimary EDPs: drift, number of load cycles and joint shear strain. The functionalEDPs, defined by Eq. 2 and Eq. 3, were calibrated to minimize the dispersion of thedata for all damage states about a line spanning the range of damage states andfunctional values from 0 to 1. Figure 1 shows damage-EDP data for the five EDPs.121212Damage State108642108642108642000.0 2.0 4.0 6.00 10 20 30 40 5000.00 0.02 0.04 0.06 0.08 0.10 0.12(a) drift (b) number of cycles (c) joint shear strain1212Damage State10864210864200.0 0.2 0.4 0.6 0.8 1.0 1.200.0 0.2 0.4 0.6 0.8 1.0 1.2(d) F(D,N) per Eq. 2 (e) F(γ,N) per Eq. 3Figure 1. Damage versus EDP.The scatter of the data in Figure 1 reinforces the need for probabilistic modelslinking EDPs with damage and repair. The variability in these data is due in part tovariability in test specimen design and loading; however, it is due also to the datacollection procedures used in the laboratory. The typical procedure used in earthquaketesting in the laboratory is as follows:1. A half-cycle of loading to a new maximum displacement demand, at whichpoint loading is paused to allow for identification of new cracks and regionsof spalling, measurement of new and existing cracks and picture taking.216