Istanbul, August 24-29, 2014

1. “Real-time” seismic vulnerability assessment of a high rise RC building using field monitoring data Sotiria Karapetrou Maria Manakou DespoinaLamprou Sofia Kotsiri Kyriazis Pitilakis Aristotle University of Thessaloniki Istanbul, August 24-29, 2014

2. Introduction • Aim of this study: “real-time” seismic vulnerability assessment of RC building using field monitoring data reflecting the actual state of the structure (degradation due to time, possible pre-existing damage, changes in geometry and mass distribution etc.) • EU project REAKT (Strategies and Tools for Real Time EArthquakeRisKReducTion): Rapid post-earthquake assessment of buildings based on field monitoring data • Target structure: high-rise RC hospital building in Thessaloniki

3. Methodological framework Fragility assessment of buildings using field monitoring data Finite element modeling (FEM) Operational modal analysis (OMA) Evaluation of MAC values Comparison between numerical and experimental modes Sensitivity in material properties Finite element model updating Selection of the “best” FE model Nonlinear incremental dynamic analysis Derivation of “real – time” fragility curves

4. Description of the hospital building in Thessaloniki Target building AHEPA Hospital complex • Target building: high-rise (8-storey) RC infilled MRF structuredesigned with low seismic code level (SYNER-G taxonomy). • It hosts both administration and hospitalization activities. • It is composed of two adjacent tall building units that are connected with a structural joint • The foundation consists of simple footings without tie-beams combined partially with a raft foundation.

5. Description of the hospital building

6. Temporary instrumentation array • February 2013: ambient noise measurements (AUTH, GFZ) N S • 36 triaxial seismometers: • → Mark short-period seismometers (L4C-3D, 1Hz natural frequency) • → EarthData recorders EDL (PR6-24) Basement • 4 stations at each floor installed along the middle corridor of the building near and far the structural joint. 4th floor UNIT 1 UNIT 2 UNIT 1 UNIT 2 • North – South (NS) ‖ longitudinal direction of the structure Top floor • 4 hour recordings • → Sampling rate 500Hz UNIT 1 UNIT 2

7. Temporary instrumentation array A A’ Structural joint between the building UNITS Basement • Section A-A’ UNIT 1 UNIT 2 UNIT 1 UNIT 2

8. System identification and Operational modal analysis (ΟΜΑ) • MACEC 3.2 software (Reynders et al. 2011) • OMA for the two adjacent building units separately (UNIT 1 and UNIT 2) and for the entire hospital building analyzed as one (BUILDING). • Grid of the models: the defined nodes correspond to the nodes that are measured. • Non-parametric and parametric identification techniques are applied. • Non-parametric: Frequency Domain Decomposition FDD (Brincker et al. 2001) • Parametric: Stochastic Subspace Identification SSI (Van Overschee and De Moor 1996)

9. System identification and Operational modal analysis Frequency Domain Decomposition – FDD  Singular values Stochastic Subspace Identification – SSI  Stabilization diagrams

10. System identification and Operational modal analysis • Modal identification results for UNIT 1, UNIT 2and BUILDING estimated using parametric (SSI) and non-parametric (FDD) identification techniques

11. Finite element model updating • “Initial” elastic numerical model of the building units: • based on the design and documentation plans provided by the Technical Service of the hospital. • → numerical modeling conducted in OpenSees (Mazzoni et al. 2009) separately for UNIT 1 and UNIT 2 • → Elastic beam-column elements to model the RC elements (beam and columns) • → Elastic truss elements to model the masonry infills: double strut model to represent the in plane behavior of the infill panel. • → Fixed base conditions are assumed for both building units

12. Finite element model updating • Sensitivity parameter: compressive strength of the masonry infill fm • Normal distribution for fm(Mosalam et al. 1997) • → mean valueμ=3MPa • → covariance COV=20% • Masonry compressive strength values calculated based on the mean and standard deviation values of the normal distribution adopted with: μ-3s≤ fm≤μ+3s, s: standard deviation • Elastic modulus in compression of masonry infills computed based on compressive strength: Em= 1000fm (Paulay and Priestley 1992)

13. Finite element model updating • Evaluation of modal assurance criterion MAC values regarding the correlation between numerical and experimental modes and the selection of the “best” updated model φj eigenvector j from numerical model φEi eigenvector i from field monitoring test • Selection of the “best model” based on MAC values (MAC>0.8) →→ Optimal scenario • Emlong1= 3GPa (fm=μ=3MPa) • Emlong2=1.8GPa (fm=μ-2σ=1.8MPa) • Emtransv1=3GPa (fm= μ=3MPa) • Emtransv2=4.8GPa (fm=μ+3σ=4.8MPa) Emlong1 Emlong1 Emtransv1 Emtransv1 Emtransv2 Emtransv2 Emlong2 Emlong2

14. Finite element model updating – UNIT 1

15. Finite element model updating – UNIT 2

16. Inelastic finite element modeling • Finite element code OpenSees(Mazzoni et al. 2009) • Inelastic force-based formulation • Geometric nonlinearity and distributed material plasticity(fiber based approach) • Concrete (confined and unconfined): Popovics concrete material model (1973) • Steel: uniaxial bilinear with kinematic hardening • Masonry infills: inelastic struts assigned with an elastoplastic force-displacement relationship • Diaphragm constraint is employed to account for the rigidity against the in-plane deformation of the floor slabs. • Fixed base conditions are assumed for both structural models

17. Seismic input motion for incremental dynamic analysis (IDA) • 15 real ground motion records from the ESMD (http://www.isesd.hi.is) referring to soil conditions at sites classified as stiff soil according to EC8 (soil type B) • → Selection criteria • moment magnitude: 5.8<Mw<7.2 • epicentral distance: 0<R<45km • average acceleration spectra of the set to be of minimal “epsilon” (Baker and Cornell 2005) at 0<T<2.0sec with respect to the acceleration spectrum adopted from SHARE for a 475 year return period.

18. Incremental dynamic analysis (IDA) • IDA (Vamvatsikos and Cornell 2002 ) initial and updated models • Intensity measure IM : peak ground acceleration PGA • Engineering demand parameter EDP : max interstorey drift ratio maxISD • Two limit states in terms of max interstorey drift maxISD • → Immediate Occupancy (IO): 0.1% according to FEMA-356 for RC infilled MRFs • → Collapse Prevention (CP): assigned on the IDA curve at a point where the IDA is softening towards the flatline but at low enough values of maxISDso that we still trust the model (Vamvatsikos and Cornell 2004). IDA curves of updated UNIT 1 IDA curves of updated UNIT 2

19. Incremental dynamic analysis (IDA) • IDA (Vamvatsikos and Cornell 2002 ) initial and updated models • Intensity measure IM : peak ground acceleration PGA • Engineering demand parameter EDP : max interstorey drift ratio maxISD • Two limit states in terms of max interstorey drift maxISD • → Immediate Occupancy (IO): 0.1% according to FEMA-356 for RC infilled MRFs • → Collapse Prevention (CP): assigned on the IDA curve at a point where the IDA is softening towards the flatline but at low enough values of maxISDso that we still trust the model (Vamvatsikos and Cornell 2004). • CP limit state maxISD values defined on the IDA curve for each building unit (initial and updated)

20. Time-dependent fragility curves • Two – parameter lognormal distribution functions: Where Φ:the standard normal cumulative distribution function IM: the intensity measure of the earthquake expressed in terms of PGA (in units of g), and β: the median values (in units of g) and log-standard deviations respectively of the building fragilities DS: the damage state.

21. Time-dependent fragility curves • Median values PGA: determined based on regression analysis of the IDA results (PGA– maxISD) for each building unit • Uncertainties are taken into account through the log-standard deviation β(t): total dispersion related to each fragility curve • Three primary sources of uncertainty: • βD: seismic demand (variability in the numerical results) • βC: structural capacity (HAZUS=0.3 for low code structures) • βds: definition of damage state (HAZUS=0.4)

22. Fragility curves • Comparative plots of the “initial” fragility curves derived for the two adjacent building units with the corresponding fragility curves provided by Kappos et al. (2003 and 2006)

23. Fragility curves • Comparative plot of the fragility curves derived for the initial and updated models of UNIT 1 and UNIT 2.

24. Conclusions • The present study provides further insight on the assessment of the “real-time” seismic vulnerability of typical RC buildings using field monitoring data, taking into account the actual state of the structure(degradation due to time, possible pre-existing damages, changes in geometry and mass distribution, etc). • The proposed updating procedure can be used to yield more reliable structural models with respect to their real conditions in terms of structural detailing, mass distribution and material properties • The applied methodology in this study may be used in the context of “real-time” risk assessmentand post-seismic fragility updating.

25. The work of this study was carried out in the framework of the ongoing REAKT (http://www.reaktproject.eu/) project, funded by the European Commission, FP7-282862. Thank you for your attention!!