Anda di halaman 1dari 21
Ground Motion Selection and Scaling

Ground Motion Selection and Scaling

Ground Motion Selection and Scaling for the analysis of the tall building case studies Farzin Zareian

for the analysis of the tall building case studies

and Scaling for the analysis of the tall building case studies Farzin Zareian &PengZhong May 7

Farzin Zareian &PengZhong May 7 th , 2010

and Scaling for the analysis of the tall building case studies Farzin Zareian &PengZhong May 7
San Andreas Sierra Madre (San Fernando) Verdugo Elsinore (Chino) Elsinore (Whittier) Newport-Inglewood-Rose
San Andreas
Sierra
Madre
(San Fernando)
Verdugo
Elsinore
(Chino)
Elsinore
(Whittier)
Newport-Inglewood-Rose Canyon
Sierra Madre
(Cucamonga)
1.5Km, Puente Hills
Santa
7.3Km, Hollywood
Monica
Raymond
Hollywood
8.8Km, Raymond
11.5Km, Santa Monica
24.5Km, Elsinore
40.0Km, Sierra Madre
56Km, San Andreas
Challenges Significance of several modes of vibration in response of the building. Similar ground motions

Challenges

Significance of several modes of vibration in response of the building.

Similar ground motions for all structures.

Five hazard levels needs to be looked at.

A large number of motions are required (we used 15) to have a reasonable estimate of the dispersion in EDP.

Disagg. for 2% / 50, T = 1.0 sec.

Disagg. for 2% / 50, T = 1.0 sec.

Disagg. for 2% / 50, T = 1.0 sec.
Disagg. for 2% / 50, T = 5.0 sec.

Disagg. for 2% / 50, T = 5.0 sec.

Disagg. for 2% / 50, T = 5.0 sec.
Site Hazard Characterization In low probability hazard levels: For long periods, hazard is dominated by

Site Hazard Characterization

In low probability hazard levels:

For long periods, hazard is dominated by two events. M = 6.6 & R = 5km (εεεε = 1.5), and M = 8 & R = 60km (εεεε =

2.5).

For short periods, hazard is dominated by single event. M = 6.6 & R = 5km (εεεε = 1.5).

In high probability hazard levels:

Domination of a single or two events is not significant. Epsilon ranges from -2 to 2 with a mean value of zero.

Record Selection and Scaling Process Selection : Used a subset of NGA database (no aftershocks

Record Selection and Scaling Process

Selection :

Used a subset of NGA database (no aftershocks & etc.)

Only two recordings from any single event was selected

No restriction on Magnitude

R min &R max at 0.0 and 100.0 Km

Min and Max shear wave velocity = 180.0 and 1200.0 m/s

Low pass filter cutoff frequency of the selected motions are less than 0.1

Record Selection and Scaling Process Scaling : Maximum acceptable scale factor = 5.0 The scale

Record Selection and Scaling Process

Scaling :

Maximum acceptable scale factor = 5.0

The scale factor, by which the smallest weighted error between the target spectrum and the geometric mean spectrum of a single recording is acquired, is computed.

Records are matched between T min &T max at 0.5 & 10.0 sec.

o

Largest T = 6.47 sec. (Bldg. IIIB)

6.47X 1.5 = 9.7 sec. X1.5 = 9.7 sec.

o

Smallest T = 4.28 sec. (Bldg IIB)

4.28X 0.2 = 0.86 sec. X0.2 = 0.86 sec.

Record Selection and Scaling Process Scaling : Maximum acceptable scale factor = 5.0 The scale

Record Selection and Scaling Process

Scaling :

Maximum acceptable scale factor = 5.0

The scale factor, by which the smallest weighted error between the target spectrum and the geometric mean spectrum of a single recording is acquired, is computed.

Records are matched between T min &T max at 0.5 & 10.0 sec.

0.2 0.15 Uniform 0.1 Variable 26% %42 32% 0.05 10% %60 30% 0 0.5 3.0
0.2
0.15
Uniform
0.1
Variable
26%
%42
32%
0.05
10%
%60
30%
0
0.5
3.0
7.0
10 Period
0
2
4
6
8
Error
Weight
Record Selection and Scaling Process Reduction in Sampling Frequency: Reduced the ground motion time step

Record Selection and Scaling Process

Reduction in Sampling Frequency:

Reduced the ground motion time step from its original to 0.04 (25 samples per second = 25Hz)

Eliminated the effect of aliasing by filtering original ground motions beyond the Nyquist frequency (12.5 Hz)

Record Selection and Scaling Process A typical case of small difference A typical case of

Record Selection and Scaling Process

A typical case of small difference A typical case of larger difference
A typical case of small difference
A typical case of larger difference

PBD, Roof displacement, N-S direction

Original ground motion

PBD, Roof displacement, N-S direction Original ground motion Filtered and down- sampled ground Motion

Filtered and down- sampled ground Motion

PBD, Roof displacement, N-S direction Original ground motion Filtered and down- sampled ground Motion
N-S, dt=original N-S, dt=.04 difference GM set OVE MCE DBE N-S E-W SLE43 median %16
N-S, dt=original
N-S, dt=.04
difference
GM set
OVE
MCE
DBE
N-S
E-W
SLE43
median
%16 th and %84 th
Individual
SLE25
earthquake
MAXIMUM IDR
Response Spectra SLE25 (25 year)

Response Spectra SLE25 (25 year)

Response Spectra SLE25 (25 year)
Response Spectra SLE43 (43 year)

Response Spectra SLE43 (43 year)

Response Spectra SLE43 (43 year)
Response Spectra DBE (475 year)

Response Spectra DBE (475 year)

Response Spectra DBE (475 year)
Response Spectra MCE (2475 year)

Response Spectra MCE (2475 year)

Response Spectra MCE (2475 year)
Response Spectra OVE (4975 year) 7 unsclaed pairs are from simulated motions (URS/SCEC)

Response Spectra OVE (4975 year)

Response Spectra OVE (4975 year) 7 unsclaed pairs are from simulated motions (URS/SCEC)

7 unsclaed pairs are from simulated motions (URS/SCEC)

Response Spectra OVE (4975 year) Matched Spectra for TBI (OVE, 3 & 0.1, 7 &

Response Spectra OVE (4975 year)

Matched Spectra for TBI (OVE, 3 & 0.1, 7 & 0.6)

3.5 Target Spectrum 3 Median Spectrum 2.5 Indv. Spectrum Rec./Scaled Indv. Spectrum Simulated 2 1.5
3.5
Target Spectrum
3
Median Spectrum
2.5
Indv. Spectrum
Rec./Scaled
Indv. Spectrum
Simulated
2
1.5
1
0.5
0
0
1
2
3
4
5
6
7
8
9
10
(Sa(T)/g [5% critical damping]

Period (T)

Response Spectra OVE (4975 year) Matched Spectra for TBI (OVE, 3 & 0.1, 7 &

Response Spectra OVE (4975 year)

Matched Spectra for TBI (OVE, 3 & 0.1, 7 & 0.6)

1.5 Target Spectrum Median Spectrum Indv. Spectrum Rec./Scaled 1 Indv. Spectrum Simulated 0.5 0 0
1.5
Target Spectrum
Median Spectrum
Indv. Spectrum
Rec./Scaled
1
Indv. Spectrum
Simulated
0.5
0
0
1
2
3
4
5
6
7
8
9
10
(Sa(T)/g [5% critical damping]

Period (T)

Summary 5 sets of 15 ground motion records representing hazard levels from 25 year return

Summary

5 sets of 15 ground motion records representing hazard levels from 25 year return period to 5000 year return period are selected for the purpose of los estimation.

Ground motion are matched to the target spectrum for the location of the buildings. (meets code requirements, and similar to procedures used by engineering seismologists)

Same ground motions are used for all buildings.

For the very low probability hazard level (OVE) a combination of recorded and simulated motions is used