THE RESPONSE SPECTRUM

A Technical Seminar on the Development and Application of the Response Spectrum Method for Seismic Design of Structures
LECTURE # 3 1-2 June 2007

NBCC 2005 Site Specific Response Spectrum

Dr. Liam Finn

Department of Civil Engineering The University of British Columbia

Dr. Liam Finn served as Professor and Head of Civil Engineering, and Dean of Applied Science at the University of British Columbia in the period 1961 to 1998. From 1999 to 2005 he was Anabuki Professor of Geodynamics at Kagawa University in Japan and Consultant to the Anabuki Construction Company. Finn is Editor in Chief of the International Journal of Soil Dynamics and Earthquake Engineering; an Overseas Fellow of Churchill College, Cambridge, UK; Honorary Professor, Metallurgical Institute, Beijing and Honorary Fellow of the Chinese Soil Dynamics Society. He is also an Honorary Member of the Japanese Geotechnical Society. Finn has published over 350 papers in geotechnical earthquake engineering and was a pioneer in the development of nonlinear dynamic effective stress analysis and Lagrangian analysis for large flow deformations. He gave the Mallet-Milne Lecture on Earthquake Engineering in London, UK, in 2005. He has international consulting experience primarily on dams, dikes, offshore platforms and underwater pipelines. Finn is a member of the committee responsible for the seismic provisions of NBCC 2005

The Canadian Society for Civil Engineering, Vancouver Section

THE RESPONSE SPECTRUM

Site Specific Response Spectra

Dr. Liam Finn University of British Columbia

A Technical Seminar on the Development and Application of the Response Spectrum Method for Seismic Design of Structures

Adrian Wightman BGC Engineering, Vancouver

Filename, 1

1-2 June 2007

Vancouver, BC

Framework for Understanding Design Motions

In what follows we use field data and simple theory

Details in attached paper (Finn & Wightman 2003)

to demonstrate what variables control ground motion amplification to explain the simple code procedure for estimating amplification factors which incorporates the control variables To construct site specific response spectra -----------------------------------------------------------------------------An overview of site response analysis

Filename, 2 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Amplification of Motions: field data and theory

Amplification of incoming motions in soft soil sites with respect to adjacent rock motions, Idriss et al. 1989. Note effect of shaking intensity and lack of field data
Filename, 3 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Amplification of Ground Motions: Loma Prieta data

Aftershock data

___ Main shock

Amplification of strong and weak motions at Treasure Island(TRI) soil site with respect to rock motions at Yerba Buena Island (YBI), Jarpe et al. 1989 Note effects of intensity and period

Filename, 4 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Effect of impedance on ground motion amplification

Elastic layer on elastic half space, Finn and Wightman 2003 Incoming wave a = A sint = A sin (2/T)t Amplification ratio = at/ao = 1/ [ + s/2] when T1= T Note effect of impedance ratio = s Vss/r Vsr
Filename, 5 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Spectral Amplification against Vs30: a 0.10g

Key finding: We can represent all short and long period amplification factors by 2 corresponding sets of factors Fa and Fv based on site conditions specified by Vs30
Filename, 6 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Spectral Amplification against Vs30: a 0.10g

Fa = (1050/ Vs30)ma

Fv = (1050/ Vs30)mv

Vref of 1050m/s is the reference velocity for California For Fa = Fv = 1.0 General Equations are

Log Fa = ma(Log Vref Log Vsite) Log Fv= m v(Log Vref Log Vsite)
Filename, 7 The Response Spectrum - CSCE Vancouver Section

(1) (2)

1-2 June 2007

Liam Finn

Defining Amplification Factors for Code Use

The Challenge How to to classify the site unambiguously develop amplification factors keep process as simple as possible

Filename, 8 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Fa and Fv for ranges in Vs30: a 0.10g

The factors given by Eqns. (1) and (2) are continuous. Too complicated for routine design. To simplify the procedure 5 site classes were selected, covering 5 different ranges in shear wave velocity, Vs30; A, B, C, D and E Fa and Fv were determined for each site class
Filename, 9 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Site Class Characteristics

10

Filename, 10 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Estimating Fa and Fv for Different Site Classes; USA

For motions 0.10g : Loma Prieta field data For site class E and mapped accelerations > 0.10g: Results of equivalent linear and nonlinear analyses Mean values for Fa & Mean + 1 std for Fv For sites C and D: use Eqns (1) and (2) below using Vref= 1050m/s appropriate Vs30 for each Site Class and Ma and Mv from Site Class E data Log Fa = ma(Log Vref Log Vsite) Log Fv= m v(Log Vefr Log Vsite) (1) (2)

11

Filename, 11 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

USA (NEHRP) Amplification Factors, Fa

12

Mapped Spectral Acceleration at Short-Periods Ss = Sa(0.2) Site Class A B C D E F Ss 0.25 0.8 1.0 1.2 1.6 2.5 * Ss = 0.50 0.8 1.0 1.2 1.4 1.7 * Ss = 0.75 0.8 1.0 1.1 1.2 1.2 * Ss = 1.00 0.8 1.0 1.0 1.1 0.9 * Ss 1.25 0.8 1.0 1.0 1.0 0.9 *

Filename, 12 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

USA (NEHRP) Amplification Factors, Fv

13

Mapped Spectral Acceleration at One-Second Period S1 = Sa(1.0) Site Class A B C D E F S1 0.1 0.8 1.0 1.7 2.4 3.5 * S1 = 0.2 0.8 1.0 1.6 2.0 3.2 * S1 = 0.3 0.8 1.0 1.5 1.8 2.8 * S1 = 0.4 0.8 1.0 1.4 1.6 2.4 * S1 0.5 0.8 1.0 1.3 1.5 2.4 *

Filename, 13 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

NBCC 2005 Amplification Factors

14

NBCC 2005 adopted the NEHRP Fa and Fv factors But decided to adopt Site Class C as the Reference Ground Condition. All factors for Site Class C are now 1.0. Factors for other Site Classes are scaled to preserve the relative amplifications of the original NEHRP factors

Filename, 14 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

NBCC 2005 Table 4.1.8.4B

Fa as a function of site class and spectral acceleration Sa at T=0.2s
Values of Fa Sa 0.25 0.7 0.8 1.0 1.3 2.1 a Sa 1.25 0.8 1.0 1.0 1.0 0.9 a

15

Site Class A B C D E F

Sa = 0.5 0.7 0.8 1.0 1.2 1.4 a

Sa = 0.75 0.8 0.9 1.0 1.1 1.1 a

Sa = 1.00 0.8 1.0 1.0 1.1 0.9 a

NOTE: Use straight line interpolation for intermediate values of Sa(0.2). aSite-specific geotechnical investigation and dynamic site response analyses shall be performed.
Filename, 15 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

NBCC 2005 Table 4.1.8.4C

Fv as a function of site class and spectral acceleration at T=1.0s
Values of Fv S1.0 0.1 0.5 0.6 1.0 1.4 2.1 a S1.0 0.5 0.6 0.8 1.0 1.1 1.7 a

16

Site Class A B C D E F

S1.0 = 0.2 0.5 0.7 1.0 1.3 2.0 a

S1.0 = 0.3 0.5 0.7 1.0 1.2 1.9 a

S1.0 = 0.4 0.6 0.8 1.0 1.1 1.7 a

NOTE: Use straight line interpolation for intermediate values of S1.0. aSite-specific geotechnical investigation and dynamic site response analyses shall be performed.
Filename, 16 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Seismic Demand in NBCC 2005

17

Site demand is specified by Spectral Acceleration Response Spectrum - a function of period T: Sa(T). Spectrum Amplitudes depend on Site Location Soil Conditions; 5 NEHRP site classes; A, B, C, D, E. Canadian RGC = Site Class C A Base Spectrum is given for Reference Ground Conditions, RGC, at the site.

Filename, 17 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Vancouver Spectrum for RGC Site Site C

18

Sa(T)
1.0

Site C
0.5

0.1 0.1 0.2 0.5 1.0 2.0

T [s]
Filename, 18 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Coping with Different Site Conditions Essential Steps for Designer Identify the Site Class of the building site Select amplification factors appropriate for the Site Class of the building Use these factors to modify the Site Class C spectrum to get the site specific spectrum

19

Filename, 19 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Rules for Construction of Spectra: Sites D & E

S(T) = Fa x Sa(0.2); T 0.2s = Fa x Sa(0.2) or Fv x Sa(0.5); select smaller one = Fv x Sa(1.0); T = 1.0s = Fv x Sa(2.0); T = 2.0s

20

= Fv x Sa(2.0)/2 T 4.0s Sa(0.2), Sa(0.5), Sa(1.0) and Sa(2.0) are for Site Class C Use values of Fa and Fv for sites D & E in Tables 4.1.8.4B and 4.8.1.4C to construct Site Class D & E spectra

Filename, 20 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Spectrum for Site Class D from Site C Data

21

Site C: Spectral Ordinates for Vancouver

Sa(0.2) 0.96 Sa (0.5) 0.66 Sa (1.0) 0.34 Sa (2.0) 0.18

Fa and Fv Values at Site D For Sa (0.2) = 0.96 For Sa (1.0) = 0.34 Fa =1.10 Fv = 1.20

Site Class D spectrum ordinates are:

(smaller of)

[0.96 x1.1], {[0.96 x1.1] or [0.66 x1.2]}, [0.34 x1.2], [0.18 x1.2] 1.06, 0.79, 0.41,
1-2 June 2007

0.22

Filename, 21 The Response Spectrum - CSCE Vancouver Section

Liam Finn

Vancouver Spectra for Site Classes C and D

Sa(T)
1.0

Site C
0.5

Site D

0.1 0.1 0.2 0.5 1.0 2.0

T [s]
Filename, 22

Site Response Analysis: SRA

23

SRA is an alternative method for determining site specific response spectra. The structural engineer needs a general knowledge of the state of the art to interact effectively with the geotechnical engineer. A general outline of the state of the art is presented here.

Filename, 23 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Elements of Site Response Analysis: SRA

T

24

Sa(T)

Soft Soil

aup

H
outcrop motion

adown
Filename, 24 The Response Spectrum - CSCE Vancouver Section

Stiff Soil or Rock

1-2 June 2007

Liam Finn

Key Steps in Site Response Analysis

25

Selection of type of analysis: Equivalent linear or nonlinear analysis Selection of input motions

Filename, 25 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Equivalent Linear vs Nonlinear Analysis

Recommended procedures for implementation of DMG Special Publication 117 Guidelines for analyzing and mitigating liquefaction hazards in California, SCEC, USC, Los Angeles, 1999, states,

26

In general, equivalent linear analyses are considered to have reduced reliability as ground shaking levels increase to values greater than 0.4g in the case of softer sols or where the maximum shear strain amplitudes exceed 1%-2%. For these cases, true nonlinear site response programs may be used. The computer program DESRA-2, originally developed by Lee and Finn (1978), is perhaps the most widely recognized nonlinear one dimensional site response program. Other nonlinear programs include MARDES (Chang et al,1991), D-MOD (Matasovich, 1993) and SUMDES (Li et al., 1992).

Filename, 26 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

Selecting Candidate Motions

Many large data bases now available such as COSMOS PEER etc Candidate motions should come from same seismic environment as the target site. Shallow crustal earthquakes Deep crustal earthquakes Subduction earthquakes Deep Puget Sound Source, PUG, contributes most to risk in the Lower Mainland yet most often shallow crustal records from California are used.
Filename, 27 The Response Spectrum - CSCE Vancouver Section

27

1-2 June 2007

Liam Finn

Other Factors Controlling Selection

Type of faulting strike-slip, normal etc Magnitude M and distance to site, R. In Canada mostly areal seismic sources. M and R are selected as the values contributing most to the hazard at the site. Use Mode Magnitude Mm and Rm to guide record selection These can be downloaded from the CGS website. Records should preferably have been recorded on site with similar velocity distributions with depth. Japanese sites yield records more compatible with Lower Mainland sites than typical California sites.
Filename, 28 The Response Spectrum - CSCE Vancouver Section

28

1-2 June 2007

Liam Finn

Scaling Candidate Motions

29

Linear scaling so that PGA of selected records site PGA Scale the motions to match the design response spectrum over a desired period range. For the BC Schools Seismic Retrofit Program input motions were scaled to match the design spectrum for Site Class C over the range 0.5s-1.5s. How is the spectral matching done? Tim Little will deal with this in his lecture.

Filename, 29 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

How many Input Motions are required ?

30

The main reason for using multiple input motions is to protect against the great uncertainty associated with ground motion predictions: For BC the range is median/2 median x 2 For the BC Schools retrofit program 10 input motions are used per site. The number of motions should be sufficient to yield reliable statistics for the response parameters: Median, mean and standard deviation of spectral values

Filename, 30 The Response Spectrum - CSCE Vancouver Section

1-2 June 2007

Liam Finn

How well does Site Response Analysis Work

Major blind prediction exercises were conducted for the Ashigari Test Site in Japan and the Turkey Flat Site in California to evaluate capability to predict site response. In both cases, despite a wealth of data on geology and soil properties and records of input motions, predictions of site responses were disappointing. Clearly the safe procedure for SRA is to use multiple inputs and try to get reliable response statistics to guide design decisions. Site response analysis is not a routine process. Go through a checklist of the essential requirements cited above and those offered by Tim Little in his lecture before proceeding with an analysis.
Filename, 31 The Response Spectrum - CSCE Vancouver Section

31

1-2 June 2007

Liam Finn