FEA of Culverts Under Embankments With Time Dependent and Non-Linear Properties

AN ABSTRACT OF THE THESIS OF
VORAVIT LERTLAKSANA for the DOCTOR OF PHILOSOPHY

(Name) (Degree)
in CIVIL ENGINEERING presented on
(Major)
Title: FINITE ELEMENT ANALYSIS OF CULVERTS UNDER
EMBANKMENTS WITH TIME DEPENDENT AND NON-

LINEAR PROPERTIES
Abstract approved: Redacted for privacy
William H. White
A computer program was developed for analyzing culverts

under earth embankments constructed of compacted cohesive soil
using the finite element method. The culvert-embankment system
is represented as a two-dimensional plane strain problem. The
plane of investigation is taken normal to the culvert axis. The soil-
culvert interface is assumed to be perfectly rough, with no possibility
of slippage. The analysis is divided into two separate parts, to deter-
mine culvert forces and deformations during construction and after
construction.
The culvert is analyzed for forces and deformations at the end
of every construction step due to the gravity effects of the fill mater-
ial. The construction pore pressure in the embankment is estimated
by the Hilf method, and zero dissipation of pore pressure during
embankment construction is assumed.
It is assumed that the soil is saturated and drainage occurs
in the vertical direction after the embankment construction is finished.
The culvert forces and deformations are calculated at any time after
construction by the modified finite element method, assuming that
changes in the culvert-embankment system occur because of consoli-
dation of the embankment.
The nonlinear, stress-dependent properties of soil are included

in the study by considering the nonlinear problem as piecewise linear
during each construction step or time increment.
The computer program can be used to analyze circular or ellip-
tical culverts under homogeneous embankments placed on a rigid foun-
dation. It has been found that for high fills the program requires
considerable computer memory locations. For complete analysis of

culverts after construction, running time for the program is large.
Finite Element Analysis of Culverts Under
Embankments with Time Dependent
and Non-Linear Properties
by
Voravit Lertlaksana
A THESIS
submitted to
Oregon State University
in partial fulfillment of
the requirements for the
degree of
Doctor of Philosophy
June 1973
APPROVED:
Redacted for privacy

Assistant Professor of Civil Engineering
in charge of major

Head of of Civil Engineering 10
Dean of graduate SChool
Date thesis is presented Fee. 22, 1973

Typed by Muriel Davis for Voravit Lertlaksana
ACKNOWLEDGMENT
The author wishes to thank the Government of Thailand for the

financial support throughout his study in the United States of America.
The author is indebted to Dr. W. H. White and Dr. W. L.
Schroeder for their encouragement, advice and suggestions provided
in the formation of this thesis.
Thanks are also extended to Dr. H. I. Laursen for his advice.
Grateful appreciation is expressed to his wife, Srisamai, for
her encouragement, suggestions and patience during the year of this
study.
TABLE OF CONTENTS
Chapter Page
I INTRODUCTION 1
1.1 Statement and Scope of the Problem 1
1. 2 Method of Solution 4
II FINITE ELEMENT FORMULATION 7
2.1 Plane Strain Matrix Equations 7

2. 2 Quadrilateral Element 7
2. 3 Curved Beam Element 13
2. 4 Non-Linear Stress-Strain Laws 20
2. 5 System Equations and Solution Process 27
III INCREMENTAL CONSTRUCTION ANALYSIS 30

3.1 Simulation of Construction 30
3. 2 Pore Pressure Calculation 31
3. 3 Initial Stresses 36
IV TIME-DEPENDENT ANALYSIS 39
4.1 An Application of the Finite Element
Method to Time-Dependent Problem 39
4. 2 Numerical. Stability and Accuracy of
Solutions 43
V TESTING OF THE COMPUTER PROGRAM 49
VI ANALYSIS OF SOIL-CULVERT PROBLEMS 57
VII DISCUSSION OF RESULTS 80
VIII SUMMARY AND CONCLUSIONS 87
BIBLIOGRAPHY 89
APPENDICES 92
Appendix A. User's Manual for Symmetrical

Soil Culvert Program 92
Appendix B. Program Description 107
Appendix C. Program Listing 110
LIST OF FIGURES
Figure Page
1.1 Finite element simulation of a soil-culvert system 6
2.1 Plane strain element 9
2.2 Quadrilateral element 13
2.3 Forming culvert elements 15
2.4 Circular beam element 15
2.5 Cantilever circular member 17
2.6 Local and global coordinate system 19
2.7 Piecewise linear curve 21
2.8 Hyperbolic stress-strain curve 22
2.9 Transformed hyperbolic stress-strain curve 22
2.10 Deviator stress-axial strain curves 25
2.11 Initial tangent modulus-confining pressure curve 26
2.12 Compressive strength-confining pressure curve 26
3.1 Simplified flow diagram for incremental

construction analysis 32
3.2 Void ratio-effective stress curve 34
3.3 Approximation of no drainage during construction 37
4.1 Interpolating parabola for difference equation 42
, 4.2 Simplified flow diagram for time-dependent analysis 44
5.1 Finite element model of circular ring 51
5.2 Deformed shape for circular ring 52

Fi_gure Page
5.3 Bending moment diagram for circular ring 52
5.4 Finite element model of soil layer 53
5.5 Numerical instability due to large time interval 54
5.6 Plot of pore pressure vs. time, linear consoli-

dation problem 55
5.7 Pore pressure distribution in soil layer 56
6.1 Properties of the simulated soil-culvert system 60
6.2 Soil-culvert finite element model 61
6.3 Initial tangent modulus-confining pressure for

undrained condition 63
6.4 Compressive strength-confining pressure for

undrained condition 63
6.5 Void ratio-effective stress curve of Bully Creek soil 64
6.6 Deformed shape at the end of construction,

flexible culvert 65
6.7 Deformed shape at the end of construction,

rigid culvert 66
6.8 Plot of vertical deflection at the crown of

culverts vs. height of fill 67
6.9 Plot of vertical pressure at the crown of

culverts vs. height of fill 68
6.10 Radial pressure on flexible culvert at the end

of construction 69
6.11 Radial pressure on rigid culvert at the end

of construction 70
6.12 Plot of bending moment during construction

period, flexible culvert 71
Figure Page
6.13 Plot of bending moment during construction
period, rigid culvert 72
6.14 Plot of maximum bending moment vs.

height of fill 73
6.15 Pore pressure contours at the end of construc-

tion, flexible culvert 74
6.16 Pore pressure contours at the end of construc-

tion, rigid culvert 75
6.17 Plot of vertical deflection at crown of flexible

culvert vs. time after construction 76
6.18 Plot of the pore pressure vs. time after

construction 77
6.19 Plot of Terzaghi's pore pressure - time

relationship 78
6.20 Plot of the approximate vertical deflection at

crown of flexible culvert vs. time after
construction 79
Appendix
Figure
A. 1 Example of finite element mesh showing node
number scheme 94
LIST OF TABLES
Table Page
1 Initial conditions for tested soil samples 62
2 Drained stress-strain parameters 62
3 Comparisons of results at the end of construction 82
4 Deflections at crown of the flexible culvert 85

LIST OF TERMS
Term Description
Area of the triangular element
Cross sectional area of culvert wall
C Coefficient of permeability
Slope of the void ratio-effective stress curve
Cy Coefficient of consolidation
Drained cohesion
E Young's modulus
E. Undrained initial tangent modulus
Drained initial tangent modulus
Et Tangent modulus
e Void ratio
Initial void ratio of the pore air
Initial void ratio of pore water
{F}' Nodal force vector
{F0}/ Nodal force vector due to body forces
Fx Horizontal body force
F Vertical body force
y
[f] Flexibility matrix
H Henry's constant of solubility of air in water
h Element thickness
I Moment of inertia
Term Description
K Modulus number
Ko Coefficient of earth pressure at rest
[k] Element stiffness matrix
{L) System load vector
N Exponent
Pa Absolute atmospheric pressure

R Radius of circular beam segment
[S] System matrix
T Time factor
t Time
u In plane horizontal displacement
Pore pressure
uo Initial pore pressure due to compaction
Volume of the soil element
v In plane vertical displacement
X Horizontal force at nodal point
Y Vertical force at nodal point
Central angle
Density of water
"Y
xy Shearing strain in x-y plane
Volumetric strain
Horizontal strain
Term Description
E Vertical strain
y
Undrained Poisson's ratio
Drained Poisson's ratio
a-
x Horizontal total stress
cr Vertical total stress
y
7x Horizontal effective stress
o- Vertical effective stress
y
T
xy Shearing stress in x-y plane
{8} Nodal displacement vector
[X.] Displacement transformation matrix
{} System unknown vector
FINITE ELEMENT ANALYSIS OF CULVERTS UNDER
EMBANKMENTS WITH TIME DEPENDENT
AND NON-LINEAR PROPERTIES
I. INTRODUCTION
To successfully design a culvert under an embankment of com-

pacted cohesive soil, it is necessary to have information about culvert
forces and deformations which develop during and after construction.
Although culverts under earth embankments have been constructed
for hundreds of years, it is only with the past few decades that
rational methods of structural design of this type of structure have
become available. These methods are based on results from field
observations and theoretical analyses. They take the time dependent
properties of soils into account only empirically.
1.1 Statement and Scope of the Problem
Analyzing the soil-culvert interaction problem is difficult be-

cause almost all cohesive soils have time-dependent and non-linear,
stress-dependent properties. Even if the soils were ideally linear,
stress-independent and time-independent, the problem would still be
a very difficult one to solve. The difficulties are threefold. First,
the soil-culvert system is a composite one; second, the geometry of
the problem is such that straight-forward solutions are not easy to
find; and third, the boundary conditions between the fill and both the
foundation and the culvert are complicated.
2
The object of this research was to combine existing theories

methods in mechanics in a way which would make them useful for
the analysis of culverts under earth embankments. The time-
dependent and non-linear, stress-dependent properties of cohesive
soils are included in the study. The step-by-step construction
process is taken into account.
To deal with the complexity of the problem, certain assump-
tions have been introduced. They are as follows:
(1) The actual three dimensional system can be represented
as a two-dimensional plane strain problem. The plane of investiga-
tion is taken normal to the culvert axis.
(2) The compacted soil is homogeneous and isotropic.
(3) The time-dependent characteristics of soil are controlled
by the dissipation of excess pore pressures.
(4) No drainage is allowed during the embankment construc-
tion.
(5) Drainage occurs in a vertically upward direction only,

irrespective of the shape of the structure.
(6) The embankment foundation is rigid, impervious and rough.
(7) The soil-structure interface is perfectly rough, with no
possibility for slip.
The scope of the study included development of a method for
analyzing soil-culvert interaction problems and preparation of a
3
computer program to make the necessary calculations.

An important method of design for culverts was introduced by
Marston and Spangler (22). The well known method was based on
arching theories and empirical formulas which resulted from experi-
ments and observations of the performance of actual culverts under
embankments. The method attempted to determine the vertical loads
to which the culverts would be subjected in service and their support-

ing strengths. There is a great deal of uncertainty regarding the
empirical factors used in the method. Most of these factors depend
on field conditions and can not be determined in advance. The pro-
posed method is limited to circular, small culverts.
White and Layer (29) proposed ring compression theory for
analyzing flexible, ring shaped culverts. The purpose of the theory
was to determine the compressive force in the ring. The ring com-
pression approach was based on the assumption that the vertical
component of the compression can be computed by taking one-half of
the span of the culvert and multiplying it by the height of cover times
the soil density.
Brown (4, 5) attempted to determine the pressure distribution
around culverts due to the gravity effects of the fill material and study
the interface conditions between the culverts and fills by using the
finite element method. The investigations were based on the assump-
tions that the problems can be represented by a plane strain system
4
and the material properties are linearly elastic. The construction

process was included in the studies. Brown found that the distribution
and magnitude of predicted pressures were not sensitive to the interface
conditions in the rigid culvert case, and the using of the interface con-
dition of no-slip in the flexible culvert case gave a good agreement be-
tween the predicted and field measured results. Brown did not con-
sider the time-dependent properties of soil and presented the compari-
sons between the predicted and field measured results only during
construction.
Trollope, Speedie and Lee (26) published field measurements
of the vertical pressures acting on a rigid culvert under an earth dam.
It was indicated that the pressures increased with time after the con-
struction. He presumed that the pressures on the culvert increased
due to continued settlement of the fill with consequently increased
arching effects.
1.2 Method of Solution
The finite element method developed by Clough (9) will be used

in the study. The basic concept of the numerical method is that a con-
tinuum may be represented by an assemblage of a finite number of
elements interconnected at the element nodal points. After the stiff-
neSs characteristic of each element is obtained, the structure is
analyzed by the standard stiffness method (15).
The soil-culvert system is not a continuum so an assumption
5
concerning the characteristics of the soil-culvert interface is needed.

It is assumed that the interface is perfectly rough, with no possibility
for slip between the culvert and the soil (4, 5). A symmetrical struc-
ture is used in the study in order to keep the number of elements
minimum. The soil-culvert system can be composed of curved beam
elements and triangular or quadrilateral elements as shown in
Figure 1.1. Face 1 is assumed fixed. For the symmetrical case,
there is no vertical shear force or deflections in the x direction
along face 2. Face 2 must therefore be kept on rollers. If face 3
is assumed to be far from the axis of symmetry oy , the deflections
in the x direction along the boundary are negligible. Face 3 can be
assumed to be on rollers.
There are two consecutive analyses in the study:
(1) Incremental construction analysis which is carried out in
terms of total stresses. It yields the distributions of stresses and
excess pore pressures in the fill, and also forces and deformations
in the culvert at any stage of construction prior to dissipation of any
excess pore pressure resulting from construction.
(2) Time-dependent analysis which starts at the end of con-
struction in terms of effective stresses by taking the results from
incremental construction analysis as the initial conditions.
6
Interface condition
Roller
Face 3
Figure 1.1. Finite element simulation of a soil-culvert

system
II. FINITE ELEMENT FORMULATION
The basic finite elements used in the study are a two-dimensional

plane strain element and a curved beam element. The embankment
is idealized by triangular and quadrilateral plane strain elements
connected at nodal points. A culvert is approximately represented
by a number of circular beam segments connected at nodal points.
Since most compacted, cohesive soils have non-linear, stress-
dependent stress-strain properties, these properties must be taken
into account.
2.1 Plane Strain Matrix Equations
\ A triangular plate element is used in the study as shown in

Figure 2.1. A global coordinate system x, y is selected as a refer-
enced coordinate system By equating the external virtual work and
the internal virtual work due to the nodal forces, body forces and
stresses acting in the element, the following matrix equation is
obtained:
[u v u
1 1 2
v2 u3v 3 x y xy ]
E 'Y 0-0-X
Ty
dV
Xy
dV (2.1)
8
The volume of the element is indicated by V. The terms X.

and Y. represent the horizontal and vertical components of nodal
force of the nodal point i which takes on values 1, 2 and 3. The body
forces Fx and F are positive if they act in the direction of positive
x and y, respectively. Admissible forms for the inplane element
displacements u and v are chosen so that displacement continuity
is maintained across element boundaries. Therefore, assume as
trial functions
u=a 1
-Fa
2
x +a 3 y (2. 2a)
and
v = a4 + a x +a by (2. 2b)
Let u. and v, refer to the x and y components of the displace-

ments, respectively, of any node i of the element. By using the
displacement boundary conditions for the horizontal displacement u
at each corner, the following matrix expression is obtained:
1 x
1 y1
1 y2 (2.3)
1 x3 y3
9
Figure 2.1 Plane strain element
The constant a, are obtained by inversion as
a x2y3-3y2 x3y1.-xly3 yz-x2yi

1
a2 2A Y -Y 2 (2. 4)
y2 -y3 y33 -y1 1
a3 x3-x2 xi-x3 ,x
-, 2 xl
where A is the area of the triangular element. If Equation (2. 4) is

rewritten as follows
al a
2
a
3
2A
bl b b (2.5)
2 3
c c c
1 2 3
and substituted into Equation (2. 23), the displacement u becomes

10
1
u= 2A.
(n
1
u +n 2 u 2 +n 3 u 3 ) (2. 6)
where
n.1 = al + blx + cly

n
2
= a2 + b2x + c2 y
n
3
= a3 + b3x + c3y .
Similarly, the displacement v may be written

1
v =
2A
(n. v
1 1
+n v
2 2
+n v
3 3
) (2.7)
The displacements at any point in the triangular element can be

written in terms of the nodal displacements as
1
0 vl
2A = [N] f-u
!Z 2. 8)
v2
u3
v
3
The chosen displacement functions in Equation (2.8) guarantees

continuity of displacements with adjacent elements because the dis-
placements vary linearly along any side of the triangular element
and, with identical displacements imposed at the nodes, the same
displacement will exist all along an interface.
By differentiating Equations (2. 6) and (2. 7) the strains can be
found by
au
ax
= 3v
E ( 2. 9)
Y ay
xy
= 3u
ay
-
av
ax
The strain-displacement relation can be written as
bl 0 b
2
0 b
3
0 ul
vl
1
0 cl 0 c
u2 (2. 10)
2A 3
2
u3
cl bl c2 b2 c3 b v3
3
where b 1 c 1 etc. are given by Equation (2.5)

Equation (2.10) can be written symbolically as
E = J (2. 11)
xy,
The stress-strain relationship for an isotropic material with

Young's modulus E and Poisson's ratio v in a state of plane strain
is
1 -v v 0
X
(l+v)(1-2v) v 1 -v 0 = [D] (2.12)
1 -2v
0 0
xy
t:r.1"
2 xy Xy,
Substituting Equation (2.9) into Equation (2.12), the stresses become

12
0
x 1
V1
1
[D] [B] (2. 13)
y V2
u3
T v3
xy
If Equations (2.8), (2.11) and (2.12) are substituted into Equa-

tion (2. 1), a relation between the nodal forces and the nodal displace-
ments is obtained
X1
Y1
X2
Y2
= [B]T[D][B]dV) [N]rr x dV (2.14)
X3
Y3
After an integration is taken over the volume of the triangular

element, Equation (2.14) can be written symbolically as
{F} [k] {6} + {F o} (2.15)
where
[k] = hA [B]T[D] [B]
is the element stiffness matrix of the triangular element with thick-

ness h and
{F} nodal force vector
nodal displacement vector
{F0) nodal force vector due to body forces.
13
2. 2 Quadrilateral, Element
One of the drawbacks to the triangular element is that if it is

very large the determined stresses may not represent the actual
stresses anywhere in the element. In order to circumvent this diffi-
culty the quadrilateral element was developed by Wilson (30),
In the quadrilateral element a fictitious nodal point is found
shown in Figure (2. 2). The coordinates of the nodal point are com-
puted as the average of the four corner point coordinates.
Figure 2. 2 Quadrilateral element
A 8-degree-of-freedom quadrilateral element matrix is formed

by first combining the four 6-degree-of-freedom triangular element
matrices into a 10-degree-of-freedom element matrix. Using a
process of static condensation (32) the two internal displacements are
eliminated, resulting in a 8-degree-of-freedom quadrilateral element
matrix.
14
The four triangular element stiffness matrices are combined by

the code number technique (15, 18, 25). If the load, vectors F0 }
for each of the triangular elements are similarly superimposed, the
partitioned matrix equation is obtained
k k
as a5 a oa
Oa 15a k55 55 Fo5
(2.16)
i
a = 1, 2, 3, 4
where a and 5 indicate the quantities associated with the external

and internal nodes, respectively.
The zero on the left-hand side of Equation (2. 16) is due to the
fact that the internal forces balance at the internal node,
By eliminating the internal unknowns {65} the quadrilateral
element matrix becomes
{Fa} = [0] {5a} + F (2.17)

oa
where
[0] = [kaa] a5][k55]-1[k5a]
- 1
oa } = {Foa} - [ka5] [k551
{F* {F05}
2.3 Curved Beam Element
The culvert is represented by a number of circular beam segments

connected at the nodal poirxts. The geometric properties of the circular
15
beam segment can be determined by using the properties of a circle

passing through three consecutive points as shown in Figure 2. 3.
y
Figure 2. 3. Forming culvert elements
The equation of the culvert element i can be written as
(x - P.)2 + (y - Q.)2 = R. (2. 18)
The circular member has 6 degrees of freedom as shown in

Y
Figure 2. 4.
y
Note: Quantities as shown are considered positive
Figure 2. 4. Circular beam element

16
The local coordinate system x, y is selected so that the x

and y axes are tangent and perpendicular to the culvert member,
respectively. The central angle (3 is found by the dot product of the
unit normal vectors at nodal points 1 and 2 and the result: can be
written in a form as
cos p = 2
[(x
1-P) (x 2 -P) + (y 1 -Q) (Y -Q)1
(2.19)
R
The desired stiffness matrix equation is written in the parti-

tioned form as
kaa kab
(2. 20)
kba kbb
i
a = 1, 2, 3
b = 4, 5, 6
where as shown in Figure 2. 4
axial force at nodes 1 and 2

q2, q5 - shearing force at nodes 1 and 2
q3, q6 - bending moment at nodes 1 and 2
and d represents the displacements corresponding to the member

forces.
The stiffness matrix is obtained by using the inversion of the
flexibility matrix [f] and the equilibrium conditions of a cantilever
circular member as shown in Figure 2.5.
17
Figure 2.5. Cantilever circular member
The matrix [k ] is obtained by the inversion of the flexibility

as
matrix [f] as
[kaa] = [f]-1 (2.21)
The flexibility coefficients are as follows

3
fll = R 13P 2 sin13+12 sin p cos (3) + R
sinPcos (3)
EI EA 2
f =
R3 -sin 213 cos(3 + 1) - sin 2p
12 =f21 EI 2 2EA
2
f =f R(P- sin (3)
13 31
f R3
22 cos (3) + (P2- - sin13 cos (3)
EI 2 2 EA
2
R
f2 =f 32 EI
(- cos +1)
f R
33 El
18
where
E- Young's modulus
-1 moment of inertia
13 central angle in radian
A. cross sectional area
By the equilibrium conditions, the matrix [kba.] is obtained
terms of the stiffness coefficients of the matrix [k aa ]. The ,coeffi- .in
cients of the matrix [kba] are given as
k
41
= -k 11 cosP+ k 21 sin
k
42 -ki2cos p + 1(22 sinP
k
43
= -k13 cos p + kn sinP
k
51
= -k 11 sinP k
21
cos P
k52
= -k 12 sinP- k 22 cos (3
k
53
= -k13 sing - k23 cos p
k
61
= -k 11 R(1-cos p) - k 21 sin(3- k 31
sin-
k62 = -k 12 R(,1-cos(3) - k 22R sinP- k 32
= -k 13R(1-cos
23R sinP- k33
k k
63
By using the symmetrical properties of the stiffness matrix and

the circular member, the matrices [k ] and [kbb] are obtained
ab
[k ] = [k ]T (2. 22)
ab ba
19
kll -k12 k13
-k21 k22 -k23 (2. 23)
k31
-k 32 k
33
In order to determine the stiffness property of the complete

structure, the stiffness matrix of the culvert element must be refer-
enced to a global coordinate system x, y as shown in Figure 2.6.
Figure 2. 6. Local and global coordinate system
The culvert stiffness matrix based on the global system, ,
is obtained by the following relationship (17)
[k*] Ex 1T [k] [x] (2. 24)
where [X ] is the displacement transformation matrix relating the

local displacements to the global displacements. The matrix [X.]
20
can be determined by calculating the direction cosines of the angles

between the local and global coordinate systems and it yields as
(yl-Q) -(x 1 -P)

0 0 0 0
R R
(x -P) (yl-Q)
1
0 0 0 0
R R
0 0 1 0 0 0
[NJ = (2.25)
2
-Q) -(x 2-P)
0 0 0
R R
(x
2
-P) (y2-Q)
0 0 0 0
R R
0 0 0 0 1
2.4. Non-Linear Stress-Strain Laws
Most cohesive soils have a stress-strain relationship that is

non-linear and depends on the state of stress. Woodward (31) pro-
posed an incremental type of non-linear analysis for estimating
stresses and deformation in earth dams during construction. The
method treats the non-linear problem as piecewise linear with the
soil properties modified after each increment of load in accordance
with the state of stress computed in the element. The tangent modulus
Et at the beginning of a load increment is calculated based on the
existing deviator stress (o- -o- 3) and confining stress o-3 and it is
21
used during the load increment as shown in Figure 2.7. Woodward

used data obtained from laboratory tests of soil samples to represent
stress-strain relation of soil. He could not express the relation in
mathematical form.
Note: Each curve performed at

different confining stress, 0-3
7t)
b
vl
Axial strain, E
Figure 2.7. Piecewise linear curve
Duncan and Chang (10) developed a simplified, practical proce-

dure for representing non-linear, stress-dependent soil stress-strain
behavior in a form which is convenient for use in incremental finite
element analysis. The non-linear stress-strain curves of soil may
be approximated by hyperbolas with a high degree of accuracy as
shown in Figure 2.8. The hyperbolic equation proposed was
(a - 6 3) - (2. 26)
1 a + bE
1
22
1
Asymptote = ( )
1 Cr 3 'ult. b
Axial strain, E
Figure 2. 8. Hyperbolic stress-strain curve
m
b
b
-
ti
0
CC:
0.)
"0
.H
Co
41)
Axial strain, E
1
Figure 2. 9. Transformed hyperbolic stress-strain curve

23
where a is the reciprocal of the initial tangent modulus E. and b

is the reciprocal of the asymptotic value of deviator stress (o- - o- )
1 3 ult.
The values of a and b may be determined if the stress-strain data
are plotted on transformed axes as shown in Figure 2.9. It is com-
monly found that the value of the asymptotic value (o-
1
)
ult. is
larger than the compressive strength of the soil (o-

1 3
)
f
by a small
amount.
(0-1 cr 3)f = R (o- o- ) (2. 27)

f 1 3 ult.
where Rf is the failure ratio which has been found to be between

0.75 and 1.00 for most of the soils.
By substituting the value of a and b into Equation (2.26), the
equation can be written as
E
1
(o-
1
- 03) -= (2. 28)
E R
1 1 f
E (o- -o- )
1 3f
If the value of the confining stress o- is constant such as is

normal in the case of triaxial shearing tests, the tangent modulus can
be written as
a(cri -o-3)
Et (2. 29)
By performing the differentiation on Equation 2.28), the tangent

modulus can be written in the convenient form
24
2
Et [1 - RI ] E. (2.30)
t( crl-cr3if
The value of compressive strength (61 -o- 3 ) f and the initial

tangent modulus E. are dependent on the confining stress 6 (16)
and they can be obtained by performing the shearing tests on the soil
samples corresponding to the field conditions. Because the confining
pressure at any point in the embankment changes during construction
and after construction and because each point will be subjected to
different stresses and deformations in general, there is no unique
stress-strain relationship for the compacted soil. What is needed
for this analysis is a series of shear tests performed at changing
confining pressures.
Since drainage is not allowed during construction, unconsoli-
dated-undrained shear tests could be used to predict stress-strain
relationship of soil during the construction period. The typical series
of curves relating the deviator stress (61-.6 3) and the axial strain
E
1
obtained from the shear test performed at several confining pres-
sures are shown in Figure 2.10,
From the curves in Figure 2.10, the undrained initial tangent
modulus E. and the undrained compressive strength (61-63)f can
approximately be obtained as a function of the confining pressure 63
as shown in Figures 2.11 and 2.12, respectively. The initial tangent
modulus and the compressive strength of the compacted soil increase
25
steadily with the confining pressure and become constant when the
confining pressure required to saturate the soil 63 is reached.
The curved portions in Figures 2.11 and 2.12 can be replaced by
straight lines without causing significant errors.
Note: Each curve performed at

different confining pressure, Cr 3
be('
Axial strain E
Figure 2.10. Deviator stress-axial strain curves
Poisson's ratio used in the undrained case will be between 0.4

and 0.5 for a nearly saturated clayey soil (31).
After the embankment construction is ended, the water is
assumed to begin to flow out of the soil mass. The drained initial
tangent modulus Ei and the drained compressive strength Fr - 7- )
1 3 f
26
w." Straight line approximation
0-
3s
Confining pressure 0-
3
Figure 2.11. Initial tangent modulus - confining pressure curve
b Straight line approximation
b0
Cs
a)
Vl
3s
Confining pressure
63
Figure 2.12. Compressive strength - confining pressure curve

27
needed in Equation (2.30) can be obtained by performing a number of

drained shear tests at different confining pressures o-3.
The relationship between the drained compressive strength and
confining pressure may be expressed conveniently in terms of the
Mohr- Coulomb failure criterion as
(o-l-o-
c cos 4) + 26 sm cl)
3
3)f (2. 31)
1 - sin (1)
where
- drained cohesion
131). - drained angle of internal friction.
The variation of the drained initial tangent modulus value with

confining pressure was represented by an empirical equation suggested
by Janbu (13)
7-1 N
Ei = Kp a pa (2. 32)
where K and N are modulus number and exponent, respectively

which are pure numbers and pa is the value of atmospheric pressure.
The typical value of drained Poisson's ratio v is 0. 3 (14).
2.5. System Equations and Solution Process
As the element stiffness matrices referenced to the global

system are generated, they are appropriately superimposed into a
system matrix [S]. The superimposition is accomplished by using
28
the code number technique (15, 18, 25).

A structural system that contains a large number of elements
will involve a good deal of input data preparation. This includes a
code number for each element. In order to reduce preliminary work
of this nature, a subroutine was written to generate the code numbers
for each element.
When the system matrix [S] and system load vector {L} are
generated using the code numbers, the total system of simultaneous
algebraic equations in the unknowns {.1)} is represented as
[S] {<1,} = {L} (2.33.)
The primary concern in the solution of this system is the con-

ditioning of the system matrix. It is a symmetrical matrix and can
be a banded or an unbanded matrix depending on the concerning prob-
lem. In the incremental construction analysis where the unknowns
are only nodal displacements, the system matrix is banded. In the
time-dependent analysis where nodal displacements and element pore
pressures are the unknowns, the system matrix is unbanded as a
result of fill-in matrices which will be explained later in Chapter IV.
By taking an advantage of the symmetry, the coefficients are stored as
an upper triangular matrix.
The solution process, contained in a single subroutine, is
Gaussian elimination. The band-width is automatically computed prior
29
to solving the system as it is required input to the solution subroutine

along with the system load vector.
30
III. INCREMENTAL CONSTRUCTION ANALYSIS
Incremental construction analysis is used to perform the follow-

ing features during the construction of a culvert under an earth fill:
(1) Calculation of stresses and deformations in the fill.
(2) Calculation of construction pore pressures.
(3) Calculation of forces and deformations in the culvert.
3.1. Simulation of Construction
The step-by-step construction procedures can be easily included

in the calculations. When a lift is added to a partially completed
embankment it forms a new structure which must be analyzed. First,
the stiffness of this new structure is determined, then the weights of
the soil elements in the new lift are calculated and the weights of the
soil elements used previously are set to zero. Finally, the displace-
ments, stresses and forces in the new structure caused only by the
weight of the new lift are calculated. These are then added algebrai-
cally to the displacements, stresses and forces existing in the par-
tially completed embankment. By repeating this procedure for each
successive construction lift, the complete history of the development
of displacements, stresses and forces can be obtained.
As mentioned in Article 2. 4 the soil is assumed to behave
linearly during each construction increment, with stiffness properties
31
in each soil element defined on the basis of the local stress state which
existed at the beginning of the increment. The accuracy of the results
that are obtained will depend, of course, on the size of the increments
which are considered, but experience indicates that relatively coarse
construction increments will yield good results in the analysis of
typical earth embankments. Woodward (31) found that by using lift
thickness of one tenth of the embankment height acceptable solutions
could be obtained.
The flow diagram for the incremental construction analysis is

shown in Figure 3.1.
3. 2. Pore Pressure Calculation
The construction pore pressures due to the gravity effects of

the fill material are predicted by the method developed by Hi lf (12).
He assumed that the pore water pressure in a partially saturated soil
could be related to the amount of compression by combining Boyle's
law for compressibility of air with Henry's law for the solubility of
air in water. The pore water pressure due to applied stresses is
computed from
pAe
U
a (3.1)
+ He - e
w0
where
u - induced pore pressure
Read input data
32
Calculate and print nodal displacement numbers

and code numbers for each element
Calculate and print the geometry of culvert elements
Add one layer and calculate tangent modulus for each

soil element based on existing stresses
Generate system stiffness matrix and system load vector
Calculate and print nodal displacements
Solve for culvert forces and stresses,

strains in soil elements
Solve for pore pressures in soil elements
Print culvert forces, stresses and pore pressures
Yes
Time-dependent analysis
Figure 3.1. Simplified flow diagram for incremental construction

analysis
33
pa - absolute atmospheric pressure
e - numerical value of change in void ratio from initial

conditions
ea - initial void ratio of the pore air
0
H - Henry's constant of solubility of lair in water

e
w
- initial void ratio of pore water
0
Equation (3.1) is correct if the assumption of equal pore air

and pore water pressures is satisfied (i. e., if surface tension can be
neglected). It is apparent that as the compaction moisture content
increases above optimum, the validity of the assumption increases
due to the higher degree of saturation (23).
For saturation due to compression, a e = , Equation (3.1)
is reduced to
pa
u (3. 2)
He
w0
The Equation (3. 2) is correct if drainage is not allowed.

The change of void ratio, i e , can be obtained by using an
iterative process and the void ratio-effective stress relationship for
the soil determined in the conventional consolidation test as shown
in Figure 3.2. Effective stress is defined by
= 0- -U (3. 3)
where
34
- effective stress
o - total stress
- pore pressure
Initial conditions
e
0 Over consolidated region
Normally consolidated region
(log) Major principal effective stress 6I
Figure 3. 2. Void ratio-effective stress curve
The calculation of the pore pressure change due to a change in

total major principal stress Ao-
1
consists of the following steps
proposed by Steward (23).
(1) Assume that the change in effective stress is equal to one-
half of the change in total stress
Ao-
1
Ao- = (3. 4)
1 2
35
The number in the bracket indicates the number of approximations.

(2) Calculate the change in void ratio due to the assumed change
in effective stress
__(1)
+ Ao-1
e = Cc log ( (3.5)
crl
The compression index Cc is the slope of the void ratio-effective

stress curve in Figure 3. 2.
(3) Add e to the sum of the previous void ratio changes and
calculate the pore pressure by using Equation (3.1).
(4) Calculate the effective stress using the calculated pore
pressure u(1)
c7-
1
= o-
1
- u (1) (3.6)
(5) Calculate the change in effective stress ,676--

1
and obtain
the second approximation of the change in effective stress X612) by
- (1)
-(2) =
(1)
o- +
1
o-
1
(3.7)
1 1 2
By repeating steps two through five until the change in the value
of Ae is very small, the change in pore pressure due to the change
in total stresses is obtained. It was found from experience that using
ten cycles of iteration gives good results.
After the pore pressure is calculated, a check is made to see if
the soil is saturated. If the soil is saturated, the pore pressure is
36
recalculated by adding the pore pressure at saturation calculated from

Equation (3. 2) to the increment of total stress from the value required
to completely saturate the soil.
It was mentioned in Article 1.1 that the assumption of no drain-
age during the construction period will be used in the study. The
assumption is justified if the following requirements are satisfied:

(1) The embankment is constructed with an impervious material
such as clay which has permeability less than 10-6 cm/sec.
(2) The drainage path is long such in the case of high fill.
(3) The construction time is short so that there is not enough
time for drainage.
The approximation of no drainage during construction is shown
in Figure 3. 3 by considering a soil element in the first lift. The
amount of reduction in pore pressure due to drainage will decrease
as the number of lift is increased, i. e., the drainage path for the soil
element is also increased.
3. 3. Initial Stresses
The initial stresses for the soil elements in the newly added
layer due to the effects of compaction can be calculated by the expres-
sions
+ uo (3.8)
l
0
37
No drainage
Drainage during construction
w
0;,.
04 / 4/ r
.31
r.,
0 / //
P4 V
/
1 2 3 4
Number of lifts
Figure 3. 3. Approximation of no drainage during

construction
38
0 = K o- (3.9)
x oy
T = 0 (3. 10)
xy
where
= vertical total stress
x
= horizontal total stress
xy = shearing stress
T
o-
1
= major principal effective stress corresponding to the
o initial void ratio as shown in Figure 3. 2.
uo .= initial pore pressure due to compaction
Ko = coefficient of earth pressure at rest.
The pore pressure due to compaction uo can be neglected when

it is compared with its final value induced by the weight of high fill,
especially when soil is compacted at the water content above the opti-
mum value (3). In the study the value of coefficient of earth pressure
at rest Ko is taken to be 1.0 which it is implied that stresses in-
duced by compaction are distributed uniformly in the newly placed
layer. The effects of compaction on the underlaid layers can be
neglected when it is compared with the effects caused by the dead
weight (31).
The initial stresses due to compaction are added algebraically

to those caused by the gravity effects of the fill material.
40
The matrix tic '} is of the form
k` } = [B]
T(11 hA. (4. 3
0
There are two sets of unknowns in Equation (4. 2), the nodal
displacement vector {6 } and pore pressure u. An additional equa-
tion is required in order to solve Equation (4. 2). The additional equa-
tion needed is the restriction on volume change AV of the soil
element. The volumetric strain En can be written as the sum of
strain in x and y directions as
En E
x
+E
y
(4. 4)
By substituting the values of E and E from Equation (2. 11),

Equation (4. 4) becomes
1 r
E
n Lb C b b C 1{6} (4.5)
2A 1 2 3 3
or
en = [1 1 0 ] [B] {6} (4. 6)
Therefore, the change in volume A V of the triangular soil element

can be written as
AV=hA[110] [B] {5} (4.7)
or symbolically as
AV = {ki}T {6} (4. 8)
41
Combining Equations (4. 2) and (4. 8) gives the generalized-

stiffness formulation for the triangular soil element
1-
k k
T 0 "-AF11)
(4.9)
k' I
'
The time-dependent analysis starts at the end of construction.

The pore pressures at time t are used to calculate the change of
volume to be expected in each soil element during an increment of
time At. This volume change AV can be used to solve for new
forces, stresses, deformations and pore pressures in the soil-culvert
system at time t + At, and so on until a steady state is reached.
The steady state is practically chosen at the time when the further
changes in pore pressures do not cause any significant changes in the
soil-culvert system.
If water is assumed to flow from the soil mass in the vertical
direction only and further that the soil is saturated after construction
due to compression, the change in volume of a soil element can be
obtained from an equation (14)
av = v c a2 (4. 10)
8t w ay2
where y is the density of water and C is the coefficient of per-

meability of soil in the vertical direction. The validity of the assump-
tion of saturation at the end of construction increases if the soil is
placed at the water content above the optimum value (2). If the
42
time increment At is taken small enough, the change of volume can

be written as
(A-v) (v) (
au ) At (4.11)
t-t+At
o- t ay 2 t
where the subscript t indicates the value at time t. If the pore pres-
sure u is assumed to vary in the vertical direction according to a
parabolic equation
2
u= 1
+ P 2Y + P 3Y (4.12)
the second derivative of u in Equation (4.12) at the center of element

i can be obtained by using the finite difference equation (19)
a2u 1 2
ay
2
)
2 a(a+1) [ auj - (1+a)u + uk] (4.13)
(Ay)
where A y and aA y are the distances from the center of element i
to the centers of element j and element k, respectively, as shown

in Figure 4.1.
2
u
2
y+ P y
3
aAY
AY
Figure 4.1 Interpolating parabola for difference equation

43
Conditions of no flow across a boundary of a soil element are

approximated by defining across the boundary a mirror image of the
center of the element and by making the pore pressure at this point
equal to that in the element inside the boundary.
The flow diagram for the time-dependent analysis is shown in
Figure 4. 2.
4. 2. Numerical Stability and Accuracy of Solution
The use of a piecewise linear analysis with tangent modulus in

each soil element defined on the basis of stress state at the beginning
of each step of analysis and the use of Equation (4. 11) to calculate the
volume change during the step of analysis will lead to the numerical
stability and accuracy problems if excessively large increments of
time are used.
If Equation (4.13) is substituted into Equation (4.11), the volume
change of a soil element i during an increment of time At can be
obtained as
(AV)i,
t+ At C At 2 t- - (1- a)u. +u (4. 14)
V.
t 2
w (Ay) a(a+1) i' t k, t]
In general, the smaller time increment and finer finite element

mesh will increase the accuracy of the solution. But as the accuracy
increases, the computer storage locations and costs also increase.
44
Print nodal displacements at the end of construction
Print culvert forces, effective stresses, and

pore pressures at the end of construction
stop
Calculate Et, and At if time steps are greater than

time steps needed to start the solution
Time steps >

time steps needed to start
the solution
Print At and time after construction
Generate system stiffness matrix and

system load vector
Calculate and print nodal displacements,

culvert forces, effective stresses and
pore pressures
Figure 4. 2. Simplified flow diagram for time-dependent

analysis
45
The conditions to ensure both stability and accuracy are not readily
obtainable for Equation (4. 14).
Scott (20) used Equation (4. 14) to solve one-dimensional consoli-
dation problems by neglecting the strain in the horizontal direction
and assuming that the change in vertical stress is equal to the change
in pore pressure. The left hand side of Equation (4.14) can be
written as
(AV)
i,t->t+At= (Au)
(1-2-75)(1+v)
(4. 15)
[
V.
t t t+At Et (1- v)
After substituting Equation (4. 15) into Equation (4. 14), the finite
difference equation for Terzaghi's one-dimensional consolidation
solution is obtained as
At 2
au. -(1-a)u +u ] (4.16)
t -.t+At a(a+1) 3, t t k, t
,
Cy
(Ay)2
where
Cy = coefficient of consolidation
(1-v)
C Et
y (1-27)(1+7)
w
Scott rewrote Equation (4. 16) in terms of time factor T and

dimensionless factor AY
(Au) AT
T -> T+AT
[auj -(1-a)u +u (4. 17)
2 a(a2 +1) , T T
(AY)
46
where
At Cv
AT =
H2
AY =
H
H = total height of the embankment
Scott proposed a condition that ensures both stability and accuracy for
Equation (4.17). He expressed that if an error E is made in the
determination of u, T
at one cycle of solution, i. e., at a time T,
from any cause, then the error in the subsequent value of ui, T+A T
should not be more than E. If this is true, the step-by-step calcula-
tion of u. as a function of time will not be divergent, since the com-
putational errors do not increase in the course of the solution. For a
stable solution, the inequality must hold
I 2M (4. 18)
a +
a -11 ) E
where
AT
M =
2.
(AY)
From Equation 4. 18), it follows that
M
a
< 0.5 (4. 19)
where a is a stability number.

By trial Scott found that the use ofa = 0.5 results in a stably
oscillating solution, which is also undesirable. In addition, the
47
smaller the value of AY selected, the more accurate will be the

solution. Barden (1) found that using AY of 0.1 will give acceptable
results.
It is necessary to emphasize that the stability condition given
by Equation (4. 19) was derived from Equation (4. 17) where the change
in pore pressure is directly calculated. In the present study the
change in pore pressure is obtained by solving the generalized-stiff-
ness equation (4. 9) which contains very large and very small coeffi-
cients such as in the case of soil-culvert problem. It is of course
desirable to carry out the solution process in double precision as
suggested by Shugar (21). This consideration must, however, be
weighed with program size and available computer storage. Because
of the large program size and limited computer storage of the CDC
3300 computer at Oregon State University, the solution process can
not be carried out in double precision at the present time (1973). In
order to keep an error in ut+At smaller than an error in ut , the
volume change of soil elements calculated from Equation (4. 14) must
be accurate enough to make the error in ut+At resulted from solving
Equation (4. 9) not larger than the error in ut and this can be done
by using the stability number smaller than the value required by
Equation (4. 19). The value of the stability number required for the
soil-culvert problem can be obtained by using the trial and error
method.
48
It is advisable to begin a solution with smaller time steps,

before advancing the solution to longer time intervals. The length of
time intervals required at the beginning of the solution can be obtained
by performing preliminary analyses with different time intervals as
explained later in Chapter V.
49
V. TESTING OF THE COMPUTER PROGRAM
In order to verify the finite element formulation of this investi-

gation and also the computer program, two different types of prob-
lems are analyzed and their results are compared with those of known
classical solutions.
First, an analysis of a circular ring subjected to a single load
as shown in Figure 5.1 is performed. Four circular beam elements
are used to model a 180 degree portion of the circular ring. The
results as shown in Figures 5.2 and 5.3 indicate the excellent agree-
ment with the exact solutions (6).
Second, a one-dimensional consolidation problem is used to
check the formulation of soil elements and the process of pore pres-
sure dissipation. Five elements are used to simulate a column of
soil as shown in Figure 5.4. Drainage is allowed at the top and no
drainage is permitted at the bottom or sides. The soil is assumed
elastic and stress-independent. The value of initial pore pressure is
constant throughout the soil column.
As mentioned in Section 4.2, the change in volume of soil ele-
ments during an increment of time is obtained by using Equation
(4.14). The use of the relation in Equation (4. 14) will lead to numeri-
cal instability if an excessively large time interval is used. The
length of time interval has to be taken small in relation to the
50
variations in the rate of change of pore pressure. The length of time

steps needed at the beginning of the solution are obtained by perform-
ing preliminary analyses with different time intervals. The variations
of pore pressure with time in element 5 which has the fastest change
of pore pressure are shown on Figure 5.5. They indicate that the
time intervals of 2 days are needed to ensure numerical stability and
accuracy. It was decided to use time steps of 2 days up to the time
of 20 days. The length of time steps used to continue the solution
are determined by using the value of stability number of 0. 4. The
variations of pore pressure with time in the soil elements shown in
Figure 5. 6 indicate a smoothly decaying pore pressure with no sign
of any superimposed oscillation at any stage. The distributions of
pore pressure over the soil column are compared with the well known
Terzaghi's solution (24). They indicate a very good agreement as
shown in Figure 5. 7.
51
4 element model
R = 4.S ft.
E = 30 x 103 ksi
. 4
I = 1. 99 M. /ft,
6 = 50. 5 degree
P=2k
Figure 5.1. Finite element model of circular ring

52
Line of symmetry
-- Exact solution (6) 8 = 50. 5 degree

o Finite element value
Displacement scale
1 inch
Figure 5. 2. Deformed shape for circular ring
Line of symmetry
Exact solution (6)

6 = 50. 5 degree
o Finite element value Moment diagram
plotted on tension
Bending moment scale side
3 k-ft.
Figure 5. 3. Bending moment diagram for circular ring

53
5 element model
C = 0.034 inch/day
= 8.65 Icsf
T, = 0.3
Initial pore pressure, uo = 1000 psf
Ground surface
/
/
/ U
/ H = 50 ft.
5 elements
Drainage
/ @ 10 ft.
direction
ag
/
///
// ////// // ////// ////
Impervious
Figure 5. 4. Finite element model of soil layer

54
{0 2 days
Time increment 6 4 days
RI 6 days
I I I
2 4 6 10 12 14 16 18 20
Time t, days
Figure 5.5. Numerical instability due to large time interval

55
1000
Element 1
Element 2
800
Element 3
600
Element 4
400
200
Element 5
10 20 30 40 50
Time,t, days
Figure 5.6. Plot of pore pressure vs. time, linear consolida-

tion problem
56
Terzaghi's solution (24)
0 Finite element value
UU
0
1. .9 .8 .6 .5 .4 .3 .2 0
y/H .4
Figure 5.7. Pore pressure distribution in soil layer

57
VI. ANALYSIS OF SOIL-CULVERT PROBLEMS
Two analyses of a soil-culvert system are presented in this

chapter in order to compare the behaviors of rigid and flexible cul-
verts under an earth embankment. As previously mentioned, the
developed computer program is applied to the analysis of symmetrical
soil-culvert system in order to use symmetrical properties to reduce
the number of unknowns. The proposed configuration for the system
is illustrated in Figure 6.1. The boundary line in Figure 6. 1 was
selected 18 ft. from the axis of symmetry. '<The selection was based
on the limiting capacity of the computer. The distance of 18 ft. is
justified by observing the deformed shapes in Figures 6. 6 and 6. 7
which indicate the uniform settlement of the soil elements along the
boundary line. This indicates that the selected boundary line is far
enough from the culvert. The circular culvert was used in the study
and the embankment was assumed to be built in six lifts. The rigid
and flexible culverts are made of concrete and steel, respectively.
Because of the assumption that drainage occurs in the vertical
direction, it is convenient to represent the embankment by a series
of vertical columns of soil in order to avoid mathematical difficulties
in obtaining the second derivative of pore pressure used in Equation
(4. 11). A 42-element model was generated for the structure and is
58
The soil proposed for the construction of Bully Creek Dam

which is on Bully Creek about nine miles west of Vale, Oregon was
used for the simulated embankment. Laboratory tests of the soil was
performed at Denver Earth Laboratory, Denver, Colorado (27). The
stress-strain parameters derived from triaxial shear tests and

mechanical properties of the soil are summarized in Tables 1 and 2
and Figures 6.3, 6.4, and 6.5. Because of the lack of test data,
some parameters have to be assumed. The failure ratio Rf used in
Equation (2. 30) can be assumed to be 0.83 (11). Poisson's ratio was
assumed to have values of 0.45 and 0.30 for undrained and drained
conditions, respectively.
The computer results are shown in Figures 6. 6 through 6.18.
One computer run was made in order to show the application of the
developed computer program to the time-dependent problem. The
flexible culvert-embankment system was used as the example of the

time-dependent analysis. The results at the end of construction were
used as an initial condition for the analysis. Because of an abrupt
change of pore pressure in the soil at the sides of the culvertas
shown in Figure 6.15, the rate of change of pore pressure in the soil
elements along the culvert walls is very fast at the beginning of the
solution. The time interval required to start the solution and the
stability number used to continue the solution were found by trial and
error method as explained in Chapter V to have the values of 0.1 day
59
and 0.1, respectively. Ten time steps of 0.1 day are used from the
end of construction up to the time of 1 day. The time intervals
required to continue the solution is determined automatically by the
computer by using the value of stability number of 0.1. Because of
the very high cost of a computer run, the time dependent analysis was
carried out only 40 days after construction. The vertical deflection at
the crown of the flexible culvert was plotted against time in Figure
6.17, The pore pressure in soil elements was plotted against time
in the dimensionless form as shown in Figure 6. 18. An attempt was

made to estimate the approximate value of the crown deflection at the
time that the pore pressure dissipation is completed. If the rate of
change of the deflection is assumed to vary linearly with the rate of
change of pore pressure in soil elements, the deflection-time relation-
ship can approximately be found by using Terzaghi's pore pres sure-
time relationship (24) shown in Figure 6. 19 and the computed deflec-
tions in Figure 6.17. The approximate crown deflection-time relation-
ship was shown in Figure 6.20.
60
Symmetrical lin 18 ft
Construction layer Ground surface Boundary line

A
5. 75 ft.
Concrete:
E = 3 x 103 ksi
I = 512 in. 4/ft.
Wall thickness = 8 inches 5.0 ft.
Steel:
Number 12 gage
6 in. x 2 in. corrugations
E = 30 x 103 ksi. 4. 5 ft.
I = 0. 725 in, 4/ft.
Wall thickness = , 1046 in.
30 ft,
3.75 ft.
5, 5 ft.
4. 5 ft.
/ riiiiii ii// it
Impervious
Figure 6.1. Properties of the simulated soil-culvert system

61
Figure 6. 2. Soil-culvert finite element model

Table 1. Initial conditions for tested soil samples
Optimum Water Degree of Coefficient
Density water content content saturation Void Specific of permeability
lb /cu. ft. % % % ratio gravity in. /day
110 26.6 29.4 86.3 .882 2.59 .0023
Table 2. Drained stress-strain parameters

-E
psf. degree
750 29.25 276 1.07 . 83 . 30

63
2000
ti
c.r.7
"ci 1000
44
a)
oo
S
4.
1.
0 100 200 300 400

Confining pressure Cr psi.
3'
Figure 6. 3. Initial tangent modulus - confining pressure
for undrained condition
60
74.4
b~ 40
ou
S
20
0 100 200 300 400

Confining pressure CT psi.
Figure 6. 4. Compressive strength - confining pressure

for undrained condition
Soil compacted to 95%
standard Proctor maximum density
.0 10.0 100.0
(log) Major principal effective stress psi
Figure 6.5. Void ratio - effective stress curve of Bully Creek soil
65
Displacement scale
10.5 inch
Axis of symmetry
Deformed shape
Figure 6. 6. Deformed shape at the end of construction,

flexible culvert
66
Displacement scale
10. 5 inch
Axis of symmetry
Deformed shape
Figure 6.7. Deformed shape at the end of construction,

rigid culvert
67
2. 5
2.0
- Insufficient data
15 Rigid culvert
0 Flexible culvert
Height of fill, ft.
-0. 5
Figure 6. 8. Plot of vertical deflection at the crown

of culverts vs. height of fill
68
-_-_ Insufficient data
Rigid culvert
Flexible culvert
Height of fill, ft.
Figure 6. 9. Plot of vertical pressure at the crown of

culverts vs. height of fill
69
1000 psf. 0
0
N
Figure 6.10. Radial pressure on flexible culvert at the

end of construction
70
1000 psf.
O.
0
O
Figure 6. 11. Radial pressure on rigid culvert at the

end of construction
71
Note: Positive moment causes tension inside culvert
Figure 6. 1 2. Plot of bending moment during construction

period, flexible culvert
72
O 1
o2
8000 A 3
Lift
® 4
7000 O5
43 6
6000
5000
3000
2000
1000
20 180
-1000 e, degree
-2000
-3000
-5000
Note: Positive moment causes tension inside culvert
Figure 6. 13. Plot of bending moment during construction

period, rigid culvert
73
8000
6000
Rigid culvert
0 Flexible culvert
4) 4000
0
00
1.4
2000
Height of fill, ft.
-2000
Figure 6.14. Plot of maximum bending moment vs.

height of fill
74
Axis of symmetry
200
500
700
850
950
1000
1050
CC)
1000 1100
2
Note: pore pressure is in lb, /ft.
Figure 6.13. Pore pressure contours at the end of

construction, flexible culvert
75
Axis of symmetry
200
500
700
850
950
1000
1050
1100
Note: pore pressure is in lb. /ft. 2
Figure 6.16. Pore pressure contours at the end

of construction, rigid culvert
76
Stability number, = 0, 1
a
10 3, 00
tv
0
ti
td
4) 2, 75
a)
-17)
U
4-4
0.)
10 20 30
Time after construction, days
Figure 6.17. Plot of vertical deflection at crown of flexible

culvert vs. time after construction
77
1.00
Stability number, = O. 1
0. 95
0,90
0. 85
0 10 20 30 40
Figure 6. 18. Plot of the pore pressure vs. time after

construction
78
Terzaghi's solution (24)
------ Computed result
100 200 300
Figure 6. 19. Plot of Terzaghi's pore pressure-time

relationship
O Approximate solution
6 Computed result
4, 0
3.0
0
U
qt
C
0
U
.2.0
cd
t'
1.0
0 t l I. I I I 1,
1.0 10.0 100. 0 1000.0
(log) Time after construction, days
Figure 6. 20. Plot of the approximate vertical deflection at crown of flexible culvert
vs. time after construction
80
VII. DISCUSSION OF RESULTS
Figure 6. 6 shows the deformed shape of the embankment-

flexible culvert system at the end of construction. The embankment
settles almost uniformly. The deflection at the crown is 0.3 inch.
The deformed shape of the embankment-rigid culvert system is shown
in Figure 6.7. The soil prism which is adjacent to the culvert settles
more than the soil prism that is directly over the culvert. This illus-
trates the arching effects explained by Spangler (22). The deflection
of the rigid culvert is very small compared with that of the flexible
culvert.
Figure 6.8 shows the vertical deflection at the crown of the cul-
verts during construction period. At the beginning of the filling
process the top of the culverts deflect vertically upward due to the
effects of lateral soil pressure acting on the culvert walls. After the
fill height is increased above the crown, the vertical deflection varies
linearly with the height of fill over the culvert. The vertical deflec-
tion of the flexible culvert is about ten times larger than the one of the
rigid culvert.
The vertical pressure at the top of the culvert is plotted in
Figure 6.9. It indicates that the vertical pressure varies linearly
with the height of fill above the culverts. Linear pressure-height
function was confirmed by field measurements (4). The vertical
pressure acting at the top of the flexible culvert is approximately equal
81
to the overburden pressure. Due to the arching effect that is caused

by unequal settlements of the central soil prism in relation to that of
the exterior adjacent soil prisms, the vertical pressure at the crown
of the rigid culvert is larger than the overburden pressure. In the
study case the vertical pressure is about 1.13 of the overburden
pressure.
The results from the study and the field measurements obtained
by Trollope (26) are compared with the ones obtained by using
Spangler's formulae and compression ring theory as shown in Table
3. The vertical pressure acting at the top of the rigid culverts ob-
tained by Trollope and the present study are about 0.65 of the values
obtained from Spangler's formulae. The calculated vertical pressure
at the top of the flexible culvert is in good agreement with the values
obtained from Spangler's formula and compression ring theory. The
deflection at the crown of the flexible culvert is ten times smaller
than the value obtained from Spangler's formula. The difference in
the deflection may be explained by the use of the approximate value
of modulus of soil reaction and the assumed pressure distribution on
culverts proposed by Spangler. The maximum axial force in the
flexible culvert is about 15 percent higher than the value obtained
from compression ring theory.
Figures 6. 10 and 6.11 show the distributions of radial pressure
on the culverts at the end of construction. The radial pressure
Table 3. Comparisons of results at the end of construction
CT Deflection at Maximum axial
Sources Culvert y at crown crown force
(cry)overburden inch. lb. /ft.
Tro llope (26) Rigid 1.2 - -
Spangler Rigid 1.8 - -
Lertlaks ana Rigid 1.13 - -
Spangler Rigid 1.75 - -

Lertiaksana Flexible 0.995 0.3 11970
Spangler Flexible 1.0 3.24 -

Ring theory Flexible 1.0 - 10400
* Field measurement
Use modulus of soil reaction of 700 psi suggested by Spangler.
83
distributes uniformly on the flexible culvert and the magnitude is

equal to the overburden pressure at the top of the culvert as proposed
by White (29). The radial pressure is maximum at the crown in the
case of rigid culvert and distributes almost uniformly on the culvert
walls with the magnitude of 0.77 of the value at the crown. The shapes
of the pressure distributions are confirmed by the results obtained by
Brown (5).
Figures 6.12, 6.13, and 6. 14 show the variations of bending
moment in the culverts during the construction period. Actually, the
bending moment diagrams should be smooth curves instead of con-
nected straight lines. The maximum bending moment in the flexible
culvert is about 25 percent of that in the rigid culvert. The bending,
moment obtained by using five culvert elements indicates the very
high and unrealistic value especially in the flexible culvert. In order
to obtain a more realistic representation of soil pressure on the cul-
vert, it is necessary to subdivide the culvert into a larger number of
elements. By increasing the number of elements, the moment dia-
gram will be smoothed and the maximum moment in the culvert should
be reduced. The translational displacements at the current nodal
points are not expected to change significantly due to the increased
number of elements. The deflected shape between current nodes is
expected tO change considerably. The slope at the current nodal
points may be changed significantly. Since the soil pressure is a
function of the translational displacements at the nodes, no significant
change in the soil pressure is expected. It should also be noticed
84
that the computed soil pressures on the flexible culvert are well con-
firmed by the field measurements (29).
Figures 6.15 and 6.16 show curves of computed pore pressures
at the end of construction. They indicate that there is almost no
variation of pore pressure in the horizontal direction except at the
area near the culvert walls. Because of the variations of pore pres-
sure, the assumption of drainage in the vertical direction can be
used without causing any significant errors.
Figures 6.17 and 6.18 show the vertical deflection at the top of
the flexible culvert and the dissipation of pore pressure in the embank-
ment, respectively, from the end of construction to 40 days after
construction. They show that the deflection increases as the results
of pore pressure dissipation.
Figure 6.19 shows the pore pressure-time relationship based
on Terzaghi's linear, one-dimensional consolidation solution. The
computed rate of change of pore pressure is slightly slower than the
one obtained from Terzaghi's solution.
The approximate crown deflection-time relationship of the
flexible culvert is shown in Figure 6.20. The approximate crown
deflection based on Terzaghi's pore pressure-time relationship is
slightly larger than the computed values at least 40 days after con-
struction. The estimations of crown deflection are summarized in
Table 4.
85
Table 4. Deflections at crown of the flexible culvert

40 days after construction End of pore
End of ft. pressure dissi-
construction, ft. Computed Predicted pation, ft.
2.50 x 10 -2 2.66x 10-2 2.75 x 10 - 2 3.86x 10-2*

Rate of change of deflection is assumed to vary linearly with rate of
change of pore pressure
The value of the crown deflection at the end of pore pressure

dissipation is approximately equal to 1.55 of the value at the end of
construction. Spangler (22) proposed a concept of deflection lag
factor to take soil time-dependent properties into account by increas-
ing the crown deflection at the end of construction by 25% - 50%.
The computed culvert forces exhibit practically no change at
40 days after construction. As mentioned earlier, the time dependent
properties of soil considered in this study is assumed to be controlled
by the dissipation of pore pressure. The skeleton creep of soil after
the consolidation process is neglected. This assumption is justified
for most cohesive soils used for embankment construction especially
when the time required to complete dissipation of pore pressure is
large (28).
The results obtained in the study are based on an assumption
that the axial deformation in the culverts is neglected by omitting the
terms corresponding to the axial deformation in Equation (2. 21). The
assumption of neglecting the axial deformation may be questioned
in the case of flexible culvert which has small cross sectional area A.
86
The results at the end of construction indicate that the computed

pressure acting on the flexible culvert is in good agreement with the
values obtained from Spangler's solution and compression ring theory.
To neglect the axial deformation in the flexible culvert may not cause
any significant errors. It is recommended to include the effects of
the axial deformation in the culverts for the future analysis.
87
VIII. SUMMARY AND CONCLUSIONS
A computer program was developed using the finite element

method for analyzing culverts under earth embankments constructed
of compacted cohesive soil. The step-by-step construction procedure
is taken into account. The construction pore pressure in the embank-
ment is estimated by the Hilf method, and zero dissipation of pore
pressure during the embankment construction is assumed. The dissi-
pation of pore pressure is assumed to start at the end of embankment
construction. The changes in the culvert forces and deformations due
to the effects of consolidation of the embankment are estimated by the
modified finite element method. The nonlinear, stress-dependent
properties of soil are included in the study by considering the non-
linear problem as piecewise linear during each construction step or
time increment.
From the study it may be concluded:
1. The pressure acting at the crowns of culverts varies linearly
with the height of fill above crowns.
2. The crown deflection varies linearly with the height of fill
above crown.
3. The vertical pressure at the crown of the rigid culvert was
found to be overestimated by Spangler's solution.
4. An equation proposed by Spangler for calculating the crown
88
deflection of the flexible culvert at the end of construction was found

to be very conservative.
5. The range of values of deflection lag factor from 1. Z5 to
1.50 suggested by Spangler is found to be reasonable.
6. No conclusions can be reached about the changes of culvert
forces due to the effects of embankment consolidation from the com-
puted results obtained at an early state of consolidation.
89
BIBLIOGRAPHY
1. Barden, L. Consolidation of compacted and unsaturated clays.

Geotechnique 15: 267-286. 1965.
2. Barden, L. and G. R. Sides. Engineering behavior and struc-

ture of compacted clay. Journal of the Soil Mechanics and
Foundation Engineering Division, American Society of Civil
Engineers 96:1171-1120. July, 1970.
3. Bishop, A. W. and L. Bjerrum. The relevance of the triaxial
test to the solution of stability problems. In: Proceedings of
the American Society of Civil Engineers on Research Confer-
ence on Shear Strength of Cohesive Soils, Boulder, Colorado,
1960, p. 437-501.
4. Brown, C. B. Forces on rigid culverts under high fills. Jour-
nal of the Structural Division, American Society of Civil
Engineers 93:195-215. Oct., 1967.
5. Brown, C. B., D. R. Green and S. Pawsey. Flexible culverts
under high fills. 'Journal of the Structural Division, American
Society of Civil Engineers 94:905-917. April, 1968.
6. Carpenter, S. T. Structural mechanics. New York, Wiley,
1960. 538 p.
7. Christian, J. T. Undrained stress distribution by numerical

method. Journal of the Soil Mechanics and Foundation Engineer-
ing Division, American Society of Civil Engineers 94:1333-
1345. Nov., 1968.
8. Christian, J. T. and J. W. Boehmer. Plane strain consolidation

by finite elements. Journal of the Soil Mechanics and Founda-
tion Engineering Division, American Society of Civil Engineers
96:1435-1457. July, 1970.
9. Clough, R. W. The finite element method in plane stress analy-
sis. In: Proceedings of the Second Conference on Electronic
Computation, American Society of Civil Engineers, Pittsburgh,
1960, p. 345-378.
90
10. Duncan, J. M. and C-Y. Chang. Nonlinear analysis of stress

and strain in soils. Journal of the Soil Mechanics and Founda-
tion Engineering Division, American Society of Civil Engineers
96:1629-1653. Sept., 1970.
11. Duncan, J. M. and C-Y. Chang. Analysis of soil movement
around a deep excavation. Journal of the Soil Mechanics and
Foundation Engineering Division, American Society of Civil
Engineers 96:1655-1681. Sept., 1970.
12. Hi lf, J. W. Estimating pore pressure in rolled earth dams.
In: Proceedings of the Second International Conference on
Soil Mechanics and Foundation Engineering, Rotterdam, 1948,
vol. 3, p. 234-245.
13. Janbu, N. Soil compressibility as determined by oedometer and

triaxial tests. In: European Conference on Soil Mechanics and
Foundation Engineering, Wiesbaden, 1963, vol. 1, p. 19-25.
14. Lambe, T. W. and R. V. Whitman. Soil mechanics. New York,
Wiley, 1969. 545 p.
15. Laursen, H. I. Structural analysis. New York, McGraw-Hill,
1969. 486 p.
16. Lee, K. L. and S. C. Haley. Strength of compacted clay at high

pressure. Journal of the Soil Mechanics and Foundation Engi-
neering Division, American Society of Civil Engineers 94: 1303-
1332. Nov., 1968.
17. Przemieniecki, J. S. Theory of matrix structural analysis.

New York, McGraw-Hill, 1968. 468 p.
X'18. Rubinstein, M. F. Matrix computer analysis of structures.
Englewood Cliffs, Prentice-Hall, 1966. 402 p.
19. Salvadori, M. G. and M. L. Baron. Numerical methods in
engineering. 2d ed. Englewood Cliffs, Prentice-Hall, 1961.
302 p.
20. Scott, R. F. Principles of soil mechanics. Reading, Addison-

Wesley, 1963. 550 p.
21. Shugar, T. A. Structural analysis of symmetrical penstock
bifurcations by the finite element method. Doctoral dissertation.
Corvallis, Oregon State University, 1971. 170 numb. leaves.
91
22. Spangler, M. G. Soil engineering. 2d ed. Pennsylvania,

International Textbook Company, 1960. 483 p.
23. Steward, J. E. Numerical method for calculating construction
pore pressure in embankments. Master's dissertation.
Corvallis, Oregon State University, 1971. 91 numb. leaves.
24. Terzaghi, K. and R. B. Peck. Soil mechanics in engineering
practice. Zd ed. New York, Wiley, 1968. 729 p.
25. Tezcan, S. S. Discussion of: Simplified formulation of stiff-
ness matrices, by P. M. Wright. Journal of the Structural
Division, American Society of Civil Engineers 89:445-449.
Dec., 1963.
26. Trollope, D. H. , M. G. Speedie and I. K. Lee. Pressure
measurements on Tullaroop Dam Culvert. In: Proceedings
of the Fourth Australia-New Zealand Conference on Soil
Mechanics and Foundation, Australia, 1963, p. 81-92.
27. U. S. Bureau of Reclamation. Division of Engineering Labora-
tories. 1962. Laboratory studies for final design on embank-
ment materials proposed for Bully Creek dam, Vale project,
Oregon. Washington, D. C., 13 p. (Soil Engineering Report
no. EM -646)
28. Walker, L. K. Secondary compression in the shear of clays.
Journal of the Soil Mechanics and Foundation Engineering
Division, American Society of Civil Engineers 95:167-187.
Jan., 1969.
29. White, H. L. and J. P. Layer. The corrugated metal conduit
as a compression ring. Highway Research Board Proceedings
39:389-397. 1960.
30. Wilson, E. L. Structural analysis of axisymmetric solids.

American Institute of Aeronautics and Astronautics Journal
3:2269-2274. Dec., 1965.
31. Woodward, R. J. Analysis of stresses and displacements in
embankments with non-linear material. Doctoral dissertation.
Berkeley, University of California, 1966. 113 numb. leaves.
32. Zienkiewicz, 0. C. and Y. K. Cheung. The finite element method
in structural and continuum mechanics. London, McGraw-Hill,
1967. 272 p.
APPENDICES
92
APPENDIX A
USER'S MANUAL FOR SYMMETRIC
SOIL-CULVERT PROGRAM
The program consists of a main program and 19 subroutines.

This appendix is included in order to explain to users the capability
of the computer program and restrictions concerning the preparations
of the finite element model and input data.
The FORTRAN IV program was written for the CDC 3300 com-
puter at the computer center of Oregon State University.
A. 1 Program Capability
The computer program is used to analyze a culvert under an

earth embankment of homogeneous, cohesive soil. The culvert and
embankment are placed on a rock foundation. Because of the com-
puter capacity, the application of the computer program is limited to
a symmetrical soil-culvert problem, The culvert can be circular or
elliptical. The program does compute culvert forces, deformations,
stresses and pore pressures in the soil at any stage of construction
due to the gravitational effects of the fill, and at any time after con-
struction due to the effects of pore pressure dissipation. The pore
pressure dissipation was assumed to begin at the end of construction.
By using piecewise linear analysis, non-linear, stress-dependent
stress-strain behavior of the soil can be accommodated.
93
The program can also analyze an arch subjected to external

forces, one-dimensional consolidation with non-linear, stress-.
dependent stress-strain behavior and one-dimensional consolidation
with linear stress-strain behavior.
Present limitation on the finite element model's size are as
follows: 52 nodal points, 5 culvert elements, 37 soil elements, and
20 constrained nodal points.
A. 2 Preparation of the Finite Element Mesh
By taking advantage of the symmetry of the soil-culvert system,

only one-half of the structure is used. The structure shown in Figure
A. 1 is used as an example to show how to prepare the finite element
model. The embankment is represented by a series of soil columns
as shown in Figure A. 1 in order to avoid mathematical difficulties in
obtaining the second derivative of pore pressure with respect to the
vertical distance used in Equation (4.11). Three kinds of elements
are used to construct the model; curved beams for the culvert, and
quadrilateral and triangular elements for the embankment. The
majority will be quadrilateral elements as they are the most efficient;
that is, the quadrilateral element is subdivided by the program into
four triangles. Therefore, a quadrilateral element will give the
same results as four triangular elements without increasing the
storage locations in the computer.
94
y
Line of symmetry
U6 4'1 Vt 45 5 1 (i.t)"
St'
41. e, t"
I'
... 3-3
7,,' 4 ',
j "?., '
ICi
6 L. 1
15 16 17 18
\-
11 12 13 14
f>i 7 / 8 70-9 10
1 x
Figure A. 1. Example finite element mesh showing node

number scheme
95
Numbering the nodal points should begin after the finite element
mesh has been established. The numbering system in this computer
program was designed in such a manner that numbering of these
nodes must start along the culvert and then proceed to the fill in a
layer by layer manner as shown in Figure A.1. The numbering system
was found to be convenient for the step-by-step analysis during con-
struction even if the obtained structural matrix may have larger band
width than that using other systems.
Once the numbering of the nodes has been accomplished, all
elements must be numbered. The numbering must start along the
culvert and then proceed to the fill in a layer by layer manner. The
culvert elements have to be numbered in the clockwise direction.
A. 3 Preparation of Input Data
The following information describes the data cards which form

the necessary input data for the program.
A. 3.1 Problem Type Card (15) LWiP)

Columns 1-5 Place a 1 for soil-culvert problem, both con-
struction and consolidation periods are needed.
Place a 2 for soil-culvert problem, only con-
struction period is needed.
Place a 3 for an arch subjected to external
forces.
96
Place a 4 for soil-culvert problem, only con-.

solidation period is needed.
Place a 5 for one-dimensional consolidation
with non-linear, stress-dependent stress-
strain behavior.
Place a 6 for one-dimensional consolidation
with linear stress-strain behavior.
A. 3. 2 Control Card (515)
Columns 1-5 Total number of elements (42 max. )

6-10 Total number of nodal points (52 max. )
11-15 Total number of constrained boundary points
(20 max. )
16-20 Total number of culvert elements (5 max. )
21-25 Total number of layers
A. 3. 3 Coordinate Cards (15, 3F15. 3)
This data set describes primarily the overall geometry of the

structure in terms of two coordinates x and y for each nodal point.
In addition the number of unknowns to be assigned the nodes are
described in this data set.
Columns 1-5 Nodal point number
6-20 The x coordinate of the node (feet)
21-35 The y coordinate of the node (feet)
97
36-50 This entry indicates the degree of freedom the

program will assign the node. A. 2 is assigned
all nodes except the nodes on the curved beam
elements which are assigned a 3.
The coordinate cards must be in order, beginning with nodal
point number 1.
A. 3. 4 Connectivity Cards (15, 5X) 615)
This data connects the element number with the nodal point
numbers on the element's perimeter. The nodal points for a quadri-
later/al and triangular element are listed counterclockwise sequentially
around the element. The nodal points for a curved beam element are
listed clockwise. Besides the above information, the element num-
bers of the elements which are directly above and below a given
element are indicated in the data set. The two element numbers are
used for calculating the distances between the center of the element
under consideration and the centers of the elements above and below.
These distances are used for calculating the second derivative of pore
pressure used in Equation (4. 11)
Columns 1-5 Element number
11-15 Nodal point i
16- 20 Nodal point j
21-25 Nodal point k. Columns are left blank if it is
98
a curved beam element.
26.-30 Nodal point 1, Columns are left blank if it is
a curved beam element or a triangular element.
31-35 Element number of the element that is below
a given element. Columns are left blank if
the given element is a curved beam element,
or it is on the foundation or the culvert.
36-40 Element number of the element that is above
the given element. Columns are left blank if
the given element is a curved beam element
or it is a soil element at the top layer.
The element cards must be in order, starting with element
number 1.
A. 3.5 The Constant Cards (415)
This data applied primarily to nodal points on the boundaries

of the finite element model. A 1 is used to indicate the constraint of
a nodal point.
Columns 1-5 Nodal point number of constrained nodal point. ON"
6-10 x direction constraint (1 or blank) 16ANA (1-1)
11-15 y direction constraint (1 or blank)

16-20 Rotational constraint (1 or blank, enter 1 if 14Nt,,,NU,44,-)
the nodal point is not on the curved beam

elements).
99
A 3. 6 Culvert Properties Card (15, 2F10. 2, 2F10. 3)
Columns 1-5 Enter a 1 if the culvert is circular.

Enter a -1 if the culvert is elliptical.
6-15 Culvert radius in feet or blank if the culvert
is elliptical. The radius of the circular beam
segments is calculated by using the properties
of a circle passing through three consecutive
points in the case of elliptical culvert as
16-25 Modulus of elasticity of the culvert (ksi)
26-35 Cross sectional area (in.2 /ft. )
4/ft.
36-45 Moment of inertia (in. )
A. 3.7 Soil Properties Card (2F10. 2, F10. 4)
Columns 1-10 Total stress coefficient of earth pressure

used for initial stresses in elements newly
added to embankment
11- 20 Unit weight of soil (lb. /cu. ft. )
21 -30 Coefficient of permeability (in. /day)
A. 3. 8 Undrained Stress-Strain Parameters Card (4F12. 2, 2F5. 2)
Columns 1-12 Initial tangent modulus at zero confining

pressure (psf.)
100
13-24 Initial tangent modulus at saturation (psf. )

25-36 Maximum deviator stress at zero confining
pressure (psf.)
37-48 Maximum deviator stress at saturation (psf. )
49-53 Undrained failure ratio
54-58 Undrained Poisson's ratio
A. 3. 9 Drained Stress-Strain Parameters Card (6F10. 2)
Columns 1-10 Drained cohesion (psf. )

11-20 Drained angle of internal friction radian)
21-30 Modulus number, K
31-40 Exponent, N
41-50 Drained failure ratio
51-60 Drained Poisson's ratio
A. 3.10 Soil Pore Pressure Properties Card (8F6. 3, 3F10. 2)
Columns 1-6 Initial void ratio

7-12 Initial air void ratio
13-18 Initial water void ratio
19-24 Void ratio corresponding to effective
preconsolidation stress
25-30 Void ratio at full saturation
31-36 Initial degree of saturation (decimal)
101
37-42 Compression index in the over-consolidation

region
43-48 Compression index in the normally consoli-
dation region
49-58 Effective stress corresponding to initial void
ratio (psf.)
59-68 Effective preconsolidation stress (psf. )
69-78 Effective stress corresponding to void ratio
at saturation (psf. )
A. 3.11 Linear Stress-Strain Parameters Card (2F12.2)
This data card gives an information about stress-strain proper-

ties of soil for a one-dimensional, consolidation problem which has
linear, stress-independent stress-strain behavior.
Columns 1-12 Undrained elastic modulus (psf.)
13-24 Drained elastic modulus (psf.)
This card is left blank if the soil is inelastic and stress-
dependent.
A. 3.12 Output Control Card (315)
Columns 1-5 Enter a 1 if tangent modulus is needed to be

printed out, otherwise, the columns are left
blank
102
6-10 Enter a 1 if the volume of soil elements

during the consolidation process is needed
to be printed out, otherwise, the columns are
left blank
11-15 Enter a 1 if the output is needed to be punched
on cards, otherwise, the columns are left
blank
A.3.13 Layer Data Cards (415)
The data cards are not needed if the given problem is a soil-
culvert problem that needs only consolidation period or it is a one-
dimensional consolidation problem.
Columns 1-5 Layer number
6-10 Number of elements up through this layer
11-15 Number of nodal points up through this layer
16-20 Element number of the first soil element
in this layer
The layer data cards must be in order, starting with layer num-
,ber 1.
A. 3.14 Arch Problem Card (15)
The card is omitted if it is not an arch subjected to external

forces problem.
Columns 1-5 Total numbers of applied forces IN,SG,
103
A. 3. 15 External Forces Cards (15, F12. 2)
The cards are omitted if it is not an arch subjected to external

forces problem.
Columns 1-5 Displacement number at the loading point cor-
responding to the direction of the applied force
6-17 External force (lb. or lb.-ft. )
A. 3.16 Initial Displacement Cards (15, 3E15. 5)
The cards are needed only if the problem is a soil-culvert prob-

lem that needs only consolidation period or if it is a one-dimensional
consolidation problem.
Columns 1-5 Nodal point number
6-20 Initial displacement in x direction (ft. )
21 -35 Initial displacement in y direction (ft. )
36-50 Initial rotation (radian). The columns are
left blank if the nodal point is not on the culvert
A. 3. 17 Initial Stress Cards (15, 3E15. 5, /, 4E15. 5)

lem that needs only consolidation process or it is a one-dimensional
consolidation problem.
The first card (15, 3E15. 5):
104
6-20 Axial force at node i (lb.) if the given ele-

ment is a curved beam element. Effective
normal stress, cr x (psf. ), if the given element
is a soil element
21-35 Shear force at node i (lb.) if the given ele-
normal stress, vy (psf. ), if the given element
is a soil element.
36-50 Bending moment at node i (lb. -ft.) if the
given element is a curved beam element.
Shear force, 'Txy (psf.), if the given element
is a soil element.
The second card (4E15.5):
Columns 1-15 Axial force at node j (lb.) if the given ele-
major principal stress, 71 (psf.), if the given
element is a soil element.
16-30 Shear force at node j (lb.) if the given ele-
minor principal stress, 7 3 (psf.), if the given
element is a soil element.
31-45 Bending moment at node j (lb. -ft.) if the
given element is a curved beam element.
105
Maximum shear force, T

max. (psf. ), if the
given element is a soil element.
46-60 Excess pore pressure (psf.). The columns
are left blank if the element is a curved beam
element.
A. 3.18 Volume Change Cards (15, E15.5)

lem that needs only consolidation period or it is a one-dimensional
consolidation problem. These data cards together with initial dis-
placement cards and initial stress cards will provide the computer
program the capability to begin the consolidation process at any time
other than at the end of construction.
6-20 Total volumn change (cu. ft.) due to consoli-
dation effects from the end of construction.
The columns are left blank if the solution
starts at the end of construction or the ele-
ment is a curved beam element.
A. 3. 19 Number of Time Increment Card (215)
The card is needed only if the consolidation process is wanted.

Columns 1-5 Total number of time increments
106
6-10 Number of time increments needed to start

the solution (see Section 4. 2)
A. 3. 20 Stability Number Card (F6. 2)
The card is needed only if the consolidation process is wanted.

The stability number is the value of M/a used in Equation (4. 19).
Columns 1-6 Stability number
A. 3. 21 Time Increment Cards (F10. 3)
The cards are needed only if the consolidation process is

wanted.
Columns 1-10 Length of time increment (days)

The number of time increment cards are equal to the number
of time steps used to start the consolidation process (see card A. 3.19)
107
APPENDIX B
DESCRIPTION OF COMPUTER PROGRAM
Nineteen subroutines comprise the body of the computer pro-

gram. They are controlled by call statements which comprise the
main program. A brief description of the function of each subroutine
is given below.
1) Subroutine INPUT. This subroutine reads into storage the
input data.
2) Subroutine SUBCOD. This subroutine generates the 4 x 6
matrix of code numbers used subsequently to combine four triangular
elements into one quadrilateral element.
3) Subroutine GENCOD. This subroutine generates a code
number for each element in the finite element model. These are
subsequently used to assemble the system matrix of algebraic equa-
tions.
4) Subroutine INIT. This subroutine initializes a given prob-
lem to the initial conditions.
5) Subroutine D2Y. This subroutine calculates the vertical
distances from the center of a given soil element to the centers of
soil elements that are just above and below it. These distances are
used to calculate the second derivative of pore pressure used in
Equation (4.11).
108
6) Subroutine MODULUS. This subroutine computes tangent

modulus of each soil element and the length of time interval used in
consolidation process.
7) Subroutine CULPRO. This subroutine computes center point,
radius and central angle of each culvert element.
8) Subroutine CULK. This subroutine generates 6 x 6 stiffness
matrix of each culvert element referenced to the local coordinate
fVuSTINV
system shown in Figure 2.4.
9) Subroutine CULTRN. This subroutine generates 6 x 6 dis-
placement transformation matrix of each culvert element. The
matrix is used to transform the culvert stiffness matrix referenced
to a local coordinate system to the one referenced to a global coordi-
nate system.
10) Subroutine TRIK. This subroutine generates 6 x 6 stiff-
ness matrix of each triangular element.
11) Subroutine QUAD. This subroutine computes the center
point of each quadrilateral element and then uses code numbers to
assemble four triangular elements into a quadrilateral element. The
10 x 10 quadrilateral stiffness matrix and associated load vector are
obtained.
12) Subroutine STIFF. This subroutine performs the static

condensation process which reduces the 10 x 10 quadrilateral stiff-
ness matrix to a 8 x 8 element matrix. The subroutine also transform
CALL di, u,41) G )
109
a culvert stiffness matrix to the one referenced to the global coordi-

cP.L.L. coLK(I.), CALL cuL:(12N C1) caLl TRItc
nate system.
13) Subroutine BGK. This subroutine uses the code numbers
to assemble the element matrices into a system matrix.
14) Subroutine FVEC. This subroutine uses the code numbers
to assemble the system load vector. iY
15) Subroutine BANWID. This subroutine computes the band

width of the system stiffness matrix in the case of incremental con-
struction analysis prior to the solution process.
16) Subroutine SOLVE. This is a subroutine used to solve the
system of simultaneous equations by Gaussian elimination.
17) Subroutine MATINV. This is a standard matrix inversion
subroutine. It is used only to invert a flexibility matrix during a
generation of culvert stiffness matrices.
18) Subroutine PORE. This subroutine computes pore pres-
sures in each soil element during construction.
19) Subroutine RESULT. This subroutine computes displace-
ments, stresses, forces and pore pressures for the elements in the
finite element model. It also computes the principal stresses.
CALL. COL
( ivt_L C UL111.N l \
(1\0, OA() LI)

110
APPENDIX C
PROGRAM LISTING
C ..*..****+*++**.*****ts,, IF(K .GT. J/J=K
PROGRAM CULVERT,,," 1 CONTINUE
IFI, .E0. 0)00 TO 200
COMMON /A1/ NUMRPT.NUMELT,NCULET.NCONP,NLAY ND=J
COMMON /A2/ COOR(53,31.1CONN(42,6) GO TO 201
COMMON /A3/ KNOVA(20,4)INODOM452,3) 200 DO 2 1=1.3
COMMON /A4/ LADOM K=NODOM(NUMNP-1,I)
COMMON /A5/ ICODE14.6) IF(K .07. J)J =K
COMMON /A6/ NCODE(42,8) 2 CONTINUE
COMMON /A7/ NL,NCASE.NUMEL,NUMNPOSEL1 ND=J
COMMON /91/ KTYPE,RADIUS,VE,VItCAREA 201 WRITE(61,105(0
COMMON /92/ ACUL( 5),BCUL(5),RCUL15),BETA(5),TLAN(6,6) 105 FORMA-H/0 THE TOTAL NUMBER OF DISPLACEMENT',
COMMON /C1/ ET(42),PR 1' UNKNOWNS = '016)
COMMON /C2/ PR1,PR2tEIC,EIU,DVSTFO,DVSTFU.RFI,RF2.
1 CEE.PHI.VMODNO,VNoVELINI,VELIN2 C** INITIALIZE THE STIFFNESS AND FORCE ARRAYS
COMMON /C3/ VK,DENS C** STIFFNESS MATRIX IS STORED IN SINGLE ARRAY
COMMON /C4/ CAYO NDT=ND*(ND+1)/2
COMMON /C5/ EVOsEAO,EWO,EVN,ESM,SIGO,SIGN,S10:, DC 3 I=1*NDT
1 CCOPCCN,SC,5V(42)
COMMON /01/ AREA(42) 3 i'')014=7:1.',ND
COMMON /02/ 516,6),SYSK110.1010(10) 4 B(I)=0.0
COMMON /03/ A(71401 IF(LPROB .NE. 3)00 TO 610
COMMON /D4/ B(119( -T READ160.7011NO
COMMON /05/ ND 1 701 FORMAT(I5)
COMMON /D6/ NB WRITE(61,905)
COMMON /07/ ALPHA142),HD131421 905 FORMAT(/,' TOTAL NUMBER OF APPLIED FORCES = ',IS)
COMMON /E1/ DISP152.31,STRESS(42,6)090REPR(42) WRITE(61,906)
COMMON /F1/ DT 906 FORMAT( /,5X,' LOAD NUMBER APPLIED FORCE,LB.OR LB.*FT. ')
COMMON /F4/ F)6,6) DO 702 I =1,NQ
COMMON /F5/ DELV(42) READ(6017031J,9(J)
COMMON /01/ KPRIN1sKPRIN2,KPRIN3 702 WRITE(61,907)I,Js9(J)
REA0(60,616)LPROB 703 FORMAT(15oF12.2)
616 FORMAT(15) 907 FORMATII3019,10X,F12.2)
CALL INPUT 610 CALL MODOLUSILPROB.11oNTIMEI,STABNO)
CALL SUBCOD DO 5 I".1.NUMEL
CALL GENCOD CALL STIFF(I)
CALL D2Y CALL BGK(I)
IF(LPROB .LE. 4)CALL CULPRO CALL FVEC(I)
CALL INIT(LPROBI 5 CONTINUE
:F(LPROB .GE. 4)00 TO 7 C** CALCULATE BAND WIDTH
NCASE =1 CALL BANWID
NL =O WRITE161,5000'0
500 FORMAN/0 BAAILXISITH
C** PLACE A NEW LAYER CALL SOLVE
6 NL=NL+1 CALL RESULT
GO TO 6
C** CHECK THE LAST LIFT 621 IFILPROB .NE. 1100 TO 615
IF)NL .07. NLAYIGO TO 621
WRITE(61.101)8IL C** TIME DEPENDENT ANALYSIS
101 FORMAT(/,' THE LAYER NUMBER =',I5) 7 WRITE(61,110)
110 FORMAT) /// /.10X,' TIME DEPENDENT ANALYSIS 'I
C** LIFT DATA NSELT=NUMELTNCULET
O
READI600102)NL,NUMEL,NH MNPSEL1 NCASE=2
102 FORMAT(415) IFILPROB .50. +1)00 TO 616_ ti"0.,1 114C "
WRITE161.908)NL.NUMEL,NUMNP.NSELI DO 900 1=1,NUMNP-1".-----
908 FORMAT(/,' LAYER NUMBER = '05./0 TOTAL NUMBER OF ELEMENTS UP', 900 READ(60,901(DISPII.JI,J=1.3)
1' THROUGH THIS LAYER = ',I5,/,' TOTAL NUMBER uF NODAL POINTS UP', 901 FORMAT(15,3E15.5)
DO 902 I=1,NUMELT- vto,
2' THROUGH THIS LAYER =',151/0 ELEMENT NUMBER OF FIRST SOIL',
3' ELEMENT IN THIS LAYER = ',I5) 902 READ(60,903)g+1STRS5117J),J=1,61.PORFPR(1)
J.0
DO I 1=1,3
K=NOCOM(NLMNP.11
903 FORMAT(15,3E15.54E15:51
DC 915 1.1,NUMELT
915 REA0(60,916)K,DELVII)
e- ,3 '
I /
916 FORMAT(15,E15.5) .,PUNCH 91).(N.DELV(NIoN=1oNVMELT)
GC TO 614 91 FORMAT(15.115.5)
C 914 STOP
C ** TRANSFORM THE TOTAL STRESSES TO EFFECTIVE STRESSES END
615 DO 111 N= 1,NUMELT C ****.************.*******************************4***************.****
STRESSIN.1)=STRESSIN.1)POREPR(N) SUBROUTINE MODULUS(LPROB.INNTIMEI.ST,BNO)
STRESS(N.2)=STRESS(N.2)POREPRIN) COMMON /A1/ NUMNPNNUMELNNCULET.NCONF,NLAY
STRESS(N.4)=STRESS(N.4)POREPR(N) COMMON /A7/ NL.ACASEsNUMEL.NUMNP.NSEL1
111 STRESSIN.5)=STRESSIN.5)POREPRIN) COMMON /61/ KTYPE,RADIUS.VEINNCAREA
IFILPROB .EQ. 2 .0R. LPROB .E0. 3)00 TO 12 COMMCA /Cl/ ET142).PR
614 TIME=0.0 COMMON /C2/ PR1s PR2.EIO.E1U.DVSTFO.DVSTFU.RF1tRF2,
WRITE(61,112)TIM6 1 CEE,PHI.VMODNCoVN.VELIN1,VELIN2
112 FORMAN/0 TIME AFTER CONSTRUCTION = .F10.30 DAYS 1 COMMON /C3/ VKI(DENS
WRITEI61,113) COMMON /C5/ EVOI,EAOsEWO.EVN,ESM.SIGO.SIGN.SIGS.
113 FORMAT( / /,' NODE 03X0 XDISP .8X0 YDISP ',6X, 1 CCO.CCN,SO.EV(42)
1' ROTATION .) COMMON /D7/ ALPHA(421.HDIS(42)
DC 400 I=1.NUMNPT COMMON /El/ DISP(52,3).STRESS(42.6).POREPR(42)
400 WRITE(61.401)I0DISP(IsJ),J=1,3) COMMON /F1/ DT
401 FORMAT(/.15.3E16.5) COMMON /01/ XPRINN,XPRIN2.KPRIN3
WRITE(61.402) IFINCASE .E0.1 2100 TO 1
402 FORMAT(//0 ELE. SIGMAXX ..5)( SIGMAYY .05X. IFILPROB .ED. 3)00 TO 2
1' SIGMAXY ../.9X0 AND ..10X0 AND '1,1030 AND ../. US=105340.0.(1.-50)/S0
27X.' MAX.S1GMA '0430 MIN.SIGMA ..4X.. MAX.SHEAR S1G3ST=SIGS+USDVSTFU
34)(10 PORE PRES. N 2 DO 3 1=1.NUMEL
DO 403 I=1.NUMELT IFII .GT. NCULET)GO TO 4
403 WRITE161,404110STRESS(I,J).J=1.6).POREPRIN ET(I)=VE
404 FORMAT1/.13.E14.5.2E15.5./.3X,E14.5.3E1505) IFIKPRIN1 .EO. 11WRITE(61.5)IvENI)
ND=LADOM+NSELT GO TO 3
NDT=ND.IND+11/2 4 SIG1=STRESSIN41
READ(60,904)NTIME,NTIME1 SIG3=STRESS11.5)
904 FORMATI215) IF(SIG3 .LT. 0.0)00 TO 600
W RITE(61.909)NTIME,NTIMEI IF(SIG3 .GE. SIG3ST)GO TO 601
909 FORMAT)!,' TOTAL NUMBER OF TIME INCRE/ENTS = ..15./. EI=E1O+SIG3*(EIUE101/SIG35T
1' NUMBER OF STARTING TIME INCREMENTS = ..15) DVSTF=DVSTFO+SIG3*(DVSTFUDVSTFO)/S/G3ST
READI60,912)STABNO GO TO 602
912 FORMATIF6.2) 601 EI=EIU
WRITEI61.913)5TABN0 DVSTF=OVSTFU
913 FORMAN/0 STABILITY NUMBER = '.F6.2) 602 RF=RF1
DO 114 11=1oNTIME IFIRF.(SIG1SIG3)DVSTF)603.600.600
CALL MODULUSILPROB.INNTIMENSTABNO) 603 ET(I)*(1.RF*(SIG1SIG2)/DVSTF)**2..f1
IF(II .LE. NTIMENREAD(60.1151IDT)-q , tec 0,444ro., GO TO 604
115 FORMAT(F10.3) 6C0 ET(I)=10.0
TIMB=TIME+DT WRITE(61,800)I
W RITE(61,112)TIME 604 IFIKPRIN1 .E0. 1)WRITE161.6)IsET(I)
DO 8 1=1.NDT 3 CONTINUE
8 A(I)=0.0 PR=PR1
DO 9 I=1.ND GO TO 7
9 B(I)=0.0 1 DTTEMP=100000.0
DO 10 I=1.NUMELT DO 8 I=1.NUMELT
CALL STIFFII) IF(1 .GT. NCULET)GO TO 9
CALL 130K(1) ET( I )=VE
CALL FVECII) IFIKPRIN1 .EO. 1/WRITE(61.5)1,ET(1)
10 CONTINUE GO TO 8
NB=ND 9 IFILPROB .NE. 6100 TO 609
CALL SOLVE ET(I)=VELIN2
CALL RESULT GO TC 10
114 CONTINUE 609 SIGI=STRESS(N41
12 IF(KPRIN3 .EQ. 0)00 TO 914 SIG3=STRESS(1,5)
PUNCH 9100N0DISP(N.MI,M=1131.N=1.NUMNPTI )F(S1G3 .LE. 0.0)00 TO 605
910 FORMAT(15.3E15.5) El=2116.8.VMODNO.(5I03/2116.8)**VN
PUNCH 9110N0STRESS(N.M),M=1.6),POREPRINNN=1.NUMELT) DVSTF=2..(CEE.COSIPH11+SIG3.SINIPHIN/)1.SINIPHIN
911 FORMAT(15.3E15.5./.4E15.5) RF=RF2
IFIRF*(SIG1SIG3)DVSTF1606,635,605 2 FORMAT(/0 TOTAL NUMBER OF NODAL POINTS 'I
606 ET(11.(1.RF*(SIG1SIG3) /DVSTF)**20.E1 WRITE(61950)NUMNPT
CO TO 10 WR1TE(61,3)
605 ET(I)=10.0 3 FORMAT( /s' TOTAL NUMBER OF LAYERS .)
WRITE(619800)1 WRITE(61,50) NLAY
10 IFIKPRIN1 9E09 1)WRITE(6196)19E711) WRITE(61,4)
IFIII .LE. NTIMEI)G0 TO 8 4 FORMAT(/,' TOTAL NUMBER OF CONSTRAINED BOUNDARY',
TEMP=ALPHA(I)*HDIS(I)*HDIS(I)/ET(I) 1' POINTS .1
IFITEMP .LE. DTTEMP/DTTEMP=TEMP WRITE(61950)NCONP
8 CONTINUE WRITE(61,51
PR=PR2 5 FORMAT( /.' TOTAL NUMBER OF CULVERT ELEMENTS ))
CONST=V0.(1.PR)/(1.-2.*PR)/(1.+PR)/748.8 WRITE(61,50)NCULET
DT=STABNO*DTTEMP/CONST 50 FORMAT(/9I5)
C WRITE(6196)
7
RETURN 6 FORMAT( /.' THE X,Y ORDINATES IN (FT.) AND THE',
5
FORMAT(/915,5X,' TANGENT MODULUS = '.F12.3.' KS!. .) 1' DEGREE OF FREEDOMS OF EACH ../.. NODAL POINT 'I
6
FORMAT(/915,5X0 TANGENT MODULUS = .0F)2.39. PSF. 'I DO 305 1=1,NUMNPT
800
FORMAT(/.. ELEMENT.,I50 FAILED') 305 WRITE(61,306)1,(COORII,J)9J=193)
END 306 F0RMAT(/93X,15,3F10631
C .**************.***..**************.*****4,**************************** WRITE(61.8)
SUBROUTINE INPUT 8 FORMAT( /,' NODAL POINT NUMBERS'OF NODES I.J,K,L',
C 1' AND IF IT IS A SOIL ELEMENT, 9/9. THE NUMBERS OF THE',
COMMON /A1/ NUMNPT.NUMELT.NCULET,NCONP,NLAY 2' SOIL ELEMENTS UNDER AND ABOVE IT ARE NEEDED 'I
COMMON /A2/ COOR(53,3).1CONN(42,6) DO 302 I=1,NUMELT
COMMON /A3/ KNOVA(2094),NODOM(529,31 302 WRITE(619303)1,11CONN( 19J),J=1,61
COMMON /81/ KTYPE9RADIUS,VEsVI,CAREA 303 FORMAT(/915,2X,6(5)
COMMON 1C2/ PR19PR2sE10,E1U,DVSTFO+DVSTFU,RF1ioRF2, WRITE(61910)
1 CEE,PHIsVMODNO,VN,VELIN19VELIN2 10 FORMAT(//,' CONSTRAINED NODE POINTS )91,
COMMON /C3/ VKOENS 113X,' NODE..6X-9. X',6X,' Y.96X,. R.)
'COMMON /C4/ CAYO DO 308 I=1,NCONP
COMM-CA /C5/ EVO,EAO.EWO,EVN9ESM,SIGO.SIGN,SIGS, 308 WRITE161.30911,(KNOVA(IsJ),J=194)
1 CCO,CCNtS0,EV142) 309 FORMAT(/915,7)(11418)
COMMON /G1/ KPRIN1,KPRIN2,KPRIN3 WRITE(61914)
READ(609102) NUMELT,NUMNPToNCONP,NCULE79NLAY 14 FORMAT(/,' THE RADIUS OF THE CIRCULAR CULVERT.FT. ')
101 FORMAT(515) WRITE(61,51)RADIUS
DC 102 I=1,NOMNPT 51 FORMAT( /,F12.3)
102 READ(60,103)K,(COOR(1,J)9J=1.31 WR1TE(61,15)
103 FORMAT(1593E15.3) 15 FORMAT( /,' MODULUS OF ELASTICITYsKSI. 'I
DO 104 I=1,NUMELT WRITE(61,51)VE
114 READ(609105)Kt(ICONNII,J)9J=116) WRITE(619161
105 FORMAT(1595X,615) 16 FORMAT) /,' THE MOMENT OF INENTIA,(INCH**4)/FT. .1
DC 106 I=1,NCONP WRITE(61951)VI
106 READ160+107)(KNOVA(IIJ),J=194) WRITE161,800)
107 FORMAT(4151 800 FORMAT( /,' THE CROSS SECTIONAL AREA ,(INCH * *2) /FT. .)
READ(6091C8)KTYPE,RADIUSINE9CAREA,V1 WRITE(61051)CAREA
108 FORMAT(15,2F10.2,2F10.31 WRITE(61938)
READ(609109)CAYO9DENS9VK 38 FORMAT(/.. TOTAL STRESS COEFFICIENT OF EARTH PRESSURE .)
109 FORMAT(2F10.2,F1004) WRITE(61,51)CAY0
READ(609110)EIO,E(U,DVSTEO,DVSTFUtRF1,PRI WRITE(61,171
110 FORMATI4F12.2,2F5.2) 17 FORMATS/9. POISSON RATIO,UNDRAINED AND DRAINED'.
READ(609111)CEE,PHI,VMODNO,VN,RF29PR2 1' CONDITIONS RESPECTIVELY '1
111 FORMAT(6F10.2) WRITE(61,52)PR19PR2
READ(609113)EVO,EA09EWO,EVN,ESM,S0,CCO,CCN,SIGOsSIGN,SIG$ 52 FORMAT(/9. PR1=',F12.3,/,. PR2=',F1293)
113 FORMAT(8F6.393F10.2) WRITE(61925)
READ(60.112)VELIN1,VELIN2 23 FORMAT( /,' COEFFICIENT OF PERMEABILITY*INCH/DAY .)
112 FORMAT(2F12.2) WRITE(61,60)VK
READ(60,900)KPRIN1,KPRIN2,KPRIN3 60 FDRMAT(/,F12.5)
900 FORMAT(315) WRITE161926)
WRITE(61911 26 FORMAT(/0 UNIT WT. OF SOIL,PCF. .)
1 FORMAT( /.' TOTAL NUMBER OF ELEMENTS .) WRITE(61,51) DENS
WRITE(61950)NUMELT WRITEI61.18)
WRITE(6192) 18 FORMAT(/,' INITIAL TANGENT MODULUS,PSF.,WHEN.,
1' SIG3=0 AND SIG3=S1G3ST RESP. .) WRITE(61.51)(CN
WRITE(61.54)E10,EIU WRITE(61.41)
54 FORMAT(/,' El0=',F12.3,/,' EIU=..F12.2.) 41 FORMAT(/0 MODULUS OF ELASTICITY,PSF.,UNDRAINED.,
WRITE(61.19) 1' AND DRAINED CONDITIONSIRESP. .)
1 FORMAT(/,' MAX. DEVIATOR STRESS,PSF.,W-IEN', WRITE(61,53)VELIN1,VELIN2
1' S)G3=0 AND SIG3=S1G3ST,RESP. 'I 53 FORMAT I/ VELIN1=',F123,/,' VELIN2= ',F12.3)
WRITE(61,55)DVSTFO,DVSIFU RETURN
55 FORMAT(/,. DVSTF0=',F12.3,/,' DVSTFU=',F12.3) END
WPATE(61020) C **********************************************************************
20 FORMAT(/,. FAILURE RATIO,UNDRAINED AND DRAINED', SUBROUTINE SUBCOD
1' CONDITIONS RESPECTIVELY .) C
WRITE.61,56)RF1,RF2 COMMON /A5/ ICODE(4.6)
56 FORMAT(/,' RF1=',F12.3./.. RF2=',F12.3) DO 199 1=1,4
WRITE. 61.21) DC 199 J=1,6
21 FORMAT(/,' EFFECTIVE COHESION .PSF. .) 199 ICODE(1,J)=0
WRITF(61,51)CEE N=0
WRITE(61.22) DO 200 1=1,3
22 FORMAT(/.. EFFECTIVE INTERNAL FRICTION ANGLE,RAD1AN ') DO 201 J=1,4
WRITE(61,511 PHI 201 ICODE(IsJ)=J+N
WRITE(61,23) 200 N=N+2
23 FORMAT(/0 MODULUS NUMBER-1,K ') 00 202 1=1,4
WRITE) 61.51) VMODNO ICODE(I.5)=9
WRITE(61.241 202 ICODE(1.6) =10
24 FORMAT(/,' EXPONENT.SMALL N .) ICODE(4.1)=7
WRITE(61,51)VN ICODE(4,2)=8
WR1TE(61,27) ICODE(4.3)=1
27 FORMAT(/,' INITIAL VOID RATIO ') ICODE(4.4)=2
WRITE(611151)EVO RETURN
WRITE(61.28) END
28 FORMAT) /,' INITIAL AIR VOID RATIO 'I *************************************************************** ****
WRITE(61.51)EA0 SUBROUTINE. GENCOD
WRITE(61.29) C
29 FORMAT) /,' INITIAL WATER VOID RATIO 'I COMMON /A1/ NUMNPT,NUMELT,NCULET.NCONP,NLAY
WRITE(61+511EWO COMMON /A2/ COOR(53,3),ICONN(42,6)
WRITE(61.30) COMMON /A3/ KNOVA(20,4),NODOM(52,3)
30 FORMAT-I/O INITIAL DEGREE OF SATURATION(DECIMALI .) COMMON /A4/ LADOM
WRITE(61,51)50 COMMON /A6/ NCODE(42.81
WRITE(61,31) DO 23I=1,NUMELT
31 FORMAT(/,' VOID RATIO CORRESPONDING TO EFFECTIVE', DO 23 J=1.8
1' PRECONSOtIDATION STRESS 1 23 NCODE(IsJ)=0
WR1TE(61,51)EVN DC 30 I=1,NUMNPT
WRITE(61.32) DO 30 J=1.3
32 FORMAT(/.. VOID RATIO AT FULL SATURATION ') 30 NO0014(1,11=0
WRITE) 61. 51) ESM K=0
WRITE(61,33) DO 12 1=1,NUMNPT
33 FORMAT(/.. EFFECTIVE STRESS CORRESPONDING TO', DO 2 N=1.NCONP
1' INITIAL VOID RATIO,PSF. .) IFIKNOVA(N.1) .EQ. I)GO TO 3
WRITE(61,51)5100 GO TO 2
WRITE(61.34) 3 DO 4 M=2.4,1
34 FORMAT( /,' EFFECTIVE PRECONSOLITICN STRESS,PSF. 1 IF(KNOVAIN,M) .E0. 1100 TO 4
WRITE(61.51)SIGN K=K+1
WRITE(61,35) J=M-1
35 FORMAII/0 EFFECTIVE STRESS CORRESPONDING TO', NODOM(I,J)=K
1' VOID RATIO AT SATURAT1ON.PSF. .) 4 CONTINUE
WRITE(61.51)SIGS GO TO 12
WRITE(61,36) 2 CONTINUE
36 FORMAT( /,' COMPRESSION INDEX IN THE OVER', IF(000R.1.3)-2.17,708
'1. CONSOLIDATED REGION '1 7 J=2
WRITE) 61,51) CCC GO TO 11
WRITE(611371 8 J=3
37 FORMAT(/,' COMPRESSION INDEX IN THE NORMALLY., 11 DO 31 L=1,J
1' CONSOLIDATED REGION ') K=K+1
NODOMAI9L)=K 13 LN=3
31 CONTINUE 14 YTEMP=0.0
12 CONTINUE DO 3 K=19LN
LADOM=K NP=ICONN11K)
DO 17 L=1INUMELT Y(K)=COORINP.2)
N=0 3 YTEMP=YTEMP+Y1K)
IFIICONN1L.3) .EQ. 0)00 TO 18 VLN=LN
IF( ICONN(L 94) .EQ. 0100 TO 19 YCENT=YTEMP/VLN
NP=4 LNMI=LN-1
GO TO 20 DO 4 K=IrLNMI
19 NP=3 101=K+1
20 K=N DO 4 L=KPleLN
DO 14 LDUM=19NP IF(Y1K) .07. Y1L))00 TO 51
NTEMP=ICONN1L9LDUM) GO TO 4
DO 15 1=1.2 51 YTEMP=Y1K)
NDUM=I+K Y(K)=YIL)
15 NCODEILOIDUM)=N0D0M(NTEMP9I) Y1L)=YTEMP
K=K+2 4 CONTINUE
14 CONTINUE IFIICONN1195) .EQ. 0 .AND. ICONN(I.6) .EO. 0)00 TO 15
GO TO 17 IF( 'COM 1 95) EO 0100 TO 16
18 NP=2 NNE=1CONN(195)
K=N GO TO 17
DO 13 LDUM=19NP 16 NNE=ICONNII96)
NTEMP=ICONN1L,LDUM) 17 IF11CONN(NNE94) EO 01GO TO 18
DO 16 1=193 J=ICONN1NNE,1)
NDUM=I+K K=;CONNINNE2I
16 NCODEILODUM)=NODOMINTEMP.I) L=ICONN(NNE3)
K=K+3 M=ICONN(NNE94)
13 CONTINUE YTEMP=1COORIJ2/+COOR(K92)+COOR(L2)+COOR(M.2))/4.
17 CONTINUE GO TO 19
C 18 J=ICONNINNE91)
WRITE161950) K=ICONNINNE.2)
50 FORMAT( // NODE '18X9-' DISPLACEMENT 1 L=ICONNINNE.3)
124X' X ',4X,' Y '9430 4 4) YTEMP=ICOOR(J2)+COORM.2)+COORIL.211 /3.
DO 51 I=1.NUMNPT 19 IFIICONN(195) .E0. 0)G0 TO 20
51 WRITE161952)I91NODOMCI,J)9J=193) YBOTT =YTEMP
52 FORMAT(/.15916X,159212X,I5)) YTOP=1YILN)+Y(LN-1)1/2.
C DYT=ABSIYTOPYCENT/
WRITE161953' DY8=ABSCYCENTY8OTTI
53 FORMAT(//0 ELEMENT '915X9' CODE NUMBER ') GO TO 5
DO 54 1=1NUMELT 20 YTOP=YTEMP
54 WRITE161955)191NCODE119J)1J=198) YEIOTT=IY(I)+Y12/1/2.
55 FORMAT 1/, 15.5X,815) DYT=ABS1YTOPYCENT)
C DYB=2.*ABS(YCENTYBOTT)
RETURN GO TO 5
END 15 YTOP=IY1LN)+Y(LN-1))/2.
C *****.*****.************************.********************************* Y3OTT=1Y(11+512))/2.
SUBROUTINE D2Y DYT=ABS1YTOPYCENT)
C DYB=2.*ABS(YCENTYBOTT'
COMMON /41/ NUMNPT.NUMELT.NCULET9NCONP9NLAY GO TO 5
COMMON /42/ COOR(5303)91CONN142961 12 NE11)=1
COMMON /07/ AIPHA142)HDIS(42) NE12)=ICONN1195)
NE13)=ICONN(196)
DIMENSION NE(3).Y14) DO 2 N=1.3
DO 50 1=1,NUMELT NNE=NE1N)
ALPHAII)=.).0 IF1ICONNINNE.4) .EO. 0100 TO 43
HDISII)=0.0 J=ICONN1NNE911
INICONN(1.3) O)GU TO 50 K=1CONN(NNE2)
/FfI0ONNI195) .NE. 0 .AND. ICONN)I,6) .NE. 0)00 TO 12 L=ICONN(NNE93)
.EQ. 0)00 TO 13 M=ICONN(NNE94)
LN=4 Y1N)=ICOOR(J92)+000R(K92)+COOR(L921+CCOR1M.2))/49
GO TO 14 GO TO 2
43 J=ICONN(NNE,1) DO 1 I=1.NCULET
K=ICONN(NNE112) II=ICONN(1.1)
L=ICONNINNE,3) JJ=ICONN(1.2)
Y(N)=(COOR(J.2)+COOR(K,2)+COOR(L.2))/2. XI=COOR(11,1)
2 CONTINUE. YI=COOR(II.2)
YTOP=Y(31 XJ=COOR(JJ.1)
Y8OTT=Y(2) YJ=COOR(JJ,2)
YCENT=Y(1) C** IF IT IS A CIRCULAR CULVERT,KTYPE =+1
DYT=ABS(YTOPYCENTI IF (KTYPE .50. 1) GO TO 3
DYB=ABS(YCENTYBOTT) C** IS THIS THE LAST CULVERT ELEMENT(
5 CONTINUE IF(I .E0. NCULET) GO TO 4
ALPHA(I)=ABS(DYT/DYB) KK=ICONN(I+1,2)
HOIS(I)=DYB KK=COORIKK.1)
50 CONTINUE YK=COOR(KK.2)
RETURN IF(XI .EO. XJ) G3 TO 5
END TEMH*(X1KK)*(X1*X1+YI*Y1XJ*XJYJ*YJ)
C ************..************.**.**************** .****** 1 (XIXJ)*(Xl*XI+YI*Y1KK*X1CYK*YK)
SUBROUTINE INIT(LPROB) TEML=2.*1 ( KIX.4)*(Y1YJ) (XIXJ )*(YIYr I )
C BCUL ( 1 ) =TEMH/TEML
COMMON /Al/ NUMNPT.NUMELT,NCULET$NCONP.NLAY TEM*1=(Xl*Xl+YI*YIXJ*XJYJ*YJ)-2.*IYIYJ)*BCUL(1)
COMMON /C4/ CAY0 TEML=2.*(XIXJ)
COMMON /C5/ EVO,EAOsEW0tEVAI,ESMoSIGO,CIGN,SIGS, ACUL(I)=TEMH/TEML
1 CCO.CCN.SO,EV(42) GO TO 6
COMMON /01/ AREA(42) 5 BCUL(I)=IYI*YJ)/2.
COMMON /E1/ DISP(52.3),STRESS(42,6),PCREPR(42) ACUL(1)=((X1*XI+YI*YIKK*XXYK*YK)-20.(mlYK)*BCUL(I))
COMMON /F5/ DELV(42) 1 /(2.*(XIXK1)
DO 1 N=IoNUMNPT 6 RCULII)=SORT((XIACULII))*IXIACULII)1+
DO 1 M=1.3 1(YIBCUL(1))*(MIBCUL(I)))
1 DISP(NsM) =0.0 GO TO 2
IF(LPROB .GT. 2)00 TO 4 3 RCUL(1)=RADIUS
DO 2 N=1.NUMELT ACUL(I)=0.0
STRESS(N.1)=CAYO*SIGO BCUL(I)=RADIUS
STRESSIN.2)=SIGO GO TO 2
STRESS(N.3)=0.0 4 RCUL(I)=RCUL(1-1)
STRESS(14,4)=STRESS(No2) ACULII)=ACUL(I-1)
STRESS(11,5)=STRESS(No1) BCUL(I)=BCUL(I-1)
2 STRESS(N,6)=0.0 2 COSBET=I(XIACULII)1*(XJACUL(I))+(YIBCUL)I))*
DO 3 N=1,MCULET 1 (YJBCUL(1)))/(RCULII)*RCUL(1))
DO 3 M=1,6 BETA(I)=ACOSF(COSBET)
3 STRESSIN,M)=0.0 WRITE(61,91I,RCUL(I).BETA(1),ACUL(1),BCUL(I)
GO TO 6 1 CONTINUE
4 DO 5 N= 1,NUMELT 9 FORMAT ( / t15,41 5X.f8.3) )
DO 5 M=1,6 RETURN
STRESS(N,M)=0.0 END
5 CONTINUE C **********************************************************************
6 DO 7 N=1,NUMELT SUBROUTINE CULK(I)
AREA(N)=0.0 C
POREPRIN1 =0.0 COMMON /81/ KTYPE,RADIUS.VE,VI,CAREA
DELV(N)=0.0 COMMON /82/ ACOL(5),BCUL(5),RCUL(5),3ETA(5),TRAN(6,6)
7 EV(N)=EVO COMMON /C1/ ET(421.PR
RETURN COMMON /D2/ S(6.6),SYSK(10,10)*Q(10)
END COMMON /F4/ F(6,6)
**.*****************.******...***********J**************************** C
SUBROUTINE COLPRO DIMENSION 18(6111)
C DO 2 [1=1.6
COMMON /AI/ NUMNPT,NUMELToNCULET,NCONF,NLAY DO 2 JJ=1,6
COMMON /A2/ COOR(53.3),ICONN(42,6) 2 S(II,JJ)=0.0
COMMON /81/ KTYPE,RADIUS,VEIVI,CAREA 5** CALCULATE THE FLEXIBILITY MATRIX OF NODE.I.
COMMON /32/ ACUL(5)+BCUL(5),RCUL(5),BETA(5),TjAN(6,6) SINB=SINIBETA(I))
WRITE(61.20) CO6B=COS(BETA(I))
20 FORMAT(//0 ELEMENT RADIUSIFT.1 ',3X, F(1,11=0.144*RCULII)**3.*(1.5*BETAII)-2.*SIN3,
I' ANGLE (1,7X0 (-ENS t7X0 CENY 'I 1 0.5*SIN3*COSB1/1ETlE)*VI)
2 +0.001*RCUL(I)*(0.5*BETA(1)+0.D.SINB.COSB)/(-2-T(I)xCAREA) TRAN(4,4)=(YJ-BCULII))/RCUL(I)
FI 1 .2 ) =0.144*RCUL ( I )**3.* (-0.5x3INB*SINB-COSE3+1.0) / TRAN(4,5)=-IXJ-ACUL(I))/RCUL(I)
IETII)*VI1-0.0003*RCULIII*SINBx5INB/CET(I)*CAREA1 TRAN(5,4)=-TRAN(4,5)
F(113)=0.144*RCUL(I)**2.*IBETAlI)-S!Nb)/(ET(I)xV1) TRAN(5,5)=TRAN(4,4)
F(211)=F(10) TRAN(6s6) =1.0
F(2.2)=0.144*RCUL(I)**3.*(0.5*BETA(I)-0.5*SIND*C058)/ RETURN
1 (ETII)*VI) END
2 +0.001*RCULiI1*(0.5.BETA(I)-0.5*SINB.(:053)/(ETII/xCAREA) C **************************************** * ** * * * * **., * * * * * ** * * ** * * ** * ** **
F 203 )=0.144*RCUL ( I 1**2.* :-COSB+1.0) / (ET( I )*VI ) SUBROUTINE TRIK(I,II.JJ,KX)
F(3,I)=F(193)
FI3o2)=F12,3) COMMON /A2/ C00R(53+3)sICONN(42,6)
F(3,3)=0.144*RCUL(I)*BETA(I)/(ET(I)*VI) 'COMMON /Cl/ ETI42)OR
C.. INVERT THE ABOVE FLEXIBILITY MATRIX COMMON /D2/ S(6s6).SYSK(10.10).0(10)
CALL MATINV(F.3.13B4OtDETERM) COMMON /F2/ SMAREA
DO 3 11=1,3
CO 3 JJ41s3 DIMENSION C(3,6I,DBA13,6)
3 SIII.JJ)=F(II,JJ) AJ=COOR(JJ11)-COOR(II.1)
5(4,1)=-5(191).005B,S(2o1)*SINB AX.COOR(KKIII)-COOR(II,I)
S( 4,2)=-5(1,2)*CO:A+3(2+2)*511B BJ=COOR(JJ.2)-COOR(II*2)
S I403)=-5(1,31 *COSB+5t2+3)*SIN3 BK=COOR(KK.21-COOR)IIs2)
S (5,1)=-511,1)*SIAB-5(2,1).COSB SMAREA.(AJ*BK-AK*Bj)/2.
SI5s2)=-5(1112)*SINB-S(2,2)*COS1 COMM=0.25ET(II/ISMAREA*(1.+PR)*(1.-2.*PR))
515.3)=-S(1,3:*SINB-S(2o3)*CCSB S(1,1)=COMMx41.-PR)
S( 6 ,11 =-S(1,1)*RCUL(I)*(1.-COSB)-SI211)*RCUI(I)*SINB- S(1.2)=COMM*PR
15(301) 5(1231=0.0
516 +2) = -5(1,2)*RCUL(I)*11.-COSH)-5(2,2)*RCULII)*SINE- S(2.1)=COMM*PR
15(302) S(2,2)=COMM*(1.-RR)
516113)=-511,3)*RCULII)*(1.-COSB)-S12,1).RCUL(I)*SINB- 5(2,3)=0.0
15(301 5I3,11=0.0
DO I 11.1,3 S(392)=0.0
DO I JJ=4.6 S(3,3)=COMM*(1.-2.00DR)*0.5
1 5(II,JJ)=5(JJ,II) C(1.1)=BJ-8K
S(494)=511.1) C(11,2)=0.0
SI4951=-S(I.2) C(1,3)=BK
5(4.6)=5(1113) C(1.4) .0.0
S(5.4)4-512,1) C(1,5) =-BJ
S(5,5)=S(2..2) C(1,6)=0.0
S (5,61=-512.3) C(2,21.0.0
S(6,4)=5(3.1) C)2,21 =AK -AJ
51515)=-5(3,2) CI20)=0.0
S(6.61=5(3,31 C(2,4)=-AK
RETURN C(2.5)=0,0
END C(2,6( =AJ
c ****.************************.*********.*.***.***4.*****11******.****** C(3,1)=AK-AJ
SUBROUTINE CULTRN(1) C(3s2)=BJ-8K
C C(3.3)=-AK
COMMON /A2/ COOR( 5313) ICONN(42 ,6 ) C(3114)=BK
COMMON /B2/ ACUL ( 5 ) IBCUL (5 ) 9RCUL ( 5 ) 'BETA ( 5) sTRAN (6, 6) C(315 ) =AJ
II=ICONN(Iol) C(3116)=-BJ
JJ=ICONN(I12) DO 40 K=1,3
XI=COORIII.1) DO 40 L=1,6
YI=C00R(11.2) DBA(KoL)=0.0
XJ.000R(JJ11) DO 40 H=1,3
YJ=COOR(JJ.2) 40 DBA(K,L)=DBA(K,L)+S((9M)*C(M,L)
DO 1 L1s,6 DO 50 K=1,6
DO 1 M=1,6 DO 5C L=1.5
I TRAN(L.M)=0.0 SlX.L)=0.0
TRAN(1.11.(YI-BCUL(1))/RCUL(1) DO 50 M71,3
TRAN(1.21=-(XI-ACUL)I))/RCULt1) 50 S(K,L)=5(K,L)+C(MIKI*DBA(M,L)
TRAN(2 sl ) =-TRAM I +2) RETURN (44 (.1)
TRAN(2,2)=TRAN(1,1) END
TRANI3,3)=1.0 r ************************************* n** *, * * * * * * * * * * ** * * * * * * * *;. ** * * * **
SUBROUTINE OUAD(I) IF(ICODEINQUAD),N) .EO. 0) 00 TO 17
C L=TCODE(NQUAD,N)
COMMON /A1/ NUMNPT,NUMELT,NCULET,NCONPoNLAY SYSK(K,L)=SYSKAE,L)+S(M,N)
COMMON /A2/ COOR(53,3),1CONN(42,6) 17 CONTINUE
COMMON /A5/ ICODE(4,6) 150 CONTINUE
COMMON /A7/ NL,NCASE,NUMEL,NUMNP,NSEL1 RETjRN
COMMON /C1/ ET(42).PR END
COMMON /C3/ VIC+DEN5
COMMON /01/ AREA142) SUBROUTINE STIFF(I)
COMMON /D2/ 5(6,6),SYSK(10,10),0(10) C
COMMON /F2/ SMAREA COMMON /A1/ NjMNPT,NUMELT,NCULET,NCONP,NLAY
C COMMCN /A2/ CODR(53.3),ICONN(42,6)
DIMENSION NP(51,5M0(6) COMMON /35/ I500614,5)
AREA(I)=0.0 COMMON /A7/ NL4NICASE,NMELoNUMNP,NSEL1
DO 10 M=1,4 COMMON /61/ KTYPE,RADILIS, VE oV 1 oCAREA
10 NPIME=ICONN(1,M) COMMON /52/ ACOL(5)26CUL(5),RCULI5),BETA(5),TRAN(6,6)
NP(5)=NUMNPT+1 COMMON /C1/ ET(42(,PR
DO 15 M=1.10 COMMON /C3/ VK,DENS
Q(M)=0.0 COMMON /01/ AREA(42)
DO 15 MM=1.10 COMMON /D2! 516,6),SYSK(10,10).0(10)
15 SYSK(MoMM)=0.0 COMMON /F2/ SMAREA
J=NP(1) COMMON /F4/ F(696)
K=NP(2) IF(ICONN(1.3) .EQ. c)p() TO 7
L=NP(3) IF(ICONNII.4) .E0. 01G0 TO 8
M=NP(4) C** THIS ELEMENT IS A QUADRILATERAL ELEMENT
KK=NP(5) CALL OUAD(I)
COOR ( KK 1 ) = (COOR ( J ) +CODR K. +COOR ( )*COOR ( M, 1 ) ) /4. DO 4 1..1.9
COORIKK,2)=ICOOR(Jt2)+COOR(Ks2)+COOR(Ls2)+500R(M,2))/4. CC=SYSK(II,10)/SYSK(1 1.10)
DO 150 NOUAD=1,4 0111)=G1I1)CC*0(10)
II=NP(NOUAD) DO 4 JJ =1.9
JJ=NP(NQUAD.1) 4 SY5K( IIIJJ)=5Y5K(11,JJ)CC SYSK(10,JJ)
IF{ NQUAD .1;. 41 JJ.NP ( 1 ) DO 5 11=1,8
KK=NP(5) CC=SYSK(11.9)/SYSKI9.9)
CALL TRIK(1,11,JJ,KK) 0(11)=O(I1)CC*0(9)
AREA(1)=AREA11)+SMAREA CO 5 JJ=108
DO 16 J=1,6 5 SYSKEII,JJ)=SYSKIII,JJ1CC*SYSK19.JJ)
16 SMO(J)=0.0 GO TO 10
C.* IF IT IS THE TIME DEPENDENT ANALYSIS, NCASE=2 7 CONTINUE
IF(NCASE .E0. 2)00 TO 18 C** THIS IS A CULVERT ELEMENT
C ** CALCULATE DEAD WT. FORCES, NEWLY ADDED SOIL ELEMENTS CALL CULK(I)
IF(1 .LT. NSELTIGO 70 19 CALL CULTRN(I)
DL=SMAREA*DENS/3. DO 2 L=1.6
SMQ(1)=0.0 DO 2 M=1,6
5M0(2)=DL F(LoM)=C.0
SMO(3)=0.0 DO 2 N=1,6
SMQ(4) =00 2 FAL,M)=FIL,M)+TRAN(NoL)*SIN,M)
SMO(5)=0.0 DO 3 L=1,6
SMO(6)=DL DO 3 M=1,5
GO TO 19 SYSK(L,M)=0.0
18 CONST= S00000. DO 3 0=1,6
SMO(1)=0.5*CONST*(COOR(JJ,21COOR(KK,2)) 3 SYSK(LIMI=SYSK(LIM)+F(L,N)*TRAN(NoM)
5M0(2)=0.5.-CONST*(COORIKK*1)000R(JJ,1)) GO TO 10
SMO(3)=0.5*CONST*(COOR(KK,2)-500R(II,2)) 8 CONTINUE
SMO(4)=0.5*CONST.(COOR(11,1iCDOR(KF,1)) C*. THIS IS A TRIANGLE
5MU15)=0.5.CONST*(COCR(11,2)GDOR(JJ,2)) :I=ICONN(1.1)
smQ( ) ,..c.5.ct)N5T.( cOG3 ( 1 )COOR ( I I ) JJ=ICONN(I,2)
C** USE COKE NO. TECHNIQUE TO ASSEMBLE QUADRILATERAL ELEM. KK=ICONNI1It3)
IC DC 17 M=1,5 CALL TRIK(I,II,JJ,KK)
IF(ICODE(NOUAD,M) .EO. 0)50 TO 17 AREA(I)=5MAREA
K=ICODEINQUADoM) DO 6 J=1,6
O(K)=0(<)+SMO(H) 6 Q(J)=0.E
DO 17 N=1,6 C ** IF IT IS TIME DEPENDENT ANALYSIS,NCALE=2
IF(NCASE .EQ. 2)60 TO 9 10 LN=6
C** CALCULATE LEAD WT. FORCES9NEWLY ADDED ELEMENT 11 L=LADOM+INCULET
1E(1 .LT. NSEL1)G0 TO 10 DO 2 M=1,LN
DL=AREA(I)DENS/3. IF(NCODE)19M) .E0. 0)00 TO 2
0111=0.0 K=NCOPE(ItM)
O)2)=PL MPACK=LK+1
0(3)=0.0 KMPACK=(K-11.(2*NDK+2)/2+MPACK
0(4) =DL A(KMPACK)=A(KAPAC)+U(M)
0(5)=0.0 2 CONTINUE
0(6)=DL 9 RETURN
GO IC 10 END
9 CONST=500000. C **.*....AKAK,,************A.***.***.***********.***********************
0(1)=0.5*CONST*(COORIJJ02)COOR(KK,2)) SUBROUTINE FVEC(I)
0(21=0.5*CONST*(C00RIKKI1)COOR(JJ,1)) C
0(5)=0.5*CAST*(COORIKK.2)COOR(11,2)) COMMON /Al/ NUMNPTsNUMELT,NCULETsNCONP,NLAY
0(4)=0.5*CONST*(COCR(II91)000R(KA,11) COMMON /A2/ COOR(53.3),ICONN(42,6)
015)=0.5*.CONST*ICCOR(II,2)COOR(JJ12)) COMMON /04/ LADOM
,/(6)=0.5*CONST*(COOR(JJ91)COOR(II91)1 COMMON /A6/ NCODE14298)
10 RETUFN COMMON /A7/ NL9NCASEPAIUMEL,NUMNP,NSEL1
END COMMON /C3/ VK.DENS
C *4-....1...*****..****.****4,1****** 4**.***M 4-************************* COMMON /P1/ AREA(42)
SUBROUTINE BGK(1) COMMON /02/ 5(6,6),SYSK(10.10)P0(10)
C COMMON /04/ 3(1191
COMMON /A1/ NUMNPT,NUMELT,NCULET,NCONP,NLAY COMMON /D7/ ALPHA(42),HDIS(42)
COMMON /02/ COOR(53,31,1CONN(42/61 COMMON /E1/ DISP(52.3),STRESS(42,6),POREPR(42)
COMMON /A4/ LADOM. COMMON /FI/ DT
COMMON /66/ NCODE(4268) COMMON /F5/ DELV(42)
COMMON /A7/ littNCASE9NUMEL,NUMNP.NSELI COMMON /61/ KPRIN1,KPRIN2,KPRIN3
COMMON /D2/ 516,6),SY3KI10,10190(10)
COMMON /03/ A(7140) IF(ICONN(193) .EO. 0)30 TO 10
COMMON /05/ ND IFINCASE .E0. 1)60 TO 11
NPEL=0 IF(ICONN(195) .NE. 0 .AND. 'CONN( I 96 ) .NE. 0 IGO TO 12
IF(ICONN(113) .EQ. 0)30 TO 3 POREC=POREPRII1
IPIICONN11.4) .00. 0)00 TO 4 IFIICONNI1,5) .10. 0 *AND. /CONNII.6) .10. 0)60 TO 13
LN=8 IFIICONN(1,5) .EO. 0)30 TO 14
GO TC 5 C** THIS SOIL ELEMENT IS ON THE TOP LAYER WITH SOIL
3 IN=6 C** ELEMENTS BENEATH IT
GO TO 5 NEB=1CONN(1.5)
4 LN=6 PORET=0.0
NPEL=3 POREB=POREPP(NE3)
5 DO 1 M=1,LN GO TO 15
DO 1 N=MoLN 14 CONTINUE
K=NCODE11,M) C** THERE 15 NO SOIL ELEMENT BENEATH THIS SOIL ELEMENT
-L=NCODE(19N) C** BUT THERE IS AT LEAST ONE SOIL ELEMENT ON TOP
IF(K .10. 0 .0R. L .10. 0)60 TO 1 NET=ICONN(I,6)
IF(K .LE. L/G0 TO 6 PORET=POREPR(NET)
TEMPO=K PORES=POREC
K=L GO TO 15
L=TEMPO 13 CONTINUE
6 MPACK=LK+1 C** THERE 15 NO SOIL ELEMENT ON TOP OR BENEATH THIS ELEMENT
KMPACK=1K-11*'(2*INDK+2)12+MPACK PORET=0.0
IF(NPEL EO 3)00 TO 7 POREB=POPFC
A(AMPACK)=A(KYPACK)+SYSK(MtN) GO TO 15
00 TO 1 12 CONTINUE
7 ACC',IPACK)=A(KMPACK)4S(MiN) C* THERE ARE SOIL ELEMENTS ON TOP AND BENEATH TMTS ELEMENT
1 CONTINUE NEC=I
IFINCASE .10. 2)60 TO 8 NET=1CONN(1,6)
0 IC- 9 NEB=ICONN(1,5)
IFIICONN(I0) .10. J)C10 TO 9 POIREC=POPPRINEC)
1F(ICONN(1,4) .10. 0)0U TO 10 PORET=PSREPRINET)
LN=8 2OREL,POREPR(NE13)
GO TO 11
15 02PORE=2..(ALPHA(1).POREB-1ALPHA(1) +1.1.POREC+PORET// E(N)=8(N) /A(N1)
I (HDIS11)*HD1511/*ALPHA(I)*(ALPMA(I( +1.)) C
VOL=AREA(I)+DELV(I) C** HECK FOR LAST EQjATION
DVOL=VOL.VK.D2PORE*DT/748.8 (FIN .EQ. ND/GO TO 700
DELV(I1=DFLV(I)+DVOL IF(NB .E2. ND)GO TO 11
IF)KPR/N2 oEClo 1)WRITE(61990)1oVOLoDELV(1) L=NDNB+1
90 FORMAT( /,IS,' ELEMENT VOLUME ',E15.5o. CL1oFTo , /,5X, IF(N .GT. L)GC TO 11
1' VOLUME CHANGE FROM THE END 3F CONSTRUCTION = ',E15.5o, UU.FT. ') NW.NB
K.LADOM+INCULET GO TO 12
CON5T=500000. 11 NW=NW-1
B(K)=CONST.DVOL
3C TO 1G C** DIVIDE N TH EQUATION BY DIAGONAL ELEMENT
11 CONTINUE 12 DO 600 K=2oNW
Co. IT IS CONSTRUCTION PERIOD NK=N1+K-1
IF(ICONN(1,4) .EQ. C(00 TO 41 F(K) =A(INK)
LN=8 6CC A(NK)=A1NK1/A(N1/
GO TO 42 C
41 LN=6 C** REDUCE REMAINING EQUATION
42 DO 1 M=loLN DO 660 L.21NW
IFINCODE(loM) oEGo 0)00 TO 1 I=N+L-1
K.NCODEII,M1 11=41-11*(2*ND-1+2)/2+1
8(K)=BIK)-0(M1 J=0
1 CONTINUE DO 650 K=LoNW
13 RETURN J=J+1
END IJ.11+0-1
####.4.41-**.#,...***.4 .*..,k**.*####*,..-.###*..-*******A-*****.******4***** NK=N1+K-1
SUBROUTINE BAWCD 650 A(1J)=6(1J1F(L).A(NK(
C 10(1)=B(1)F(L)*E(N)
COMMON /A6/ NCODE(42,8( 66C CONTINUE
COMMON /67/ NLoNCASE,NUM.ELINUMNP,NS L: GO TO 500
COMMON /D6/ N8 C
J=0 C ** BACK SUBSTITUTION
DC 259 N=1oNUMEL 700 NW=1
DO 258 1=1,8 701 N=N-1
IF(NICODE(No1) *EU. 0)06 TC 258 'FIN .F:0.. 0(30 TO 900
DO 257 L=1,8 N1=(N-1)*(2.NDN+21/2+1
IF(NCODE'NoL) oEQ. 0/20 TO 257 (FINS oE0o ND1GO TO 13
KK=1A0S(NCODE(No1)NCODE(N,L)/ L=NDNE+1
FIKKJ1257,257,256 IF(N .GT. L)G0 TO 13
256 J=KK NW=NB
257 CONTINUE GO TO 14
258 CONTINUE 1E NW=NW+1
259 CONTINUE 14 DC 800 K=2,NW
NB=J+1 L=N+KI
RETURN NK=N1+K-1
END Ei(N)=B(N)A(NK1*B(L)
c *
800 CONTINUE
SUBROUTINE SOLVF_ GO TO 701
C 900 RETURN
COMMON /D3/ A(7140/ END
COMMON /04/ 0(1191
COMMON /05/ ND SUBROUTINE RESULT
COMMON /06/ NB C
C CCWDON /Al/ 1.LI9NPT,NUMELToNCULETINCONP NLAY
DIMENSION F(I35) 00134013 /A2/ 0001(53,31,10JNN)42.53
NW =ND +1 COMMON /A3/ KNOVA(20,41oNiD0F(52*3)
N=0 COMMON /A4/ LADOM
50G N=N+1 COMMON /A5/ IITODE(4.6)
C.. FIND THE FIRST TERM OF 9312 N CO:19LN /AA/ NCODE(42,8)
N1=(N-1) .:2*NDN+2)/2+1 COkAION /67/ NLoNCASE,NUM:L,NC'9NPoNSEL1
CC''WON /81/ KTYPEoRADIUSoVE.VI,CAPEA
C.. DIVIDE RIGHT SIDE BY DIAGONAL ELLMEIIT 20:IDON /02/ ACQL(5)o3CUL(5),RCUL(5),ATAL,I,TRP '6,6'
00MON /01/ ET(42)s99 105 TDIST(N)=0.0
COMMON /03/ VA.FOENS 12 CONTINUE
COMMON /05/ EV3 sEA3sE,EVN,,,,A161(10,SIGNsSIGSs I; INCASE EQ. 2)G3 TO 106
1 CCO.CCN,SO,EV(42) GO TO 107
COMMON /31/ AREA(42) 106 11LADOM+1N0ULET
COMMON /02/ 5(6,6)sSYOK(10,10),3(20) DO 13 L=1,10
COMMON /D4/ 8(119) 13 0(L)=Q(L)+8(K)
COMMON /101/ DISP(52,3),STRESS(42,6),POREPR(42) 107 DO 14 KA1.2
COMMON /F3/ ST(6) ST(K)=Q(K+8)
COMMON /F4/ F(6,6) DO 14 L=1,8
14 GT(K)=ST(K)SYSK(K+8,L).TDISP(L)
DIMENSION TDISP(10) COMM=SYSK (9, 9 ) *SYSK ( s 10 5111K )10s9 ) *SYSK ( 9 IP 10 )
WRITE(61.502) TDISP(9).(SYSK(10,10)*ST(1)-6YSK(10+9)*ST(2))/COMM
502 FORMAT(//,' NODE 0s8Xs, XDISP ,.850 YDISP (s6X, TOISP(10)=(SYCK(9110) .5T(11+CYSK(9,9)*STi2/1/00MM
1' ROTATION ') DO 15 N=1.6
IFINCASE .50. +1)00 TO 301 15 ST(N)=0.0
NUMNPANUMNAT DO 16 NOUAD=1,4
NUMEL=NUMELT II=ICONN(IsNQUAD)
301 DO 50C 1=1,NJMNP JJ=ICONN(IsNOUAD+1)
DO 4 J=1.3 IF(NQUAD .E0. 4)JJ=ICONN(1,1)
IF(NODOM(IsJ) .EQ. 0)00 TO 4 KK=NUMNPT+1
K=N300M(1,1) AJ=COORIJ.1011-0009(1I,1)
DISP(IsJ)=DISD(IsJ)+8(K) AK=C00R(KKs1)000R(11s1)
4 CONTINUE EfJ=COORIJJ.2)COOR)11,2)
I,(DISP(1.0) BK=C00RIKKs2)COOR(11.2)
501 FORMA-T( /.1513E16.5) U=NOUAD
50C CONTINUE K=NOUAD+1
'WRITE(61s503) IF(NOUAD .EQ. 4)K=1 .
503 FORMAT(//r, ELE. SIGMAXX ,,5X0 SICMAYY ,v5Xs L=5

1' SIGMAXY ../s9X0 AND ,s10)(1.' AND ,s10110 AND '. /. EPX=I(BJ-8K).TDISP(2.J-1)+EK*TDISP(2*K-1)
27)(0 MAX...SIGMA ,s4X0 MINsSIUMA ,s4Xs, MAX.SHEAR )1 1 E1J*TDISD(2*L-11)/(AJ*BKAK*BJ)
34X,' PORE FRES. EPYAI(AKAJ)*TDISP(2*J)AX*TDISP(2*K)+
DO 201 I=1sNUMEL 1 AJ*TDTSP(2*L))/(AJ*BKAK*3J)
IF(I .GT. NCULET)G0 TO 102 GAM=C(AKAJ)*TDISP(2.J-1)AK*TDISP42*K-1)+
CALL CULK(I) 1 AJ*TDISP(2*L-1)+(BJBK)*TDISP(2AJ)+
CALL CULTRA(1) 2 EiK*TDISP(2*K)BJ*TDISP(24-L))/(AJ*15K-40.8J,
DO 6 N=1,6 ST(1).ST11)+((1sFR)*EPX+PR*EPY).(ET(1)/((1.+CR1+
IF(NCODE(IsN) .ED. 0)00 TO 103 1 11.-26*PR)))
K=NCODE(IsN) ST(2)=ST(2)+(PR*EPX+(1.PR) *EPY).(ET(1)/((1.+PR)*
TDISPIN)=13(K) 1 (1.-2.*PR1))
GO TO 6 18 ST(31=ST(3)+(0.5*ET(I)/(1.+PRI)*GAM
103 TDISP(N)=0.0 DO 17 N=1.3
6 CONTINUE 17 STRESS(IsN)=STRESS(IsN1ST(N) /4.
DO 7 L=1.6 GO TO 18
DO 7 M=1116 104 DO 10 N=1.6
FIL,M)=0.0 IF(NCODE(10) .50. 0)00 TO 110
DO 7 N=1,6 K=NCODE(IsN)
7 F(L,M)=F(LsM)+G(L,N)+TRAN(NsM) TDISPIN)=B(K)
DO 8 L=1,6 GO TO 1C
E,T(L)=OsC 110 TDISP(N)=0.0
00 8 N=1,6 13 CONTINUE
b ST(L)=STIL)+F(L,N)+TDISP(N) II=10ONN(1.1)
DO 9 LA1s6 33=10048(1.2)
9 STRESS('.L)= STRE.SS(1IL) +ST(L) KK=ICONN(1.3)
00 TC 5 AJ.000R(JJ,1)COORII1s11
132 11(100NN(114) .10. 0)00 TO 104 AK=COOR( KK s 1 ) COOS ( I s 1)
CALL QUAD(I) OJ.COOR(JJ,2)-000R(11,2)
DO 12 N=1,8 BK=COOPAKK,21-000R(11,2)
IF(NCODE(IsN) sEOs )))G0 TO 105 EPX=((8J-8K)*TDIEF(11+3K*TD15?(3)BJ.TOISP(5)1/
K=NCODE(IsN) 1 (4,1*BKAK+8J)
TOISP(N),,E(K) EPYA((AKAJ)*TDISP(2)-46.TDISP(4) +4J.TDISP(6))/
GO TO 12 1 (A.1.15KA*0J)
GAM=I(AKAJ)*TDISP(1)AK=TDISP(3)+AJ*DISP15). C** IN NC RANGE
1 (BJBK)*TDISPI 2 )+BK*TDISF( 4 )BJ*TL ISPI 6) ) / Y3= SIGP2 /SIG2I
2 IAJ*BKAK*BJ) Y3=A8S(Y3)
ST( 1).(11.PR)*EPX+PR*EPY)*(ET(1)/1(1.+PR)* DELE1=CCN*ALOG10(Y3)
1 (1.-2.*PR))) 13 EP=El+DELF1
ST(2)=IPR*EPX+11.PRI*EPY)*(ETII)/(11.+PR/* DELE=EVOEP
1 (1.-2*PR))/ U=2116.8 *DELE/(EA0+0.02*EWODELE/
ST(3)=I0.5*ET(I)/(1.+PR))*GAM U1 =(U1 +U1/2.
DO 11 N=193 1 CONTINUE
11 STRESS(I,N)=STRESSII,N)STINI IF(DELE .LT. EAO)GO TO 4
18 CONTINUE 2 CONTINUE
CMOHR=ISTRESSI1021 +STRESS(I,1)1 /2. WITE(61,21)1
DMOHF=ISTRESSI192)STRESS(191)//2. 21 FORMAT(' ELEMENT.,150 SATURATED')
RMOHR=SORT( STRES5(193)*STRES5(1,3)+DMCHR*DMOHR) C** IT IS SATURATED
5TI4)=CMOHR+RMOHR U5=105340.0*(1.-50)/50
ST(5)=CMOHRRMOHR SIG1S=SIGS+US
IFINCASE .EQ. 2)G0 TO 108 DLSIGS=ST(4)SIG1S
CALL PORE(I) DELOS=DLSIGS
GO TO 109 US=US+DELUS
108 K=LADOM+INCULET POREPRII)=US
CONST=500000. EV(I)=EWO
DPORE=BIK)*CONST GO TO 3
P OREPR(1)=POREPRIIIDPORE 4 POREPR(1)=U
109 STRESS(194)=ST(4) EVII)=EP
STRESS( .5 )=ST I 5 1 3 RETURN
STRESS I ,6 I =RNOHR END
5 WRITE(61,200)1,ISTRESS(I,J),J=1.6),POREPR(1) C ******************+.*********.******+**+************************rn***
201 CONTINUE SUBROUTINE MATINN'(AoN,B.M,DETERM)
200 FORNATI/$13,E14.5,2E15.59/.3X9E1A.593E15.5) C MATRIX INVERSION WITH ACCOMPANYING SOLUTION OF LINEAR
RETURN C EOUATICNS OF THE FORM AX = B. JORDON'S METHOD
END C A IS THE ARRAY TO BE INVERTED.
C ....,.....,.....*+.+++.+*.+*******+++*******..****+.++********+********* B IS THE COLUMN 01- CONSTANTS FOR LINEAR EQUATIOA SOLUTION.
SUBROUTINE POREII) N IS THE ORDER OF A
C C M IS THE INDICATOR FOR SPECIFYING INVERSION OR SOLUTION
COMMON /C5/ EVO,EAO,E409EVN,ESM,S160,SIGN13105, OF LINEAR EGUAT1UN5.
1 CCOICCN,SO,EV(42) C M=09 INVERSION IS PERFORMED.
COMMON /E1/ DISP(52,3),STRESSI4296),PORERR(421 C 1=1, SOLUTION OF LINEAR E0OAT IONS IS PERFORMED.
COMMON /F3/ ST(6) C AT THE RETURN TO THE COLLING PROCR,',M, A INVERSE
E 1=EV(I) C IS STORED AT A AND X AT 0.
S IS21=STRESS(1,4)POREPRII) C NOTE.. IF USED SOLELY FOR INVERSION, THE CALL SATTEMENT
DELSIG=ST14)STRESS(194/ C MUST STILL CONTAIN AN ENTRY CORRESPONDING TO B.
C** CHECK THE SATURATION OF THE ELEELNT C DETERM IS THE LOCATION IN WHICH THE DETERMINANT IS STORED.
IEIE1 .LE. EU0)150 TO 2
OI=POREPRIJ)+OEL510/2. DIMENSION IPIVOT(6)9A(6,6),B(6,1),INDEX(6,2),
DO 1 NUM=1,10 1RIVC7(6)
SIGP2=5TI4,U1 .
DETERM=1.0
IFISIG21 .LT. SIGN S1GP2 Si )GU 11) 11 DO 20 J=1,N
IF(SIG21 .GE. SIGN)GO TO 12 20 1PIVOT(J)=0
C.* IN OC AS), NC RANGE DO 550 1=19%
51=51:3N/01621 C SEARCH FOR PIVOT ELEMENT
Y1=ABS(Y1) AMAX=0.0
Y2=5IGP2/SIGN DC 105 J=1.N
Y2=ABS(Y2) IF (IPIVOT(J)-1) 60, 105, 60
DELE1=CCO=ALOG10(Y1)CCN=ALOG1C(Y2) 63 DO 100 K=1,N
GO TO 13 IF IIPIVOT(K)-1) 809 100, 740
11 CONTINUE SO IF IABSE(AMAX)ABOF(AIJ,K))) 65. 100, 100
C.* IN DC RANGE 65 IROW=J
Y3=SIGP2/SIG21 ICOLUM=K
Y3=ABSIY3) AMAX=AIJ,X)
DELE1=CCO*ALOG10(Y3) 100 CONTINUE
GO TO 13 105 CONTINUE
12 CONTINUE IPIVOTIICOLUM)=IPIVOTIICOL-f,)+1
C INTERCHANGE ROWS TO PUT PIVOT ELEMENT ON DIAGONAL
IF (IROdICOLUMI 140. 260, 140
140 DETERM=DETERY
DO 200 L=1,N
SWAP=A(IROW,L)
AIIROW,L)= ACICOLUM,L)
200 AfICOLUM,LI.SMAR
IF(V) 260, 260, 210
210 00 250 L=1, M
S(JAP=MIIROW,L)
C(IROWIPL)=BIICOLUM,L)
250 3IICOLUM,L)=SWAP
260 INDEX(I,1)=IROW
INDEXII,2I=ICOLUM
FIVOT111=A(ICOLUM,ICOLUM)
DETERM=DETERM.PIVOTII)
DIVIDE PIVOT ROW BY PIVOT ELEMENT
A(ICOLDM,ICOLUM)=1.0
DO 35D L=lioN
350 AlICOLUM,L)=ACICOLUM,LUPIVOTIII
IF(M) 380, 360, 360
360 DO 370 L=/,M
370 e(lcoLum,L)=e(tcaum,L)/PivoT(1)
REDUCE NONPIVOT ROWS
380 DO 550 L1=1,N
IFILlICOLUM) 40C, 550. 400
400 T=AILIOCOLUMI
A(LI,IcoLumr=c.n
00 450 0=1,N
450 AIL] ,L ) =A I Ll ,L ) A ICOLUN,L
IF(M) 550, 550, 460
46C DO 500 0=1,M
500 BILI,L)=BIL1,LIBIICOLTI,L).T
550 CONTINUE
INTEP0HANGE GOLDY2IS
00 710 1 -1,N
L=N+3-1
IF (INDFX(L,1)INDEXIL,2)) 630, 710, 130
630 JR04=INDEX(1,1/
JCOLCM=INDFX(L,2(
LO 705 K=I,N
.SP=A(K,JFON)
A(K,JROW)=A(K,JC3LUM)
AIK,JCOLDM)=SWAP
7C5 CONTINUE
710 CONTINUE
740 RETURN
END
*, 4*. =c,,,.****

FEA of Culverts Under Embankments With Time Dependent and Non-Linear Properties

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

FEA of Culverts Under Embankments With Time Dependent and Non-Linear Properties

Uploaded by

Copyright:

Available Formats

AN ABSTRACT OF THE THESIS OF

VORAVIT LERTLAKSANA for the DOCTOR OF PHILOSOPHY

EMBANKMENTS WITH TIME DEPENDENT AND NON-

A computer program was developed for analyzing culverts

The nonlinear, stress-dependent properties of soil are included

considerable computer memory locations. For complete analysis of

Redacted for privacy

Redacted for privacy

Redacted for privacy

Dean of graduate SChool

Date thesis is presented Fee. 22, 1973

The author wishes to thank the Government of Thailand for the

1.1 Statement and Scope of the Problem 1

II FINITE ELEMENT FORMULATION 7

2.1 Plane Strain Matrix Equations 7

III INCREMENTAL CONSTRUCTION ANALYSIS 30

V TESTING OF THE COMPUTER PROGRAM 49

VI ANALYSIS OF SOIL-CULVERT PROBLEMS 57

VII DISCUSSION OF RESULTS 80

VIII SUMMARY AND CONCLUSIONS 87

Appendix A. User's Manual for Symmetrical

2.1 Plane strain element 9

2.2 Quadrilateral element 13

2.3 Forming culvert elements 15

2.4 Circular beam element 15

2.5 Cantilever circular member 17

2.6 Local and global coordinate system 19

2.7 Piecewise linear curve 21

2.8 Hyperbolic stress-strain curve 22

2.9 Transformed hyperbolic stress-strain curve 22

2.10 Deviator stress-axial strain curves 25

2.11 Initial tangent modulus-confining pressure curve 26

2.12 Compressive strength-confining pressure curve 26

3.1 Simplified flow diagram for incremental

3.2 Void ratio-effective stress curve 34

3.3 Approximation of no drainage during construction 37

4.1 Interpolating parabola for difference equation 42

, 4.2 Simplified flow diagram for time-dependent analysis 44

5.1 Finite element model of circular ring 51

5.2 Deformed shape for circular ring 52

5.4 Finite element model of soil layer 53

5.5 Numerical instability due to large time interval 54

5.6 Plot of pore pressure vs. time, linear consoli-

5.7 Pore pressure distribution in soil layer 56

6.1 Properties of the simulated soil-culvert system 60

6.2 Soil-culvert finite element model 61

6.3 Initial tangent modulus-confining pressure for

6.4 Compressive strength-confining pressure for

6.5 Void ratio-effective stress curve of Bully Creek soil 64

6.6 Deformed shape at the end of construction,

6.7 Deformed shape at the end of construction,

6.8 Plot of vertical deflection at the crown of

6.9 Plot of vertical pressure at the crown of

6.10 Radial pressure on flexible culvert at the end

6.11 Radial pressure on rigid culvert at the end

6.12 Plot of bending moment during construction

6.14 Plot of maximum bending moment vs.

6.15 Pore pressure contours at the end of construc-

6.16 Pore pressure contours at the end of construc-

6.17 Plot of vertical deflection at crown of flexible

6.18 Plot of the pore pressure vs. time after

6.19 Plot of Terzaghi's pore pressure - time

6.20 Plot of the approximate vertical deflection at