Академический Документы
Профессиональный Документы
Культура Документы
METHODOLOGY FOR
KEYSTONE WALLS
(AASHTO Based Methodology)
by
KEYSTONE Retaining Walls and other mechanically stabilized earth (MSE) retaining structures
have shown considerable resistance to seismic forces based on the lack of reported failures or
severe distress being noted. This superior performance is based on the visual investigation of
many existing structures that have been exposed to major seismic events in California in recent
years as well as in Japan and Central America. Many of the structures observed had no special
seismic requirements incorporated in the design yet have still performed satisfactorily without
experiencing any significant damage during major seismic events.
The lack of seismic performance problems with earth retaining structures has resulted in little
advancement in the current seismic design approach since those proposed by Mononobe and
Okabe in the 1920's and later discussed by Seed and Whitman - ASCE 1970. The psuedo-static
approach to the seismic analysis of retaining walls is referenced in the AASHTO Standard
Specifications for Highway Bridges (1999), Division I-A - Seismic Design and Section 5 -
Retaining Walls. While MSE earth retaining structures appear to be earthquake resistant
(adjacent foundations and slopes may be subject to seismic instability), a rational psuedo-static
design analysis has been established to satisfy seismic engineering standards when required.
This document is intended to provided the appropriate seismic design procedure for KEYSTONE
Retaining Walls utilizing earth reinforcement. The external stability analysis is based on the
Mononobe-Okabe psuedo-static approach referenced in Division I-A and Section 5 of the
AASHTO design manual. The internal stability analysis is based on the methodology stated in
Section 5 of the AASHTO Design Manual for mechanically stabilized embankment structures.
Design Approach
A peak ground acceleration coefficient, "A", is location and site specific and is often chosen
based on prevailing design code criteria or from the seismic acceleration contour map contained
within AASHTO specifications which represents a 90% probability of not being exceeded in 50
years. Typically, any seismic zone with A < 0.10g can safely ignore seismic analysis for
retaining wall structures as the calculated effects are negligible.
AASHTO further refines the horizontal ground acceleration coefficient, A, to reflect a maximum
retaining wall acceleration, Am, at the center of mass above ground as follows:
Note:
1) Set Am = A if A > 0.45g
2) Deformation analysis recommended if A > 0.29g
Am, the maximum structure acceleration, is intended to simulate the greater average acceleration
that occurs above the foundation level, approximately at the centroid of the retained mass and
soil behind the mass during a seismic event.
"For free standing abutments or retaining walls which may displace horizontally
without significant restraint, the psuedo-static Mononobe-Okabe method of analysis
is recommended for computing lateral active soil pressures during seismic loading.
A “psuedo-static” seismic coefficient equal to one-half the acceleration coefficient
(kh =0.5A) is recommended. The effects of vertical acceleration may be omitted".
This criteria is based on the permissible displacement approach discussed by Richards and Elms
(1979) based on the work of Newmark which compares residual displacement, dr versus the ratio
of transmittable acceleration, N, to peak ground acceleration, A. The formula developed is:
1/4
0.087 V2
dr = 0.087 (V2/A*g) * (N/A)-4 or N/A =
dr A g
where:
dr = is the displacement
V is the peak velocity of earthquake record
(In absence of information, V = 760 (A) (mm/s))
N is the structure yield acceleration coefficient
A is the peak ground acceleration coefficient
“For most design purposes, it has been shown (Elms and Martin, 1979) that a
design value of Kh = 0.50A is adequate, provided that the wall can accommodate
an outward displacement of up to about 250A mm”(10A in inches).
AASHTO also states that the dynamic horizontal thrust be "...applied a 0.6 H2 above the base on
the back surface of the effective mass".
3) Set the seismic coefficients, kh = Am/2 and kv = 0 for the external analysis.
The author believes that it is acceptable to use kh = Am/2 for a flexible retaining structure and
50% of the dynamic component as suggested in AASHTO Section 5. However, it may be
questionable to use the 50% reduction on the dynamic component as setting kh = Am/2 already
accounts for a force reduction due to the yielding nature of the system.
Since the Mononobe-Okabe seismic analysis was intended for rigid cantilever structures and not
flexible reinforced soil structures, some refinements have been made to the analysis by AASHTO
to provide a more appropriate methodology. Externally, the dynamic and inertial forces are
assumed to only act on an effective reinforced mass that is "H" high by "0.5H" deep [not the full
depth of the reinforced structure [typical MSE structure is "H" high by "0.7H" deep]. For a
sloping backfill case, the effective mass is defined as having an effective height "H2" and base of
"0.5H2" as defined by the following equation:
W3
H'
Ev Estatic
H Eh δ or ι
W1 W2
Wall batter, β, is typically
H'/3
cL assumed = 0 in highway
projects.
e
d R L
khM3
M3 W3
Effective
H' zone Ev Edynamic @ 50%
H2 Eh
δ or ι
khW1 khM2 Estatic
Ev
H Eh
M2
W1 W2 0.6H2 Note: The maximum Estatic is
H'/3 calculated at the end of
cL reinforcement or L = H, which
e ever is less, regardless of
d R final reinforcement design
H2/2 length.
L
Fig. 2 - Static + Dynamic Design Case
1/12/05 Seismic Design 5-16
The Mononobe-Okabe analysis is an extension of Coulomb wedge theory that provides a total
earth pressure acting on the structure due to static and dynamic forces. It is customary to break
this into a static and dynamic component with the static component acting at H/3 (or H'/3 for
sloping backfill) from the bottom, and with the dynamic component taken to act at a height of
0.6H (or 0.6H2 for an infinite slope as shown Figure 2). The total pressure coefficient is
represented by the following equation:
cos2(φ−θ+β)
Kae = (eq. 3)
ψ cos(θ)cos2(β)cos(δ−β+θ)
where:
2 Note: this expression may
ψ= 1 + sin(φ+δ)sin(φ−θ−ι) become negative with steep
cos(δ−β+θ)cos(ι+β) backslopes so it is customary to
set equal to zero in those cases.
φ = angle of friction of soil
θ = arc tan(Kh/(1-Kv))
Am = A(1.45-A)
Kh = horizontal acceleration coefficient = Am/2
Kv = vertical acceleration coefficient = 0 assumed
ι = backfill slope angle
δ = angle of friction between soil & mass (δ = ι for Rankine)
β = slope of wall face (positive wall batter) = 0 assumed
The dynamic component is determined by calculating the total seismic earth pressure coefficient,
Kae, and subtracting the static earth pressure coefficient, Ka:
Ka is the calculated Coulomb or Rankine external earth pressure coefficient for the section
geometry and soil properties. The external earth pressure forces are calculated as follows:
In addition to the above, the lateral inertial forces of the wall facing, reinforced mass, and
backslope must be included (W’s x Kh). All forces are applied as shown in Figure 2 and factors
of safety in sliding and overturning are calculated in the conventional manner. Increased bearing
pressure and eccentricity under seismic conditions is typically ignored due to the transient nature
of the loading but may be calculated and checked. The calculated "static + dynamic" safety
factors are compared to 75% of the static-only minimum safety factors for design compliance
Note: live loads are typically not considered in the seismic retaining wall analysis under the
"combination of loads" criteria in AASHTO - Group VII.
1/12/05 Seismic Design 6-16
Internal Stability
The earth reinforcement behind a Keystone wall face must be designed to resist the lateral earth
pressure as well as the horizontal forces generated by the inertial force of the facing and the
failure zone. The total inertial force is equal to the mass times the maximum wall acceleration
coefficient, "Am".
ι 0.3H
ι
AmW3
Am W3
AmW2 W3 W3
H’/2 AmW2
H'
Am W1 W2 Am W1
H H W2
W1 Le H’/2 W1 Le
45+φ/2
d d B
B
Extensible Reinforcement Inextensible Reinforcement
The inertial force must be distributed in addition to the static forces to the earth reinforcement in
some manner. AASHTO suggests that the inertial force be distributed over the height of the wall
proportionally to the resistant area (embedment length, Le) of each reinforcement level. This
distribution tends to weight the seismic forces towards the bottom of the wall where the
embedment lengths are longest.
The inertial seismic force distribution by embedment length is described by the following
relationship:
Le of reinforcement level
Add'l Load/reinf = (Pi) (eq. 8)
Sum of Le for all reinforcement levels
Lei
Tmd = Pi
Σ Lei
"Factors of safety under the combined static and seismic loads for pullout and
breakage of reinforcement may be reduced to 75% of the factors used for static
loading."
Pullout resistance must be re-calculated under AASHTO guidelines due to the possibility of
reduced pullout resistance during seismic excitation. Excerpted from AASHTO:
"For seismic loading conditions, values of f*, Np and fd shall be reduced to 80%
of the values used for static design."
This statement requires that the tension in each reinforcement level be increased due to seismic
forces while at the same time reducing the available pullout resistance to 80% of the static
condition. The result is then compared to the reduced safety factor.
Seismic Design Commentary
Given the limited information that is typically available regarding seismic design parameters, site
response factors, and the empirical nature of the psuedo-static seismic procedure presented, it is
the author's opinion that AASHTO-Section 5 tries too hard to be precise without being more
accurate. Ground acceleration coefficients are typically guessed at or approximated for
geographical areas by design codes so it makes little sense to guess at "A" then arbitrarily
increase "A" to "Am", then decide whether to use Kh = Am/2 or 100% of Am in determining the
dynamic component of which we are only going to apply at 50% for some other logical reason.
Internally, it is not obvious why the seismic loading would be distributed proportionally by
embedment lengths. It appears that distributing the load equally to the number of reinforcement
levels would accomplish a similar distribution with the same level of precision.
Objective: Perform seismic analysis, A=0.20g, on 10' wall with 20 degree backslope, Keystone Compac
units, Rankine earth pressure, near-vertical wall assumed
A. EXTERNAL STABILITY
i = 20 degrees Soil Properties
φ Angle = 30 degree
Unit Wt = 120 pcf
W3 Cohesion = 0 psf
ψ = 1.474
Kae = cos2(φ−θ+β)
ψcos(θ)cos2(β)cos(δ−β+θ)
cos2(30−7.13+0)
Kae =
1.474cos(7.13)cos2(0)cos(20−0+7.13)
Kae = 0.652
Es = 1/2(120pcf)(12.91')2(.414) Ed = 1/2(120pcf)(11.78')2(1-0)(.652-.414)
[1200+9600+1397+(4140+991)(0.342)](0.577)
SF =
(4140+991)(0.940)+(1200+5868+522)(0.125)
Note: Sliding reduction (Cds)along
8,055 lbs/lf reinforcement levels is not checked in
SF = = 1.40 > 1.1 OK > 75% of 1.5 this example due to transient nature
5,772 lbs/lf of seismic loading.
SF = Resisting Moments
Overturning Moments
72,182 ft-lbs
SF = = 2.54> 1.5 (OK > 75% of 2.0)
28,429 ft-lbs
R
σv =
L-2e
σv = 13,952
9'-2(1..36')
B. INTERNAL STABILITY
i = 20
degrees
W3 3.00'
Hw 4.15'
Hw = H/(1-tan(ι)/tan(ρ)) = 12.66'
10' W2 5.30'
W1 6.46'
ρ = 45+φ/2 = 60 degrees
7.61'
1' 5.77'
9'
Am = (0.20)(1.45-0.20) = 0.25g
Pi = Am x (W1+W2+W3)
Pi = 0.25(1200+3462+921)
Pi = 1,396 lbs/lf
The loads must be distributed to the reinforcements proportionally to their resistant areas (Le lengths) and
compared to 75% of static safety factors for tension, connection and pullout. Use KeyWall analysis w/o liveload
for data:
2. Tensar UX1400, LTDS = 1333plf (889plf * 1.5) for seismic safety factor calculation.
3. Connection strengths reflect minimum 1.5 safety factor so multiply KeyWall allowable peak connection
strengths times 1.5 for seismic safety factor calculation.
4. Static pullout resistance can be determined in KeyWall by multiplying the calculated tension w/o
liveload times the calculated pullout safety factor. This pullout resistance is taken at 80% to calcluate
seismic safety factor.
OK - Calculated reinforcement tensile and connection capacity exceed the 1.1 minimum static + dynamic safety
factor criteria. Note: Some specifications permit a reduction in the creep factor on geosynthetics for a seismic
analysis but the actual method of analysis differs.
Summary of Pullout Resistance
The pullout resistance of each geogrid level must be checked at 80% of the static pullout resistance against the
combined static + dynamic loading previously determined. A minimum pullout safety factor of 1.1 (75% of 1.5) is
required.
From Pullout From above
From KeyWall KeyWall Resistance Static +
Static tension FSpo Tension x FSpo Dynamic
SUMMARY
The example problem meets the seismic design criteria stated in the AASHTO design manual. It should be noted
that a static-only design for this structure might only require reinforcement lengths of 7 or 8 ft, and meet
existing design standards. This example utilized 9 ft. reinforcement lengths and five levels of medium strength
reinforcement to insure seismic compliance which is confirmed in this example.