Duncan - Comparison of Computer Programs for Analysis of Reinforced Slopes

7/21/2019 Duncan - Comparison of Computer Programs for Analysis of Reinforced Slopes

1/155

1\

0

Table 3: Program Ratings

UTEXAS4

SLOPE W

SLIDE

XSTABL

WINSTABL

Accuracy

5

4.5

4.5 4

3.5

Program

5 5

5 5

4 1 *

Com_Q_utation Time

Time

for Learning

3 5 5 4

3.5

Curve

Time to Enter Data

3 5

5

4.5 4

Complete Analysis

Ease of Reinforced

5

initial only -

Slope Design

1.5

2.5

2.5

no final design

3

capabilities

Ease of

Unreinforced Slope

3.5 5 5 4 3.5

Data Entry

Ease of Soil Nail

No

Provision for

Data Entry

2.5 3.5 3.5

Reinforcement

3.5

Ease of Tieback

No

Provision for

Data Entry

2.5 5 5

Reinforcement

4

Ease of Geogrid

2.5 3.5 3.5

No

Provision for

4.5

Data Entry

Reinforcement

Time Req'd to Make

Output Report 4 5

5

3 2

Ready

Quality of Graphical

4

5 5 3

2

Output

*

In

WINSTABL, Spencer's Method has a computation t1me

of

up to several m1nutes.

1 -

Poor

2 - Fair

3 - Average 4 - Good

RSS

4

4

3.5

3.5

5 horizon

reinforcem

only

3

5 horizon

reinforcem

only

5

horizon

reinforcem

only

5

3

3

5 Ex


2/155

Virginia Polytechnic Institute

and

State

University

The

harles

E Via Jr.

Department

of

ivil

and

Environmental Engineering

CENTER FOR

GEOTECHNICAL PRACTICE AND RESEARCH

COMPARISON OF COMPUTER PROGRAMS

FOR ANALYSIS OF REINFORCED SLOPES

by

Michael Pockoski

and

J

Michael Duncan

Report of a study performed by the Virginia Tech Center

for

Geotechnical Practice and Research

Center for

GeotechnicaJ Practice

and

Research

200 Pattoo Hall.

Blacksburg V 24061

December 2000

Virginia

Tech


3/155

COMPARISON

OF

COMPUTER PROGRAMS

FOR ANALYSIS

OF

REINFORCED SLOPES

Program Review

1 Objective &

Method

1 Program Highlights

- Comments on important program features.

2 - UTEXAS4

3 -SLOPE/W

4 -SLIDE

6 -

XSTABL

7 - WINSTABL

8 -RSS

9 -SNAIL

10 -

GoldNail

11

-Summary

13

Summary

Table

of

Program Features

- Compare the programs side

by

side

15

Table

of Analysis Methods

-Conditions of equilibrium, assumptions, and

comments.

16

Program Ratings

-Discussion

of

program performance in key areas.

16

-

Accuracy

16 - Computation Time

17 - Learning Curve

17 -Data Entry/Analysis Time

18 - Reinforced Slope Design

18

- Unreinforced Slope Data Entry

18 - Soil Nail Data Entry

19 - Tiedback Wall Data Entry

19 - MSE Wall Data Entry

19 -Output Time/Quality

20

Summary

Table

of

Program

Ratings

-Which

program

will

suit your needs?

- Lessons Learned -

21 Analysis Difficulties

- The calculated solution may be incorrect

21 - Causes of Difficulties

22 - Tips fol' Coping with Difficulties

r .. _75'

Sandy Clay

' = 120 pd

c = 3CO psf

' - 30

- - $ - E . I r - . .l lO : _

Scndy

Ciay

OJ JOVJ XJVJ XX;: ) XJVJ

XJVJ

JOOJ }()OJ

JOOJ

XX?J

X V J

XX? J >VOY XX?:> ) j )J XX?J : . > < 7 T X

Sandy

C ay

1

Y

= :20

pcf

c

= 300

psf

'

30

0 . . .75

Firm

Stratum

Elcttc"n al orcc

( - C ~ )

Sandy

Clay

V>JVVJ XX J JVVJ XJVJ XIVJX>VJ X>OJ NVJ

XXi?

;v..;;; X?OJXXJJ

XJVJ

X X J KX/J

N I J J X ~

Firm S:ratum

Sandy

CICJy

Highly Plostie Cloy

)

=

1.30 pcf

c -

:300

psi

4l =

;sQ

r

:o

1 30 pcf

c 0

I>

' 25

8 o ~ t c m

o1

croo T :

o

-

r

............_

C:


9/155


10/155

The Slide package

is

comprised

o

three programs: Slide

is

for

definition

o

the problem, Compute

performs the analysis, and Interpret

displays results. As Figure 6

illustrates, the graphics and screen

icons are intended to make the

program as

user-friendly as possible.

The program

is

easy to learn because

it has many features common

to

other programs. The top icons on

the left

o

Figure 6 illustrate the

typical open, save, and print

commands, in addition to the

familiar zoom icons used in other

CAD programs. The top icons

displayed on the right

o

the screen

are for defining the slope, adding,

deleting, and moving points, and for

drawing soil boundaries, tension

cracks, and water tables. The four

icons on the top left are for

specifying the grid

o

rotation

centers, specific slip surfaces, and

search focus items. These icons

change depending upon the type

o

search to be performed. The lower

icons

in

the center

o

the screen are

for applying distributed loads, line

loads, single anchors, or sets

o

anchors. The four icons on the

lower right are for assigning

properties to the soil layers, anchors,

and tension cracks. As these icons

are pressed, pop-up windows appear

requesting the necessary

information. One

o

the major

highlights

o

the program is the

method

o

assigning soil types by

Figure

:

The graphical interface in Slide makes it

easy to learn and to use.

clicking in the region where the

selected soil type applies. Anchors

are assigned in the same manner.

The process makes defining the

problem very fast and helps to

avoid errors.

Slide can search for a critical

circular, non-circular, or composite

slip surface, using specified points

or windows

to

focus the search on

problem areas

o

the slope. The

user can specify any number

o

individual surfaces

to

be

investigated, define a fixed grid

o

search centers, or allow the

program to define its own grid.

Multiple grids can be analyzed in a

single analysis, which allows the

user to quickly refine the critical

areas

o

an initial search grid, while

keeping previous grids displayed for

reference. With every run, the

program performs analyses using the

Ordinary method, Bishop s

Simplified method, and Janbu s

Simplified method. As Figure 7

illustrates, it can also perform

analyses with Spencer s method, the

Corps

o

Engineer s method, Lowe

and Karafiath s method, and General

Limit Equilibrium with numerous

force functions. (see Table 2.) The

program can perform analyses with

all

o

the methods at the same time.

Compute displays the current lowest

value for each method, shows the

progress

o

analysis though a

percentage bar and cumulative

number

o

surfaces searched, and a

run time in the lower left corner.

Figure

7:

Slide can perform analysis with several

different methods for every run.

Output

is

displayed by Interpret, the

third program

o

the package. The

program displays factor

o

safety

contours in addition

to

the factor

o

safety, and allows the user options

to

display either the global minimum,

the minimum from each grid point,

or every surface searched. Results

with any

o

the aforementioned


11/155

analysis methods can be displayed.

The user also has control over how

the contours are displayed and

labeled, and can add text and

common drawing features like

arrows and circles to the slope. The

program also offers auto-text, which

can automatically display data such

as

the file name, title, soil properties,

analysis methods, load descriptions,

and anchor properties. The user can

display the factor

of

safety, center

coordinates, and radius for selected

surfaces, as illustrated in figure

8

lide Update

During the course

of

this study, the

developers

of

Slide worked out

several bugs

in

the program, and

added features

to

make the

program easier to use. Major

improvements were made

in

the

computation time by allowing the

user

to

select the amount

of

data

saved for each analysis.

Additional soil models were

implemented, and an option to

reject surfaces that become

inverted, or to continue using them

with an assumed tension crack was

also added. The creators are

receptive to comments and

suggestions, and are currently

working

to add shortcuts to

improve the program. Major

advancements are planned for

improved functionality for

esign

of

all types of reinforced slopes.

XST BL

Although XSTABL doesn t use a

windows interface, it

is

interactive,

and it

is

one

of

the easier programs

to

use

of

those included in this

study. (A Windows version,

however,

is

currently being

developed.) The main feature that

makes it user friendly

is

the

information the user receives during

the analysis. As the screen capture

shown in Figure 9 illustrates, the

XST ABL screen displays a lot

of

useful information. Units are

displayed in the top right

of

the

screen, the file name

is

highlighted

Figure

8:

Interpret plots contours

of

factor of safety, and lets the user

quickly see information about other surfaces searched.

in the center

of

the screen, and the

function keys that can be used with

this screen are highlighted on the

bottom with clear descriptions. On

the screen shown in Figure 9,

Fl

brings

up

a help file, F2 shows a

graphical representation

of

the

problem, ESC exits the program,

and F4 changes units from English

to

SI. As the cursor bar

is

moved

around the screen, a message line

in the center

of

the screen explains

the function of the highlighted text.

The message on this screen

explains that the program can search

for the critical circular surface using

the Bishop or Janbu methods.

XSTABL can also perform analyses

on a single circular or non-circular

surface using Spencer s method,

General Limit Equilibrium, Janbu s

Generalized Procedure

of

Slices, and

force equilibrium procedures such

as

Janbu s Simplified method, the

Army Corps

of

Engineers method,

or Lowe and Karafiath s method.

(See Table 2.)

Slope Stdbilitp

Progral'l

- XSTABI

5.2

r e p a r ~

Slope Odtd

> PROFILE SOil

W TER .NIILYSIS

LOADS/l IHITS

.

Cirnl


12/155

Although XSTABL does not have

the capability

to

input

reinforcement, a slope can be

analyzed to determine the magnitude

of external load required to achieve

a specified factor of safety. This is

an

important feature, useful in

preliminary analysis or design. (See

sidebar, Minimum Required Force

for Stability. )

The graphical slope representation

displayed in Figure 10 is particularly

helpful, and can be displayed at any

time. While it isn 't as polished as

the graphics

in

some of the other

programs, it contains useful

information such

as

the method of

analysis, number of soil types, water

surfaces, and boundary loads,

seismic coefficients, and search

extents. A similar screen

is

displayed during the search process,

allowing the user to see where on

the slope the slip surfaces are being

drawn.

Minimum Required Force

for

Stability

XSTABL does not have an option

to

input reinforcement directly,

but it does have an option to

compute the minimum horizontal

force to be applied for a required

factor of safety. In this study,

Slope No. 4 is highly over

reinforced. Examination of the

failure surfaces shows that the

most critical surface extends past

the top two anchors, and only

intersects the bottom anchor at the

very end where the reinforcing

force

is

small. Analysis with

XST ABL illustrates that the same

factor

of

safety can be achieved

with much less force. (The force

predicted by XSTABL was

similar to the force in the bottom

reinforcement at the location of

slip surface intersection.) This

option can be a useful preliminary

design tool, giving the user a

ballpark idea of the magnitude of

the reinforcing force required for

stability.

C:\XSTABL\SCHIIABEL\LAYRNAIL.IPT

I

Sol i

Nailed

Wall In Layered Sol i i rcular search - JANBU I

3

Soi l

units 1 Water

surfaces

125

TERI11NATION

100

59

25

IN

T

AT Oft

I

J

a _ _

___

.

_____ _

___, J.

______________________

________

_

o zs so

75

tee

2s se 175

zoo

Figure 10. XSTABL graphical representation isn't elaborate,

but it is clear and contains useful information.

Output for XSTABL consists of an

output text file, and simple sketch

of the slope, containing the title,

minimum factor of safety, and the

ten most critical surfaces. The

quality

is

similar to that displayed

in Figure

10

The text output

contains all problem definition

information, coordinates and

factors

of

safety for the ten most

critical circles, and warnings about

possible problems with the analysis

performed.

WINSTABL

Purdue University's DOS computer

program, PCSTABL6, has been

recently re-created in a Windows

environment by engineers at GSS.

The new program

is

called

WINSTABL. As the screen

capture in Figure 11 illustrates, the

windows based program

is

intended

to

make a user friendly interface for

the popular PCSTABL program.

One advantage of the new program

is

that the slope

is

displayed

as

it

is

created. During analysis, slip

surfaces can be displayed on the

slope to help control the search

area. The program can be run

using either conventional or SI

units, and has separate data entry

windows for soil nails, tiebacks,

7

and geosynthetics reinforcement.

Anisotropic soil, boundary loads,

and seismic loads can also be

utilized. The icons on the left of the

screen bring up data entry windows,

where information

is

entered

in

spreadsheet format.

The program can use Bishop's

Simplified method to search for the

critical circular slip surface, or

to

analyze a specified circular surface.

It can also use Janbu's Simplified

method and Spencer's method

to

perform a circular search, block

surface search, random path search,

or to analyze a specific non-circular

surface.

Because the program

is new, the

program's creators have been

working

to

remove bugs that have

been identified during the course

of

this report. For example, the

accuracy

of

the program has been

improved so that the program

is

generally reliable. Other advances

have been made in the data entry for

reinforcement, and for the reference

grid printed with the output. The

creators of the program are eager to

improve the functionality of the

program, and are very receptive

to

comments and suggestions.


13/155

Figure

11.

WINSTABL requires data

to

be entered in tables, but

immediately displays the problem geometry.

SS

RSS, like XST ABL, is another

program that

is

not Windows based

but

is

interactive and user friendly.

As

Figure

12

illustrates, the program

uses six pull down menus that are

accessed with Alt commands. The

cursor

is

then used to select data

entry screens. The program was

created for design and analysis

of

mechanically stabilized earth walls.

It

has two modes

of

operation. The

analysis mode can perform slope

stability analysis on reinforced and

unreinforced slopes. The design

mode can be used

to

calculate the

strength or length

of

reinforcement

to

achieve a desired factor

of

safety. As illustrated in Figure

12

RSS can perform analysis on

circular failure surfaces using

Bishop s Modified Method and

Help Ho

Id press

H. Then press Enter.

File

Edit

ndlyze

Ouign

Options

Uiew

See StepBvStep under Help

for

steps

to

use this

progra

..

load s a ~ e print clear

data. Exit progra...

Input slope

and

soi l data.

Calcolote fctor of

safety.

Find reinforce .ent

for

a given required factor of safety.

Control output

de

ta I

Uiew input data.

Figure

12:

RSS

is

interactive, and uses six pull down

menus to navigate among its data entry screens.

8

Janbu s Simplified Method, or on

Block surfaces and noncircular

surfaces using Janbu s Simplified

Method. (See Table 2.)

One

of

the best features

of

the

program

is

the exhaustive search

performed on slopes with

reinforcement. Analyses using

several different types

of

failure

surfaces are performed. The first are

circular surfaces passing through the

toe

of

the wall. The second are hi-

linear failure surfaces passing

through the toe

of

the wall and

extending

to

the back

of

the

reinforcement. The last two are hi-

linear failure surfaces at the bottom

third and top third

of

the slope,

as

illustrated in Figure

13.

These

surfaces are the most common

failure surfaces for reinforced

slopes. More importantly, RSS

computes the factor

of

safety

in

the

upper parts

of

the slope. This makes

the reduction

of

reinforcement

in

the

upper parts

of

the slope simple and

accurate. The same surfaces are

used in the design mode to

determine the most efficient spacing

and strength

of

reinforcement.

Output for RSS consists

of

a text file

and simple graphical output, which

displays the problem geometry,

factor

of

safety, title, and a set

of

~

Circular

- - - - ' ~ B i - L i n e a r

Figure

13:

RSS uses four types

of

failure surfaces

to

analyze the

stability

of

a reinforced slope.


14/155

axes as illustrated in Figure 14. The

text file echoes the input parameters,

and lists center coordinates and

widths, weights, and locations

of

slices. It also contains pore water

pressure, normal stress, and shear

stress for each slice. Any errors

encountered in the analysis are also

listed at the end

of

the output text

file.

RSS Reinforcement

Limitations

Only horizontal reinforcement

can be used, which seriously

limits the applicability

of

the

program for many reinforced

slope applications. It is,

however, well suited for MSE

walls, which typically have

horizontal reinforcing layers.

Reinforcement properties are

specified in data input. This

makes fast design changes

possible, because the program

performs the time-consuming

calculations

of

determining the

force along the length

of

the

reinforcement.

SNAIL

SNAIL is a useful tool for design

of

soil nail walls. Early versions

of

the

program

we

DOS based, but a new

Windows version was released near

the end

of

this study. Although only

reviewed briefly, the Windows

version appears easier to use and

navigate, and accepts input files

from the DOS-version. One

of

the

most appealing aspects of the

program is that it is free. It can be

downloaded from CALTRANS s

Division

of

Materials and

Foundations website. Although the

searches for critical failure

mechanisms that the program

performs aren t as rigorous as some

other programs, it can provide

valuable design information much

faster than conventional slope

stability programs. This is because

reinforcement properties can be

changed very easily. The program

Title ; Mechanically Stabilized

arth Wall

Description

;

Reinforcement Analysis ~ o s t Critical S u r f a c e ~

160

Minimum Reinforced Factor

of

Safety : 1

1 81

140

120

de

100

ea

00

...

00

80 100 120 140 10D

100

200

Figure 14: The RSS graphical output is simple, but contains helpful

information. A similar slope can be displayed during data entry to

ensure correct input.

calculates forces and coordinates

for use in stability computations

automatically, replacing the most

time consuming step in the process

of

designing a soil nailed slope. As

the input screen in Figure 12

illustrates, the program was

intended to make soil nail wall

design quick and easy. At each

input location, units and a

description are provided, thereby

reducing errors and increasing

efficiency. The input required

deals specifically with soil nail

walls, though the program can also

be used without reinforcement.

With effort, input parameters can

be

modified to apply the program to

tiedback wall and MSE wall design

cases (See Appendix D). The

program allows up to seven different

soil types, although soil boundaries

must extend to the left and right

extents

of

the slope. This often

requires that a simplified soil profile

be used

in

analysis. Distributed

loads and earthquake accelerations

can also be applied to the slope.

H 6

f t - ---- : -Uert ical

~ a l l Height.

8 13.6lo O e g r e e \ ~ a l l Batter tro01 Uert ical l ine.

11= 0 Degree S1= 0 ftht Slope Angle and Distance.

12= 0 Degree S2= 0 ft 2nd Slope Angle and Distance.

13= 0

Deg1

1 SJ= 0

f t - - - l rd

Slope Angle

and

Distance.

14= 0 Deg1 1

S4=

0 f t - --4th Slope Angle and Distance.

15= 0 Degeel S5= 0 ft---5th Slope Angle and Distant


15/155

SNAIL uses its own force

equilibrium procedure to compute a

factor

of

safety on a surface defined

by

a two-part or three-part wedge

mechanism. See Table 2.) A typical

three-part wedge mechanism

is

illustrated in the output screen in

Figure

13.

The graphical output

displays the factor

of

safety, title,

date, file name, and soil parameters

in a simple but clear way. SNAIL

does not have an option to print the

output, however, and effort

is

required to copy and paste the screen

into a separate program for printing.

In addition to this form

of

graphical

output, a text file containing input

parameters and information about

the ten most critical failure surfaces

is

produced. For each

of

the failure

surfaces reported, the program

computes the reinforcement stress at

each reinforcement level, and

determines whether pullout, yield

stress, or punching shear

is

the most

critical failure mechanism.

GoldNail

GoldNail is one

of

the most

powerful design programs included

in the study. It was written

specifically to aid in design and

analysis

of

soil nail walls. The

program operates in three modes:

Design mode, Factor

of

Safety

mode, and Nail Service Load mode.

In the design mode, GoldNail

changes nail lengths, strengths, and

MSE Walls - Include The

Bottom Layer?

An important issue in analysis of

MSE walls

is

that the critical slip

surface often passes just above

the lowest layer

of

reinforcement.

It

is

important to define the slope

geometry and search limits so that

this type

of

surface will be

analyzed. GoldNail includes the

bottom layer

of

reinforcement in

the analysis of toe circles. The

strength

of

this reinforcement

layer should be set to zero.

PROJECT TITLE

: Tiedbdcl< Wall in

Layered Soil

Date:

99 ll9 ZOee

I H n i - Factor of Safety = 1.33

75

.e f t Behind Wall Crest

27

.e

f t Be low Wall

Toe

File:

layrt l l

Pp=107 .1

k/f t

LEGEI tD:

GAll

PHI

COH SIG

~ ; i ~ ~ ~ .1

p s ~

-

.....

~ ~ ; ; ; z 1

1

f f i

n

I / 1 117.8 9 1985

r r

- - -1- : -59- - - - .cc- \ - . . . . . . l

............. .......................................................................

\ /

.

,_/

Soli Bound. f>l Water

SCALE

= 19

f t .

Ext. Force-P D

Press: Q= Quit. T= Toe.

S=

Screen. Z= Zoo... R= Report.

Figure 16. SNAIL can use a three-part wedge analysis

to

c ; P . : : t r ~ h

for f::tilnrP.

c ; n r f : : t ~ P . c ;

hP.Iow thP. toP.

spacings

to

achieve a specified

factor of safety. A trial design

is

required for input, but the program

carries out the time-consuming

process

of

changing nail strengths

and lengths, which is the heart of

soil nail wall design. Factor

of

safety mode

is

used for analysis

of

completely defined slopes. Nail

service load mode

is

intended for

research purposes.

It

is

used

to

predict individual nail service loads.

In addition to analysis with soil nail

walls, the program can be run using

different face pressure distributions,

or with no reinforcement. Tiedback

walls can also be analyzed with little

extra effort since the program only

requires tendon strengths and pullout

resistances for input, rather than nail

properties and hole diameters. The

Figure 17. GoldNail uses six pop-up windows for data entry for

organization and ease.

10


16/155

program also features seismic

loading, Service Load Design or

Load and Resistance Factor Design

options, and English or SI units.

GoldNail uses its own slope stability

analysis method for determining the

factor

o

safety. A method

o

slices

is

utilized to determine an initial

estimate

o

the normal stress

distribution on circular failure

surfaces. Iterations proceed to

change the normal stress distribution

until force and moment balance

is

achieved. The program only

analyzes slip surfaces that pass

through the toe

o

the wall, or

through a specified point on the face

o

the wall.

As the screen capture in Figure

17

illustrates, the program was

designed

to

be very efficient. Six

data entry windows define the

problem and search limits. These

windows are quickly accessed with

the toolbar at the top

o

the screen.

Data

is entered into tables that are

clearly defined and well organized.

During data entry, a simple sketch

o

MS Reinforcement in

GoldNail

The development

o

pullout

resistance

is

defined in GoldNail

for each soil layer, rather than for

each layer

o

reinforcement. This

is

convenient for soil nails,

because the soil type defines the

development

o

pullout along the

nail. However, pullout resistance

for MSE wall reinforcement is

determined by the effective stress

acting on the reinforcement layer.

GoldNail therefore cannot

accurately model the development

o

pullout resistance along

reinforcement layers unless each

layer

is

within a separate soil

layer. Development lengths for

deeply covered layers are

typically short (several inches) for

walls 30 feet or higher, and one

average value

o

development

length for several lower layers of

reinforcement should yield

satisfactory results.

Figure

18

GoldNail s graphical output

is

very simple, and displays only

thP l n ~ l l t i o n o

thP

~ r i t i ~ l l l ~ i r d P llnrl thP. tit P

the slope can be displayed to ensure

that information is

entered

correctly. The sketch

is

similar to

that displayed in Figure

18

Output for the program consists

o

a text file and graphical output,

though information contained

within the output varies depending

on the type

o

analysis run. The

graphical output, as illustrated in

Figure 18,

is

extremely simple.

Only the critical circle and title are

displayed. This output may be

printed, but the minimum factor

o

safety must be recorded from the

results screen because it

is

not

displayed on the graphical output.

Information contained in the text

file includes center coordinates and

factors

o

safety for every circle

analyzed.

Summary

Some

o

the programs have

features that allow the user a great

deal

o

freedom during problem

definition, and others are strict

in

their data entry requirements.

Confusion often occurs when

program manuals or data entry

screens are vague. This issue

is

more apparent during the analysis

o

reinforced slopes, because they

often contain features that aren t

common in unreinforced slopes,

such

as

vertical walls or horizontal

line loads. Programs vary

in

their

abilities to handle these unique

features, and program manuals often

do not describe program limitations.

Table I was created to clarify some

o

this confusion, and to provide a

side-by-side summary

o

program

features for a set

o

criteria

important to analysis

o

reinforced

slopes.

For example, vertical walls are

common in reinforced slopes, yet

many programs require all surfaces

to

be inclined slightly. Some

programs allow negative coordinates

and some do not. Some programs

only analyze slip surfaces that rotate

in one direction. Time can be lost

during problem redefinition due

to

confusion about features such

as

these.

In addition

to

saving time, Table 1

can also identify programs that are

well suited for particular needs.

Tension cracks are very common,

but some programs do not have an

option for their input. Graphics

during input, and error checking


17/155

significantly reduce errors and

debugging time. Program manuals

aren't always clear whether a

piezometric or phreatic surface is

used during analysis. As the

sidebar, Piezometric or Phreatic

Surface illustrates, this definition

results in an important difference in

the factor

o

safety.

Equivalent Tension rack

f a program does not have an

option to enter a tension crack,

the soil above the bottom o the

crack should be represented as

a surcharge pressure, which

provides an accurate

representation

o

an empty

crack. There is no simple

measure or representing a

water-filled crack, but

representing the soil above the

base

o

the crack as a soil with

C=O and = provides a

somewhat

approximation.

conservative

Two Ways o Defining Factor o

Safety for Reinforced Slopes

Most

o

the programs reviewed use the same definition

o

factor

o

safety--

F factor by which the soil strengths must be divided

to

bring the

slope to a barely stable state

o

equilibrium

With this definition, the factor

o

safety is not applied to reinforcing

forces. The reinforcing forces input are allowable forces that reflect

considerations such as tensile strength, creep behavior, damage during

installation, stiffness, corrosion, etc.

Many engineers prefer this definition

o

factor

o

safety because

different considerations are involved

in

defining acceptable values

o

factor o safety for soil strength and allowable reinforcement forces.

This definition is used in UTEXAS4, SLOPE/W, Slide, XSTABL,

WINSTABL, and GoldNail.

The computer program RSS uses a different definition o factor o

safety--

F factor by which oth the soil strengths and reinforcement

forces must be divided to bring the slope to a barely stable

state

o

equilibrium

For the same soil strengths and the same input reinforcement forces, this

second definition results in a lower value o F

SNAIL provides a user option to select either o these factors o safety.

2


18/155

Table : Summary of Program Features

UTEXAS

SLOPEIW

SLIDE XSTABL WINSTABL ASS

Vertical Walls?

yes no

yes no no no

Tension Crack Option? yes

yes

yes yes

no no

Search Below Toe? yes yes

yes

yes

yes yes

Graphics During Input? no

yes yes yes yes yes

Seismic Option? yes yes

yes yes yes yes

Error Checking?

yes yes

yes yes

no yes

On Screen Help? yes

yes yes yes no yes

of Soil Types?

Infinite Infinite

500 20

Infinite

128

Slope Face Direction?

Right

r

Left Right or Left

Right r Left Right r Left Left Only

Left Onl

Distributed Loads?

Tangential and

Vertical or Normal

Horizontal, Horizontal, or Horizontal,

r

Horizontal

Normal

Vertical,

r

Normal Vertical Vertical

Vertica

Horizontal and

Normal

r

Vertical

Horizontal and Use Distributed Use Distributed

Use Distrib

Line Loads?

Components,

r

Magnitude and

Vertical Load Option, Load Option,

Load Opti

Direction Conponents, r

Magnitude and

Magnitude and Magnitude

Magnitude and

Direction

Magnitude and

Direction

Direction Directio

Direction

Circular Search? yes yes yes yes yes yes

Non-circular Search?

yes

yes yes yes yes yes

Composite circular-

no

yes yes

no

no no

noncircular) Search


19/155

Table 1: Summary of Program Features Continued

UTEXAS

SLOPEIW

SLIDE XSTABL WINSTABL RSS

Piezometric

r

Phreatic

Piezometric

Piezometric

Either Either Phreatic Piezomet

Surface?

Coordinate System

First

First

First First First First

Quadrant

Quadrant

Quadrant Quadrant Quadrant Quadran

Negative Coordinates

yes

yes yes

no yes

no

Allowed?

More than one

Piezometric or Phreatic

yes

yes yes yes

yes

yes

Surface?

Axes on output?

Numbers without Numbers with

Numbers with Numbers with Numbers without Numbers w

units units

units units units units

Plot F Contours? yes in TexGraf4

yes

ves

no

no no

*Mohr-Coulomb C - Phi



*Undrained

*Su=Linear increase

*No Strength

*Undrained

below profile line

*Very Strong (bedrock)

*No Strength

*Su=Linear increase

*Si-linear Envelope

*Infinite Strength

below datum

*Su=Function of Depth

Anisotropic Strength


*Constant clp

*Su=Function of

*User-defined Shear-

*Su=Function of s'v *Mohr-Coulomb C - Phi

Ways to Model Strength

*Anisotropic Strength

Overburden

Normal Stress Function

*Si-Linear Envelope (Isotropic and *Mohr-Coulomb

*User-definied

*Nonlinear, Curved

*Su=Function of Datum

Anisotropic Function

(Each may be Isotropic Anisotropic)

Envelope

Reference

*Su=Function of Depth

r Anisotropic)

su grid interpolation

Anisotropic Strength

*Su=Function of Datum

*Two-Stage Linear

*User-defined Normal

Reference

*Two-Stage Nonlinear

Stress Function

*Hoek-Srown

*Very Strong

(Isotropic and

*Gen. Hoek-Srown

Anisotropic)

Constant

Piezometric Line

Phreatic Surface

Constant

u

u

Coefficients Piezometric Surfaces

Phreatic Surface

Phreatic Piezometric

Ways to Input Pore

Piezometric Line

u

Contours

r u Coefficients

Piezometric Surface

Average

Interpolation from grid

Pore Pressure Grid

Piezometric Su

Water Pressure

Interpolation of u from

Heads

Grid of Total Head

u Coefficients

u

Coefficients

Finite Element Grid of

Grid of Pressure Head

Pressure Head

grid

Pressures

Grid of Pore Pressure

Constant

Negative Allowed


20/155

Table 2 Descriptions of Methods of Analysis

h

0

~ (10

0 0 ~ 0 ~ 0 ~

A tG . ~

f.:' )

~

0

~ li

's ~ o ~

Method

~

ssumptions Com

.:'

Swedish Circle

Yes No No

No Circular Slip Surface Only

Ordinary Method of Slices

Yes

No

No

No

Circular Slip Surface Conse

I (Fellenius 1927)

Side Forces Parallel to Base

Very inaccurate for hiQ

Bishop's Modified Method

Yes

No

No

Yes

Circular Slip Surfaces

Very inaccurate for hig

(Bishop 1955)

Side Forces Horizontal

Morgenstern and Price's

Slip surface of any shape

Much engineering tim

Method (Morganstern and

Yes Yes Yes Yes

Price 1965)

Pattern of Side Force Orientations

force ass

01

Spencer's Method (Spencer

Yes

Yes

Yes

Yes


Simples

1967)

Side Forces Parallel

Corps of Engineers

No No Yes Yes


High factoModified Swedish

(1970)

Side Forces Parallel to Slope


Lowe Karafiath (1960)

No No Yes Yes Side Force Orientations Average of

Best side forc

Slope and Slip Surface

Janbu Simplified (Janbu

No No Yes Yes


Low Fact

1954)

Side Forces Horizontal

GLE - General Limit

Yes Yes Yes Yes


Much engineering tim

Equlibrium

Pattern of Side Force Orientations force ass

GoldNail Method* (Golder)

Yes * Yes

Yes


Toe cir

Normal Stress Distribution

SNAIL Method


(CALTRANS) No No Yes Yes

Two or three wedges, with side

Limited shapes

force angle = 4>


21/155

Program Ratings

In an effort to quantify and compare some important

aspects

o

the different programs, a rating system

was developed. The programs were evaluated on a

scale

o

one to five, l=poor, 2=fair, 3=average,

4=good, and 5=excellent. The aspects

o

the

programs that were rated are discussed below. Note

that with exception

o

learning curve , the ratings

were assigned to each program from the standpoint

o a user with competent skills in using the

program.

Accuracy

A program that does not provide reasonable and

correct solutions is unreliable, and

o

little use. In

addition to being able to compute the correct value

o F for a slip surface, accuracy also requires that a

program be able to locate the critical slip surface

effectively. Programs with limited search methods

are limited in how accurate their search can get.

UTEXAS4 receives a very high rating because o

the way that the search

is

performed. During the

search for the critical circle, the spacing o the grid

used for circle centers is reduced as the grid moves,

searching for a circle with a lower factor o safety.

SLOPE/W and SLIDE received high ratings

because they both allow the user good control o the

search area and radius by utilizing options to focus

the search within the slope. SNAIL received a low

score because its search routine is minimal, and it

does not allow the user to refine the search

effectively. While this project was underway, the

developers

o

WINSTABL made changes that

significantly improved the accuracy

o

their

program. However, the program produced slightly

higher factors

o

safety for some o slopes analyzed

during this study.

Computation Time

Valuable engineering time can be lost i a program

takes too long to run. In order for a program to

receive a high score in this category, it must be able

to analyze a large number

o

surfaces in a

reasonable amount o time. During the period when

data files are being debugged, or when parametric

studies are being performed, a great deal

o

engineering time can be consumed waiting for

results. Most

o

these programs take only a few

seconds to perform an analysis, while some took

minutes. SNAIL analyzes only 560 surfaces in a

search, and was one

o

the slowest programs. Other

programs analyze over 10,000 surfaces in the same

time. WINSTABL runs with adequate speed for the

Simplified Bishop and Janbu Methods, but analysis

using the Spencer Method can take several minutes.

Piezometric or Phreatic Surface

Some computer programs use piezometric surfaces

to

characterize pore pressures, others use phreatic surfaces, and

some can use either. As shown in the sketch below, the

relationship between pore pressures and these two types of

surfaces are not the same. Mistakenly defining a piezometric

surface as

a phreatic surface will result in pore pressures that

are too low. Mistakenly defining a phreatic surface as a

piezometric surface will result in pore pressures that are too

high. The larger the slope o the surface (the value o 8 , the

greater the difference. Analyses performed on example

slopes included in this study indicate that defining a phreatic

surface as a piezometric surface reduces the factor o safety

by 3% to 6% for cases where the average inclination

o

the

water table

is

4: 1 or flatter. The difference will be greater if

the water table is inclined more steeply than 4: l.

In a strict sense, piezometric surfaces only correspond to a

single slip surface (except for hydrostatic conditions).

However, in practice there is little inaccuracy involved

in

using the same piezometric surface for all slip surfaces

during a search.

6

Typical Slice

Piezometric

Surface

Pore Water

Pressure

ead

hw)

Piezometric Surface Pore

Pressure Calculation

Equipotential

Line

Typical

Slice

Phreatic

Surface

Pore Water

Pressure

Head

hwcos

2g

Phreatic Surface Pore

Pressure Calculation


22/155

Learning Curve

Some of the programs included in the study

were designed so that they are user

friendly, and thus have a very short

learning curve. For example, SLOPE/W

and SLIDE are Windows-based programs

that provide a lot

of

feedback to the user

because all of the programs functions are

displayed as icons on the screen. Being a

Windows-based program doesn't guarantee

a short learning curve. WINSTABL

is

a

windows program that has a longer

learning curve because there is

no

manual

or on-screen help, and the guidance

provided is often ambiguous and

confusing. XST ABL,

is

not windows

based, but still provides the user with a

large amount of explanation and direction.

It displays all the different available

options for data entry and analysis on the

screen, which makes it easy

to

learn. RSS

is

interactive, however, not to the extent of

XSTABL. UTEXAS4 has a slower

learning curve, because the user doesn't get

much initial feedback from the program,

and because the program can perform so

many functions that remain hidden

within the manual until the user discovers

them. It should be noted that the score

given

to

SNAIL

is

for the DOS version

of

the program, which was utilized for this

study. A Windows version released at the

end of this study appears to be easier to

learn and use than the DOS version.

Time to Enter Data and Complete

an Analysis

This criterion is intended to capture the

whole process of entering the data to define

a slope stability problem, defining search

limits, and refining the search to an

accurate factor of safety. SLOPE/W and

Slide received the highest scores because

their graphical methods of defining the

problem are very efficient. Also, the

search areas are easily refined using similar

graphical methods. WINSTABL scored

lower because the method

of

data entry

is

not efficient. Although it uses a scheme

of

connected line segments similar

to

XSTABL, the tables used for data entry are

not as efficient. In XSTABL, the tables for

data entry are partly automated. The end

of one segment automatically appears

as

the start of the next segment. This

feature cuts the data entry time

considerably. Methods

to refine

Language Barrier

The development of force along the length of a soil nail

is

usually

calculated using nail properties such

as

bond stress, drill hole diameter,

punching shear capacity, nail diameter, and yield stress. Tiebacks are

usually described

in

terms

of

maximum force and bonded length. The

development of force for MSE wall reinforcement

is

related to the

overburden pressure, and is usually described by safe design strength and

an interface friction angle. Because these different terms are commonly

used to describe the capacities of reinforcement, a language barrier

is

encountered in programs designed for only one type.

For example, SNAIL

is

intended for analysis and design of soil nail walls,

and requires data entry

in

terms

of

nail properties. In order to enter

tiebacks and MSE wall reinforcement, equivalent nail properties must be

calculated from the bonded length and maximum force. On the other

hand, SLOPEIW

is

designed for tieback data entry, and the user must

calculate the reinforcement force diagram, and enter the bonded length

and maximum force. In cases where soil nails are defined as tiebacks, a

check must be performed to ensure adequate head capacity, if the failure

surface falls close to the head of the lower nails.

Head

Capacity

Pullout Tensile Capacity

Resistance

---------- ---:____...___

Typical Nail Force Diagram

Pull out

Resistance

....__ ...____ __.. ______. . _____. _______.. ____.

~ r - @ ' : f # l * M ? r - ? N b h N ' ' * r N 1 b 3 ' V 3 : ~ . ? @ r - M 1 d * ' N ? t Y - % t - J x q t a p ; z r , w t : ; , : ; t q % ; - & x r : - 8 x % ' - d < M N ? f 8 ? . - & a c t c 9

Head

Capacity

...----- ---,... --,... -- ...

---,...

-- ...

Tensile Capacity

Typical

Tiebock

Force

Diagram

Pull out

Resistance

_____..

____. ______.. _ ______.. _

_____..

W2-&,P,t:-,.:-/,;r,..r:.P,r.M?*IW.&-2kht#?b@;;.,w-?,'W-hx&.f?tt1h0hr:t)'l.di&-&i1-&8Zl?r.f-?.z-( %Z/,[email protected])3

Tensile Capacity

Typical

Force

Diagram

Pullaut

Resistance

~ ~ g g i t y For Shallow Geogrid Reinforcement

__ _ _...... ________. _ _______. _ -

N < h ? i t ~ - w N M / 4 t * ' Y - : z w h ? . w t < - 1 - ? : - t r - 1 - ' W f a ~ ~ . w r : - e w : ~ . - a x a - c r : . . , . . . w r : - n w r . - & n 7 . - W i & : ~ - a - z . . : t - ? % 1 P . ~

Head

Capacity

--......________....---------,... - .....________. ..

Tensile Capacity

Typical

Force

Diagram

For Deep Geogr 1d Reinforcement

Pullout

f ~ s i s t o n e

____..

...

w'%r:Miz;::-2i{i?, .-. / f } @ Y , ; , % ' - & & Z r - r . . @ % t 7 . - ; , 0 ; 9 ; f X ; { ( . t ' l . ' ? Z ; t . 0 t k x i i , ' / % i Z ~

~ - . . ~

17

Available in:

UTEXAS4

WINSTABL

SNAIL

Available in:

UTEXAS4

SLOPE W

SLIDE

WINSTABL

RSS

SNAIL

GoldNail

Available in:

UTEXAS4

SLOPE W

SLIDE

WINSTABL

RSS

SNAIL

GoldNail

Available in:

UTEXAS4

SLOPE W

SLIDE

WINSTABL

RSS

SNAIL

GoldNail


23/155

the search in both programs are

very similar. UTEXAS4 received

an average score not only because

is takes time to look through the

manual

in

order to create the data

file, but also because it can take

time to debug the data file.

Several attempts are often

required

in

order to get the format

o

the input file correct. Once the

program is running, however, it

almost always finds the critical

factor

o

safety in a very short

time. GoldNail received a low

score due to the coordinate system

in which the program operates.

Typical engineering designs are

drawn in the first quadrant.

Coordinates have

to

be transferred

into the fourth quadrant before the

program can be utilized, which is

a time consuming process.

Ease o Reinforced

Slope esign

There are many different aspects

o

reinforcement that change often

during reinforced slope design.

The reinforcement length,

spacing, pattern, size, strength,

and bond strength are all related,

and a change in one can often

affect the others. This category is

intended to characterize the ease

with which these parameters can

be changed during the design

process. Some

o

the programs

investigated are intended as design

tools, and reinforcement changes

can be made quickly and easily.

Others were not, and

reinforcement design is labor

intensive. GoldNail contains

within the program, a mode o

operation specifically intended for

design. It will recalculate required

nail lengths and strengths

necessary for a required factor o

safety. It received the highest

score in this category because

o

this efficiency, and because data

for all types

o

reinforcement can

be easily entered. SNAIL also

received a high score in this

category because it was created

specifically for the purpose o

analyzing soil nail walls, and it

calculates nail forces based on

simple property input values.

However, SNAIL doesn't have a

design mode like GoldNail, and

input data for tendons must be

converted into soil nail

parameters . (Refer to Appendix

B for example calculation.)

Although WINSTABL gets past

the language barrier using three

separate input windows for

anchors, geosynthetics, and soil

nails, it still requires extensive

calculations and data input to

make changes in reinforcement.

SLOPE/W and SLIDE scored

poorly for design capabilities,

because they require extensive

calculations prior to data input,

and because the programs are not

designed to handle all types o

reinforcement. The creators

o

Slide are currently working on

improvements for better design

functionality. UTEXAS4 scored

very low because longitudinal and

transverse reinforcement forces

must be entered at x-y coordinates

along the length o the nail. The

user has to calculate both the

forces along the nail, and the

coordinates where they act. This

process requires extensive

calculations, and a large amount

or data entry for each nail. RSS

only allows for horizontal

reinforcement, and was only rated

with MSE wall reinforcement

under consideration. The program

does contain a design mode,

which reduces engineering time

considerably. XSTABL contains

no

provision for reinforcement, so

it could not be rated

in

this

category. t does, however, have

the ability to calculate the

magnitude

o

a single external

force required to achieve a

specified factor o safety. This

ability is useful in initial analysis

or design to get a ballpark

estimate

o

the magnitude

o

the

forces required.

8

Ease o Unreinforced

ata Entry

Some

o

these programs are

designed to be very user-friendly,

and the slope geometry, pore

pressures, loads, and soil properties

can be input with very little effort.

SLOPE/W and SLIDE are

Windows-based programs, where

nearly all data input can be

performed with the mouse.

Because the user can draw the

slope, errors are minimal and

problem definition is easy.

XSTABL, although not a graphical

interface program, is interactive

and automated. These two qualities

make problem definition clear and

simple. WINSTABL, RSS, and

GoldNail define the slope using a

method

o

line segments similar to

XSTABL. Since they aren't

automated, more problems tend to

arise during data entry, and

problem definition becomes more

complicated and labor intensive.

Problem definition with UTEXAS4

Is difficult because the user

receives no feedback until analysis

has begun. Although the methods

used

in

the program to define

geometry, soil boundaries, and pore

pressures are relatively simple,

errors can occur when creating the

data file that can cost valuable

engineering time. SNAIL received

the lowest score

in

this category for

several reasons. The program

cannot handle complex geometry,

and the user must redefine the

problem to fit the restrictions

o

the

program. Also, the sign convention

used to define the wall is different

from that used to define the slopes

above and below the wall, which

can cause confusion and input

errors.

Ease o Soil Nail

ata Entry

This category is intended to

highlight the programs that allow

soil nails to be entered with little

effort. XSTABL was not rated

because it has

no

provision for


24/155

reinforcement. GoldNail and

SNAIL are intended for analysis

and design of soil nail walls, and

therefore received the highest

scores. SNAIL received the

highest score because it allows the

user the most flexibility for nail

geometry and pattern, and because

units are displayed at entry

locations. When entering

reinforcement, units are very

helpful, and significantly reduce

errors. GoldNail requires less data

entry but requires that all nails

have the same horizontal spacing

and declination, and units are not

displayed. Although WINSTABL

has a separate window for

entering soil nails, data entry

requires a great deal

of

effort.

Fourteen numbers must be entered

to input a single nail. SLOPE/W,

SLIDE, and UTEXAS4 require

the magnitude of the force along

the nail for data entry. The user

must calculate the nail force

diagram See Appendix B for

example calculation) before

entering data for a nail.

SLOPE/W and SLIDE scored

slightly higher than UTEXAS4

because the nail can be entered

using the mouse, and force

magnitudes can be keyed into

pop-up windows rather than

entered into a separate data file.

UTEXAS4 requires the user to

calculate the coordinates at which

the nail forces act, which

consumes valuable engineering

time.

Ease of Tieback

Data Entry

The reinforcement abilities of

SLOPE/W and SLIDE were

designed with tiebacks in mind, so

they received the highest scores in

this category. The tieback can be

drawn with the mouse, and the

bonded length and maximum

magnitude are keyed into pop-up

windows. WINST ABL has a

separate window for tieback data

entry, which is simple, and allows

the user any geometry required.

However, units are not provided

and data entry can be confusing.

Reinforcement entry in GoldNail

is relatively simple, because it was

designed with all types of

reinforcement in mind. However,

all tiebacks must have the same

declination and horizontal

spacing, and units are not

provided. Because SNAIL was

intended for analysis of soil nail

walls, reinforcement

is

entered in

terms of nail properties. To enter

tiebacks, the user must calculate

equivalent soil nail properties for

input. Refer to Appendix B for

example calculation.) UTEXAS4

received the lowest score for the

same reasons as previously

discussed. Although it can handle

any type of reinforcement, it

requires time-consuming

calculation of forces and

coordinates. XSTABL was not

included because it cannot

evaluate reinforced slopes.

aseofMS

Reinforcement Data Entry

RSS is designed for analysis and

design of mechanically stabilized

earth walls and slopes, and

therefore received the highest

score. WINSTABL also received

a high score, because common

layers of reinforcement may be

entered as a single group to reduce

input time. Data input is similar

in GoldNail; however, values for

each layer must be entered

separately. SLOPE/W, SLIDE,

and UTEXAS4 all require the

force along the length of the

reinforcement layer. The user

must compute this separately for

each layer prior to data entry.

f

these three programs, SLOPE/W

and SLIDE allow the user

to

enter

the reinforcement graphically,

which helps to speed up entry and

reduce errors. UTEXAS4

received the lowest score because

it requires significant effort to

calculate both forces and

19

coordinates along reinforcement,

and also to enter the data.

Time Required to Make

Graphical Output Report-

Ready and Quality

of

Output

n order for output to be report

ready, it

is

important that items

such as title, project number, soil

parameters, axes, analysis method,

and other user comments can be

included on graphics. Some of the

programs produce high quality

output that is suitable for reports.

Others produce output

of

very poor

quality. SLOPE/W and SLIDE

produce very high quality output.

The programs allow the user to add

text and other drawing features

anywhere on the screen. Both

programs also include an auto-text

function, where selected parameters

can be easily displayed, and are

updated when the problem

definition is changed. UTEXAS4

output is well-designed, and

contains much of the desired

information. Although

TEXGRAF4 does not allow the

user

to

add text or drawing figures,

the graphical output can easily be

exported into a CAD program for

editing. Although XST ABL, RSS,

and SNAIL also provide some

important information, the quality

is much lower than other programs,

and the user has

no

option to add

text. WINST ABL output is very

crude, contains little useful

information, and does not allow the

user to add text. GoldNail received

the lowest score because the output

is

not suitable for use in a report.

The graphics are extremely simple,

and no information is displayed,

even the minimum factor of safety.


25/155

1\

0

Table 3: Program Ratings

UTEXAS4

SLOPE W

SLIDE

XSTABL

WINSTABL

Accuracy

5

4.5

4.5 4

3.5

Program

5 5

5 5

4 1 *

Com_Q_utation Time

Time

for Learning

3 5 5 4

3.5

Curve

Time to Enter Data

3 5

5

4.5 4

Complete Analysis

Ease of Reinforced

5

initial only -

Slope Design

1.5

2.5

2.5

no final design

3

capabilities

Ease of

Unreinforced Slope

3.5 5 5 4 3.5

Data Entry

Ease of Soil Nail

No

Provision for

Data Entry

2.5 3.5 3.5

Reinforcement

3.5

Ease of Tieback

No

Provision for

Data Entry

2.5 5 5

Reinforcement

4

Ease of Geogrid

2.5 3.5 3.5

No

Provision for

4.5

Data Entry

Reinforcement

Time Req'd to Make

Output Report 4 5

5

3 2

Ready

Quality of Graphical

4

5 5 3

2

Output

*

In

WINSTABL, Spencer's Method has a computation t1me

of

up to several m1nutes.

1 -

Poor

2 - Fair

3 - Average 4 - Good

RSS

4

4

3.5

3.5

5 horizon

reinforcem

only

3

5 horizon

reinforcem

only

5

horizon

reinforcem

only

5

3

3

5 Ex


26/155

nalysis Difficulties

This investigation has shown

clearly that analysis o reinforced

slopes is much more difficult than

analysis

o

slopes without

reinforcement. In some cases, it is

difficult to determine the factor

o

safety for a reinforced slope even

after the most exhaustive analysis.

Example Slope No. 6 in this report

is such a case. This section

attempts to explain some

o

the

reasons why evaluating the stability

of reinforced slopes is difficult.

Also included

is

a set o tips that

can make coping with these

problems easier.

Causes o Difficulties

Small changes in the location or

shape

o

the slip surface can result

in large changes in the factor o

safety, because the stabilizing

forces change depending on where

the slip surface cuts across the

reinforcement. In Figure 19 the

volumes

o

soil within the two

surfaces are similar. However, the

factors

o

safety for the surfaces are

significantly different due to the

location where the failure surfaces

cross the reinforcement. Although

both surfaces cut the reinforcement

in the bonded zone, the circle with

the higher factor of safety crosses

the reinforcement at locations

where the forces in the soil nails

are slightly higher. This small

difference makes a large difference

in the factor o safety.

In some cases, analyzing more

closely-spaced slip surfaces will

not solve the problem, because

solutions may not converge for

some slip surfaces, leaving

important holes in the search

pattern. Figure 20 shows a set o

contours drawn for very closely

spaced circle centers. As the error

message in the pop-up window

indicates, non-convergence is the

reason for the obvious holes in the

factor

o

safety contours. The

critical circle could exist within

.Figure 19: A small difference in the location

o

the circle

center results in a large difference in the factor o safety.

these holes, but remain hidden by

non-convergence.

One case has been found where there

were two solutions that satisfied all

conditions

o

equilibrium, with

factors

o

safety that differed by 20%

(Slope No. 6). It was not possible

to

determine which factor o safety was

more reasonable.

Seemingly unimportant factors

related to the design

o

the

software, such as details

o

the

search method, limits on the

acceptable ranges

o

side force

inclinations, method

o

including reinforcement forces,

Figure 20: Non-convergence can be a big

problem, leaving holes

in

the search grid.

21


27/155

and convergence criteria can result

in significant differences in the

minimum factor

of

safety that

is

found. Convergence criteria are

especially important. The solution

converges when the moment and

force equilibrium factors

of

safety

are equal. On a plot

of

factor

of

safety versus side force inclination,

as

shown in Figure 21, the curves

of

factor

of

safety for moment and

force equilibrium values should

intersect. In slope stability

programs, convergence

is

achieved

when the values are within a

specified tolerance.

f

the tolerance

isn't strict enough, convergence

may be incorrectly reported,

because the values

of

the moment

and force equilibrium factors

of

safety are close, but the curves

have not intercepted,

as

shown in

Figure 21. For example,

SLOPE/W has a loose default

tolerance (0.01) for convergence

between the force and moment

factors

of

safety. This

is

acceptable

on most unreinforced slopes with

normal side force inclinations, but

may not be strict enough for

reinforced slopes where analysis

is

complicated due to large

reinforcement forces and side force

inclinations that are not within the

normal range for unreinforced

slopes. Figure

21

is

an example

where

in

incorrect factor

of

safety

would be reported

if

the tolerance

was equal to 0.01.

F J c / o I O / ~ , , I o l y v s l . t< IIUI > th ly emo

FactorofSafefot''f'S Lambda

I

_)

-

Figure 21: A strict tolerance

is

necessary

to

avoid incorrect solutions.

\ \ \ 1 \ ~ \ 1 \ t \ \ \ \ \ \ l

\\o

u\O)o\lo \p 5 13

' '=

Figure 22: There are often many local minima with reinforced

slopes.

f

the search routine

is

not sufficiently thorough,

the results can be misleading.

With reinforced slopes there are often

several local minima in factor of

safety contours, which can mislead

search routines and result in incorrect

values

of

minimum factor

of

safety.

Figure 22 illustrates such a case.

f

the search routine does not search a

wide enough area, a local minimum

may be mistaken for the true factor

of

safety and the search may be

terminated too quickly.

Tips

For

Coping With

Difficulties

Some simple things can help

to

understand where problems can arise,

and may help

to

avoid them.

One such step

is to

compute the

factor

of

safety

of

the slope with

no

reinforcement. This provides and

idea

of

what has

to

be achieved by

reinforcing the slope. Slopes that are

marginally stable without

reinforcement are generally easier

to

analyze, because reinforcement forces

are often smaller, and the stability of

the slope doesn't depend entirely

upon the reinforcement. Slopes that

depend entirely upon reinforcement

forces for stability are generally more

22

difficult to analyze, due

to

increased numerical problems and

non-convergence.

A difficulty index can be

defined

as

DJ FReinj >rced

FU11rei1 fln-ced

The degree

of

difficulty

to

be

expected in the analysis

is

indicated

roughly by the difficulty index.

Value

of

Degree

of

DJ

Difficulty

1.0

to

1.5 Minimal

1.5 to

6

Moderate

6 or

Maybe

larger

impossible

The value

of

DJ gives some

indication

of

difficulty, but

is

not a

perfect predictor,

as

shown by the

fact that Slope No.

6

with a

DJ '

10

was much more difficult

to

analyze than Slopes 9,

I

0, and 1

which had approximately the same

value

of DJ

These three slopes,

however, do have a high degree

of

non-convergence. With a lot

of

non-convergence, it

is

difficult

to


28/155

have confidence in the reported

factor

o

safety. (See Appendix A,

Difficulty Index. )

It is

often helpful

to

estimate a

reasonable amount

o

force to make

the slope stable. One simple means

to so this is to estimate the at-rest

earth pressure on a plane from the

crest

o

the slope down to the

elevation o the toe o the slope.

The area under this at-rest pressure

diagram represents an upper limit

on the total amount o

reinforcement force required to

make the slope stable. The flatter

the slope, the smaller is the fraction

o

the at-rest force required.

During analysis, use different

methods (Spencer's Method,

Bishop's Modified Method, etc.)

to

compute the factor o safety. When

one method has a high degree

o

numerical problems and non

convergence, another method

having more simple side force

assumptions may provide a more

reliable estimate o the factor o

safety.

f possible, use different computer

programs to compute the factor o

safety. Because computer analysis

o

reinforced slopes is a relatively

new topic, there

is

no accepted

convention

o

applying the

reinforcement forces

to

slice

boundaries and slip surfaces, and

different computer programs

handle reinforcement forces

differently. Different methods o

applying the reinforcement forces

result in different side force

inclinations, and sometimes better

convergence.

During the actual search, be sure

to

search thorou hly for the most

critical slip surface (minimum

factor

o

safety).

The first step should be to search

a wide area for the most critical

circle center. Use a grid spacing

small enough

to

give a complete

picture

o

the search area, but

large enough that many analyses

can be performed quickly.

Refine the grid size in a second

phase

o

the search in the area

around the lowest factor

o

safety

o the initial search. However, do

not immediately jump to a small

grid spacing directly around the

initial lowest grid point. Local

minima may be present, and can

mislead the search. Instead, reduce

the size o the grid in several steps,

re-centering it on the minimum i

necessary.

f

the convergence criterion in the

program can be controlled by the

user, use a small final tolerance

(0.000 1 . A coarser tolerance may

result in a false indication that

convergence has been reached, and

erroneous search results.

Examine the program output

carefully for warnings o problems

with convergence or search results,

Program Comments

on Contours

UTEXAS4 refines the search

area automatically when using

the floating grid search option,

but contours can only be

viewed i a fixed grid search

is

performed. Locate the lowest

factor

o

safety with the

floating grid search, and center

a fixed grid on the center

o

that

surface. Run the analysis with

the fixed grid to view the

contours to see i non

convergence or local minima

may have been a problem.

SLOPE/W and SLIDE both

illustrate contours easily, but

the search grid must be moved

manually. These programs do

not have convergence

tolerances that are as strict as

UTEXAS4, and do not allow

the user to select the method of

application

o

the reinforcement

force

as

UTEXAS4 does.

XSTABL, WINSTABL, RSS,

SNAIL, and GoldNail do not

have search grids, nor do they

olot contours.

23

and consider carefully how these

may have influenced the search.

Realize that some reinforced

slopes may be virtually

impossible to analyze by limit

equilibrium methods. Example

Slope No. 6 in this report is such

a case.

Example Slope nalyses

A set o example slopes, as defined

by Schnabel Foundation Company,

was used to provide a basis for

comparison

o

the programs. The

example slopes include three

unreinforced slopes, three tiedback

walls, three soil nailed walls, and

two MSE walls

in

varying soil

conditions and geometries.

A minimum factor

o

safety was

calculated for each slope using

each program. In most cases

(UTEXAS4, XST ABL,

WINSTABL, RSS, and GoldNail)

only circular slip surfaces were

analyzed. SLOPE/W and Slide

analysis included with composite

(circular+ planar) slip surfaces that

are generated automatically by

those programs where slip circles

intersect rock. All

o

the SNAIL

analyses were performed using the

two-part or three-part wedge

mechanisms analyzed by that

program.


29/155

ppendix

Tables

Results for Example Problems

See Appendix B for figures showing

example slopes and critical slip surfaces

4


30/155

1\)

01

J


31/155

1\

n

Table 5: Factors of Safety by Spencer's Method

Slope Slope No 1

Slope No 2

Unreinforced

Unreinforced

Title Homogenous

Homogenous

Slope

Slope with a

Crack

UTEXAS4

1.30

1.29

Spencer

SLOPEIW

1.30

1.29

Spencer

SLIDE

1.31

1.30

Spencer

XSTABL

1

l

Spence

+

+

WINSTABL

1.34

1.32

Spence

RSS

2

l

1.29

1.28

SNAIL

3

l

1.22

1.18

GOLDNAIL

3

l

1.32

1.30

-Solution did not converge.

+ XSTABL has no provision for reinforcement.

J:

RSS only allows for horizontal reinforcement.

Slope No 3

Unreinforced

Layered Slope

1.42

1.40

1.40

+

1.45

1.41

1.39

1.40

Slope No 4

Tiedback Wall

in Layered

Soils

1.14

1.14

1.15

+

1.20

J:

1.03

1.19

Notes: (1) XSTABL does not search using

th

Spencer Method.

Slope No S

Slope No 6

Tiedback Wall

Tiedback Wall

in

in Fill Over

Homogenous

Residual Soil

Sand

0.65

0.98

0.60

0.78

0.62

1.48

+

+

0.59

1.23

J:

J:

0.65

1.16

0.62 0.69

2) ASS does not use the Spencer Method. Factors of safety for Bishop's Modified Method

are shown here for comparison with the Spencer Method values.

Slope No 7

Soil Nailed Wall

in Homogenous

Clay

1.02

1.02

1.02

+

0.99

J:

0.84

0.91

(3) The Programs SNAIL and GoldNail do not use the Spencer Method. They use their own unique methods.

Factors of safety are shown here for comparison with the Spencer Method values.

Slope No 8

Soil Nailed Wall

in Layered Soil

1.12

1.09

1.10

+

1.06

J:

1.26

1.20


32/155

Table

6:

Factors of Safety by Bishop's Modified Method

Slope

Slope

No 1

Slope No 2

Unreinforced

Unreinforced

Title Homogenous

Homogenous

Slope with a

Slope

Crack

UTEXAS4

1.29 1.28

Bishop

SLOPEIW

1.29 1.28

Bishop

SLIDE

1.29 1.29

Bishop

XSTABL

1.29 1.28

Bishop

WINSTABL

1.34 1 31

Bishop

ASS

1.29 1.28

Bishop

SNAIL

l

1.22 1.18

GOLDNAIL

l

1.32

1.30

. .

+ XSTABL has no prov1s1on for reinforcement.

i

RSS only allows for horizontal reinforcement.

Slope No 3

Slo_l _e

No 4

Tiedback Wall

Unreinforced

in Layered

Layered Slope

Soils

1.41 1.14

1.39 1.14

1.39 1.15

1 41

+

1.39

1.16

1 41

i

1.39 1.03

1.40 1.19

Slope No 5

Slo_l _e

No 6

S l ~ e N o 7

Tiedback Wall in Tiedback Wall in Soil Nailed Wall

Homogenous Fill Over in Homogenous

Sand Residual Soil Clay

0.56 1.45 1.00

0.60 0.67 1 01

0.58 1.50 1.00

+ +

+

0.74 0.63

1.06

i i i

0.65 1.16 0.84

0.62 0.69

0 91

Notes: (1) The Programs SNAIL and GoldNail do not use the Bishop Method. They use their own unique methods.

Factors of safety are shown here for comparison with the Bishop Method values.

Slope

No B

Soil Nailed Wall

in Layered Soil

1.20

1.20

1.21

+

1.13

i

1.26

1.20


33/155

1\

CD

Table 7: Factors of Safety by Janbu s Simplified Method

Slope

Slope No 1 Slope No 2

Slope No 3

Slope No 4 Slope No S Slope No 6 Slope No 7

Unreinforced

Unreinforced

Tiedback Wall Tiedback Wall Tiedback Wall in Soil Nailed Wall

Title

Homogenous

Homogenous Unreinforced

in Layered

in Homogenous Fill Over in Homogenous

Slope

Slope with a Layered Slope

Soils Sand Residual Soil

Clay

Crack

UTEXAS4

1.15 1.14 1.20

1.13 0.64

1.35

1.08

Simp. Janbu

SLOPE/W

1.15

1.14

1.21

1.05

0.61

0.76

1.07

Simp. ~ n b u

SLIDE

1.15 1.14 1.22

1.06 0.62 1.35 1.06

Simp. Janbu

XSTABL(

2

1.24 1.23

1.34

Simp. Janbu

+ + + +

WINSTABL

1.20

1.18

1.23

1.12 0.76 0.30 1.10

Simp. Janbu

ASS

1.15 1.13 1.24

p p p p

Simp. Janbu

SNAIL(

1

l

1.22

1.18 1.39 1.03 0.65 1.16 0.84

GOLDNAIL(

1

l

1.32 1.30

1.40 1.19 0.62 0.69

0.91

+

XSTABL has no prov1s1on for remforcement.

Notes: (1) The Programs SNAIL and GoldNail do not use the Simplified Janbu Method. They use their own unique methods.

Factors of safety are shown here for comparison with the Simplified Janbu Method values.

(2) XSTABL reports the factor of safety using Janbu s correction factor. Uncorrected values are provided

for comparison with other uncorrected values.

Slope Slope No 1

Slope No 2 Slope No 3

XSTABL Simp.

1.241 1.231 1.344

Janbu*

Correction

1.079 1.080 1.086

Factor

Janbu s

Fo

1.15 1.14 1.24

Slope No S

Soil Nailed

Wall in

Layered Soil

1.07

0.99

0.98

+

1.08

p

1.26

1.20


34/155

Table

8:

Difficulty Index (DI)

Slope Slope No.1

SlopeNo 2

SlopeNo 3

Slope No.4 Slope No.5 Slope

Documents

Duncan - Comparison of Computer Programs for Analysis of Reinforced Slopes