You are on page 1of 23

DEPARTMENT OF THE ARMY

U.S. Army Corps of Engineers


Washington, DC 20314-1000

CECW-ED
Technical Letter
No. 1110-2-321

ETL 1110-2-321

31 December 1993

Engineering and Design


RELIABILITY ASSESSMENT OF NAVIGATION STRUCTURES
STABILITY OF EXISTING GRAVITY STRUCTURES

Distribution Restriction Statement


Approved for public release; distribution is unlimited.

DEPARTMENT OF THE ARMY


U.S. Army Corps of Engineers
Washington, DC 20314-1000

CECW-ED
Technical Letter
No. 1110-2-321

ETL 1110-2-321

31 December 1993

Engineering and Design


RELIABILITY ASSESSMENT OF NAVIGATION STRUCTURES
STABILITY OF EXISTING GRAVITY STRUCTURES

1. Purpose
This engineer technical letter (ETL) supplements
ETL 1110-2-532 and provides guidance for assessing the reliability of existing gravity structures
founded on rock, as a basis for rehabilitation
investment decisions.
2. Applicability
This ETL applies to HQUSACE elements, major
subordinate commands, districts, laboratories, and
field operating activities having responsibilities for
the design of civil works projects.
3. References
a. EM 1110-2-2602, Planning and Design of
Navigation Lock Walls and Appurtenances.
b. ETL 1110-2-310, Stability Criteria for
Existing Concrete Navigation Structures on Rock
Foundations.

for new structures. The use of these criteria may be


overly conservative for older structures that have
performed satisfactorily. Therefore evaluation
criteria were developed for existing structures,
Reference 3b.
b. While most concrete gravity structures have
performed well in the past, there have been notable
exceptions that cause concern. A 600-foot length of
Dam 26 on the Ohio River failed by sliding in
1912. The concrete sill for the wicket dam was
founded in shale. Fifteen monoliths of TVAs
Wheeler Lock slid along a weak clay shale seam in
limestone into an adjacent excavation for a new
lock on 2 June 1961. Both of these structures slid
along weak seams that were undetected by exploratory borings and escaped notice during construction. Stability of older structures is a concern
because they were designed and constructed to
lesser standards than are used today, and foundation
explorations were generally limited. In addition,
project modifications have increased the loading on
many of those older structures, and information on
their design and construction is frequently lacking.

c. ETL 1110-2-338, Barge Impact Analysis.


d. ETL 1110-2-532, Reliability Assessment of
Navigation Structures.
e. Memorandum. CECW-OM-O dated 21 July
1992, Subject: Guidance for Major Rehabilitation
Projects for Fiscal Year 1995.
4. Background
a. The stability of existing gravity structures
has historically been evaluated using design criteria

c. Reliability analyses provide a good means


of evaluating stability of existing structures. The
analyses combine what is known about a structure
with reasonable limits for the unknowns to assess
reliability. Thus the reliability index reflects the
uncertainties associated with older projects and
gives a good indication of the level of confidence
that should be placed on the structure. The principles of reliability analyses are presented in Reference 3d. This ETL applies those principles to
assessing the stability of existing gravity structures
on rock foundations.

ETL 1110-2-321
31 Dec 93
d. Reliability assessments of the stability of
navigation structures will be used in rehabilitation
investment decisions, Reference 3e.

as a function of the base dimensions and


eccentricities of the resultant. This relationship is
derived in Appendix A.

5. Objectives

d. Integration of the performance function,


C/D, for the performance modes of sliding, foundation bearing, or resultant location may be approximated by either the point estimate or Taylors series
methods described in Reference 3d.

This guidance is intended to provide:


a. A consistent method of evaluating the
stability of concrete gravity structures on rock in
terms of the reliability index, (Ref 3d).

e. General guidance for applying reliability


analyses to the stability of gravity structures is
given in Appendix A of this ETL. An example
application of reliability assessment methods to
stability of a gravity structure is in Appendix B of
this ETL.

b. A consistent method for prioritizing the


investments needed to assure an acceptable level of
reliability in navigation systems.
6. Methodology

7. Actions

a. Reliability indices, , are used as a relative


measure of reliability or confidence in the stability
of concrete gravity structures.

a. Reliability assessments should be made of


gravity navigation structures that fail to meet current design criteria for stability. Reports recommending project rehabilitation or replacement
should include reliability assessments of deficient
gravity structures.

b. Stability will be evaluated for sliding,


foundation bearing, and resultant location.
c. The reliability indices, , will be calculated
using the performance function, capacity, C, divided
by demand, D. For sliding analyses, capacity will
be the resistance along a sliding plane beneath the
structure, plus the resistance of earth and rock in
which the structure is embedded. Demand will be
the applied loads on the structure tending to cause
sliding. For foundation bearing, capacity will be
taken as the bearing capacity of the foundation.
Demand will be the maximum bearing pressure at
the toe of the structure. For resultant location, the
ratio of capacity divided by demand will be taken

b. Reliability assessments of navigation structures failing to meet current design criteria for
stability should be included in periodic inspection
reports for projects inspected after the date of this
ETL.
c. This ETL supplements ETL 1110-2-310,
Reference 3b, for investment decisions. Recommendations for major rehabilitation involving
gravity navigation structures will be based on the
results of reliability analyses described in this ETL.

FOR THE DIRECTOR OF CIVIL WORKS:

PAUL D. BARBER, P.E.


Chief, Engineering Division
Directorate of Civil Works

2 Appendices
APP A - Guidance for Applying Reliability
Analysis to the Stability of Gravity
Structures
APP B - Example Application for Guide
Walls

ETL 1110-2-321
31 Dec 93

APPENDIX A
GUIDANCE FOR APPLYING RELIABILITY ANALYSIS TO THE STABILITY
OF GRAVITY STRUCTURES

1. Introduction

3. Loading Conditions

a. The objective of this ETL is to provide a


methodology for reliability assessment of the
stability of existing gravity structures founded on
rock. These reliability assessments are intended to
be sufficiently consistent and uniform to be used in
prioritizing structural rehabilitation investments.
ETL 1110-2-532 should be consulted for details of
the principles of reliability analyses.

Loadings of interest in reliability analyses will


generally be the same as those used in the design of
the structures. Conventional deterministic analyses
may be used to limit the number of conditions
analyzed to controlling cases that are typical of
normal operation, maintenance, extreme operation
such as higher pools where applicable, and
catastrophic events such as the loss of lower pool or
an earthquake.

b. Traditionally, stability has been expressed in


terms of factor of safety against sliding or excessive
foundation bearing pressure, and location of the
resultant with respect to the kern or quarter point of
the base. For sliding and excessive bearing pressure, the factor of safety can be easily expressed in
terms of capacity divided by demand. For location
of resultant, an alternate relationship is presented in
paragraph 4c of this appendix.

a. Normal operating condition. Navigation


projects usually have either gated or fixed crest
dams. Pools for gated dams are normally controlled
within close tolerances in accordance with operating
rules and are predictable. The normal operating
condition for gated dams will be taken as the combination of normal upper pool controlled by the
dam gates and the minimum seasonal lower pool
that can reasonably be expected on an annual basis.
The normal operating conditions for fixed crest
dams will be taken as the more critical of: (1) the
upper pool as determined by the fixed crest elevation of the dam and the lower pool as controlled by
the downstream dam, or (2) the most critical combination of pools that routinely occur due to fluctuations in river flow. This may be taken as a set of
fixed values where the pools of structural significance can be determined with certainty and as variables where there is uncertainty. Additional discussions of pool levels for reliability analyses are
found in paragraph 5c(1) of this appendix.

c. Reliability analyses are ideally suited to


assessments of the stability of older structures.
While there are uncertainties as to design and construction, reasonable limits can usually be assigned
to these variables based on available information.
Thus the uncertainties that are reflected in the reliability indices indicate the level of confidence that
should be placed in the stability of the structure.
2. Representative Structures
Reliability analyses should be performed on a
limited number of structures which are important to
the operation of the project and are representative of
other similar project structures. Conventional deterministic analyses may be used to select typical
piers, sills, lockwall monoliths, or other structures.
Structures which are shown by analysis or evidence
of distress to be marginally stable will be of
primary interest. Structures which appear adequate
should also be analyzed for comparison.

b. Maintenance condition. The extreme


maintenance condition for most structures will be
the dewatered condition which places the structure
under maximum differential water loading. The
pool used in conjunction with dewatering should be
considered a fixed value where it can be determined
with certainty, and as a random variable where
there is uncertainty. The pool levels that affect

A-1

ETL 1110-2-321
31 Dec 93
stability during dewatering are controlled by the
elevation of the maintenance bulkheads, the operation of gated dams, the hydraulic rating curves for
fixed crest dams, and hydrologic conditions.
Hydrographs that are kept as a part of a projects
records will be useful in assigning a value or a
distribution to pool elevation.

Nevertheless, it is prudent that bearing capacity be


evaluated as a performance mode. An appropriate
method based on Corps criteria should be chosen to
meet the condition of the foundation rock.
Meyerhofs bearing capacity factors, including
reductions for inclined loads, are used in this ETL
to assess the reliability of gravity structures for the
performance mode of foundation bearing. These
bearing capacity factors were developed for homogeneous, isotropic soils but may be used for rock
foundations with closed near vertical joints and
other conditions that approximate homogeneous
isotropic conditions. In applying bearing capacity
factors, it may be necessary to use the mean values
of rock strength as constants. Bearing capacity
factor equations are sensitive to large variations of
which result from the combination plus or
minus .

c. Extreme operating conditions. Some navigation projects are multipurpose projects and may
retain high pools under extreme operating conditions. Reliability analyses should be made for a
sufficient number of expected headwater and tailwater conditions to allow the reliability index to be
reported as a function of the level and frequency of
the pool.
d. Catastrophic loading conditions. Reliability
analyses should include conditions such as loss of
lower pool where a probability of occurrence can be
placed on these events and lower levels of reliability would be anticipated.

c. Resultant location.
(1) Two-dimensional stability. Resultant location is the performance mode selected herein to
replace overturning analyses. Overturning is
unlikely as a pure mode of failure, as foundation
bearing and/or sliding failure would occur before
the resultant reached the toe. In practice, the location of the resultant on the base is used to determine the percentage of the base in compression
which is then used as a measure of stability. Reliability analyses, however, require that stability be
expressed in terms of capacity divided by demand,
i.e., a performance function. This ratio can conveniently be represented by the equation

4. Performance Modes
The three performance modes that describe stability
are sliding, foundation bearing, and resultant
location.
a. Sliding stability. Sliding may occur at the
base of a structure, but experience has shown that
sliding along weak seams in the rock is more likely.
For sliding along the base, the shear strength used
in the analyses is the lesser of the contact strength
of concrete on rock and the rock shear strength as
defined by an angle of internal friction, , and
cohesion, c. Sliding may or may not involve a
passive wedge, and will depend on embedment of
the structure in rock and orientation of weak seams
within the foundation. Where a passive wedge is
considered, the magnitude of the passive force or
factors affecting the force must be treated as random variables. These factors are discussed in detail
in paragraph 5k below.

C
D

B
B 2XR

The derivation of this equation follows:

b. Bearing capacity. Foundation bearing failure for structures founded on rock is unlikely.

A-2

ETL 1110-2-321
31 Dec 93

where
e
M = summation of moments about centroid of
base

M
Rv

If moment M about the centroid of the base is taken


as demand and Rv B/2 is taken as capacity, the
equation for capacity divided by demand can be
derived as follows:

RV = resultant of vertical forces


RH = resultant of horizontal forces

C
D

XR = distance from toe to N


e = eccentricity of resultant forces about centroid
of the base

A-3

RV B/2
M

RV B/2
RV B/2

XR

B
B 2XR

ETL 1110-2-321
31 Dec 93
certainty and should be treated as a random variable. Research under the REMR program and
guidance in Engineer Technical Letters and Engineer Manuals indicate that earth pressure will
normally be between active and at-rest values.
Earth pressure for structures founded on firm rock
will tend to be nearer at-rest values. Earth pressures for walls founded on piling or earth foundations will tend to be nearer active earth values.
Small movements of walls with dense backfills will
allow the establishment of active earth pressures.
This movement may be in the form of angular
rotation due to deformation of the foundation or it
may be due to sliding. Rotational movement of
walls founded on firm rock, however, will normally
be insufficient to develop active pressures.

(2) Three-dimensional stability. Structures such


as dam piers and miter gate monoliths are normally
analyzed for stability about two axes. The equation
for capacity divided by demand for problems of
three-dimensional stability can be expressed as
follows:
C3D

D3D
B

2XR

B 1

2ZR

where the eccentricities and base dimensions are as


shown in Figure A-1.
5. Forces

Sliding movement of structures founded on rock


may be possible; however, it should be noted that
any reduction in earth pressure attributed to sliding
may be accompanied by a reduction in foundation
shear strength due to the movement. Additionally,

a. Earth pressure.
(1) Lateral driving earth forces. Earth pressure
on navigation lock walls is not known with great

Figure A-1. Plan view of base

A-4

ETL 1110-2-321
31 Dec 93
any reduction in earth pressure due to either angular
rotation of the base or lateral displacements should
be applied equally to analyses of sliding stability,
resultant location, and foundation bearing. Distributions of earth pressure values for gravity structures
should have a mean value that is weighted toward
the more likely state of stress in the backfill and
has a dispersion that considers the possible extremes
due to uncertainties in the densities of the fills and
movements that may occur in the structures. Additional discussion of earth pressure on walls is in
paragraph 3a(1) of Appendix B.

c. Water loads.
(1) Pool levels. Pool levels at navigation
projects are controlled by project features,
operating procedures, and hydrological conditions.
Knowledge of these factors is essential in defining
the certainty or uncertainty of pool levels selected
for analyses. Operating procedures are dictated by
project purposes, i.e. flood control, hydropower
generation, recreation, water supply, and navigation.
Pools are controlled for the various project purposes
as much as hydrological conditions will allow. At
fixed crest dams, pool levels are a result of hydrological conditions, weir discharge coefficients, and
possible control at upstream or downstream
projects. Gated spillway dams operated for navigation alone have pools that are closely regulated by
both design features and operating procedure. Pool
combinations of structural interest are predictable
and can be verified by review of project records.
The pools at fixed crest dams are not regulated but
will fluctuate predictably in accordance with
hydraulic rating curves. Pool combinations that are
of structural significance can be found to occur a
large part of the time and can usually be assigned
single values. However, it should be noted that, the
nominal upper and lower pool combination for
some fixed crest dams is not the critical load. The
critical load should be determined for the structure.
This critical pool combination usually occurs with
moderate rises in the river causing the upper pool to
rise a few feet and the tailwater to build a moderate
amount. The nominal design pool combination of
minimum upper and lower pools, which is associated with no flow in the river, is a rare event.
Projects which include flood control, and some with
hydropower as a project purpose, have been
designed for a probable maximum flood. Reliability assessments of these projects will normally
be based on the probable maximum pool, or some
other designated pool of interest.

(2) Lateral resisting earth forces. Consideration


must be given to the strain compatibility of materials offering passive resistance to sliding. If peak
strengths of these materials are not reached at the
same strain as the foundation rock, their combined
peak strengths would not be mobilized at the limit
state of imminent sliding. The full passive capacity
of earth backfill may not be available due to stressstrain incompatibility of earth and rock, and the
available resisting force is likely to be closer to the
at-rest value.
(3) Vertical earth pressure. Moist and saturated
soils resting on the structure offer resistance to
sliding and overturning and affect bearing pressure
and capacity. The unit weights of these materials
may be known or can otherwise be estimated with
reasonable accuracy from available information.
Common values and coefficients of variation for
soils are found in engineering textbooks referenced
in ETL 1110-2-532.
b. Wall friction. While lateral movement of
concrete gravity walls founded on rock may not be
sufficient to develop wall friction, settlements which
occur in most backfills would produce wall friction.
Wall friction would depend on the capacity of the
soils involved to sustain friction forces. If the soils
are not silts or soft clays, it would be possible to
develop wall friction up to a maximum of the angle
of internal friction of the backfill although values of
60-70 percent would be more likely. Values of
wall friction for reliability analyses should be based
on evaluation of available project data. For granular
backfills, a mean friction angle = 50 percent of
and a coefficient of variation V = 25 percent of the
mean would be reasonable.

(2) Saturation of backfill. The water pressure


in backfill is one of the most significant loads on
lock walls, and water level in backfill has been
found to vary substantially with projects. Remedial
work has been undertaken at numerous projects to
lower the saturation level. Special effort should be
made to obtain data on the saturation level, if it is
not provided. The design elevation will not always

A-5

ETL 1110-2-321
31 Dec 93
forces on flexible structures. Coefficients of variation of 20 percent applied to these values computed
by rigorous analyses should be adequate.

be a good indicator of the water level in the backfill. If drains were incorporated in the backfill, the
saturation level would be more likely be near lower
pool. If random backfill were used, the saturation
level could be expected to be nearer the upper pool
level. Without piezometer readings or other information, the water level should be assumed to have a
uniform probability distribution with values between
the two extremes.

h. Hawser pull. Hawser pulls are from lines


attached to line hooks, check posts, or floating
mooring bitts. They are used to help stop the forward momentum and control the alignment of tows
using the locks. Hawser pulls can vary substantially with the size of tows using the locks and river
conditions that affect the way tows approach or
leave a lock. Hawser pulls may be limited only by
the strength of lines used for checking but are generally taken as 100 to 130 kips. A VI of 15 to
20 percent will be adequate.

d. Uplift. Uplift is one of the more uncertain


variables in stability analyses. Reliability analyses
need to consider realistic values as well as probable
extremes. For reliability assessment, the project
data should be closely examined to determine a
realistic mean value and distribution for the uplift
variable. Where project data do not include uplift
measurements, pressure assumptions should be
made in accordance with EM 1110-2-2602.

i. Concrete. When available, project data


should be used for the determination of mean unit
weight and standard deviation for concrete. In lieu
of project data or other local sources of data, typical
values and coefficients of variation may be found in
engineering textbooks referenced in ETL 1110-2532.

e. Seismic forces. Performance under seismic


loading is usually a safety issue and not a major
rehabilitation investment decision. Therefore, seismic forces should not be considered in reliability
analyses.

j. Shear strength of foundation rock. Foundation shear strengths are not known with great
certainty and are treated as random variables.
Uncertainty falls into three categories: (1) uncertainties due to sampling and testing representative
weak rock strata, (2) uncertainty as to the correlation between the peak shear strength of small rock
samples and prototype rock foundations, and (3) uncertainties due to past geological events.

f. Ice loads. Ice forces will vary with climate


and pool conditions. If ice forces are significant for
the structure being evaluated, they should be treated
as a random variable. It would be reasonable to
assume the mean as the value generally used for
design in any given locality. Coefficients of variation of 15 to 20 percent may be reasonable
depending on the site conditions.

(1) Uncertainties due to sampling and testing


representative weak rock strata. Samples of intact
rock recovered for testing are normally representative of the stronger rock in the foundation, as the
sampling operations often destroy the weaker rock
strata. Clay seams and open bedding planes may
not only remain unsampled, they may remain
undetected.

g. Impact. Expected values of impact forces


should be calculated by methods contained in
ETL 1110-2-338, "Barge Impact Analysis." Impact
forces vary by project with the size of tows, their
speed as they approach the structures, and river
conditions that affect maneuverability. Impact
forces due to tows landing on guide and guard walls
may be large, 600 to 800 kips and highly variable.
A coefficients of variation VI = 30 percent is considered appropriate. Impact forces on lock walls
within the chamber will usually be small enough
not to affect the analyses and can generally be
assigned a single value, typically 100 kips. Impact
forces on structures such as gates can be large, and
may be critical depending on the performance mode
being analyzed. ETL 1110-2-338 places an upper
bound of 1950 kips for impact forces on rigid structure and presents methods for calculating impact

(2) Uncertainty of correlation between the


peak shear strength of small rock samples and prototype rock foundations. As a general rule, shear
strengths based on the peak strength of small intact
rock samples or small samples containing the natural discontinuity will overpredict prototype foundation strength. Such tests, however, are useful in
establishing the upper limits of foundation strength.
The extent of discontinuities in the foundation rock
will need to be evaluated in establishing a mean

A-6

ETL 1110-2-321
31 Dec 93
value for the variable, shear strength of the foundation rock.

embedment in rock. Where information on the


nature of the rock jointing is unavailable, analyses
must reflect this uncertainty.

(3) Uncertainties due to past geologic events.


Past geologic events may have destroyed the intact
strength of the rock, introduced stress changes in
the rock mass, created weak seams or joints, or
produced fault zones in the mass. Refer to appropriate Corps guidance for a discussion of some of
these events.

(2) Methods of excavation. Excavation by


blasting will frequently damage rock in the faces of
excavations. The upper layers of rock are
especially vulnerable to damage as they lack confinement and are subject to uplift by explosive
gases. The upper layer of rock is also likely to be
the weaker rock and thus more likely to be damaged. Where construction records are unavailable,
analyses must reflect this uncertainty.

(a) Uncertainties due to geologic events generally reduce to evaluating the shear strength of weak
seams and evaluating the influence of jointing and
bedding upon uplift distributions and/or the formation of unstable blocks. Both of these considerations require the determination of whether the
shear strength along potential sliding planes is at the
intact or residual strength of the rock.

(3) Placement of concrete. Unless construction records show the concrete to be placed against
rock, analyses again must reflect uncertainty as to
available passive resistance due to embedment in
rock.

(b) The question of using residual versus peak


strengths can usually be resolved during construction by extensive explorations and investigations
while the structure excavation is being made. The
determination of appropriate rock strength for older
projects will be more difficult. Where adequate
information for this determination is not available,
engineering judgment will be required.

6. Reliability Analyses
Reliability analyses for stability of gravity structures
are basically the same as analyses for steel structures addressed in detail in ETL 1110-2-532. The
main difference is the limit states that define
stability differ from those for steel structures, and
material properties of soil and rock are more variable. The steps to be taken are described briefly
below and demonstrated in the example in Appendix B of this ETL.

(c) Residual strengths represent a lower bound


for prototype foundation strengths. The use of the
residual strength test values for sliding assumes
sliding to be unimpaired by any irregularities or
discontinuities along the sliding surface. Adjustments in test values are necessary to account for
undulations and discontinuities where they exist.

a. Stability model. Develop a model which


describes the performance mode and limit state to
be evaluated. The model for a lock wall may be a
cross section of the wall complete with loading
diagrams and stability analysis.

k. Passive resistance of rock. Embedment in


competent rock of only a few feet can offer
substantial resistance to sliding. The passive
resistance in sliding analyses, however, is normally
used with much caution due to uncertainties about
the rock mass structure and possible damage to the
rock mass during construction which may have
negated the benefits of embedment. The following
uncertainties need to be considered in evaluating
sliding stability.

b. Limit state. Determine the limit states.


Limit states for stability are defined in this ETL as
the ratios of capacity to demand which result in
unsatisfactory performance as the ratios approach
unity. These ratios are defined for location of
resultant, sliding, and bearing capacity.
c. Parameters. Identify material properties
and structure geometric parameters that enter into
the analyses of stability. Determine parameters
which may be considered constant and those which
should be treated as random variables. Generally
parameters with variances that account for 5 percent
or more of the loads or resistances acting on the
structure will be treated as random variables.

(1) Joint patterns in rock. Rock in which


structures are embedded will frequently have
adversely oriented inclined joints. Consequently,
rock strengths taken from tests on intact rock may
overestimate the passive resistance available due to

A-7

ETL 1110-2-321
31 Dec 93
E [C] or C, and E [D] or D of the individual
capacity and demand functions, by the equation

d. Variables. Assign means, standard deviations, and coefficients of correlation to random


variables based on available or obtainable information. These may be derived from distributions,
calculated directly from data, or assigned
judgmentally.

VC

ln [1

ln C/D

VD

D
D

ln C / D

(1-3)

ln C / D

ln

C / D / C / D

and
ln C /D
2

(1-1)

ln C/D

2
C/D

ln C / D

7. Interpretation of Results

where the coefficient of variation


VC/D

These moments are related to the mean, C/D, and


the standard deviation, C/D, of the probability
distribution of C/D by

(1) The reliability index, , may be expressed


as a function of the mean, C/D or expected value,
E [C/D], and standard deviation, C/D, of the performance function as

E [ln (C/D)]
ln (C/D)

(3) The reliability index, , may also be


expressed in terms of the mean, , and the standard
deviation, , of the logarithm of the performance
function C/D as

g. Calculate reliability index, . Calculate the


reliability index, , by either the point estimate
method or by the Taylors series method. Analyses
by the Taylors series method will show which
variables significantly affect the reliability index, .
Once the moments of capacity, C, and demand, D,
are determined by either the point estimate or the
Taylors series methods, the reliability index, ,
may be calculated from equation 1-1, 1-2, or 1-3 as
follow:

VD

where the coefficients of variation of capacity and


demand are defined as the ratio of the standard
deviation of the variable to its mean value as

f. Means and standard deviations of performance function. Determine the means and standard
deviations of the performance function, capacity, C,
divided by demand, D, for the performance modes
of sliding, foundation bearing, and resultant
location.

(1-2)

VC

e. Performance function. Formulate the performance function, C/D, in terms of constants and
variables that describe the performance mode of
interest.

E [C / D]
ln
1 V2
C/D

ln (E [C] / E [D])

C/D

a. Compare structures. The results of


reliability analyses may be used to identify deficient
structures in need of stabilization and to prioritize
investment decisions. Target reliability indices that
may be used in evaluating and comparing structures
are given in ETL 1110-2-532.

C/D

(2) The reliability index, , may also be


expressed in terms of the expected or mean values,

A-8

ETL 1110-2-321
31 Dec 93
b. Define investigative program. The results of
reliability analyses may be used to identify problem
areas and areas of uncertainty that require exploratory investigations to better define reliability.
c. Establish monitoring programs. The results
of reliability analyses may be used to identify
potential problem areas that need to be monitored to
determine rates of degradation.

A-9

ETL 1110-2-321
31 Dec 93

APPENDIX B
EXAMPLE APPLICATION FOR GUIDE WALLS
The foundation rock for the guide wall is described
in five borings taken in 1977 and two taken in 1973
as a highly fractured and weathered siltstone. The
three borings through the concrete indicated poor
contact between concrete and rock. This may have
been due to the highly fractured nature of the rock,
and possibly a result of the drilling process. Nevertheless, there is no indication of bond between the
concrete structure and its foundation. The guide
wall was constructed within a low cofferdam consisting of parallel wood sheeting with earth fill and
earth berms. It is therefore reasonable to assume
there was some foundation preparation that would
have led to better contact between the structure and
the foundation than could be inferred from the
foundation boring data. This discrepancy leads to
much uncertainty as to the appropriate foundation
strength.

1. Introduction
This example application is taken from analyses of
guide wall monolith L-8 at Locks and Dam 3 on the
Monongahela River. The locks and dam were
constructed between 1905 and 1907 and placed in
operation in May 1907. The guide wall is typical
of structures built during that period in that most
have relatively narrow bases and consequently are
marginally stable.
2. Description of Structure
a. Gravity wall. Monolith L-8 is taken to be
representative of a section of the upper guide wall
monoliths. It is approximately 27 feet in height and
is founded on or near top of rock which is horizontally bedded and slopes toward the river. A cross
section of the wall is shown in Figure B-1.

Figure B-1. Example problem. Guide wall stability evaluation

B-1

ETL 1110-2-321
31 Dec 93
periods that the maximum differential water loads
on the wall occur. Based on observed piezometric
data, a mean differential hydrostatic load of 2 feet
with a standard deviation of 1.0 foot will be used to
evaluate monolith L-8. This loading occurs often
enough to be considered a normal loading
condition.

b. Backfill. The guide wall was backfilled with


available excavated materials. This fill consists of
highly variable mixtures of sands, silts, clays, and
rock fragments. Standard penetration values for the
backfill range from a low of 3 to a high of 21. The
rock fragments may account for some of the higher
values. The backfill slopes upward approximately
2 feet to the top of a low berm immediately behind
the wall. This has the effect of inducing a surcharge loading on the backfill.

(3) Hawser pull. A hawser pull with a mean


value of 1 kip per linear foot of wall and a standard
deviation of 0.5 kip is used in the reliability analyses for resultant location. A constant value of
1.0 kip per linear foot is used in the sliding analyses. The 1.0-kip force accounts for about 30 percent of the total moments in resultant location
analyses but only about 5 percent of the forces in
sliding analyses. The hawser pull thus has significant impact on resultant location and negligible
impact on sliding analyses. The hawser force is
applied 5 feet above pool level but no higher than
1.0 foot above the top of wall.

c. Basis for structure selection. Monolith L-8


was selected for example purposes only and because
project data is available. The monolith has a flat
base while others have stepped bases, making
analysis more difficult. Some of the monoliths with
stepped bases extend to lower elevations and stability of these may be more critical than for monolith
L-8.
3. Forces

(4) Wall friction. A mean value for wall


friction = 12o, or about 40 percent of , and a
standard deviation = 3o are used in the reliability analyses. These values are chosen to be a little
conservative because of the random nature of the
backfill and the presence of silts and clays.

a. Random variables. The following loads


have been identified by preliminary analyses as
significant random variables. Other loads will be
treated as constants.
(1) Lateral earth pressure coefficient, backfill.
Lateral earth pressure is taken to be normally distributed with a mean value of the coefficient k =
0.42 and a standard deviation K = 0.08. The standard deviation K = 0.08 represents a significant
degree of uncertainty due to the variable nature of
the backfill materials. The k = 0.42 is slightly less
than the full at-rest value as the structure is founded
on rock of poor quality and has a rather narrow
base. Full at-rest pressure is considered to be near
k0 = 0.47 which is based on assumed backfill
strengths of = 32 and c = 0 psi.

(5) Foundation rock strength. The foundation


strengths for Monongahela River Locks and Dam 3
have been derived from direct shear tests on intact
samples of the foundation rocks. Residual values
were obtained from direct shear tests of smooth
sawn samples. Results of the direct shear testing
are summarized below.
E[c] VC
Vtan

ksf percent Tan %


tan,c
Peak strength

(2) Saturation level in backfill. Monongahela


River Dam 3 is a fixed crest dam and the upper
pool varies with the flow in the river. A single
piezometer within the backfill shows a saturation
level that is frequently below river level. This
suggests that the seepage gradient around the locks
starts near the upper end of the guide wall. During
periods of rapid drawdown following high river
stages, the saturation level may exceed the river
level by as much as 2 feet. Readings are sparse
following lock-wall overtoppings which occur on
slightly less than an annual basis. It is during these

Residual strength

11.0 70.0

1.500 45.0 -0.70


0.632 50.0 ------

The foundation for L-8 is described in boring logs


as a weathered, highly fractured siltstone. The peak
strengths tabulated above are based on intact samples, and therefore, will be used as an upper limit
of foundation strengths. The residual strengths
should be taken as a lower limit on foundation
strengths. Extremely weak seams in siltstones are
less common than in materials such as clay shales.

B-2

ETL 1110-2-321
31 Dec 93
Analyses for the example problem in this enclosure
will be based on peak strengths. Statistical values

for the results of the direct shear tests on intact rock


samples are shown in Table B-1.

Table B-1
Example for the Determination of a Coefficient of Correlation for Tan and C and the Probability
Distribution Factors for Peak Strengths of Foundation Rocks

Test No.

deg

1
2
3
4
5
6
7
8
9
10

60.8
68.9
54.7
67.8
58.7
30.4
44.5
56.2
36.1
46.3

Mean
Std. Dev.

= 52.4
= 12.9

Tan

C
ksf

Tan-tan

C-c

(Tantan)*
(c-c)

1.79
2.59
1.41
2.45
1.64
0.59
0.98
1.49
0.73
1.05

9.1
3.8
15.3
2.9
6.9
29.3
11.1
6.3
15.3
10.4

0.29
1.09
-0.09
0.95
0.14
-0.91
-0.52
-0.01
-0.77
-0.45

-1.90
-7.20
4.30
-8.10
-4.10
18.30
0.01
-4.70
4.30
-0.60

0.55
-7.86
-0.39
-7.70
-0.57
-16.65
-0.01
0.05
-3.31
0.27

tan = 1.5
tan = 0.675

c = 11.0
c = 7.7

COV(tan,c) = -3.56

The covariance of tan and c, COV(tan,c), is calculated using the equations:

COV (tan,c)

E [(tan
1
N

tan) (c

(tani

c)]

tan) (ci

c)

i 1

where N = the number of pairs of tan and c.


The coefficient of correlation for the peak shear strength of the foundation rock is determined by:

COV (tan,c)
tan C

tan,c

3.56
0.675 x 7.7

0.68 0.7

The probability concentration factor for the correlated variables, tan and c are based on:

Ptan

,c

Ptan

,c

1
[ 1
2n

tan,c ]

1
(1
22

( 0.7))

1
(0.3)
4

0.075

Ptan

,c

Ptan

,c

1
[ 1
2n

tan,c ]

1
(1
22

( 0.7))

1
(1.7)
4

0.425

where n = the number of correlated variables, and tan, c = -0.7.

B-3

ETL 1110-2-321
31 Dec 93
Preliminary analyses have also shown this to be the
controlling stability case for this structure. Values
for loads and resistances are shown in Tables B-2
and B-3. The values in Table B-2 are used as constants in the analyses. The loads and resistances in
Table B-3 are used as variables in the analyses.
Stability analysis of the monolith using mean values
of these loads and resistances is shown in
Table B-4.

b. Constants. The following loads are not


precisely known but preliminary analyses have
indicated they can be treated as constants without
significantly affecting the results of reliability
analyses.
(1) Unit weight of concrete. Measurements of
the weights of concrete at upstream and downstream
projects show mean weights of about 148 pcf with
standard deviations of less than 3 pcf. A constant
unit weight of 145 pcf is used in the analyses of
monolith L-8.

Table B-2
Example Problem. List of Constants

(2) Unit weight of moist backfill. The moist


backfill is assigned a mean unit weight of 130 pcf
based on densities estimated from standard penetration tests. Preliminary analyses have shown that
changes in weight affect both vertical and lateral
forces which tend to offset each other and have no
significant effect on stability of this particular
structure.
(3) Unit weight of saturated backfill. The saturated unit weight of the backfill is also assumed to
weight 130 pcf and is treated as a constant for the
same reason given above.

Constant

Value

Pool elevation
Unit weight of concrete, CONCRETE
Unit weight of soil, MOIST
Unit weight of soil, SATURATED
Phi angle of rock foundation,
(for bearing capacity)
Cohesion of rock foundation, c
(for bearing capacity)
Hawser pull (for sliding stability)
Uplift

732.9
145 pcf
130 pcf
130 pcf
52.4
11 ksf
1 kip
100 percent

Table B-3
Example Problem. List of Variables

(4) Uplift. Uplift is assumed to vary linearly


from full saturation head at the heel and portions of
the base not in compression to pool level at the toe.
The differential head is so small that variations in
its application have no significant effect on the reliability analyses of this particular structure.
(5) Lateral resistance of riverside overburden.
The overburden riverward of the guide wall provides less than 5 percent of the lateral resistance to
sliding or overturning and is therefore treated as a
constant. Strains in the wall which is founded on
rock will not be sufficient to develop passive
resistance in this overburden. Therefore, a coefficient of lateral earth pressure k = 0.5, which corresponds to an at-rest condition, is assigned to this
material.

Variable

Mean

Backfill saturation level


Backfill pressure coefficient, k
Wall friction angle
Tan of rock foundation, tan1
Cohesion of foundation rock, c1
Hawser pull (for overturning)

734.9
0.42
12o
1.50
11 ksf
1.0 kip

Std. Dev.
1 foot
0.08
3o
0.675
7.7 ksf
0.5 kip

Coefficient of correlation, tan,c = -0.7

5. Reliability Analyses
Reliability evaluations are made for the following
performance modes of resultant location, sliding,
and foundation bearing.
a. Resultant location. The point estimate
method is used to approximate the integration of the
resultant location performance function, which is
defined by the equation

4. Stability Analyses
A cross section of the wall and loads for which the
monolith will be analyzed are shown in Figure B-1.
These loads correspond to rapid drawdown following high water and overtopping which occur on an
annual basis. These loads are considered to be a
normal loading condition for the guide wall.

FRESULTANT LOCATION

B-4

B
2XR

ETL 1110-2-321
31 Dec 93
Table B-4
Stability Analysis of Example Guidewall Using Mean Values of Variables Shown in Figure 2-1

Item

Vert
Kips

Value

Horz
Kips

Arm
Ft

MomentTOE
(clockwise +)

Concrete

0.145

31.78

4.75

+150.96

Earth, VE&W

0.130

21.49

10.19

+218.98

Earth, HE1

k = 0.42

-14.07

10.27

-144.50

Earth, HE2

k = 0.50

0.42

1.67

+0.70

Hawser Pull

-1.00

27.90

-27.90

Water, HW1

0.0625

-20.80

8.60

-178.88

Water, HW2

0.0625

17.70

7.93

+140.36

2.99

14.00

+41.86

Wall Friction,

12O

Uplift, U1

0.0625

-20.82

7.00

-145.74

Uplift, U2

0.0625

-1.52

7.85

-11.93

Totals

33.92

-17.75

------

RESULTANT LOCATION, XR

XR

MomentTOE
vertical forces

FRESULTANT LOCATION

B
2 XR

43.91
33.92

1.30

14
2 x 1.30

14

1.23

SLIDING STABILITY

FSLIDING

FSLIDING

Sliding Resistance
Horizontal Forces

V Tan

Horizontal Forces

[33.92][1.50] 3 [1.30][11.0]
17.75

B-5

3 XR c

93.78
17.75

5.28

+43.91

ETL 1110-2-321
31 Dec 93
Table B-4 (Concluded)
BEARING CAPACITY

FBEARING

CicNC

Ult. Bearing Capacity


Found. Pressure

0.5 iB rock N

0.667 x
V

XR

qiqONq

(0.48 11 392) (0.48 0.34 510) (0.5 0.224 3 1.3 0.0975 1703.4)
0.667 x 33.92
1.30

FBEARING

(2069.76 83.23 72.54)


0.667 x 33.92
1.30

2225.54
17.4

127.9

Where Nq, NC, and N are bearing capacity factors with values as follow:

[ e ( tan )] tan2 ( 45o

Nq
NC

( Nq

1 ) cot

NC

5.14

( Nq

/2 )

( when

( when

0)

0 )

1 ) tan ( 1.4 )

and, c i, q i, and i are inclination factors with values as follow:

q i

c i
i

( 1
( 1

o/90o )2
o/o )2

and,
qo = effective overburden pressure at the bearing level.
= effective unit weight of the foundation material.
= angle of internal friction of the foundation materials.
c = cohesion of the foundation materials.
= inclination of the resultant force in degrees.
B = effective width of base in bearing.

The four random variables identified in paragraph 3


above are earth pressure, differential hydrostatic
pressure, hawser pull, and wall friction. The reliability analysis requires repetition of the conventional stability analyses to include the 16 possible
combinations of mean plus and minus one standard
deviation of each of the four random variables. The
combinations of random variables and results of the
reliability analysis are shown in Table B-5.

b. Sliding stability. The sliding stability evaluation is based on the limit state safety ratio
F

sliding resistance
1.0
driving forces

Both the point estimate method and the Taylors


series method are used to approximate the integration of the performance function, capacity divided

B-6

ETL 1110-2-321
31 Dec 93
Table B-5
Reliability Analysis of Example Guide Wall By the Point Estimate Method
Performance Mode: Resultant Location Under the Loading Condition of Drawdown, from Figure 2-1
Probability
pi

1
2
3
4

% Comp

Capacity
Demand
B/(B-2 XR)

pi (C/D)

pi(C/D)2

1.09
2.29
1.94
3.09

23.4
49.0
41.7
66.2

1.18
1.49
1.38
1.79

0.07
0.09
0.09
0.11

0.09
0.14
0.12
0.20

3
1
3
1

0.61
1.83
1.49
2.66

13.0
39.2
31.9
57.1

1.10
1.35
1.27
1.61

0.07
0.08
0.08
0.10

0.08
0.11
0.10
0.16

1.5
1.5
0.5
0.5

3
1
3
1

0.10
1.17
0.91
1.94

2.0
25.0
19.6
41.5

1.01
1.20
1.15
1.38

0.06
0.08
0.07
0.09

0.06
0.09
0.08
0.12

1.5
1.5
0.5
0.5

3
1
3
1

-0.65
0.45
0.21
1.26

0.0
9.6
4.5
27.0

0.92
1.07
1.03
1.22

0.06
0.07
0.06
0.08

0.05
0.07
0.07
0.09

Summations

1.26

1.64

KFILL

deg

Hawser
kips

H
ft

XR
ft

0.0625
0.0625
0.0625
0.0625

0.34
0.34
0.34
0.34

15
15
15
15

1.5
1.5
0.5
0.5

3
1
3
1

5
6
7
8

0.0625
0.0625
0.0625
0.0625

0.34
0.34
0.34
0.34

9
9
9
9

1.5
1.5
0.5
0.5

9
10
11
12

0.0625
0.0625
0.0625
0.0625

0.50
0.50
0.50
0.50

15
15
15
15

13
14
15
16

0.0625
0.0625
0.0625
0.0625

0.50
0.50
0.50
0.50

9
9
9
9

Run

E[C/D]

= C/D = pi(C/D) = 1.26

E[(C/D)2] = pi(C/D)2 = 1.64


Var(C/D) = 2C/D = E[(C/D)2] - (E[C/D])2 = 1.64 - (1.26)2 = 0.052
C/D = 0.23,

VC/D = C/D/C/D = 0.18,

= 1.21

where the reliability index, , is computed by the equation

E [C/D]
ln
1 V2
C/D

[ ln C/D ]
ln C/D

VC/D]

ln[1

by demand. The five variables identified in paragraph 3 above are earth pressure, differential hydrostatic pressure, wall friction, foundation phi angle,
and foundation cohesion. The first three are
random variables. The last two are correlated variables, i.e., c and tan vary together. The reliability
computations are much the same as for resultant
location, except probability concentration factors are
introduced into the analysis. These account for the
probability that values of the mean plus or minus
one standard deviation of the correlated variables
will be paired. Probability concentration factors for

this example problem are calculated in paragraph 3a(5) above. The combinations of variables
and results of the reliability analysis are shown in
Table B-6 for the point estimate method and
Table B-7 for the Taylors series method.
c. Bearing capacity. The bearing capacity
evaluation is based on the limit state safety ratio
F

B-7

foundation bearing capacity


1.0
foundation bearing pressure

ETL 1110-2-321
31 Dec 93
Table B-6
Reliability Analysis of Example Guide Wall By the Point Estimate Method
Performance Mode: Sliding Stability Under the Loading Condition of Drawdown, from Figure 2-1
Probability
pi

Tan
rock

CROCK
ksf

H
ft

KFILL

deg

XR
ft

Capacity
Demand

pi (C/D)

pi(C/D)2

1
2
3
4
5
6
7
8

0.00937

2.175

18.7

3
3
3
3
1
1
1
1

0.50
0.50
0.34
0.34
0.50
0.50
0.34
0.34

15
9
15
9
15
9
15
9

0.50
-0.22
1.52
1.05
1.55
0.86
2.69
2.25

4.84
3.31
9.72
7.93
8.46
6.26
16.19
14.22

0.05
0.03
0.09
0.07
0.08
0.06
0.15
0.13

0.22
0.10
0.89
0.59
0.67
0.37
2.46
1.89

9
10
11
12
13
14
15
16

0.00937

0.825

3.3

3
3
3
3
1
1
1
1

0.50
0.50
0.34
0.34
0.50
0.50
0.34
0.34

15
9
15
9
15
9
15
9

0.50
-0.22
1.52
1.05
1.55
0.86
2.69
2.25

1.56
1.26
2.62
2.27
2.31
1.88
3.96
3.57

0.01
0.01
0.02
0.02
0.02
0.02
0.04
0.03

0.02
0.01
0.06
0.05
0.05
0.03
0.15
0.12

17
18
19
20
21
22
23
24

0.05313

2.175

3.3

3
3
3
3
1
1
1
1

0.50
0.50
0.34
0.34
0.50
0.50
0.34
0.34

15
9
15
9
15
9
15
9

0.50
-0.22
1.52
1.05
1.55
0.86
2.69
2.25

3.74
3.31
5.40
4.93
4.82
4.25
7.33
6.81

0.20
0.18
0.29
0.26
0.26
0.23
0.39
0.36

0.74
0.58
1.55
1.29
1.23
0.96
2.85
2.46

25
26
27
28
29
30
31
32

0.05313

0.825

18.7

3
3
3
3
1
1
1
1

0.50
0.50
0.34
0.34
0.50
0.50
0.34
0.34

15
9
15
9
15
9
15
9

0.50
-0.22
1.52
1.05
1.55
0.86
2.69
2.25

2.66
1.26
6.97
5.27
5.97
3.90
12.82
10.98

0.14
0.07
0.37
0.28
0.32
0.21
0.68
0.58

0.38
0.08
2.58
1.48
1.89
0.81
8.73
6.41

5.65

41.72

Run

Summations
= C/D = pi(C/D) = 5.65

E[C/D]

E[(C/D) ] = pi(C/D)2 = 41.72


2

Var(C/D) = 2C/D = E[(C/D)2] - (E[C/D])2 = 41.72 - (5.65)2 = 9.79


C/D = 3.13, VC/D = C/D/C/D = 0.55,

= 3.09

where the reliability index, , is computed using the equation

[ln C/D]
ln C/D

E [C/D]
ln
1 V2
C/D

ln [1

B-8

VC/D]

ETL 1110-2-321
31 Dec 93
Table B-7
Reliability Analysis of Example Guide Wall by Taylors Series Method
Performance Mode: Sliding Stability Under the Loading Condition of Drawdown, from Figure 2-1
Run

TanROCK

CROCK
ksf

H
ft

Ksoil

deg

XR

Fsliding

1
2
3
4
5
6
7
8
9
10

2.175
0.825
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50

11.0
11.0
18.7
3.3
11.0
11.0
11.0
11.0
11.0
11.0

2.0
2.0
2.0
2.0
3.0
1.0
2.0
2.0
2.0
2.0

0.42
0.42
0.42
0.42
0.42
0.42
0.50
0.34
0.42
0.42

12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
15.0
9.0

1.28
1.28
1.28
1.28
0.69
1.81
0.70
1.87
1.56
0.99

6.54
3.97
6.94
3.57
3.86
6.67
3.65
7.48
5.86
4.65

1.28

1.69

1.41

1.91

0.61

1.50

11.0

2.0

0.42

12.0

1.23

5.26

SD-FD

MEAN VALUES
11

[Fsliding] = 5.26
[Fsliding] = 3.24

Mean
Standard Deviation

tan,c = -0.7
Var[tan,c] = 7.46
[tan,c] = 2.73
V[Fsliding] = 0.52

Coefficient of correlation,
Variance - correlated,
Standard Deviation - correlated
Coefficient of variation,

= 3.15

Reliability Index,
where the correlated variance, VAR[tan,c], is defined as:
Var[tan,c] = ([Fsliding])2 + 2(FDtan)(FDc)(tan,c)
= (3.24)2 + 2(1.28)(1.69)(-0.7)
= 7.46
The correlated standard deviation, [tan,c], is calculated from:

[tan,c]

VAR [tan,c]

7.46

2.73

The coefficient of variation, V[Fsliding], is determined by:

V[Fsliding]

[tan,c]
[Fsliding]

B-9

2.73
5.26

0.52

ETL 1110-2-321
31 Dec 93
Table B-7 (Concluded)
Where the reliability index, , is computed using the equation:

[ln Fsliding]
ln F

E [F
sliding]
ln

1 V[2F

sliding

ln[1

sliding

V[F

sliding

will vary each year. The limit state obviously has


not been reached. However, if the variables which
are consequential to stability are examined, it is
reasonable to assume that the limit state may be
reached in time. Foreseeable events that could
jeopardize the wall are (1) unprecedented overtopping and rapid drawdown, (2) emergency operations and equipment that would place additional
surcharge loads on the backfill, and (3) changes in
operating conditions or tows using the lock that
would increase hawser pulls.

The point estimate method is used to approximate


the integration of the performance function, capacity
divided by demand. The four random variables
identified in paragraph 3 above are earth pressure,
differential hydrostatic pressure, wall friction, and
hawser pull. Mean values of the foundation phi
angle and foundation cohesion as shown in
Table B-2 are used as constants in the reliability
analysis. This simplification was used to reduce the
computational effort with the knowledge that foundation bearing capacity is not critical for this structure. The direct shear tests used to determine rock
strengths for sliding analyses conservatively estimate the cross-bed shear strength applicable to
foundation bearing capacity computations. The
reliability computations are much the same as the
resultant location analysis. The combinations of
variables and results of the reliability analysis are
shown in Table B-8.

b. Sliding stability. The reliability index for


sliding is 3.09 by the point estimate method and
3.15 by the Taylors series method. These are
indications of an expected above average performance level correpsonding to a probability of
unsatisfactory performance of about 0.001 (Ref 3e).
The wall is only slightly less vulnerable to problems
of sliding than to unacceptable resultant location.
Events which affect earth pressure and differential
hydrostatic loads will have significant influence on
the sliding stability of the wall.

6. Conclusions
a. Resultant location. The reliability index for
resultant location is 1.21. Table 1-1 in ETL 11102-532 indicates this would result in an expected
unsatisfactory performance level corresponding to a
probability of unsatisfactory performance of about
0.12. The wall is subjected to this type of loading
on an annual basis, and the severity of the loads

c. Bearing capacity. The reliability index for


bearing capacity is 6.75. The expected performance
level under this performance mode is high.

B-10

ETL 1110-2-321
31 Dec 93
Table B-8
Reliability Analysis of Example Guide Wall By the Point Estimate Method
Performance Mode: Bearing Capacity Under the Loading Condition of Drawdown, from Figure 2-1

Run

Probability
pi

KFILL

deg

Hawser
kips

H
ft

XR
ft

1
2
3
4

0.0625
0.0625
0.0625
0.0625

0.34
0.34
0.34
0.34

15
15
15
15

1.5
1.5
0.5
0.5

3
1
3
1

1.09
2.29
1.94
3.09

5
6
7
8

0.0625
0.0625
0.0625
0.0625

0.34
0.34
0.34
0.34

9
9
9
9

1.5
1.5
0.5
0.5

3
1
3
1

9
10
11
12

0.0625
0.0625
0.0625
0.0625

0.50
0.50
0.50
0.50

15
15
15
15

1.5
1.5
0.5
0.5

13
14
15
16

0.0625
0.0625
0.0625
0.0625

0.50
0.50
0.50
0.50

9
9
9
9

1.5
1.5
0.5
0.5

E[C/D]

FP
ksf

pi(C/D)2

115.0
277.0
223.2
407.9

7.2
17.3
14.0
25.5

827
4796
3114
10399

0.61
1.83
1.49
2.66

35.1
12.3
14.2
8.4

2214
2696
2404
2941

63.2
219.7
169.4
349.2

4.0
13.7
10.6
21.8

250
3017
1794
7621

3
1
3
1

0.10
1.17
0.91
1.94

241.0
20.7
25.3
12.5

1849
2158
1998
2325

7.7
104.0
79.0
185.5

0.5
6.5
4.9
11.6

4
676
390
2151

3
1
3
1

-0.65
0.45
0.21
1.26

51.1
103.9
18.3

1745
2010
1834
2171

0.0
39.3
17.7
118.7

0.0
2.5
1.1
7.4

0
97
20
881

Summations

148.6

36,045

Var(C/D) = 2C/D = E[(C/D)2] - (E[C/D])2 = 36,045 - (148.6)2 = 13,963


= 6.75

where the reliability index, , is computed by the equation

pi (C/D)

2338
2831
2527
3080

E[(C/D)2] = pi(C/D)2 = 36,045

VC/D = C/D/C/D = 0.80,

Capacity
Demand

20.3
10.2
11.3
7.6

= C/D = pi(C/D) = 148.6

C/D = 118.2,

Ult FP
ksf

[ln C/D]
ln C/D

E [C/D]
ln

1 V2
C
/
D

ln[1

B-11

VC/D]

You might also like