12.07.2015 Views

Report - PEER - University of California, Berkeley

Report - PEER - University of California, Berkeley

Report - PEER - University of California, Berkeley

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

The four-storey reinforced concrete frame building presented in Section 4.1 is thesubject <strong>of</strong> this example. The structure was modeled numerically and was assumed tobe in a High-seismicity zone in Europe with a hazard compatible with its designseismic action. Response simulations (non-linear models under earthquake inputmotions) were obtained for several input intensities (each using 5 artificiallygenerated accelerograms). The response curve was approximated by an analyticalfunction (average values <strong>of</strong> the simulation results) and a constant c.o.v. <strong>of</strong> 25% wasassumed for the sectional lognormal distribution <strong>of</strong> the response. Performance curveswere obtained for a few different cases (using the same number <strong>of</strong> ‘experimental’points but distributed differently along the intensity ranges to approximate theresponse curves) and subsequently Annualized Earthquake Losses (AEL) werederived. It was concluded that the approximation <strong>of</strong> the structural response curvesrepresents a key component <strong>of</strong> the risk assessment process, with very significantimplications on the values <strong>of</strong> the expected earthquake losses. Approximation shouldbe based on well-distributed ‘experimental points’ covering low, medium and highinput intensities.0.2AEL [o/oo ]0.100 0.1 0.2 0.3 0.4 0.5 0.6 0.7 . 0.8 0.9Damage index rangeInjuries & FatalitiesContents & EconomicRepair / replacement/Figure 1. Contribution <strong>of</strong> damage ranges to total expected AEL.There is another important aspect to take into consideration in the risk assessmentprocess, which is concerned with the contribution <strong>of</strong> the damage ranges to the totalexpected losses. Fig. 1 shows the partial contribution <strong>of</strong> the damage ranges (0–0.05,0.05–0.1, 0.1–0.15, …) to the total expected losses. It is noted that damage states inthe vicinity <strong>of</strong> 0.1 are predominantly contributing to the repair and economic losseswhereas ‘human losses’ are practically constant for all damage ranges other than forthe damage values lower than 0.1, for which they are very limited.The key concluding note is that reduction <strong>of</strong> economic losses is effective in thezones corresponding to low – low/medium damage indices, which can be addressedby the reduction <strong>of</strong> damage-inducing demands (e.g., drifts) corresponding to lowmediuminput intensities with high probability <strong>of</strong> occurrence.284

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!