MOTIONS for SPECIF IFIED EARTHQUAKE SOURCE and SIT ITE - - PowerPoint PPT Presentation

β–Ά
motions for specif ified earthquake
SMART_READER_LITE
LIVE PREVIEW

MOTIONS for SPECIF IFIED EARTHQUAKE SOURCE and SIT ITE - - PowerPoint PPT Presentation

Nov 18, 2022 306 likes β€’706 views

SIM IMULATION of NEAR-FAULT GROUND MOTIONS for SPECIF IFIED EARTHQUAKE SOURCE and SIT ITE CHARACTERISTICS Mayssa Dabaghi Visiting Assistant Professor, American University of Beirut (AUB) Abdurrahman Almikati Graduate Student, American

slide-1
SLIDE 1

SIM IMULATION of NEAR-FAULT GROUND MOTIONS for SPECIF IFIED EARTHQUAKE SOURCE and SIT ITE CHARACTERISTICS

Mayssa Dabaghi

Visiting Assistant Professor, American University of Beirut (AUB)

Abdurrahman Almikati

Graduate Student, American University of Beirut (AUB)

Armen Der Kiureghian

President, American University of Armenia Taisei Professor of Civil Engineering Emeritus, UC Berkeley

Funding: Caltrans, PEER, AUB

8/17/2015 - ESG5, Taipei, Taiwan

slide-2
SLIDE 2

Motivation

  • Characteristics of near-fault (NF) ground motions (GM) are different

from far-field GMs

2

(figure courtesy of Y. Bozorgnia, adopted from Somerville et al. [1997]). (after Somerville et al. [1997]).

Rupture directivity effect Fling Step

slide-3
SLIDE 3

Motivation

  • NF effects not properly represented in modern codes and current

ground motion prediction equations (GMPEs).

  • May underestimate:
  • Elastic demands on long period structures
  • Damage potential to ductile short-period structures
  • Attempt to develop NGA West2 directivity models (Spudich et al.,

2014)

3

NF records from 2014 South Napa EQ vs. code spectra and UHS (directivity and/or site effects)

(source: Bray et al., 2014).

slide-4
SLIDE 4

Motivation

  • Recorded NF GMs remain scarce
  • Ongoing effort to understand, model, and simulate NF GMs and their

effects on the response of structures

  • Proposed Methodology:
  • Site-based stochastic model of NF GMs in 2 orthogonal horizontal directions
  • Simulation procedure for specified EQ source and site characteristics

References:

  • Dabaghi M, Der Kiureghian A. (2016). Stochastic model for simulation of near-fault ground motions, submitted manuscript.
  • Dabaghi M, Der Kiureghian A. (2014). Stochastic modeling and simulation of near-fault ground motions for performance-

based earthquake engineering, PEER Report No. 2014/20, Pacific Earthquake Engineering Research Center, University of California, Berkeley, CA.

  • Rezaeian S, Der Kiureghian A. (2010). Simulation of synthetic ground motions for specified earthquake and site
  • characteristics. Earthquake Engineering & Structural Dynamics; 39 (10): 1155-1180.

4

slide-5
SLIDE 5

Outline

  • Overview of NF Ground Motion Model
  • Overview of Simulation Framework
  • Illustrative Examples
  • #1: Forward vs. Backward Directivity Sites
  • #2: Hypothetical M6.5 Strike-Slip Earthquake Event
  • Summary and Conclusions

5

slide-6
SLIDE 6

Overview of f Stochastic Model of NF Ground Motion

Pulse-like and Non-pulse-like

6

slide-7
SLIDE 7

Puls lse-like GM Model

  • Pulse in at least one horizontal direction, due to rupture directivity…
  • Model formulated in:
  • direction of the largest pulse: 𝑏𝑄𝑄(𝑒)
  • corresponding orthogonal direction: 𝑏𝑄𝑃(𝑒)

10 20 30 .5 .5

Simulated

10 20 30 40 .5 .5

𝑏𝑄𝑄(𝑒) 𝑏𝑄𝑃(𝑒)

slide-8
SLIDE 8

10 20 30 40 .5 .5

Puls lse-like GM Model

  • Pulse in at least one horizontal direction, due to rupture directivity
  • Model formulated in:
  • direction of the largest pulse: 𝑏𝑄𝑄(𝑒)
  • corresponding orthogonal direction: 𝑏𝑄𝑃(𝑒)

8

10 20 30 40 .5 .5

Simulated

10 20 30 40 .5 .5

mMP: modified Mavroeidis & Papageorgiou (2003) pulse model (π‘Š

π‘ž, π‘ˆ π‘ž, 𝛿, πœ‰, 𝑒𝑛𝑏𝑦,π‘ž)

MFW: modified Rezaeian & Der Kiureghian (2010) modulated and filtered white noise model (𝐽𝑏, 𝐸5βˆ’95, 𝐸0βˆ’5, 𝐸0βˆ’30, 𝑔

𝑛𝑗𝑒, 𝑔′, πœ‚π‘”)

𝑏𝑄𝑄 𝑒 π‘π‘žπ‘£π‘š(𝑒) 𝑏𝑠𝑓𝑑(𝑒)

10 20 30 .5 .5

Simulated

10 20 30 40 .5 .5

𝑏𝑄𝑄(𝑒) 𝑏𝑄𝑃(𝑒)

slide-9
SLIDE 9

10 20 30 .5 .5

Simulated

Puls lse-like GM Model

  • Pulse in at least one horizontal direction, due to rupture directivity
  • Model formulated in:
  • direction of the largest pulse: 𝑏𝑄𝑄(𝑒)
  • corresponding orthogonal direction: 𝑏𝑄𝑃(𝑒)

9

MFW: modified Rezaeian & Der Kiureghian (2010) modulated and filtered white noise model (𝐽𝑏, 𝐸5βˆ’95, 𝐸0βˆ’5, 𝐸0βˆ’30, 𝑔

𝑛𝑗𝑒, 𝑔′, πœ‚π‘”)

𝑏𝑄𝑃 𝑒

10 20 30 .5 .5

Simulated 10 20 30 .5 .5

Simulated

10 20 30 40 .5 .5

𝑏𝑄𝑄(𝑒) 𝑏𝑄𝑃(𝑒)

slide-10
SLIDE 10

Puls lse-like GM Model

10

Table 1. Complete list of the parameters 𝛽𝑄,𝑗 of the pulse-like model, 𝑗 = 1, … ,19. Pulse 𝛽𝑄,1 𝛽𝑄,2 𝛽𝑄,3 𝛽𝑄,4 𝛽𝑄,5 π‘Š

π‘ž(cm/s)

π‘ˆ

π‘ž(s)

Ξ³ Ξ½/Ο€(rad) 𝑒𝑛𝑏𝑦 ,π‘ž(s) Residual 𝛽𝑄,6 𝛽𝑄,7 𝛽𝑄,8 𝛽𝑄,9 𝛽𝑄,10 𝛽𝑄,11 𝛽𝑄,12 𝐽𝑏,𝑠𝑓𝑑(cm/s) 𝐸5βˆ’95,𝑠𝑓𝑑(s) 𝐸0βˆ’5,𝑠𝑓𝑑(s) 𝐸0βˆ’30,𝑠𝑓𝑑(s) 𝑔

𝑛𝑗𝑒 ,𝑠𝑓𝑑 (Hz) 𝑔 𝑠𝑓𝑑 β€² (Hz/s) πœ‚π‘”,𝑠𝑓𝑑

Orthogonal 𝛽𝑄,13 𝛽𝑄,14 𝛽𝑄,15 𝛽𝑄,16 𝛽𝑄,17 𝛽𝑄,18 𝛽𝑄,19 𝐽𝑏,𝑄𝑃(cm/s) 𝐸5βˆ’95,𝑄𝑃(s) 𝐸0βˆ’5,𝑄𝑃(s) 𝐸0βˆ’30,𝑄𝑃(s) 𝑔

𝑛𝑗𝑒 ,𝑄𝑃(Hz)

𝑔

𝑄𝑃 β€² (Hz/s)

πœ‚π‘”,𝑄𝑃

Complete list of the parameters 𝛽𝑄,𝑗 of the pulse-like model, 𝑗 = 1, … , 19.

π‘π‘žπ‘£π‘š 𝑒 : 𝑏𝑠𝑓𝑑 𝑒 : 𝑏𝑄𝑃 𝑒 :

slide-11
SLIDE 11

Non-pulse-like GM Model

  • No pulse in any horizontal direction
  • Formulated in:
  • Major principal direction: 𝑏𝑂𝑄1(𝑒)
  • Intermediate principal direction: 𝑏𝑂𝑄2(𝑒)

11

30 40 50 60 .5 .5 30 40 50 60 .5 .5

MFW: modified Rezaeian & Der Kiureghian (2010) modulated and filtered white noise model (𝐽𝑏, 𝐸5βˆ’95, 𝐸0βˆ’5, 𝐸0βˆ’30, 𝑔

𝑛𝑗𝑒, 𝑔′, πœ‚π‘”)

𝑏𝑂𝑄1(𝑒) 𝑏𝑂𝑄2(𝑒) 𝑏𝑂𝑄1(𝑒) 𝑏𝑂𝑄2(𝑒)

slide-12
SLIDE 12

Non-pulse-like GM Model

12

Table 1. Complete list of the parameters 𝛽𝑂𝑄,𝑗 of the non-pulse-like model, 𝑗 = 1, … ,14. Major 𝛽𝑂𝑄,1 𝛽𝑂𝑄,2 𝛽𝑂𝑄,3 𝛽𝑂𝑄,4 𝛽𝑂𝑄,5 𝛽𝑂𝑄,6 𝛽𝑂𝑄,7 𝐽𝑏,𝑂𝑄1(cm/s) 𝐸5βˆ’95,𝑂𝑄1(s) 𝐸0βˆ’5,𝑂𝑄1(s) 𝐸0βˆ’30,𝑂𝑄1(s) 𝑔

𝑛𝑗𝑒 ,𝑂𝑄1(Hz) 𝑔 𝑂𝑄1 β€²

(Hz/s) πœ‚π‘”,𝑂𝑄1 Intermediate 𝛽𝑂𝑄,8 𝛽𝑂𝑄,9 𝛽𝑂𝑄,10 𝛽𝑂𝑄,11 𝛽𝑂𝑄,12 𝛽𝑂𝑄,13 𝛽𝑂𝑄,14 𝐽𝑏,𝑂𝑄2(cm/s) 𝐸5βˆ’95,𝑂𝑄2(s) 𝐸0βˆ’5,𝑂𝑄2(s) 𝐸0βˆ’30,𝑂𝑄2(s) 𝑔

𝑛𝑗𝑒 ,𝑂𝑄2(Hz) 𝑔 𝑂𝑄2 β€²

(Hz/s) πœ‚π‘”,𝑂𝑄2

Complete list of the parameters 𝛽𝑂𝑄,𝑗 of the non-pulse-like model, 𝑗 = 1, … , 14.

𝑏𝑂𝑄1(𝑒): 𝑏𝑂𝑄2 𝑒 :

slide-13
SLIDE 13

Overview of f Simulation Framework

Simulation of 2 orthogonal NF GM components

13

slide-14
SLIDE 14

Predictive Models

  • Input:
  • EQ source and site characteristics (𝐺, 𝑁π‘₯, π‘Žπ‘ˆπ‘ƒπ‘†, 𝑆𝑆𝑉𝑄, π‘Š

𝑑30, 𝑑𝑝𝑠𝑒, πœ„π‘π‘ πœš)

  • Probability of occurrence of a pulse (Shahi and Baker, 2011)
  • Empirical predictive relations:
  • for both pulse-like and non-pulse-like model parameters
  • Orientation of the simulated components with respect to fault strike

14

slide-15
SLIDE 15

In Input Parameters (o (or Predictor Vari riables)

  • EQ design scenario:
  • description of the EQ source and site characteristics
  • including directivity parameters

15

Strike-Slip Reverse Reverse-Oblique

(𝐺, 𝑁π‘₯, π‘Žπ‘ˆπ‘ƒπ‘†, 𝑆𝑆𝑉𝑄, π‘Š

𝑑30, 𝑑𝑝𝑠𝑒, πœ„π‘π‘ πœš)

𝑑𝑝𝑠𝑒, πœ„π‘π‘ πœš) For any style of faulting, we use:

𝑑𝑝𝑠𝑒 = max 𝑑, 𝑒 If 𝑑𝑝𝑠𝑒 = 𝑑, πœ„π‘π‘ πœš = πœ„ If 𝑑𝑝𝑠𝑒 = 𝑒, πœ„π‘π‘ πœš = 𝜚

slide-16
SLIDE 16

Puls lse Probability Model

  • Input: (𝐺, 𝑆𝑆𝑉𝑄, 𝑑𝑝𝑠𝑒, πœ„π‘π‘ πœš)
  • Output: Probability of occurrence of a directivity pulse in at least one

direction at a site (Shahi and Baker, 2011)

16

Pr pulselike|𝐺, 𝑆𝑆𝑉𝑄, 𝑑𝑝𝑠𝑒, πœ„π‘π‘ πœš =

1 1+exp (0.642+0.167π‘†π‘†π‘‰π‘„βˆ’0.075𝑑𝑝𝑠𝑒),

𝑗𝑔 𝐺 = 0 =

1 1+exp (0.128+0.055π‘†π‘†π‘‰π‘„βˆ’0.061𝑑𝑝𝑠𝑒+0.036πœ„π‘π‘ πœš),

𝑗𝑔 𝐺 = 1

slide-17
SLIDE 17

Empirical Predictive Relations for r Model Parameters

  • Input: (𝐺, 𝑁π‘₯, π‘Žπ‘ˆπ‘ƒπ‘†, 𝑆𝑆𝑉𝑄, π‘Š

𝑑30, 𝑑𝑝𝑠𝑒)

  • Using: results of regression and correlation analyses
  • Output: simulated parameters of pulse-like and non-pulse-like GM

models

  • model parameters 𝑨𝑄,𝑗, 𝑨𝑂𝑄,𝑗 in the normal space (correlated, and with

natural variability)

  • transformed to original parameter space: 𝛽𝑄,𝑗, 𝛽𝑂𝑄,𝑗

17

Functional forms chosen to be consistent with seismological theory E 𝑨 = 𝛾0 + 𝛾1𝑁π‘₯ + 𝛾2 𝑁π‘₯ βˆ’ 6.5 𝕁 𝑁π‘₯ > 6.5 + 𝛾3𝐺𝑔

π‘”π‘šπ‘’,π‘Ž

+ 𝛾4 ln 𝑆𝑆𝑉𝑄

2

+ β„Ž2 + 𝛾5𝑁π‘₯. ln 𝑆𝑆𝑉𝑄

2

+ β„Ž2 + 𝛾6 ln(π‘Š 𝑑30) + 𝛾7𝑑𝑝𝑠𝑒

slide-18
SLIDE 18

Sim imulation Procedure

18

, , , ,

30, ,

, , , Pulse Probability Model Pulselike? Yes No

30

Pulselike GM Model Non-Pulselike GM Model

30

( ) ( ) ( ) ( ) ( ) ( ) , , , ,

30, ,

e? Yes

30 30

( ) ( ) ( ) ( ) ( ) ( )

slide-19
SLIDE 19

Sim imulation Procedure

19

, , , ,

30, ,

, , , Pulse Probability Model Pulselike? Yes No mMP Pulse Model MFW Model Non-Pulselike Predictive Relations , , , ,

30,

Pulselike GM Model Non-Pulselike GM Model , , , ,

30

Pulselike Predictive Relations ( ) ( ) ( ) MFW Model ( ) ( ) ( )

Model Parameters

slide-20
SLIDE 20

Sim imulation Procedure

20

, , , ,

30, ,

, , , Pulse Probability Model Pulselike? Yes No mMP Pulse Model MFW Model Non-Pulselike Predictive Relations , , , ,

30,

Pulselike GM Model Non-Pulselike GM Model , , , ,

30

Pulselike Predictive Relations ( ) ( ) ( ) MFW Model ( ) ( ) ( )

10 20 30 40 .5 .5

Simulated

10 20 30 40 .5 .5 10 20 30 .5 .5

Simulated

10 20 30 40 .5 .5

Pulse Motion Residual Motion Orthogonal Motion Motion in Direction of Largest Pulse Model Parameters

slide-21
SLIDE 21

Sim imulation Procedure

21

, , , ,

30, ,

, , , Pulse Probability Model Pulselike? Yes No mMP Pulse Model MFW Model Non-Pulselike Predictive Relations , , , ,

30,

Pulselike GM Model Non-Pulselike GM Model , , , ,

30

Pulselike Predictive Relations ( ) ( ) ( ) MFW Model ( ) ( ) ( )

β€œFling step” Model β€œFling step” Model

slide-22
SLIDE 22

Scope: : Predictive Equations & & Sim imulation Procedure

22

𝑁π‘₯ π‘Žπ‘ˆπ‘ƒπ‘† (km) 𝑆𝑆𝑉𝑄 (km) π‘Š

𝑑30

(m/s) 𝑑𝑝𝑠𝑒 (km) πœ„or𝜚 (Β°) min 5.74 0.07 139 4.97 0.1 max 7.90 5.92 30.49 2016 101.5 67.4

Recorded Pulselike GMs Recorded Non-Pulselike GMs

𝑁π‘₯ π‘Žπ‘ˆπ‘ƒπ‘† (km) 𝑆𝑆𝑉𝑄 (km) π‘Š

𝑑30

(m/s) 𝑑𝑝𝑠𝑒 (km) πœ„or𝜚 (Β°) min 5.50 0.21 361 1.20 0.15 max 7.90 14.50 30.9 1428 135 84.4

6 ≀ 𝑁π‘₯ ≀ 7.5 5 < 𝑆𝑆𝑉𝑄 ≀ 25 km 400 < π‘Š

𝑇30 < 1000 m/s

Shallow Crustal EQs Subsets of PEER NGA-West2 Database Recommended Ranges

slide-23
SLIDE 23

Il Illustrative Example # 1

Forward vs. Backward Directivity Sites

23

slide-24
SLIDE 24

Forward vs. . Backward Dir irectivity Sit ites

24

𝑺𝑺𝑽𝑸 = 𝟐𝟏 km 𝑺𝑺𝑽𝑸 = 𝟐𝟏 km 𝒕 = πŸ•πŸ km

FD site BD site

Directivity 7.0 10 760 FD 60 9.5 0.90 BD 90 0.09

slide-25
SLIDE 25

Sim imulated Tim ime Series - Acceleration

25

30 40 50 60

  • 0.5

0.5 Largest Pulse Component 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5 Time (s) 30 40 50 60

  • 0.5

0.5 Orthogonal Component 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5

Acceleration (g)

30 40 50 60

  • 0.5

0.5 Time (s) 30 40 50 60

  • 0.5

0.5 Largest Pulse Component 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5 Time (s) 30 40 50 60

  • 0.5

0.5 Orthogonal Component 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5 30 40 50 60

  • 0.5

0.5

Acceleration (g)

30 40 50 60

  • 0.5

0.5 Time (s)

Major Component Intermediate Component

Directivity 7.0 10 760 FD 60 9.5 0.90 BD 90 0.09

slide-26
SLIDE 26

30 40 50 60

  • 100

100 Largest Pulse Component 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100 Time (s) 30 40 50 60

  • 100

100 Orthogonal Component 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100

Velocity (cm/s)

30 40 50 60

  • 100

100 Time (s) 30 40 50 60

  • 100

100 Largest Pulse Component 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100 Time (s) 30 40 50 60

  • 100

100 Orthogonal Component 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100 30 40 50 60

  • 100

100

Velocity (cm/s)

30 40 50 60

  • 100

100 Time (s)

Sim imulated Tim ime Series - Velocity

26 Major Component Intermediate Component

Directivity 7.0 10 760 FD 60 9.5 0.90 BD 90 0.09 Pulselike Non-Pulselike

30 40 50 60 30 40 50 60

  • 100

100 30 40 50 60 30 40 50 60

  • 100

100 30 40 50 60

Pulselike

Major Component Intermediate Component Largest Pulse Component Orthogonal Component

30 40 50 60

Non-Pulselike

slide-27
SLIDE 27

30 40 50 60

  • 50

50 Largest Pulse Component 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50 Time (s) 30 40 50 60

  • 50

50 Orthogonal Component 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50

Displacement (cm)

30 40 50 60

  • 50

50 Time (s) 30 40 50 60

  • 50

50 Largest Pulse Component 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50 Time (s) 30 40 50 60

  • 50

50 Orthogonal Component 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50 30 40 50 60

  • 50

50

Displacement (cm)

30 40 50 60

  • 50

50 Time (s)

Sim imulated Tim ime Series - Dis ispla lacement

27 Major Component Intermediate Component

Directivity 7.0 10 760 FD 60 9.5 0.90 BD 90 0.09

slide-28
SLIDE 28

Il Illustrative Example # 2

Hypothetical M6.5 Strike-Slip Earthquake Event

28

slide-29
SLIDE 29

Earthquake Desig ign Scenarios

  • Hypothetical earthquake event
  • 𝐺 = 0, 𝑁π‘₯ = 6.5, π‘Žπ‘ˆπ‘ƒπ‘† = 0
  • 𝑀R = 29 km (Wells and Coppersmith, 1994)
  • Hypocenter locations:
  • uniformly distributed along the strike (buffer zones of length

0.1 𝑀R at the ends)

  • clustered into 4 discrete locations (A, B, C, and D)
  • Site locations:
  • 𝑆𝑆𝑉𝑄 = 10 km, π‘Š

𝑑30 = 760 m/s

  • 5 sites (1,2,…, 5) uniformly distributed around fault rupture

29

slide-30
SLIDE 30

Earthquake Desig ign Scenarios

  • 20 equally probable scenarios (source-site combinations):
  • 𝐺 = 0, 𝑁π‘₯ = 6.5, π‘Žπ‘ˆπ‘ƒπ‘† = 0 km, 𝑆𝑆𝑉𝑄 = 10 km, π‘Š

𝑑30 = 760 m/s

  • For each scenario, calculate:

𝑑𝑝𝑠𝑒 = 𝑑, πœ„π‘π‘ πœš = πœ„ Pr pulse βˆ’ like|𝐺, 𝑆𝑆𝑉𝑄, 𝑑𝑝𝑠𝑒, πœ„π‘π‘ πœš

30 Source A B C D Site 1 𝑑𝑝𝑠𝑒 = 2.9 km πœ„π‘π‘ πœš = 35.5Β° 𝑄 = 11.0% 𝑑𝑝𝑠𝑒 = 10.6 km πœ„π‘π‘ πœš = 21.7Β° 𝑄 = 18.0% 𝑑𝑝𝑠𝑒 = 18.4 km πœ„π‘π‘ πœš = 15.5Β° 𝑄 = 28.3% 𝑑𝑝𝑠𝑒 = 26.1 km πœ„π‘π‘ πœš = 12.0Β° 𝑄 = 41.2% π‘ΈπŸ = πŸ‘πŸ“. πŸ•% 2 𝑑𝑝𝑠𝑒 = 13.4 km πœ„π‘π‘ πœš = 36.7Β° 𝑄 = 21.3% 𝑑𝑝𝑠𝑒 = 5.7 km πœ„π‘π‘ πœš = 60.4Β° 𝑄 = 13.2% 𝑑𝑝𝑠𝑒 = 2.1 km πœ„π‘π‘ πœš = 78.3Β° 𝑄 = 10.4% 𝑑𝑝𝑠𝑒 = 9.8 km πœ„π‘π‘ πœš = 45.6Β° 𝑄 = 17.1% π‘ΈπŸ‘ = πŸπŸ”. πŸ”% 3 𝑑𝑝𝑠𝑒 = 26.1 km πœ„π‘π‘ πœš = 4.7Β° 𝑄 = 41.2% 𝑑𝑝𝑠𝑒 = 18.4 km πœ„π‘π‘ πœš = 6.0Β° 𝑄 = 28.3% 𝑑𝑝𝑠𝑒 = 10.6 km πœ„π‘π‘ πœš = 8.3Β° 𝑄 = 18.0% 𝑑𝑝𝑠𝑒 = 2.9 km πœ„π‘π‘ πœš = 13.2Β° 𝑄 = 11.0% π‘ΈπŸ’ = πŸ‘πŸ“. πŸ•% 4 𝑑𝑝𝑠𝑒 = 20.7 km πœ„π‘π‘ πœš = 25.8Β° 𝑄 = 31.8% 𝑑𝑝𝑠𝑒 = 12.9 km πœ„π‘π‘ πœš = 37.7Β° 𝑄 = 20.7% 𝑑𝑝𝑠𝑒 = 5.2 km πœ„π‘π‘ πœš = 62.7Β° 𝑄 = 12.7% 𝑑𝑝𝑠𝑒 = 2.5 km πœ„π‘π‘ πœš = 75.7Β° 𝑄 = 10.7% π‘ΈπŸ“ = 𝟐𝟘. 𝟏% 5 𝑑𝑝𝑠𝑒 = 2.9 km πœ„π‘π‘ πœš = 70.6Β° 𝑄 = 11.0% 𝑑𝑝𝑠𝑒 = 10.6 km πœ„π‘π‘ πœš = 41.6Β° 𝑄 = 18.0% 𝑑𝑝𝑠𝑒 = 18.4 km πœ„π‘π‘ πœš = 27.7Β° 𝑄 = 28.3% 𝑑𝑝𝑠𝑒 = 26.1 km πœ„π‘π‘ πœš = 20.5Β° 𝑄 = 41.2% π‘ΈπŸ” = πŸ‘πŸ“. πŸ•% 𝑸𝑩 = πŸ‘πŸ’. πŸ’% 𝑸π‘ͺ = 𝟐𝟘. πŸ•% 𝑸𝑫 = 𝟐𝟘. πŸ”% 𝑸𝑬 = πŸ‘πŸ“. πŸ‘%

slide-31
SLIDE 31

Forward vs. . Backward Dir irectivity Scenarios

31

Geometric mean spectra at 5% damping

  • NF Median and Median +/- 𝜏 of 20x300 simulations
  • A3 Median of 300 simulations
  • D3 Median of 300 simulations
slide-32
SLIDE 32

Near-Fault Sim imulations vs. . GMPEs (N

(NGA West t 2)

32

Forward directivity scenario Backward directivity scenario

Geometric mean spectra at 5% damping

  • A3 Median and Median +/- 𝜏 of 300 simulations
  • D3 Median and Median +/- 𝜏 of 300 simulations

RotD50 spectra at 5% damping

  • NGA-West2 GMPEs Median and Median +/- 𝜏
slide-33
SLIDE 33

Near-Fault vs. . Far r Fie ield Sim imulations (R

(Rezaeia ian and Der r Kiu Kiureghia ian, , 2010)

33

Geometric mean spectra at 5% damping

  • NF Median and Median +/- 𝜏 of 20x300 simulations
  • FF Median and Median +/- 𝜏 of 300 simulations (Rezaeian and Der Kiureghian, 2010)

RotD50 spectra at 5% damping

  • NGA-West2 GMPEs Median and Median +/- 𝜏
slide-34
SLIDE 34

Summary ry and Conclusions

34

slide-35
SLIDE 35

Summary ry and Conclusions (1/3)

  • Parameterized stochastic model of NF GM in 2 orthogonal horizontal

directions

  • Empirical (site-based) simulation method
  • practical and simple, computationally efficient
  • input can be readily available to the design engineer
  • forward directivity (FD) vs. backward directivity (BD) conditions (through 𝑑𝑝𝑠𝑒

and πœ„π‘π‘ πœš)

35

slide-36
SLIDE 36

Summary ry and Conclusions (2/3)

  • Resulting ensemble of synthetic NF GMs
  • capture the important characteristics of recorded NF GMs (rupture directivity,

pulse-like and non-pulse-like, intensity, duration, frequency characteristics)

  • replicate the natural variability
  • can be used in PSHA and PBEE
  • Simulation procedure was illustrated for 20 NF EQ scenarios
  • same (𝐺, 𝑁π‘₯, π‘Žπ‘ˆπ‘ƒπ‘†, 𝑆𝑆𝑉𝑄, π‘Š

𝑑30)

  • range of different rupture directivity conditions

36

slide-37
SLIDE 37

Summary ry and Conclusions (3/3)

  • FD vs. BD simulations:
  • median GM level at a site depends on the directivity configuration
  • forward directivity sites: highest amplitudes in the long-period range
  • NF GM simulations vs. NGA-West2 GMPEs
  • GMPEs ~ backward directivity scenarios
  • at forward directivity sites, GMPEs predict lower GM levels at longer periods
  • GMPEs do not adequately represent the rupture directivity effect
  • NF vs. FF GM simulations
  • far-field model is not well constrained at a distance of 10 km

37

slide-38
SLIDE 38

Thank You!

mayssa.dabaghi@aub.edu.lb

38