The importance of strongmotion data in engineering seismology

The importance of strong-motion data in engineering seismology and earthquake engineering Roberto PAOLUCCI Department of Structural Engineering Politecnico di Milano, ITALY

Outline Influence of strong motion processing on numerical simulations of soil-structure interaction problems Selection of real accelerograms based on displacement-spectrum compatibility Roberto Paolucci 2

Influence of strong motion processing on numerical simulations of soil-structure interaction problems ü A benchmark problem Seismic response analysis of a diaphragm wall (Foti and Paolucci, 2012) Roberto Paolucci 3

Influence of strong motion processing on numerical simulations of soil-structure interaction problems Selection of input records for non-linear time-history analyses Corrected acceleration time histories from: 1) 2) 3) European Strong Motion Database ITalian ACelerometric Archive PEER strong motion database Roberto Paolucci 4

Influence of strong motion processing on numerical simulations of soil-structure interaction problems ATINA NS record Roberto Paolucci 5

Influence of strong motion processing on numerical simulations of soil-structure interaction problems ATINA NS record ITACA “pad-strip” procedure to safely remove zero-padding and ensure compatibility of SM records tapering + detrend on displacements Roberto Paolucci 6

Influence of strong motion processing on numerical simulations of soil-structure interaction problems Roberto Paolucci 7

Influence of strong motion processing on numerical simulations of soil-structure interaction problems Atina Roberto Paolucci 8

Influence of strong motion processing on numerical simulations of soil-structure interaction problems Bagnoli Roberto Paolucci 9

Influence of strong motion processing on numerical simulations of soil-structure interaction problems Displacement time histories at top of the wall – Atina record Roberto Paolucci 10

Influence of strong motion processing on numerical simulations of soil-structure interaction problems peak values of displacement and bending moment – Atina record Roberto Paolucci 11

Influence of strong motion processing on numerical simulations of soil-structure interaction problems Atina Bagnoli Roberto Paolucci 12

Influence of strong motion processing on numerical simulations of soil-structure interaction problems 13 Summary ü Consequences of processing procedure § Limited on bending moments § Relevant (20 -30%) on displacements (important for PBD) ü Differences observed on a relatively rigid system (conservative design with PS approach of EC 8). Likely higher effects on more flexible systems ü “Engineering” rules to avoid gross errors in the use of real accelerograms used as input motion for non-linear dynamic soil-structure interaction analyses: § do not manipulate the corrected record provided by the database; § prefer records corrected by acausal filtering; § prefer digital records; § check, before the numerical simulation, that velocities and displacements resulting by integration of the input acceleration are not affected by unphysical drifts. Roberto Paolucci

Outline Influence of strong motion processing on numerical simulations of soil-structure interaction problems Selection of real accelerograms based on displacement-spectrum compatibility Roberto Paolucci 14

Selection of real accelerograms based on displacementspectrum compatibility 15 Introductory works at Politecnico from 2000 to 2010, for characterization of long period ground motion ü displacement spectra at long periods (→ Faccioli et al. , 2004) ü Study on the reliability of long period spectral ordinates from digital accelerograms (→ Paolucci et al. , 2008) ü GMPE at long periods (→ Cauzzi and Faccioli, 2008) ü PSHA at long periods for Italian sites (→ Faccioli and Villani, 2009) Roberto Paolucci

16 Reliability of long-period response spectral ordinates from digital accelerograms Morge Sep 8, 2005 Pennine Alps (MW 4. 4, Re=17 km) Zihuatanejo Jan 11, 1997 Michoacán (MW 7. 1, Re=143 km) After Paolucci et al. , 2008 "the elastic spectra from the most basic processing, in which only the pre-event mean is removed from the acceleration time series, do not diverge from the baseline-corrected spectra until periods of 10– 20 sec (. . . ) Akkar and Boore (2009) Roberto Paolucci

17 Probabilistic seismic hazard studies in Italy Seismic Hazard Map of Italy (0 – 2 s) DPC-INGV Project S 1 – 20052007 http: //esse 1. mi. ingv. it Roberto Paolucci

Probabilistic seismic hazard studies in Italy 18 Long period PSHA in Italy: maps of D 10 DPC-INGV Project S 5 – 20052007 Faccioli and Villani, 2009 Roberto Paolucci

Probabilistic seismic hazard studies in Italy Long period PSHA in Italy: map of TD Roberto Paolucci 19

Probabilistic seismic hazard studies in Italy How to match short and long period PSHA results and put them in a format for engineering applications ? → towards a target displacement spectrum for Italian sites (TDSI) Roberto Paolucci 20

Selection of real accelerograms based on displacementspectrum compatibility Joint research activity of Politecnico di Milano & Università Federico II Napoli Target displacement spectra for Italian sites SIMBAD: a database for engineering analyses of long period ground motion Software REXEL-DISP v 1. 1 Examples of application Roberto Paolucci 21

A Target Displacement Spectrum for Italian Sites Broadband displacement spectrum for design, matching the Italian NTC 08 regulations at short periods with the long period PSHA D 10, TD from Project S 5, while ag, , S, Cc, F 0, TC, TE, TF come from NTC 08 : factor introduced to match short and long periods ( =1 for constant velocity) F: long period site factor (from S 5 project) Roberto Paolucci 22

A Target Displacement Spectrum for Italian Sites Italian norms NTC 08 Long period PSHA Connecting branch T-α Roberto Paolucci 23

24 A Target Displacement Spectrum for Italian Sites = 0. 85 -1. 4 Roberto Paolucci

25 A Target Displacement Spectrum for Italian Sites Comparison of TDSI with NTC 08 (Italian seismic regulations) TDSI NTC 08 A: VS 30 = 800 m/s; B: VS 30 = 580 m/s; C/E: VS 30 = 270 m/s; D: VS 30 = 140 m/s Roberto Paolucci

27 A Target Displacement Spectrum for Italian Sites Site factors NTC 08 TDSI Roberto Paolucci

Selection of real accelerograms based on displacementspectrum compatibility Target displacement spectra for Italian sites SIMBAD: a database for engineering analyses of long period ground motion Software REXEL-DISP v 1. 1 Examples of application Roberto Paolucci 31

EC 8 Site classes Worldwide regions SIMBAD: Selected Input Motions for displacement-Based Assessment and Design Roberto Paolucci 32

Selection of real accelerograms based on displacementspectrum compatibility Target displacement spectra for Italian sites SIMBAD: a database for engineering analyses of long period ground motion Software REXEL-DISP v 1. 1 Examples of application Roberto Paolucci 33

34 Software REXEL-DISP v 1. 1 (www. reluis. it) available at http: //www. reluis. it Roberto Paolucci

Selection of real accelerograms based on displacementspectrum compatibility Target displacement spectra for Italian sites SIMBAD: a database for engineering analyses of long period ground motion Software REXEL-DISP v 1. 1 Examples of application Roberto Paolucci 35

Examples of application Dependence on the target spectrum (NTC 08 vs TDSI) Aquila, TR = 475 years Roberto Paolucci 36

Examples of application Dependence on seismicity level (Aquila vs Udine, TR = 475 yr, TDSI) Roberto Paolucci 37

Examples of application Broadband compatibility (Aquila, TR = 475 years, TDSI) Roberto Paolucci 38

39 Practical hints for using REXEL-DISP Type of application Hints Search for 7 one- or two- - preference to unscaled records component displacement - use of wide magnitude and distance intervals (e. g. , default -spectrum compatible values: 5 -7 and 0 -30 km) and any site class (due to the accelerograms limited number of records on some soil types in the SIMBAD database) - ensure spectral matching over a rather broad range of vibration periods (e. g. , default values: 0. 5 -8 s) - when searching for scaled records the use of limited magnitude and distance range is found to be more feasible. Search for displacement- - preference to unscaled records spectrum compatible - use of wide magnitude and distance intervals (e. g. , default individual records values: 5 -7 and 0 -30 km) and any site class - limit spectral compatibility to relatively small period ranges (e. g. , 1 -3 s) Search for 30 - preference to unscaled records when selecting 30 onedisplacement-spectrum component records but the use of scale factors is advisable compatible when searching for 30 two-component records. accelerograms - use of wide magnitude intervals (e. g. , default values: 5 -7) and any site class - ensure spectral matching over a rather broad range of vibration periods (e. g. , default values: 0. 5 -8 s) - when searching for scaled records the use of limited Roberto Paolucci

40 Concluding remarks Why using spectral displacements as a target for ground motion selection? ü the target magnitude range is “naturally” satisfied; ü no need to scale accelerograms; ü a broadband spectral compatibility is easily achieved (→ NLTHA of MDOF systems – non-linear dynamic SSI – soil stability problems). . . but. . . ü the accelerograms should be selected from high-quality strong-motion databases, covering the seismic hazard levels and site conditions of interest; ü the target spectrum should be carefully defined based on seismic hazard studies at long periods Roberto Paolucci

41 Thank you ! Roberto Paolucci