bray ucanterbury seismic displ

Upload: raul-contreras

Post on 14-Apr-2018

239 views

Category:

Documents


1 download

TRANSCRIPT

  • 7/27/2019 Bray UCanterbury Seismic Displ

    1/46

    Simplified Seismic SlopeDisplacement Procedures

    Jonathan D. Bray, Ph.D., P.E.

    Univ. of California at Berkeley

    Thanks to Dr. T. Travasarou, Prof. E. Rathje, & others,

    with support from PEER & the Packard Foundation

  • 7/27/2019 Bray UCanterbury Seismic Displ

    2/46

    Simplified Seismic Slope Displacement Procedures

    OUTLINE

    I. Introduction

    II. Seismic Slope Displacement Analysis

    a. Earthquake Ground Motion

    b. Dynamic Resistance

    c. Dynamic Response

    d. Seismic Displacement Calculation

    III. Simplified Slope Displacement Procedures

    IV. Conclusions

  • 7/27/2019 Bray UCanterbury Seismic Displ

    3/46

    I. Seismic Slope Stability

    4th Ave. Slide, 1964 Alaska EQ 1999 Chi-Chi, Taiwan EQ

    EERC Slide Collecti on

    Landslide Stabilization

    CSASolid-Waste Landfi ll

  • 7/27/2019 Bray UCanterbury Seismic Displ

    4/46

    Two Critical Design Issues

    1. Are there materials that will lose significantstrength as a result of cyclic loading?

    Flow Slide

    2. If not, will the system undergo significant

    deformations that may jeopardize system

    performance?Seismic Displacements

  • 7/27/2019 Bray UCanterbury Seismic Displ

    5/46

    Seismic Slope Displacement

    Slip along a distinct surface

    Distributed shear deformation

    Add volumetric-induced deformation, when appropriate

    Newmark-type

    seismic displacement

  • 7/27/2019 Bray UCanterbury Seismic Displ

    6/46

    FEA of Shaking Table Test of Clay Slope

    0 2 4 6 8 10-2.5

    -2

    -1.5

    -1

    -0.5

    0

    0.5

    Time (sec)

    Displacement(inches)

    Horizontal displacement (D2)

    PLAXIS

    RECORD

    0 2 4 6 8 10-3

    -2

    -1

    0

    1

    Displacement(inches)

    Horizontal displacement (D4)

    0 2 4 6 8 10-1.5

    -1

    -0.5

    0

    0.5

    1

    Time (sec)

    Horizontal displacement (D5)

    0 2 4 6 8 10-1.5

    -1

    -0.5

    0

    0.5Vertical displacement (D1)

  • 7/27/2019 Bray UCanterbury Seismic Displ

    7/46

    II. Seismic Displacement Analysis:

    Critical Components

    a. Earthquake Ground Motion

    b. Dynamic Resistance

    c. Dynamic Response

    d. Seismic Displacement Calculation

  • 7/27/2019 Bray UCanterbury Seismic Displ

    8/46

    a. Earthquake Shaking:

    Acceleration Time History

    acceleratio

    n(g)

    -0.50

    -0.25

    0.00

    0.25

    0.50

    time (s)

    0 5 10 15 20 25 30

    Izmit (180 Comp) 1999 Kocaeli EQ (Mw

    =7.4) scaled to MHA = 0.30 g

    PGA = 0.3 g Tm = 0.63 s & D5-95 = 15 s

  • 7/27/2019 Bray UCanterbury Seismic Displ

    9/46

    Acceleration Response Spectrum(provides response of SDOF of different periods at 5% damping,

    i.e., indicates intensity and frequency content of ground motion)

    0 1 2 3 4Period s

    0.0

    0.5

    1.0

    1.5

    Spe

    ctralAcceleration(g)

    5% Damping

    Sa(0.5)

    Sa(1.0)

    PGA

    Sa(0.2)

  • 7/27/2019 Bray UCanterbury Seismic Displ

    10/46

    b. Dynamic Resistance:Simplified Estimates of Yield Coefficient (ky)

    (seismic coefficient that results in FS=1.0 in pseudostatic stabili ty analysis)

    H

    c = cohesion

    = friction angle

    kc

    Hy= +

    + tan( )

    cos ( tan tan )

    2 1

    1S2

    1 H

    L

    S1

    kFS S H

    H S S Ly

    static=

    + +

    ( ) cos sin

    ( )

    1 2

    2

    1 1 1

    1 2

    ( )FS

    S H L S H

    S Hstatic

    = + +

    tan cos

    cos sin

    1

    2

    1 2

    1 1 1

    2 2

    2

    with 11

    11=tan ( )S

    Shallow SlidingDeep Sliding

    Shewbridge 1996Bray et al. 1998

  • 7/27/2019 Bray UCanterbury Seismic Displ

    11/46

    Factors Affecting

    Dynamic Strength of Clays

    Rate of loading: Crate > 1

    Number of significant cycles: Ccyc < 1

    Progressive failure: Cprog < 1

    Sdynamic = Sstatic (Crate) (Ccyc ) (Cprog)

    Also, post-peak strength reduction if EQ-induced strain exceeds strain at peak stress

  • 7/27/2019 Bray UCanterbury Seismic Displ

    12/46

    Vane Shear and Direct Shear tests

    Small vane 2.5 cm diameter

    Large vane 5.7 cm diameter

    Large direct shear box 30 cm

    by 30 cm square

    0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2-10

    0

    10

    20

    30

    40

    50

    60

    70

    displacement (in)

    shearstress(p

    sf)

    original vane shear data

    derived from vane shear

    direct shaer data

  • 7/27/2019 Bray UCanterbury Seismic Displ

    13/46

    Seed and Martin 1966

    c. Dynamic Response of Sliding Mass

  • 7/27/2019 Bray UCanterbury Seismic Displ

    14/46

    Dynamic Response: Equivalent Acceleration Concept

    accounts for cumulative effect of incoherent motion in deformable slid ing mass

    Horz. Equiv. Accel.: HEA = (h/v) g MHEA = max. HEA value

    kmax = MHEA / g

    H vh

    Seed and Mart in 1966

  • 7/27/2019 Bray UCanterbury Seismic Displ

    15/46

    0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

    Ts-waste/Tm-e

    0.00.2

    0.4

    0.6

    0.8

    1.0

    1.2

    1.4

    1.6

    1.8

    2.0

    M

    HEA/(MH

    A,rock*NR

    F) Rock site median, and 16th and 84th

    0.1 1.350.2 1.200.3 1.09

    0.4 1.000.5 0.920.6 0.870.7 0.820.8 0.78

    MHA,rock (g) NRF

    probability of exceedance lines

    (Bray & Rathje 1998)

    FACTORS AFFECTING MAXIMUM SEISMIC LOADING

    kmax

    /(MHA

    roc

    k

    *NRF/g)

    Ts-sliding mass / Tm-EQ

  • 7/27/2019 Bray UCanterbury Seismic Displ

    16/46

    kmax depends on stiffness and geometry of the

    sliding mass (i.e., its fundamental period)

    Ts,1-D = 4 H / Vs

    Ts , 1-D = Initial Fundamental Period of Sliding Mass

    H = Height of Sliding Mass

    Vs = Average Shear Wave Velocity of Sliding Mass

  • 7/27/2019 Bray UCanterbury Seismic Displ

    17/46

    d. Seismic Displacement CalculationNewmark (1965) Rigid Sliding Block Analysis

    Assumes:

    Rigid sliding block

    Well-defined slip surface develops

    Slip surface is rigid-perfectly plastic

    Acceleration-time history defines EQ loading

    Key Parameters:

    ky - Yield Coefficient (max. dynamic resistance)

    kmax - Seismic Coefficient (max. seismic loading)

    ky /

    kmax (if > 1, D = 0; but if < 1, D > 0)

    C

  • 7/27/2019 Bray UCanterbury Seismic Displ

    18/46

    Rigid

    Sliding

    Block:

    use

    accel.-

    time hist.

    kmax = PGA/g

    Cover

    Base

    Rock

    Deformable

    Sliding

    Block:

    use equiv.

    accel.-time

    hist.

    kmax = MHEA/g

    Mid-Ht

  • 7/27/2019 Bray UCanterbury Seismic Displ

    19/46

    Decoupled & Coupled

    Deformable Sliding Analysis

    Earth Fill

    Potential Slide Plane

    Decoupled

    Analysis

    Coupled

    Analysis

    Flexible System

    Dynamic Response

    Rigid Block

    Sliding Response

    Flexible SystemFlexible System

    Dynamic Response and

    Sliding Response

    Max Force at

    Base = ky W

    Calculate HEA-

    time history

    assuming nosliding along base

    Double integrate

    HEA-time

    history given kyto calculate U

    0.1 1 10 100 1000

    U (cm)

    -40

    -20

    0

    20

    40

    60

    80

    100

    DisplacementDifference(cm):

    U

    -U

    decoupled

    decoupled

    co

    upled

    k = 0.05

    k = 0.1

    k = 0.2

    y

    y

    y

    (b)

    Decoup

    ledConserva

    tive

  • 7/27/2019 Bray UCanterbury Seismic Displ

    20/46

    Calculated Seismic Displacement

    Expected ky

  • 7/27/2019 Bray UCanterbury Seismic Displ

    21/46

    SAFE

    UNSAFE

    ?

    Think About It as a Cliff

    Calculated Seismic Displacement is an Index of Performance

  • 7/27/2019 Bray UCanterbury Seismic Displ

    22/46

    Evaluate Seismic Performance

    Given seismic displacement estimates:

    Small (e.g., D 15 cm)

    Moderate

    Large (e.g., D 1 m)

    Evaluate the ability of earth structure and systems founded on it to

    accommodate this level of deformation.

    Consider:

    Consequences of failure and conservatism of hazard assessmentand stabili ty analyses, e.g., the consequences of shallow slid ingare usually less severe than deep sliding & more easily repaired.

    Defensive measures that provide "ductility" or " shield" criticalcomponents from damage, e.g., slip layer above critical linersystem.

  • 7/27/2019 Bray UCanterbury Seismic Displ

    23/46

    III. Simplif ied Seismic Slope Procedures:

    Makdisi & Seed (1978)

    1. Evaluate materials strength loss potential

    If significant, do not use this method

    If slight, use slightly reduced strength (in many

    cases, use ~ 10% to 20% strength reduction)

    2. Calculate ky

  • 7/27/2019 Bray UCanterbury Seismic Displ

    24/46

    3. Estimate max. crest acceleration

    MHA at Top vs. Base Rock MHA for Some Solid-Waste Landf ills(Bray & Rathje 1998)

    MHA,top = ?

  • 7/27/2019 Bray UCanterbury Seismic Displ

    25/46

    4. Estimate kmax for sliding mass using plot of

    max. accel. ratio vs. (y/h)

    Kmax varies with Ts

  • 7/27/2019 Bray UCanterbury Seismic Displ

    26/46

    5. Estimate seismic displacement (U) basedon ky/kmax & EQ Magnitude

    WARNING: Do not use

    figure in original

    Makdisi & Seed (1978)

    paper; it is off by anorder of magnitude

  • 7/27/2019 Bray UCanterbury Seismic Displ

    27/46

    Limitations It must be noted that the design curves are derived from alimited number of cases. These curves should be updated and

    refined as analyt ical results for more embankments are obtained.Makdisi & Seed (1978)

    Very limited number of earthquake ground motions used.

    Estimating MHA at the crest to estimate kmax is problematic.

    Relatively simple analysis employed, e.g., shear slice method.

    Decoupled analysis to calculate seismic displacement.

    Upper and lower bounds are not t rue upper and lower bounds.

    No estimate of uncertainty.

    B & T (2007)

  • 7/27/2019 Bray UCanterbury Seismic Displ

    28/46

    INTENSITY MEASURES, e.g.:

    amplitude: PGA, PGV, PGD, Sa, Svfrequency content: Tp, Tm

    duration: D5_95,combination: Arias Intensity

    Housner Spectral Intensity

    Bray & Travasarou (2007)

    1. SLOPE MODEL

    nonlinear soil response

    fully coupled deformable stick-slip

    stiffness (Ts) & strength (ky)

    8 Ts values & 10 ky values Tsky

    D2. EARTHQUAKE DATABASE41 EQ - 688 records

    Over 55,000 analyses

    EFFICIENCY

  • 7/27/2019 Bray UCanterbury Seismic Displ

    29/46

    EFFICIENCY

    of Ground Motion

    Intensity Measures

    STIFF WEAK SLOPE

  • 7/27/2019 Bray UCanterbury Seismic Displ

    30/46

    PGASA

    SA(1.3Ts)

    SA(1.5Ts)

    SA(2.0Ts)

    arms

    AriasIc

    PGVEPVSI

    PGDTm

    D5_

    95

    Intensity Measure

    0

    1

    2

    S

    tandardDeviatio

    n

    0

    1

    2

    StandardD

    eviation

    ky = 0.10 Ts = 0.5 s

    ky = 0.20 Ts = 0.5 s

    0

    1

    2

    PGASA

    SA(1.3Ts)

    SA(1.5Ts)

    SA(2.0Ts)

    arms

    Arias Ic

    PGV

    EPVSI

    PGDTm

    D5_

    95

    Intensity Measure

    0

    1

    2 ky = 0.20 Ts = 1.0 s

    ky = 0.10 Ts = 1.0 s

    ST

    R

    O

    NG

    E

    R

    MORE FLEXIBLE

    EFFICIENCY RESULTS

    Sa (1.5Ts) Arias Intensity Housner Spectral Intensity

  • 7/27/2019 Bray UCanterbury Seismic Displ

    31/46

    (1) Model for D = 0 P (D = 0 ) = f(Sa(1.5Ts), ky, Ts)

    Note: D = 0 is negligible displacements, i.e. D < 1 cm

    MIXED RANDOM VARIABLE APPROACH

    0.01 0.1 1

    Yield Coefficient

    0.00

    0.25

    0.50

    0.75

    1.00

    P(D

    ="0")

    0 1 2 3

    Fundamental Period (s)

    0.00

    0.25

    0.50

    0.75

    1.00

    P(D="0")

    0.001 0.01 0.1 1 10

    SA(1.5Ts) (g)

    0.00

    0.25

    0.50

    0.75

    1.00

    P(D="0")

    (a) (b) (c)

    MIXED RANDOM VARIABLE APPROACH

  • 7/27/2019 Bray UCanterbury Seismic Displ

    32/46

    (2) Model for D > 0 Ln (D) = f(Sa(1.5Ts), ky, Ts, M) + [ = 0.66]MIXED RANDOM VARIABLE APPROACH

    B & T (2007) S i i Di l t

  • 7/27/2019 Bray UCanterbury Seismic Displ

    33/46

    1) Zero Displacement Estimate

    ( ) ++= ))5.1(ln()ln(566.0)ln(333.0)ln(83.210.1)ln( 2 sayyy TSkkkD

    ))5.1(ln(52.3)ln(484.0)ln(22.376.11)"0"(sasyy TSTkkDP +==

    2) NonZero Displacement Estimate

    ( ) ++ )7(278.050.1))5.1(ln(244.0))5.1(ln(04.3 2 MTTSTSssasa

    is the standard normal cumulative distribution function (NORMSDIST in Excel)

    Bray & Travasarou (2007) Seismic Displacement

    is a normally-distributed random variable with zero mean and standard deviation = 0.66

    *Replace first term (i.e., -1.10) with -0.22 for cases where Ts < 0.05 s

  • 7/27/2019 Bray UCanterbury Seismic Displ

    34/46

    MODEL TRENDS

    0 0.5 1 1.5 2

    Ts (s)

    0

    0.5

    1

    SA(1.5Ts)(g) 5% damping

    Scenario Event:

    M 7 at 10 km

    Soil SS fault

    0 0.5 1 1.5 2

    Ts (s)

    0

    0.2

    0.4

    0.6

    0.8

    1

    P(D="0")

    ky= 0.3

    ky= 0.2

    ky= 0.1

    ky= 0.05

    0 0.5 1 1.5 2

    Ts (s)

    1

    10

    100

    Displac

    ement(cm) ky= 0.05

    ky= 0.1

    ky= 0.2

    ky= 0.3

  • 7/27/2019 Bray UCanterbury Seismic Displ

    35/46

    PROBABILITY OF EXCEEDANCE

    0 0.5 1 1.5 2

    Ts (s)

    0

    0.2

    0.4

    0.6

    0.8

    1

    P(D

    ="0")

    ky= 0.3

    ky= 0.2

    ky= 0.1

    ky= 0.05

    0 0.5 1 1.5 2

    Ts (s)

    1

    10

    100

    Displace

    ment(cm) ky= 0.05

    ky= 0.1

    ky= 0.2

    ky= 0.3

    0 0.5 1 1.5 2

    Ts (s)

    0

    0.2

    0.4

    0.6

    0.8

    1

    P(D

    >30cm)

    ky= 0.3ky= 0.2

    ky= 0.1

    ky= 0.05

    Scenario Event:

    M 7 at 10 km

    Soil SS fault

    Validation of Bray & Travasarou Simplif ied Procedure

  • 7/27/2019 Bray UCanterbury Seismic Displ

    36/46

    System EQ

    Obs.

    Dmax

    (cm)

    Bray & Travasarou 2007

    P (D = 0 ) Est. Disp (cm)

    Makdisi &

    Seed 1978

    D (cm)

    Bray &

    Rathje 1998

    D (cm)

    BuenaVista LF LP None 0.75 0.8 - 3 0 0

    Guadalupe LF LP Minor 0.95 0.3 - 1 0 0 4

    Pacheco Pass LF LP None 1.0 0 - 0.1 0 0

    Marina LF LP None 0.9 0.5 - 2 0 0

    Austrian Dam LP 50 0.0 20 - 70 1 30 20 100

    Lexington Dam LP 15 0.0 15 - 65 0 10 30 110

    Chiquita Canyon C LF NR 24 0.0 10 - 30 1 40 3 20

    Chiquita Canyon D LF NR 30 0.0 6 - 22 0 10 2 15

    Sunshine Canyon LF NR 30 0.0 20- 70 0 0

    OII Section HH LF NR 15 0.1 4 - 15 3 30 2 25

    La Villita Dam S3 1 0.95 0.3 - 1 0 0

    La Villita Dam S4 1.4 0.5 1 - 5 0 0

    La Villita Dam S5 4 0.25 3 - 10 0 1 0

    Validation of Bray & Travasarou Simplif ied Procedure

  • 7/27/2019 Bray UCanterbury Seismic Displ

    37/46

    Hypothetical Example: 57 m-High Earth Dam

    Located 1.1 km from Hayward Fault

    s

    sm

    m

    V

    HT

    s

    s 33.0

    /450

    576.26.2

    =

    =

    No liquefaction or soils that will undergo significant strength loss

    Undrained Strength: c = 14 kPa and = 21o so ky = 0.14Triangular Sliding Block with avg. Vs ~ 450 m/s & H = 57 m

    Deterministic Analysis:

  • 7/27/2019 Bray UCanterbury Seismic Displ

    38/46

    Deterministic Analysis:

    Mean Max. Mw ~ 6.9 at R ~ 1.1 km; Ts = 0.33 s; & ky = 0.14

    Using 1.5 Ts = 1.5 (0.33 s) = 0.5 s & Abrahamson & Silva (97) &Sadigh et al. (97) rock attenuation relations for strike-slip fault:avg. Sa(0.5 s) = 1.07 g

    Probability of Zero Displacement (i.e., D < 1 cm):

    Nonzero Seismic Displacement Estimate:

    Design Seismic Displacement Estimate (16% and 84 %):

    D ~ 0.5 to 2 times median D = 20 cm to 80 cm

    ( ) 0)07.1ln(52.3)33.0)(14.0ln(484.0)14.0ln(22.376.11)"0"( =+==DP

    ( ) ++= )07.1ln()14.0ln(566.0)14.0ln(333.0)14.0ln(83.210.1)ln( 2D

    ( ) ++ )79.6(278.0)33.0(50.1)07.1ln(244.0)07.1ln(04.3 2

    cmDD 40)77.3exp())exp(ln( ==

    For fully probabilist ic implementation see Travasarou et al. 2004

  • 7/27/2019 Bray UCanterbury Seismic Displ

    39/46

    Pseudostatic Slope Stability Analysis

    1. k = seismic coefficient; represents earthquake loading

    2. S = dynamic material strengths

    3. FS = factor of safety

    Selection of acceptable combination ofk, S, & FS

    requires calibration through case histories or

    consistency with more advanced analyses

  • 7/27/2019 Bray UCanterbury Seismic Displ

    40/46

    A Prevalent Pseudostatic Method

    Seed (1979)

    appropriate dynamic strengths

    k = 0.15

    FS > 1.15

    FS > 1.15 does not mean the system is safe!

    BUT this method was calibrated for earth dams where1 m of displacement was judged to be acceptable

    WHAT about other systems?

    WHAT about other levels of acceptable displacement?

    IS k = 0.15 reasonable for all sites?

  • 7/27/2019 Bray UCanterbury Seismic Displ

    41/46

    Seismic Coefficient

    k should depend on level of shaking, i.e.,distance to fault, earthquake magnitude,

    dynamic response of earth structure

    k should also depend on crit icality of structure

    and acceptable level of seismic performance, i.e.,

    amount of allowable seismic displacement

    How does one then select k?

    S i i C ffi i f All bl Di l

  • 7/27/2019 Bray UCanterbury Seismic Displ

    42/46

    Calculate k as function ofDa, Sa, Ts, Mw, &

    Seismic Coefficient from Allowable DisplacementBray &Travasarou (2009)

    +

    Seismic coefficient, k:

    Seismic Coefficient depends on:

  • 7/27/2019 Bray UCanterbury Seismic Displ

    43/46

    Ts= 0.2s

    0.00

    0.10

    0.20

    0.30

    0.40

    0 0.5 1 1.5 2

    Sa(@1.5Ts) (g)

    S

    eismicCoefficient

    Seismic Coefficient depends on:

    Da, EQ (Sa & Mw), & Ts

    5 cm

    15 cm

    30 cm

    M=7.5 M=6

    Pseudostatic Slope Stability Procedure

  • 7/27/2019 Bray UCanterbury Seismic Displ

    44/46

    1. Are there materials that could lose significant

    strength? If so, use post-shaking reduced strengths.Otherwise, use strain-compatible dynamic strengths

    2. Select allowable displacement: Da and select

    exceedance probability (e.g., use = 0 for 50%)3. Estimate initial period of sliding mass: Ts

    4. Characterize seismic demand: Sa(1.5T

    s) & Mw

    5. Calculate seismic coefficient: k = f(Da, Sa, Ts, Mw & )& perform pseudostatic analysis using k. If FS 1,

    then Dcalc < Da at specified level of exceedance

    Pseudostatic Slope Stability Procedure

  • 7/27/2019 Bray UCanterbury Seismic Displ

    45/46

    Simplified Seismic Coefficient Estimates(used in code applications, e.g., B.C., Canada)

    Stiff Slopes: Ts 0.1 s & Using: Da = 15 cm

    For M 7.5:

    k15 cm 0.25 Sa 0.25 (2.5 PGA) 0.65 PGA

    Conclusions

  • 7/27/2019 Bray UCanterbury Seismic Displ

    46/46

    Conclusions

    First question: will materials lose strength?

    If not, evaluate seismic slope stabil ity interms of seismic displacements

    Bray & Travasarou (2007) approach withdeformable sliding mass captures:

    a. Earthquake shaking - Sa(1.5 Ts) & Mw

    b. Dynamic resistance of slope - ky

    c. Dynamic response of sliding mass - Tsd. Coupled seismic displacement - D

    Earthquake and material characterization

    are most important