Finite element analysis of circular cell bulkheads
Transcript of Finite element analysis of circular cell bulkheads
AN ABSTRACT OF THE THESIS OF
LIKHIT KITTISATRA for the degree of DOCTOR OF PHILOSOPHY
in CIVIL ENGINEERING presented on 16e.A.A4seiv- 13. lg7Y
Title: FINITE ELEMENT ANALYSIS OF CIRCULAR CELL
BULKHEADS
Abstract approved:
Redacted for Privacy
Dr. Laursen
A mathematical model is developed for the analysis of circular
cell bulkheads subjected to gravity loads of cell fill, backfill, lateral
pressure, and surcharge load using the finite element method. The
finite element circular cell model consists of axisymmetric triangular
and/or quadrilateral ring elements for soil and cylindrical shell ele-
ments for the steel cell. To formulate and derive the governing
equations for such elements, the Ritz displacement functions and the
theorem of minimum potential energy are used. The loadings of the
bulkhead system and corresponding displacements are expanded in
Fourier series. The soil fill and foundation are assumed to be
elastic, isotropic and non-homogeneous materials. It is assumed that
there is no slippage in the soil-steel shell interface.
The element matrices and the Fourier harmonic analysis are
verified by comparing results for several structures for which
classical solutions are available.
The circular cell bulkhead is analyzed for stresses in the soil
elements and shell elements and deformations for the isolated filled
cell and backfilled cell cases. Results of the study are presented,
discussed and compared with field measurement data obtained by
other investigators.
Finite Element Analysis of Circular Cell Bulkheads
by
Likhit Kittisatra
A THESIS
submitted to
Oregon State University
in partial fulfillment ofthe requirements for the
degree of
Doctor of Philosophy
June 1976
APPROVED:
Redacted for Privacy
Professor of Civil Engineeringin charge of major
Redacted for Privacy
Associate Professor of Civil Engineering
Redacted for Privacy
Head gf Department of Civil Engineering
Redacted for Privacy
Dean of Gradu4te School
Date thesis is presented NioVeyn0-er ICl'75
Typed by Clover Redfern for Likhit Kittisatra
ACKNOWLEDGMENT
The author wishes to express sincere appreciation and gratitude
to his major professor, Dr. H.I. Laursen, for his guidance, under-
standing and encouragement during the preparation of this thesis.
Thanks are due to Dr. W. L. Schroeder for providing the author
with invaluable guidance and philosophical influence throughout the
study.
A special thanks is expressed to his wife, Supaporn, for her
patience, encouragement and understanding during the year of this
study.
Throughout the author's entire academic career, he has
received the most enthusiastic support from his parents and his
brother. This thesis is dedicated to them.
TABLE OF CONTENTS
Chapter Page
I. INTRODUCTION 1
1.1 Statement and Scope of the Problem 2
1.2 Method of Solution 4
II. FINITE ELEMENT FORMULATION 102.1 Triangular Axisymmetric Ring Element Matrix
Equations 122.2 Quadrilateral Element Matrix Equations 272.3 Shell Element Matrix Equations 292.4 System Equations and Solution Process 40
III. TESTING OF COMPUTER PROGRAM 42
IV. ANALYSIS OF CIRCULAR CELL BULKHEADS 564.1 Isolated Circular Cell 584.2 Circular Cell Bulkhead 58
V. DISCUSSION OF RESULTS 1055.1 Discussion of the Isolated Circular Cell 1055.2 Discussion of Circular Cell Bulkhead 108
VI. SUMMARY AND CONCLUSIONS 118
BIBLIOGRAPHY 121
APPENDICES 124Appendix A: Element Matrices 124Appendix B: Fourier Harmonic Coefficients 133Appendix C: Gaussian Quadrature Numerical
Integration Procedure 135Appendix D: Modulus of Elasticity of Soil 138Appendix E: User's Manual for Circular Cell Bulkhead
Program 141Appendix F: Description of Computer Program 158Appendix 0: Program Listing 167
LIST OF FIGURES
Figure
Typical circular cell bulkhead structure.
Typical triangular soil element and shell element.
Finite element simulation of a circular cell bulkheadsystem.
Triangular axisymmetric element.
Quadrilateral element.
Shell element and coordinates.
Finite element model of rigid base soil system.
Vertical stress in soil due to surface load.
Radial stress in soil due to surface load.
Tangential stress in soil due to surface load.
Shear stress in soil due to surface load.
Finite element model of circular tank.
Radial displacement of circular tank.
Hoop force and longitudinal moment of circular tank.
Edge load applied to the cooling tower shell.
Approximate loading diagram for cooling tower.
Cooling tower stress resultants.
Radial and axial displacements for cooling tower at 0 =
Radial and axial displacements for cooling tower at0 = 22.5°.
An isolated circular cell.
Page
1. 1.
1.2.
1.3.
2.1.
2.2.
2.3.
3.1.
3.2.
3.3.
3. 4.
3.5.
3.6.
3. 7.
3. 8.
3.9.
3. 10.
3. 11.
3.12.
3.13.
4.1.
0°.
5
7
8
13
28
30
44
45
46
47
48
49
49
50
51
52
53
54
55
59
Figure Page
4.2. A circular cell bulkhead. 60
4.3. Circular cell finite element model. 61
4.4. Radial displacement, hoop force and sheet stressesdue to gravity load of cell fill (axisymmetric loadingcase). 62
4.5. Vertical displacement of circular cell due to gravityload of cell fill. 63
4.6. Contours of vertical stress in soil due to gravityload of cell fill.
4.7. Contours of radial stress in soil due to gravity loadof cell fill.
64
65
4.8. Contours of circumferential stress in soil due togravity load of cell fill. 66
4.9. Contours of shear stress in soil due to gravity loadof cell fill. 67
4.10. Principal stresses in soil due to gravity load of cell fill. 68
4.11. Equivalent loading diagrams due to backfill. 69
4.12. Assumed arc tension loads. 71
4.13. Fourier harmonic loading expansions for arc tensionload Ta (Case I). 72
4.14. Fourier harmonic loading expansions for uniformdistributed pressures pa and qb (Case II).
4.15. Reference diagram for displacements and stresses.
4.16. Horizontal displacements of cell due to arc tension loadsat levels H5 and top of cell (Case I).
73
76
77
4.17. Horizontal displacements of cell due to arc tensionloadings at levels H6 and H8 (Case I). 78
Figure Page
4.18. Horizontal displacements of backfilled cell at levelsH5 and the top of cell.
4.19. Horizontal displacements of backfilled cell at levelsH6 and H8.
4.20. Radial displacement of backfilled cell at 0 = 0 and180 degrees.
4.21, Radial displacement of backfilled cell at 0 = 30 and150 degrees.
79
80
81
82
4.22. Radial displacements of backfilled cell at 0 = 60 and120 degrees. 83
4.23. Radial displacement of cell at 0 = 90 and 270 degrees. 84
4.24. Deformed shape of circular cell bulkhead. 85
4.25. Vertical and hoop stresses of steel cell at 0 = 0 degree. 86
4.26. Vertical and hoop stresses of steel cell at 0 = 30 degrees. 87
4.27. Vertical and hoop stresses of steel cell at 0 = 60 degrees. 88
4.28. Vertical and hoop stresses of steel cell at 0 = 90 degrees. 89
4.29. Vertical and hoop stresses of steel cell at 0 = 120 degrees. 90
4.30. Vertical and hoop stresses of steel cell at 0 = 150 degrees. 91
4.31. Vertical and hoop stresses of steel cell at 0 = 180 degrees. 92
4.32. Hoop force in cell vs. position. 93
4.33. Shearing stress in cell vs. position. 94
4.34. Contours of vertical stress in soil. 95
4.35. Contours of radial stress in soil. 96
4.36. Contours of shear stress (TrZ) in soil. 97
Figure Page
4.37. Contours of circumferential stress in soil. 98
4.38. Radial and vertical stresses in soil vs. depth of cell. 99
4.39. Radial and vertical stresses in soil vs. depth of cell. 100
4.40. Coefficients of lateral earth pressure inside the cell fill. 101
4.41. Comparison of radial deformation of backfilled cellat level H5. 102
4.42. Settlement of the top of the steel sheet piles vs. position. 103
4.43. Comparisons of hoop force in steel cell. 104
Appendix
E. 1. Example of finite element mesh showing node and elementnumber scheme. 143
E. 2. Boundary pressure sign convention. 152
F.1. Flow diagram for circular cell bulkhead analysis. 159
LIST OF TABLES
Table Page
4.1. Mechanical properties of soil. 57
5.1. Comparisons of results. 114
Appendix
B. 1. Fourier force coefficients. 134
C. 1. Gaussian weighting functions and stations. 137
D. 1. Modulus of elasticity of soil. 140
LIST OF TERMS
A Cross sectional area of triangular element
{A} Generalized displacement coordinate in soil
an Fourier coefficient for harmonic number n
ar Radial acceleration
a Axial acceleration
{B} Generalized displacement coordinate in shell
C Shell contour (circle)
[C], [Ciiki] Material constant tensor field
D Diameter of cell
d Distance from the mid -plane to the surface of shell
E Young's modulus
e Void ratio
{F} System load vector
{f} Element load vector
{Fb}, {fb} Body force vector field
{Fs}, {fs} Surface traction force vector field
G Shear modulus of elasticity
Gs Specific gravity
H Height of cell above dredge line
Hd Embedment depth of cell
h Depth of soil above point
Volume integrals
K Coefficient of lateral earth pressure
Ka Rankine coefficient of active earth pressure
[K] System stiffness matrix
[k] Element stiffness matrix
Length of shell element
M,. Shell moment tensor field
M Longitudinal moment in shellzz
00Hoop moment in shell
Mz0 Twisting moment in shell
N,. Shell stress resultant tensor field
N Longitudinal force in shellzz
00Hoop force in shell
Nz0 Shearing force in shell in z-0 plane
n Harmonic number
Lateral earth pressure due to backfillPa
PiPrescribed surface traction on surface S
qb Vertical pressure due to backfill
r Global radial coordinate (in cylindrical coordinate system)
ra Radius of connecting arc
rc Radius of circular cell
S Portion of surface on which stresses are prescribed
[5] Stiffness matrix of soil element in generalized coordinates
[SS] Stiffness matrix of shell element in generalized coordinates
Ta Tension in connecting arc per unit length
t Shell thickness
U Strain energy
{U} Nodal point displacement for entire structure
{u} Displacement vector field
u Radial displacement
V Volume of the soil (solid) element
v Circumferential (tangential) displacement
W Strain energy density
w Axial displacement
Z Global axial coordinate (in cylindrical coordinate system)
z Local longitudinal coordinate of shell
a Rotation of the shell surface in longitudinal axis
13 Transverse shell coordinate
Total unit weight of soil
Buoyant unit weight of soil
Yd Dry unit weight of soil
Unit weight of water
ZShearing strain in r-Z plane
Yr0 Shearing strain in r -8 plane
ZOShearing strain in Z-0 plane
5 Variation symbol
E, {E..} Strain tensor field
E rr' E E00 Radial, vertical and circumferential strains,
respectively
{o-..} Stress tensor field
o-o Average confining pressure
err' ZZ' o-00
Radial, vertical and circumferential stresses,
respectively
Cr Zi' CrZoVertical stresses in shell on inside surface and outside
surface, respectively
Lei' 0-00 Hoop stresses in shell on inside surface and outside surface,
respectively
Shearing stress in r-Z planeTrZ
Shearing stress in r-0 planeTr0
Shearing stress in Z-0 planeTZO
X Shell curvature change tensor field
Shell curvature change in longitudinal directionXzz
X 00Shell curvature change in hoop direction
Xz0 Shell curvature change due to twisting moment
v Pois son's ratio
p Density
Angular velocity
[X 0Shell coordinate transformation matrix
[],{(1)} Displacement expansion tensor field
[v]
(i)
(i)
Strain generalized coordinate tensor field
Inclination angle between local and global coordinates of shell
element
Angle of internal friction of soil
[f0 ] [10] Displacement transformation matrix
Total potential energy of the entire systemTrp
Total potential energy of element eTrpe
[(POI' ['POO] Shell displacement transformation matrix
0 Circumferential coordinate (in cylindrical coordinate system)
FINITE ELEMENT ANALYSIS OF CIRCULARCELL BULKHEADS
I. INTRODUCTION
A cellular bulkhead is a waterfront retaining structure formed
from a series of interconnected straight web steel sheet pile cells and
filled with soil, usually sand or sand and gravel. The combination of
steel and soil fill, which individually are unstable, forms a stable unit
offering resistance to its own gravity loads, lateral loads of water and
earth and surcharge loads. Current cellular bulkhead design methods
have adapted methods used for cellular cofferdams which are still
essentially empirical. Although various theories have been suggested
to derive analytical solutions for the stresses in the cell, so far most
designers in this field still rely heavily on past practice and experi-
ence.
Terzaghi (23) proposed an important method for design. In
using this method, a cellular cofferdam on a rock foundation is first
considered a rigid gravity structure. He proposed essentially a
simple bending theory in which a linear distribution of normal stress
on the base of the cofferdam is assumed and lateral pressures on
vertical shearing planes are obtained from an assumed coefficient of
earth pressure. The shearing stress in the fill and friction in the
interlocks was considered as a critical factor in cell stability.
2
Terzaghi also included other possibilities of failure: interlock pullout,
sliding on the base, and foundation bearing capacity failures.
The Tennessee Valley Authority (22) follows the same general
method of design as proposed by Terzaghi with some modifications.
In the TVA method it is assumed that the maximum lateral pressure
for interlock design occurs at a point one-fourth of the exposed height
of the cell above the dredge line. The full value of pressure at the
dredge line is not used because of the restraint provided by
embedment of the sheet piles.
Cummings (6) has proposed a method of cellular cofferdam
analysis known as the interior sliding theory where the resistance of a
cell to failure by tilting is gained largely through horizontal shear in
the cell fill. Cummings' conclusions are based on model tests which
indicate that the plane of rupture goes from the top of the pressure
side to the bottom inner corner (toe of cofferdam). The cell fill in
the rupture region acts essentially as a surcharge and only the soil
below the failure plane will develop shear resistance. The Cummings'
method for determining interlock tension and sliding on the base is
the same as the Terzaghi method.
1. 1 Statement and Scope of the Problem
Analyzing the soil-steel sheet pile interaction of the cellular
bulkhead problem is difficult because the structure consists of two
3
very different materials. In addition, the bulkhead system is
subjected to poorly defined non-axisymmetrical loading and the
boundary conditions of the system are complicated.
The objective of this research is to develop a rational approach
for stress analysis of circular cell bulkheads and to compare
analytical results with data obtained in full scale field studies by other
investigations. To do so would not only provide an aid for predicting
the most likely mode of failure, but would also provide an insight to
the elastic stress distribution throughout the structure.
To deal with the circular cell bulkhead problem, certain
assumptions have been made. Specifically, they are as follows:
1) The soil is elastic and isotropic.
2) The thickness of steel sheet piles is uniform throughout the
whole structure and no slippage occurs along the interlocks.
3) The soil-structure interface is perfectly rough with no
possibility for slip.
4) The actual continuous circular bulkhead system can be
represented by a single circular cell which is considered as
an axisymmetric structure subjected to non-axisymmetrical
loading.
5) At some finite distance beneath the bottom edge of the sheet
piles, the soil foundation is rigid and rough.
4
The scope of the study includes development of a method for
analyzing circular cell bulkhead problems and the development of a
computer program to make the necessary calculations. Two load
conditions are studied:
1) the gravity load of the cell fill and
2) external loading due to backfill.
These loading conditions are considered as axisymmetrical and
non-axisymmetrical loads, respectively.
1. 2 Method of Solution
A typical circular cell bulkhead is shown in Figure 1.1 wherein
the structure is exposed to water on the river side and soil fills on the
shore side. The system is assumed an axisymmetric structure sub-
jected to non-axisymmetrical loading. The loading, however, is
symmetric about a vertical plane containing the axis of the cell.
Because the bulkhead system is not a continuum, an assumption
concerning the characteristics of the soil-sheet pile interface is
needed. It is assumed that the interface is perfectly rough, with no
possibility for slip between the soil and the sheet pile. Also, the cell
itself is assumed to be a continuous circular cylindrical shell. The
structure is thought of as being composed of two substructures:
solid soil and the cylindrical steel shell enclosing it.
0°. -
Back
I) APlan view
Steel sheet pile
- Inside fill
Dredge line
Section A-A
Backfill
Figure 1.1. Typical circular cell bulkhead structure.
5
6
The basic concept of the finite element method is the idealization
of the actual continuum as an assemblage of discrete structural ele-
ments, interconnected at a finite number of joints or nodal points.
For structures that are physically axisymmetric, i. e. , geometrically
axisymmetric and possessing material properties that are axisym-
metric, the nodes are actually circles and are called nodal circles.
Figure 1.2 illustrates a typical triangular element and a shell ele-
ment. The generalized displacements and loadings of a nodal circle
can be expressed in terms of finite Fourier-series and the problem is
uncoupled in each harmonic (27).
For each harmonic the system stiffness [Kn ] and system load
vector {Fn} are formed by summing appropriately the stiffness
[kn] and force {fn} for the discrete elements of the structure,
where subscript n denotes the nth harmonic. After the system
stiffness and load vectors are obtained, the structure is analyzed by
the standard stiffness method (14). Since the bulkhead system is
assumed as linear and elastic, the principle of superposition allows
a solution for the specified load by simply adding the separate solu-
tions that are obtained from the separate Fourier harmonic terms of
separate load components.
The circular cell bulkhead can be represented by shell elements
and quadrilateral elements as shown in Figure 1.3. Boundary 1 is
assumed to be a firm foundation and can be considered fixed. Boundary
7
Z (Axis of
(a) Triangular ring elenlent
Figure 1.2. Typical triangular soil element and She element.
0
4-Axis of symmetry
Shell element
Soil element
Interface condition
Boundary 2 --,...,.
il
8
r01"Boundary 1
Figure 1.3. Finite element simulation of a circular cell bulkheadsystem.
9
2 is assumed to be far from the axis of symmetry OZ and beyond
the zone of failure. Therefore, boundary 2 can be assumed to be on
rollers.
The finite element analyses yield the distribution of stresses
in the soil elements, the steel sheet stresses and nodal point displace-
ments.
10
II. FINITE ELEMENT FORMULATION
The basic finite elements used in this study are quadrilateral
axisymmetric ring elements for the soil and cylindrical shell-of-
revolution elements for the steel sheet pile with constant cross -
section. The quadrilateral element is composed of four sub-triangles.
To formulate such elements, the Theorem of Minimum Potential
Energy is used.
To apply the minimum potential energy theorem it is necessary
to assume a displacement field in terms of a set of unknown Ritz
parameters (coordinate functions) that satisfy the hypotheses of the
theorem (5). The restrictions on assumed displacement functions are
that they be continuous over the entire body and possess piecewise
continuous first partial derivatives. The method of analysis is to
subdivide the domain into an assemblage of discrete elements and
assume appropriate kinematic functions within each element such that
the compatibility conditions across the element interfaces are
satisfied.
The contribution to the total potential energy of one element
can be written as
Tr pe = U .SSS fiu.dV p.u.dS (2. 1)
11
where
U= irSicWdV
V
U is the element strain energy and W is the strain energy density
which is assumed to be positive definite. The body force vector field
f. and the prescribed surface traction p. on the portion of the
surface S are positive if they act in the direction of positive coordi-
nate axes. u. is the displacement vector field in the volume V.
The strain energy density can be written as
W 1 (2.2)3
-
where o-.. is the stress tensor field and E.. is the strain tensor
field.
The total potential energy of the entire system is the sum of the
potentials of the individual elements. Thus, for a system of L
elements
L
IT = ITp pe
1
(2. 3)
An absolute minimum potential energy is sought by taking the
variation of the potential energy function with respect to the discrete
displacement variables and set it equal to zero
8Tr = 0P
12
(2.4)
Then a set of matrix equations are formed accordingly.
2. 1 Triangular Axisymmetric Ring Element Matrix Equations
A triangular axisymmetric ring element is used in the study as
shown in Figure 2.1. Matrix notation and cylindrical coordinates are
used in the analysis, that is, radial distance r, axial distance Z,
and circumferential angle 0. The right handed system is used in
the coordinate system. Since the soil is assumed a linear elastic
material, the stress-strain relationship may be expressed in the
form of the constitutive equation (7)
IT = Ci. Elj jki k/
or, rewritten in matrix form
(2. 5a)
{a} = [C] {E} (2.5b)
where Cijki is a fourth-order material constant tensor.
If Equations (2. 2) and (2. 5b) are substituted into Equation (2. 1),
the element potential energy becomes
13
Figure 2.1. Triangular axisymmetric element.
14
Trpe = S S S 2 {E(r, Z, 0)}T[C]r{E(r, Z, 0)}dV
V
gir {u(r, Z, 0)}T{f(r, Z, 0)}dV
V
SS {u(r, Z, 0)}T{P(r, Z, 0)}dS (2. 6)
in which
{ } denotes a column or a row matrix
[ ] denotes a rectangular or square matrix
The superscript T denotes the transpose of the matrix.
For an isotropic material with Young's modulus E and Poisson's
ratio v , the constant C will be
1-v v
1-v
0 0
0 0
0 0
v
v
1-v
0
0
0
0
0
0
1-2v
0
0
0
0
122v
0
0
0
0
0
1-2v
2. 7)[C] = [C]T (1 +v)(1-2v)2
0
0 02
The displacements within each triangular element are assumed
to be a linear function of the coordinates r and Z (27).
un = Aln + A2nr + A3nZ
wn A4n + A5nr + A6nZ
vn = A7n + A8nr + A9nZ
15
(2. 8a)
(2. 8b)
(2. 8c)
where u, w, and v are radial, axial and circumferential displace-
ment components, respectively. Subscript n is an integer called
the harmonic number of the Fourier series. As are constant coef-
ficients that represent the generalized displacement coordinates of
the element.
In general cases, an axisymmetric structure subjected to
arbitrary loadings, displacements and the body and surface loads are
represented by Fourier series as follows:
N
u(r, Z, 0) = un r, Z)cos n0 +) un(r, Z)sin nO (2. 9a)
n=0 n=1
N N
w(r, Z, 0) = wn
r, Z)cos ne + wn(r, Z)sin ne (2. 9b)
n=0 n=1
N N
v(r, Z, 0) = vn r, Z)sin nO + r, Z)cos nO (2. 9c)
n=1 n=0
body forces are:
N N.....
fr (r, Z, 0) = f (r, Z)cos nO + f (r, Z)sin nern rnn=0 n =1
fz(r, Z, 0) =
fe(r, Z, 0) =
n=0
N
n(r, Z)cos nO + f Zn(r, Z)sin nO
n=1
r, Z)sin nO +
n=1
N
n=0
16
(2. 10a)
(2.10b)
r, Z)cos nO (2. 10c)
and surface loads are:
pr(r, Z, 0) = prn(r, Z)cos nO + prn(r, Z)sin nO
n=0 n=1
(2. 11a)
N N
pZ
(r, Z, 0) = pZn (r, Z)cos no + / pzn(r, Z)sin nO
n=0 n=1
p0(r, Z, 0) = L pen(r, Z)sin nO + TiOn (r, Z)cos no
n=1 n=0
(2.
(2.
11b)
11c)
where the single barred ( - - ) and double barred ( -) quantities
represent functions of r, Z and n only but not 0.
Since the circular cell bulkhead system as shown in Figure 1. 1
has symmetric loading with respect to the rZ plane at 0 = 0, the
double barred series in Equations (2.9), (2. 10) and (2. 11) will not be
17
used in this study. The axially symmetric case is represented by use
of only the n = 0 term of the single barred series.
Substituting Equations (2.8) into Equations (2. 9), the
displacements become
u(r, Z, 0)
w(r, Z, 0)
v(r, Z, 0)
n=0
where
{4)(r, Z) }Tcos nO 0 0
0 {4(r, Z)}T cos nO 0
0 0 {4)(r, Z) }Tsin nO
{4(r, Z)}T = {1 r Z}
Equation (2. 12) can be written symbolically as
N
(2. 12)
i{u(r,Z, 0)} = [l(r, Z, 0)] {A. = 1, 2, , 9 (2. 13 )in
n=1
Thus the displacement vector field for each Fourier term may be
written as
{u (r, Z)} = ['(r, Z)]{A. } (2. 14)
18
Let u., w., and v. refer to the r, Z and 0 direction
displacement components of any corner node j of the element as
shown in Figure 2.1. Using the displacement boundary conditions for
u at each corner, the following matrix expression is obtained:
1 r Aln
1 r2 Z2 A2n (2. 15a)
Similarly, the displacements w and v can be expressed in terms
of generalized coordinates A. as follows:in
and
l
{f2nrnv
v3n
rl
r2
r3
Z2
Z3
5n
A6n
(2. 15b)
1
r2
r3
Zr1
Z2
Z3 A9n
(2. 15c)
Equations (2. 15) can be combined together and written in symbolic
form as
fu on} [ Y{Ain} (2. 16)
where uOn is nodal displacement and /0
A.
19
is as shown in Appendix
The generalized coordinates A. are obtained by inversionin
of Equation (2. 16) and expressed in terms of nodal point displacements
as
{Ain} = [ 0 1}{uOn}
1 is shown in Appendix A.
(2. 17)
If Equation (2. 17) is substituted into Equation (2. 14), the
displacements at any point in the triangular element can be written
in terms of the nodal displacements as
w
n1
0
0
0
n1
0
0
0
n1
n2
0
0
0
n2
0
0
0
n2
n3
0
0
0
n3
0
02A
where
n1 al + blr + c1Z
n2 a2 + b2r + c2Z
n3 = a3 + b3r + c3 Z
(2. 18)
(2. 19a)
20
al = r2Z3 - r3Z2, b1 = Z2 - Z3, c1 = r3 r2
a2 = r3Z1 - riZ3, b2 = Z3 Z1, c2 = r1
r3
(2. 19b)
a3 = riZ2 r2Zi, b3 = Zi Z2, c3 = r2
- r1
and A is the cross-sectional area of the triangular element.
The element strains are obtained by differentiating Equation
(2,18) to obtain (15)
au
rr araw
'zz az1 , aV
+E u)OA r ae
= 2E =au
+8w
rZ rZ 8Z ar
1 au2E = v) + av
r8 r0 r 80 8r
= 2e ay 1 awze ze az + ae
(2. 20)
The strain tensor can be expressed in terms of generalized
coordinates
{E (r, Z, 0)} = (r, Z, 9)1{A. } (2. 21a)
n=0
where Is' is shown in Appendix A.
If Equation (2. 17) is substituted into Equation (2. 21a) a relation
between element strain tensors and the nodal displacements is
obtained
{E (r, Z, e)} =
n=0
P (r, Z, 0)Dvillu }On
21
21b)
Substituting Equation (2.21b) into Equation (2.6), the total
potential energy of a single triangular element can be written in terms
of its nodal displacements
N M
ape= OiSic(uif
2 On L 0 n(r,Z,O)]T[C][Vm(r,Z,0)][fiNu
Om}
0
n=0 m=0 V
IJS
{ti0n}T[0-1]r[tt.n(r, Z, 0)]T{fm(r, Z, 0)})dV
u n}T
-1 ]T['/n(r, Z, {pm (r, Z, 0)}dSj (2.22)
0
where n and m are harmonic numbers.
Since the 0-dependence in the integrals in Equation (2. 22) is
known explicitly from variables r and Z it can be carried out
directly. At the same time by making use of the orthogonal proper-
ties of the harmonic functions the integrals are integrated from
0 = -Tr to 0 = Tr. These integrals are a well-known type in Fourier
analysis. The general forms are stated as follows:
22
sin m0 sin nOdO =-Tr
for m = n 0
for m n and for m = n = 0
2Tr for m = n = 0
.51 Trcos me cos n0d0 =-Tr
1T for m = n 0 (2. 23)
0 for m mn
iTsin m0 cos n0d0 = 0 for all m and n-Tr
Thus, the sum in Equation (2. 22) exists only for m = n and it
becomes
N
wpe US'S( 2 Onfu }TW0 lir[V (r, Z)]T[CPn'(r Z)P0 1] {u0n}
n=0 V
Let
and
TTS
{uOn
T0
I1]T[n(r, Z) ]T {fn(r, Z)})dV
n}T -1 T Z)]T{pn(r, Z)}dS)
{fbn(r, Z)} = [n(r, Z)]T{fn(r, Z)}
{fsn(r, Z)} = [n(r, Z)]T{Pn(r, Z)}
[sn(r, Z)] = [nt(r, Z)]T[C][P (r, Z)]
Substituting Equation (2. 25) into Equation (2. 24) we obtain
(2. 24)
(2. 25a)
(2. 25b)
(2. 25c)
T -1 T -1IT ssy(1, } [s (r, Z)P, 1{u }
pe 2 On 0 u On
n=0 V
ffS
= {uOn}T
[/-1 ]T{fbn(r, Z)})dV
0
T 1 Tn}
[.10 ] If (r, Z) }dSsn
The body force vector will be (21)
{f(r, Z, 0)}T = {prw2-par cos 0, -pa -par sin 0}
where p is the mass density of the body,
23
(2. 26)
(2. 27a)
w is the angular velocity,
and ar and az are radial and axial acceleration, respectively.
After substituting Equation (2. 27a) into Equation (2. 25a) and
dropping the 0-dependence, the body force components of the triangu-
lar element are
and
{fbn(r, Z)} T{prw
2, pr2 w2,
prZw2, -pa z' -prat, pZa Z' 0, 0, 0}
for n = 0 (2.27b)
{fbn(r, Z)}T = {-par r r r r, -pra, -pZa, 0, -pa, -pra, -pZa
for n = 1 (2. 27c)
{fbn(r' Z)} = 0 for n > 2 (2. 27d)
24
The body force components in the direction of r and 0 are
included only for generality and will not be used in the analysis of
circular cell bulkheads. The surface traction vectors will be inte-
grated explicitly for each loading because they are arbitrary. The
area integration is performed over the triangular area in the r-Z
plane by denoting
[Sn] =Cg [sn(r, Z)]dV
Matrix [Sn
] is shown in Appendix A.
The system body force and surface traction vectors are
SIS {fbn(r, Z)}dV
V
{F sn} = g{fsn(r, Z)}dS
The following notation is used for the various integrals:
11 ,acciv, 16 as dV
V V
12 =ffidV, 17 = siscrdVV V
(2. 28a)
(2. 28b)
(2. 28c)
(2. 29)
25
I3
= S.11c dV, I8
= ZdV
V
I4
= .17 dV, 19
= ygr ZclV (2.29cont. )
V
15 SYS dV, 110 = SSSrZdV
V
After the integration in Equation (2. 28b) is carried out, the body
force vector may be written as
{Fbn}T = {pw2I p2I9'
pw2I10'
Z-pa_I
7, -pa
ZI
8' 0, 0, 0} (2. 30a)
for n= 0
{Fbn}T = {-pa rIF -parI 7'-par 0, 0, 0, -pa rIV -par - parI8} (2. 30b)
for n = 1
For a single finite element the potential energy becomes
N
pen=0
On}T L 0 [ 5 ] 0 ] {u0n}
{uOn}T 0 {Fbn}+{Fsn} }) (2. 31)
Thus for the whole system of L such elements, the total
potential energy in Equation (2.3) is
Tr
p
L N
ff =1 n=0
{Un 0 0}T[OVn[S(1 )][0)]{Un}
{U }T[ 0(i)]rf {F" ")}+{F)} })n bn sn
26
(2. 32)
where {un} is the discretized displacement vector for the entire
system and [1. (/) ] is the generalized coordinate transformation
matrix of element f to the displacements in the entire system.
Taking the variation of the stationary potential energy with
respect to the discrete displacement variables as stated in Equation
(2.4) we obtain
where
L N
=1 n=0
([%-1(1 )]T[S(1)P0-1(1)1{Un}-[%-i"V{F")}) = 0 (2. 33)
{F(i)} = {FM} + {Fsn)}bn
From Equation (2. 33) a set of governing simultaneous equations
can be written for each Fourier harmonic term as
[Kn] {lin} = {Fn} for n = 0,1,2, N (2. 34)
where Kn is the stiffness of the assembled system and Fn is the
system load vector. They may be expressed as follows:
and
L
[Kn] = [0-101)]TisCPn I- 0
f =1
{Fn} =
L
1=1
-1(.e) T ( )] {Fn
)}0
in which [1.01(/)] is shown in Equation (A. 2) in Appendix A.
From Equation (2. 35a), the stiffness of element
Fourier harmonic n can be expressed as
[k(rif )] 01(1)]r[s(f)][
)1(.1 )]
2.2 Quadrilateral Element Matrix Equations
27
(2. 35a)
(2. 35b)
in any
(2. 36)
The use of the quadrilateral as the discrete element to idealize
the system is desirable since it reduces the required input in the
computer program and the resulting set of equilibrium equations has
fewer unknowns for a given number of triangular elements. A typical
quadrilateral element is composed of four triangles as illustrated in
Figure 2.2. The coordinates of the center node are computed as the
average of the four corner point coordinates.
In the case of non-axisymmetric loads, the 12 degrees of
freedom quadrilateral element matrix is formed by first combining
the four 9-degree-of-freedom triangular element matrices into a 15
degree of freedom element matrix. Using a process of static
condensation (5, 28) the three internal displacements are eliminated,
resulting in a 12 degree of freedom quadrilateral element matrix.
4(10, 11, 12)
1(1,2,3)
28
Numbers in bracketscorrespond to unknownat each node
Figure 2.2. Quadrilateral element.
The four triangular element stiffness are combined by the code
number technique (14, 24). If the load vectors {F0} for each of the
triangular elements are similarly superimposed, a partitioned matrix
equation is obtained
kaaS
ab
lc_b a kbb_
(2.37)
where subscript a is associated with nodal points 1 to 4 and
subscript b is associated with point 5. Equation (2. 37) may be
written as two matrix equations
{Fa} [kaa]{ua/ kabilub/ {F0a}
{Fb} = [kba] {ua} + [kbb] {ub} + {F0b}
Equation (2. 38b) can be solved for the displacements ub:
{ub} -[kbb]-l[kba]1
}- {F0b }}
29
(2. 38a)
(2. 38b)
(2. 38c)
If Equation (2. 38c) is substituted into Equation (2. 38a), an
expression is found relating the forces at points 1 to 4 to the
unknown displacements at points 1 to 4 and the known loads as
where
{Fa} [k*aa] lua}{4-a}
J.
[k ] = [k]
[k ilkaa aa ab bb ha]
the quadrilateral element stiffness matrix, and
{F0a} = {F0a} [kab][kbb] -1{FOb
}
the modified load matrix.
2.3 Shell Element Matrix Equations
Finite element analysis of shells of revolution has been
(2. 39a)
(2. 39b)
(2. 39c)
developed and used for axisymmetrical loadings (9, 10, 19). For non-
axisymmetrical loadings, the displacements and loadings are expanded
in Fourier series (12, 18, 21, 27). A typical shell element is
illustrated in Figure 2.3.
Z, w
p,
30
Note: v is positivewhen directedinto the paper
9,v r, u
Figure 2.3. Shell element and coordinates.
For thin shells of revolution, the potential energy of a single
element may be written as (7)
where
1Trpe = {
2(N E
j+1\4..X
1.)ds} p.1
1u.dC
1 13 3( 2.40a)
t /2N.. = o-..df3 (2. 40b)
-t/2
t/2
"
M.. = cr..Pd13 (2. 40c)13 -t /2
31
N,., M.., E.. and x.. are stress resultant, moment,13 13 13 13
extensional strain and curvature tensor fields, respectively, acting
on the portion of shell surface S. p. and u. are nodal loads and
nodal displacements along the nodal circle C, and t is the shell
element thickness assumed as constant throughout the whole cell, p
is the local transverse coordinate (see Figure 2.3).
The displacement field of the conical shell may be assumed as
w'(z) = B1 + B 2z
v'(z) = B3 + B4z
u'(z) = B5 + B 6z + B z 2+ B8z3
3
7
(2. 41a)
(2.41b)
(2. 41c)
The shell displacements of non-axisymmetrical loading case
also can be expanded into Fourier series:
w'(z, 0) =
n=0
yr' (z)cos nOn(2. 42a)
v'(z, 0) =
n=0
v' (z)sin non
(2. 42b)
u'(z, 0) = un(z)cos no (2. 42c)
n=0
where v' and w' are the transverse, circumferential or
tangential, and longitudinal shell displacements, respectivey, with
32
respect to the load element coordinates. z is the longitudinal shell
coordinate and n is the Fourier harmonic number.
The coefficients B1, B2, ..., B8 represent the generalized
displacement coordinates of the shell element. The number of con-
stants B are assumed equal to the number of degrees of freedom
of the element. Each nodal point has four degrees of freedom that are
denoted by u', v', w' and a. Translation u', v', w' have been
defined previously; a is the rotation of the meridian in the shell
surface in a plane which passes through the nodal point and contains
the axis of revolution of the shell. The rotation a is positive if it
corresponds to a positive value of aultaz along the meridian.
By using the boundary conditions, the nodal displacements in
Equation (2.41) can be written as
un' (1)-1
wi(1)n
v.' (1)n
au' (1)
az.
u' (2)n
WI (2)n
vI (2)n
au' (2)n
az
0 0 0 0 1 0 0
1 0 0 0 0 0 0 0
0 0 1 0 0 0 0 0
n
2n
3n
n
n
n
7n
n
(2. 43a)
or in symbolic form
{uon} = [90]{Bin} i = 1 2, , 8
where I is the length of the shell element.
The constants {B.}
are obtained by inversion asin
33
(2. 43b)
{Bin} = [(po- l]{uol
n} (2.43c)
where [cp0
1] is shown in Equation (A. 12) of Appendix A.
Introducing a shell displacement transformation matrix, the
relationship between local and global coordinate displacements are
where [x0]
Appendix A.
{uon} = [xo]{uon)(2. 44a)
is the shell transformation matrix and is shown in
{uOn}
is the global coordinate displacements.
If we substitute Equation (2. 44a) into Equation (2. 43c), the
generalized coordinates become
{Bin} = [(P0 1][X 0i{u
On}
or
{Bin} [`P00]{u0n}
where
[(p00] [C201][X0]
(2. 44b)
(2. 44c)
[9001]is shown in Equation (A. 15) of Appendix A.
According to the Novozhilov theory (17) of thin shells, the
strain-displacement relationships are
aw'E -
ZZ aZ
1 , ayae 'E00 r ( + L.1+w'1 cosCOS +w' sin (0)
.Yz0
ay' 1 aw'az + (r De
a2u,Xzz =
az 2
- sin 4:v')
1 ay' a2 12'X = (cos
ae00 r 2 4)-30 2 )
1 au'- r az sin (I)
1 au' 1 a 2u' lay' 1Xz0 = 42 sin 4) ae
- 1-7azae +(r.az
-r r
34
(2.45)
sin 4)Ocos 4)]
where E zz' E 00' and Ez0 are the longitudinal, normal and shear
strains; xzz , X00 and Xze are the curvature changes of the mid-
dle surface of the shell.
The general stress-strain relationships for a linear isotropic
elastic thin shell are (17)
Nzz
00
Nz0
M zz
MOO
Mz0
Et
1
v
0
0
0
0
v
1
0
0
0
0
0
0
1-v
0
0
0
t 2
0
0
0
vt2
2
0
0
0
21-v 122
vt
122t
12
0
12
0
or written symbolically as
where
{o} = [c]{ E}
0
0
0
0
35
(2.46a)
(2. 46b)
The potential energy in Equation (2. 40a) can be expanded as
13 13
N N
{Sln("T[Nnm(0)] {Slm(z)}
N
3M13 ..X1... =
m=0
N
n=0 m=-0
n(z)}T[Mnm(0)] {S2m(z)}
avI v'{S (z)}
T n n n n nln az az ' r r
a2 av' au' V'
{s(z)}T{n 1 n 1 nnni2n az2 r az r az r2 r2
[Nnm(0)] and [Mnrn(0)] are developed in Appendix A.
(2. 47a)
(2. 47b)
(2.47c)
(2. 47d)
36
Due to the orthogonal properties of the Fourier harmonic
functions in the period of -Tr < 0 < Tr, the sum of the potential
energy exists only for m = n.
If Equation (2. 41) is differentiated and substituted into Equation
(2.47c), {S in(z)}
{S1(z)} =
will become
0 1 0 0
0 0 0 1
10 0r r1
0 0r r
0 0 0 0
0
0
0
0
1
r
0
0
0
0
zr
0
0
0
0
z2r
0
0
0
0
z3
In
B2n
B3n
B 4n
B5n
B6n
B7n
8n
(2.48a)
r_
or written in symbolical form
{S1(z)} = [Xi(z)1{Bid (2. 48b)
Substituting Equation (2. 44b) into Equation (2.48b), {S1(z)}
may be written in terms of nodal displacements as follows
{S1(z)} [X1(z)][4900]{ti0n} (2.48c)
Similarly, if Equation (2.41) is differentiated and substituted into
Equation (2. 47d), {S2n(z)} can be expressed in terms of
generalized coordinates and nodal displacements, respectively, as
{s 2(z ) }=
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
r0
0
0
0
0
1
0
0
1
r
0
z
2
0
2z
6z
0
3z 2
r
0
z2
0
z3
2r0
2r0
2r
2r 2r 2r
or in symbolical form
and
{S2(z)} = [X2 i n
(z)] {B }
)} [x2(z)][900] lu0n}
37
(2. 49a)
(2. 49b)
(2. 49c)
The terms [X 1(z)] and [X2(z)] in Equations (2. 48b) and
(2. 49b) can be expanded further as
where
X1 (z).= G.. x.(z)
X2(z). = Hijkx (z)j
(2. 50a)
(2. 50b)
2, T 1 z z z
31 z z
2z
3{ zx.( )1 =
r,z, 2 2 2 2
, (2.50c)r r r r
38
and
G112 = G214 = G321 = G332 = G423 = 0434 = G525 = G536
= G547 = G558 = 1
H224 = H326 = H473 = H484 = H575 = H586 = H597 = H5108 = 1
H117 = H337 = 2
H348 = 3
H168 = 6
All other G's and H's are zero.
After substituting stress and strain vector fields into Equation
(2. 40a) and using the advantage of 0-dependence being known explicitly,
the 0-integration can be performed directly and the potential energy
becomes
apepen=0
({110n}
T[90-013T [ [X1 z)]
T [Nn][X i(z)]
+ [X2(z)3T[Mn][X2(z)1][9003 {uOn})di
{u0n}T [C900}T{Fzn}(2. 51)
where Q is the integration over the length of the shell element
which is carried out by using numerical integration of the Gaussian
Quadrature Formula (4, 16).
39
Introducing
[Ssn] = S([X1(z)]T[Nn][X1(z)] +[X2(z)]T[Mn][x2(z)])d/ (2.52a)
Equation (2.. 51) can be rewritten as
is
N
/Trpe 2(1 {u
On}T[900-1] T[SS
n ][900-1
] {u n}-{uOn
}T [yo001
] {Fzn})
n=0 (2. 52b)
For the assemblage of L elements, the total potential energy
L N
( fun
}T[q)00
-1(0]T[ss(nikp00 -1(/ )llu2
1=1 n=0
{un}T[P00(fT
{Fzn} ) (2.53)
The governing equations are obtained by performing a variation
on the total potential energy of the system with respect to the discrete
displacements and setting them equal to zero. Accordingly
Ear = 0p
which yields
N L
([901(1 )]r[SSn(1 )1[490-01(1)]{un) = [(P001(1)}{Fzn )})
n=0 1=1
(2. 54)
40
For each Fourier harmonic term
[Kn]{Un} = {Fri}, n = 0,1,2, . . . , N (2. 55a)
where [Kn
] is the system stiffness matrix of the shell element and
is given by the expression
L
[Kn] [(P 1" )]T[SSn" )1[9010" )]00.Q=1
[Fn] is the system load vector of the shell element
L
n1(1) T Fzn}
L 1-1' 001 = 1
The corresponding shell element stiffness matrix is
(1)[k ] [q)
-1( )][SS(1)][0-1()1n 00 n '00
2.4 System Equations and Solution Process
(2. 55b)
(2. 55c)
(2. 55d)
As the element stiffness matrices [kn] for each Fourier
harmonic term are generated, they are appropriately superimposed
into a system matrix [Kn]. The superposition is accomplished by
using the code number technique (14, 24).
A structural system that contains a large number of elements
will involve a large amount of data preparation. This includes a code
41
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.
The load vector {Fn} for each harmonic is assembled at the
same time as the structural stiffness matrix is formed. They are
generated using the code numbers. The total system of simultaneous
algebraic equations in the unknown {Un} is represented as
[Kn] {un} =n}
(2. 56)
The primary concern in the solution of this system is the
conditioning of the system matrix. It is a banded symmetrical matrix
where the band width is dependent on the direction of numbering the
nodal points. The nodal points should be numbered in such a way as
to minimize the difference between the largest and smallest nodal
point numbers for any element. By taking advantage of symmetry,
the coefficients are stored as an upper triangular matrix.
The solution of the equations is obtained using the linear
equation solver BANSOL. This subroutine uses the Gaussian elimina-
tion method. The band-width is automatically computed prior to
solving the system as it is required input to the solution subroutine
along with the system load vector.
42
III. TESTING OF COMPUTER PROGRAM
The purpose of this chapter is to present verifications of the
finite -21c.rresnt formulation of this investigation and also the computer
program. T3ecause the circular cell bulkhead system consists of
quadrilateral ring elements (4 triangles) and cylindrical shell elements
and Fourier harmonic series are used, three different types of prob-
lems are analyzed and their results are compared with those of
known classical solutions.
First, an elastic soil layer which overlies a rough rigid base is
analyzed. The soil system is subjected to a constant surface loading
q which is applied uniformly over a circular area of radius a as
shown in Figure 3.1. The soil is assumed to be a linearly elastic,
isotropic and homogeneous material. Thirty-five rectangular ring
elements are used to model the system and the results are shown in
Figures 3.2, 3.3, 3.4, and 3.5. It is seen that there is excellent
agreement with the classical solutions proposed by Burmister (3).
Second, a circular cylindrical tank filled with water is
considered in demonstrating the applicability of the shell element
formulation. The shell is idealized as 12 cylindrical elements and 13
nodes as shown in Figure 3. 6. The tank is subjected to hydrostatic
pressure only. The displacements and stress resultants are shown
in Figures 3.7 and 3.8, respectively. It is seen that there is good
agreement with the exact solution (25) except that the displacements
43
and hoop force at the top of the cylinder are somewhat on the high
side. This is due to the assumed point load at the top nodal point of
the tank.
The third structure to be analyzed is a circular cylindrical thin
shell cooling tower. It is included in this investigation in order to
verify and demonstrate the power and versatility of using the Fourier
harmonic analysis of the shell portion of the finite element method.
The cooling tower rests on 8 columns and supports its own weight.
The finite element model is shown in Figure 3.9(a). Five elements
are shown in Figure 3.9(a). A second model, not shown, with ten ele-
ments was also considered in which each element was one-half the size
of the first model. The second model was prepared because the first
model proved to be very coarse. Flugge (8) has developed a classical
solution for this problem with the total edge load as shown in Figure
3. 9(b). The approximate loading diagram which is based on a 5 term
Fourier expansion of the given stress function is shown in Figure 3.10.
The results for force resultants and displacements are shown in Fig-
ures 3.11, 3.12 and 3.13, respectively.
From the results it can be seen that the hoop force of the
5-element model is on the high side but the corresponding axial force
is in good agreement with the classical solution. A significant
improvement is achieved with the ten element model. Therefore it is
concluded that the results are quite dependent on the number of elements.
10'
It.
H
H = 10 ft
a = 2 ft
q = 1 ksf.
E = 30 ksi
v = 0.4
///AW//4Y/WwW7ACTUA
Rigid base with rough surface
35 element model
Roller
r
Fixed20'
Figure 3.1. Finite element model of rigid base soil system.
r
44
4-1
2
4
Vertical stress in percent of applied pressure20 40 60 80 100
8
10
Numbers on curves indicate radialdistances in radii
Figure 3. Z. Vertical stress in soil due to surface load.
Radial stress in percent of applied pressure10 20 30 40 50 60 70
0.75 0.25
Numbers on curves indicate radialdistances in radiiExact solution (3)
O a D e Finite element
Figure 3.3. Radial stress in soil due to surface load.
0Tangential stress in percent of applied pressure
10 20 30 40 50 60 70
0.75
0.25
Numbers on curves indicate radialdistances in radiiExact solution (3)
0 p Finite element
Figure 3.4. Tangential stress in soil due to surface load.
Shear stress in percent of applied pressure4 8 12 16 20 24
Nkinabers on curves indicate radialdistances in radii
Figure 3. 5. Shear stress in soil due to surface load.
D
13
12
Figure 3.6. Finite element model ofcircular tank.
9
8
7
6
5
4
32
Node number
4-3a)(1)
30
25
20
15
10
Exact solution (25)
0 Finite element
1 10-30 1 2 3 4 xRadial displacement, feet
Figure 3. 7. Radial displacement ofcircular tank.
50
Longitudinal moment, MZZ
in kips -ft/ft
-0. 4 -0. 2 0 0. 2 0. 4
Exact solution (25)
0Finite element
Note: Positive momentcauses compressioninside tank
-40 -30 -20 -10 0 10 20 30
Hoop force, N00, kips /ft40
Figure 3. 8. Hoop force and longitudinal moment of circular tank.
3a
V
a
Nodalnumber
5-element model
4t
5
43
21
Section C-D
(a) Cylindrical cooling tower
51
Section A-B
E= 30 x 103 ksiv = 0t = 0.24 incha = 3 ftP= 0.4 k/ft
0° 45°
1412°).1 33°
(b) Total edge load at the bottom
90° 135°
2. 75P
Figure 3.9. Edge load applied to the cooling tower shell.
A loading
- Approximation by 5non-zero Fourier terms
2.75P\rn=8
n=32
-P
/
rn.16 /
..I. -\ /
/ =40 // /// /
/ n.24 // /
\ .e-\ / I, ,-, // "4"\ '\ 4/ ` \ i ; \ ;,/ ,r\N I
is .' Ii \ 'i i I r 0 / / V
% i,0 1..... / \ 4eii---4,, \\/ ,,/ 4. iI\i' -/1 /I \\I A', \iAZI,i'..,4t -.4.1./ 11-,_,,,, \I .......,- 1
' /Ni
i
\ I, \ I\ Ai i' 1 ' /\l/ 1 I ...,/ \ / .... I v
_...-- \ 1 / % / ....1
% ,' I
\,A //III I
WVA\
lir Alik VidaV/
\ /\ /\ /\ /
Figure 3.10. Approximate loading diagram for cooling tower.
9 0 °
0 = 22.5° N/P
Exact solution (8, p. 231)
A 5 Element solution
0 A 10 Element solution
0.5
0.4
0.3
0.2
0.1
0
0 = 0°
Figure 3.11. Cooling tower stress resultants.
N /P
4
0.7
0.6
0.5
0.4
0.3
0.2
0.1
w (axial displacement)
100 200 300 400 500
Eu/P
Figure 3.12. Radial and axial displacements for coolingtower at A = 0°.
600
54
55
0. 7
0. 6
0. 5
0. 4
0.3
0. 2
0. 1
u (radial displacement
-w (axial displacement)
100 200 300 400 500 600
Eu/P
Figure 3. 13. Radial and axial displacements for coolingtower at 0 = 22. 5°.
56
IV. ANALYSIS OF CIRCULAR CELL BULKHEADS
Because the mathematical model and computer program are
applied to the analysis of axisymmetrical cell structures whose con-
figurations are shown in Figures 4.1 and 4.2, no difficulty was
experienced in generating the necessary input data. A user's manual
for the program which describes all necessary operations for gen-
erating the data is included in Appendix E. Appendix F describes the
function of the primary subroutines and Appendix G contains a listing
of the program.
Two analyses of the circular cell are presented in this chapter.
They are:
1. An isolated circular cell subjected to gravity load of the cell
fill which simulates the behavior of the cell after the inside
is filled completely as shown in Figure 4.1.
2. In addition to the weight of cell fill, the circular bulkhead is
subjected to the weight and lateral pressure due to backfill.
The general configuration is shown in Figure 4.2.
The circular cell structure used in this study was of similar
overall dimensions to one constructed at Port of Portland Terminal
No. 4 on the Willamette River near Portland, Oregon. In that bulk-
head, the front sheet piles were 18 feet longer than the back ones.
The web thickness of the front sheet piling was 0.532 inch and of the
57
back was 0.406 inch.
Since the computer program was developed for geometrically
axisymmetric structure, the average web thickness of 0. 469 inch was
assumed to be the uniform shell thickness. The average height of the
cell (from the dredge line) was assumed to be 58.5 feet and the depth
of embedment was uniformly 29.25 feet. The water level was
assumed to be at the top of the cell both inside and outside.
Because of the lack of soil test data, some parameters were
assumed and these properties are summarized in Table 4.1.
Poisson's ratio for the soil was assumed to have a constant value of
0.35 throughout the system. The modulus of elasticity was assumed
constant only within each individual layer of elements. Values were
obtained by using Richart's formula (20) as shown in Appendix D.
The modulus of elasticity of the steel was assumed to be 30 x 10 3 ksi.
and Poisson's ratio was assumed to be 0.3.
Table 4.1. Mechanical properties of soil.
TotalDensity
pcf(1)
Degree
Degree ofSaturation
%
VoidRatio
SpecificGravity
Poisson'sRatio
117.5 34 100 0.94 2.71 0.35
Soil is classified as medium dense to dense fine sand.
A 48 soil element and 9 shell element model was generated for
the structure as shown in Figure 4.3. The vertical boundary line and
58
the base boundary line were selected 60 feet and 29.25 feet from the
cell, respectively. The selection of these boundary lines and the
number of elements were based on the storage limitations of the
computer and estimated zone of influence for the cell.
4.1 Isolated Circular Cell
The isolated circular cell as shown in Figure 4. 1 is subjected
internally to the gravity load of the cell fill and externally by water
pressure. Since the water table inside the cell is at the same level
as the adjacent water, the submerged unit weight of the soil is used to
determine the net pressure loading. The system is obviously axisym-
metrically loaded, so only one Fourier harmonic term of n equal
to zero is necessary for the analysis.
The computer results for the isolated circular cell are shown
in Figures 4. 4 to 4. 10.
4. 2 Circular Cell Bulkhead
The general configuration of the circular cell bulkhead is shown
in Figure 4.2. The phreatic line both inside the cell fill and backfill
was observed at the same' elevation of the outside water (11). In this
study, it was assumed at the top throughout the bulkhead system.
Figure 4.11 illustrates the equivalent loadings due to backfill.
The quantity pa is the active earth pressure and can be calculated
Water
60'
Circular cell
D =65. 8'1.4
Plan
Axis of symmetry
59
60'
H=58.5'
H/2
H12
Boundary line1
/VANYZAN,
Cell
)0,y^NY,,,,ANyA
Steel sheet piles
Dredge line-,
ffy-,/,/,//m/,e,//mrt /7//f/f/ ////////////7 i/77/////7///////////////////////4Boundary line Soil foundation
Figure 4.1. An isolated circular cell.
F
r-Water--)
60'
Circular cell
D=65.8'
'Backfill--4
Axis of symmetry
Plan
60'
60
Water --)
H =58. 5'
Dredge line
H/2
H 2
'>/'/".K
1
1
(-Cell fill)
Soil foundation
Y/A \ssr/Ass
sheet piles
Boundary lines
Figure 4.2. A circular cell bulkhead.
//7/
61
Axis of symmetry
O00
Element layer number
Shell element
Soil elements cRoller
r"
r/7/7 /./ 7.7/7/7////////7/7/7/7/,'-o"
Fixed
Figure 4. 3. Circular cell finite element model.
Hoop force, N kips/in.-2 0 00 2
Radialdispl.
Dredge line
-1 0
0
19.5
39.0a)
a)
58.5
78.0
10 outsideo- ZZ1 inside
( : o. utsdidessufraface
surfacesurface
OR.
-
Li4
to
M1/4St1 \-
A/
I_ _ I _______-%-7.5 -5.0 -2.5 0 2.5 5.0
Displacement, inch Sheet stress, kips per square inch(a) (b)
Figure 4.4. Radial displacement, hoop force and sheet stresses due to gravity load of cell fill(axisymmetric loading case).
ts.)
63
0.1 ftDisplacement scale
---- Deformed shape
- -'I
Figure 4. 5. Vertical displacement of circular cell dueto gravity load of cell fill.
64
-0. 5
-1. 5
-z
- 3
-4
- 5
- 6
-7
Axis of symmetry
N.N. \
..... \.... N...... \ N.....
7-Steel sheet pile
Numbers on curves indicatestress contour in ksf.
5
-- - -- =i N.___ ....... `\- .
, \.....--
.... --
...... \
-1.0
\ ....
\N \ \
s. \ \..... \ \
\ \\
1\ \1 \ \
I %
% 1 \I 1 1 1
-7.0 -4.0 -3.0 -2.0
Figure 4. 6. Contours of vertical stress in soil due togravity load of cell fill.
-0. 9
- 0. 6
-0. 4
- 0. 5
- 1. 0
-2. 0
- 3. 0
Axis of symmetry
i Ir if
I
I/ 1
/ r \,.._
/. i
Steel sheet pile
-0. 3
65
Numbers on curves indicatestress contour in ksf.
-1.0 -1.5
4.0
.----- - , \i, -1.0 , 1,
1
...-- ----, N.../ ../ ...,..
-4. 0
0
Figure 4.7. Contours of radial stress in soil due togravity load of cell fill.
Axis of symmetry
Steel sheet pile
66
-1. 0
-0.7
-0. 4
-0. 5
-1.0
-2. 0
-3.0
Figure
- .A.
I/
//.../
,/
/
A
II
I/
,/ /
I-
I
II
I/
////
//
//
/
//
I
I
1
I
%
A
AA
///
/rI
I1
1
I
I
I
I
1
1
i
1
1
I
I
I
I
s
-0 3
Numbers onstress contour
-0. 1
-0.
curves indicatein ksf.
5
in soil
..,"---',./ ....
/ I
-0. 7 // //....
... /...,____.... /I ...
...
/
/'
5'
I/
///
4.8.
-3.0 -1.0
Contours of circumferential stressdue to gravity load of cell fill.
5Akis of symmetry
0 -------
0.20.5
1.0
1.4
0
67
Steel sheet pile
Numbers on curves indicatestress contour in ksf.
I // I
//
//I
1. 0
Figure 4. 9. Contours of shear stress in soil due togravity load of cell fill.
68
Axis of symmetry
A
A
Steel sheet pile
Compression (-)
10 ksf
4-4" .Arrow indicates5 ksf tens ion (+)
k k 4
Figure 4.10. Principal stresses in soil due to gravityload of cell fill.
Front
Steel sheet piles-)Boundary line
Dredge
(Axis of symmetry
Back
69
pa
qb
VV
//////////7777////////////////7/ /////////////////////////////7/Side view
T
Backfill) Ta
Plan view
Figure 4.11. Equivalent loading diagrams due to backfill.
by the following equations (13)
where
pa = Y'hKa
Ka = tan2(45°-)
2
Ka is the Rankine coefficient of active earth pressure
y' is submerged unit weight of soil
h is depth of soil
(I) is angle of internal friction of soil.
70
(4. 1)
(4.2)
The quantity qb in Figure 4.11 is assumed to be the surcharge
load which is equivalent to the dead load of the backfill. pa and qb
are uniformly distributed loads applied from one connecting arc to
the other on the back side of the cell.
The connecting arc tension load was assumed to be Ta and
can be obtained by
Ta = pra a
where ra is the radius of the connecting arc.
(4.3)
In order to represent the arc tension load in a Fourier expansion,
the arc tension load was assumed as a trigonometric loading function
which subtended a small angle at the center of the cell. The area of
the assumed loading must equal to Ta as shown in Figure 4.12.
71
Figure 4. 12. Assumed arc tension loads.
Since the Fourier coefficients for the connecting arc tension
load and backfill distributed loads are different, the computer
analyses were done separately and the results superimposed. The
loadings are divided into two cases in the following manner:
Case I due to connecting arc tension loads. The approximate
loading diagrams are represented by Fourier expan-
sions and are shown in Figure 4. 13.
Case II - includes the gravity load of the system itself and the
distributed load of pa and qb. The Fourier expan-
sion of these loads is shown in Figure 4.14.
The corresponding Fourier coefficients used in the analyses are
shown in Appendix B. Because of the very high cost of a computer
Actual loading
Approximate loading
\ / .........
,... ./ /V, 1C:...'
/."--- s_-,.........4,.<4:41,,,,N-,..., ...--.4 .,z..1-7....:),....:"'
\ / '....... s./ /././ N./ \
',......._ ....,'
30 60
0 in degrees
\
90 120
Line of symmetry
--4 - -\<
150 180
Figure 4. 13. Fourier harmonic loading expansions for arc tension load Ta (Case I).
Approximate loading
n=1
n=4
/A
0 60
Actual loading
8 in degrees
120
" / 1>,
/ / \-.."- ic./_,, ).0\ .... .
180
,
240
,
- -**
--"
300
V
360
Figure 4.14. Fourier harmonic loading expansions for uniform distributedpressures pa and qb (Case II).
74
run, the Fourier analyses were carried out using only 11 harmonics
for Case I loading and 8 harmonics for Case II loading.
An extensive amount of data is obtained (see Appendix E) from
the computer output for each run. In order to present it clearly, it
has been divided into two separate categories. First, the data per-
taining to the shell elements (sheet piles) is presented. This data
includes displacements and stresses in the steel sheet piling cell.
Second, the data pertaining to the soil portion of the structure is pre-
sented. A reference diagram for subsequent graphs is given in Fig-
ure 4.15.
Figures 4.16 and 4.17 show the horizontal displacement of the
cell due to connecting arc tension loads (Case I). The horizontal dis-
placement due to gravity load and uniform distributed pressures
(Case II) and combined displacement are shown in Figures 4.18 to
4.23. The deformed shape of the bulkhead is shown in Figure 4.24.
Figures 4.25 to 4.31 show the vertical stress and hoop stress
on both outside and inside surfaces of the cell at various angles.
Figures 4.32 and 4.33 show the variation of hoop force and
shear stress in the sheet pile along the circumferential direction of
the cell for each level as indicated.
The contours of stresses in soil elements are plotted in Figures
4.34 to 4.37. The radial stress and vertical stress distribution vs.
75
the depth of the cell in sections V1, V2 and V3 are shown in Figures
4.38 and 4.39.
Figure 4.40 is a plot of the coefficients of lateral earth pressure
K in the cell fill against the depth of the cell. The K values were
obtained directly from the ratio between the computed radial stress
and vertical stress in soil elements. The radial deformation, cell
settlement and hoop forces in the cell are plotted in Figures 4.41,
4.42 and 4.43, respectively, to compare with the values obtained in
the field (11).
Top
H1
H2
H3
H4
H5
H6
H7
H8
H9
Center line of cell
V1 V2 V3
I
Arc 0=90°
120° 60°Cell
Plane of
Front
symmetry 1,°Back0=0
150° °
0 =180°
Soil
.4Shell
0=270°
Plan
V1 V2 V3
Vertical section
76
Figure 4.15. Reference diagram for displacements and stresses.
0 9 Harmonics
0 11 Harmonics
Deformed cellOriginal cell
0=0°
At level H-5
Acr
0=180°
Displacement scale
0.2 inch0=0°
At top of cell
Figure 4.16. Horizontal displacements of cell due to arc tension loads at levels H5 and topof cell (Case I).
0=180°
0- 9 Harmonics
0 11 Harmonics
Deformed cellOriginal cell
0=0°
At level H-8
Displacement scale
0.1 inch0=0°
At level H-6 (dredge line)
Figure 4.17. Horizontal displacements of cell due to arc tension loadings at levels H6 andH8 (Case I).
0=1180°
Original cell------ Case II-- Final deformed cell (Case I + Case II)
0=180°
0=0°
At level H-5
Displacement scale
1 inch 0=0*
At top of cell
Figure 4.18. Horizontal displacements of backfilled cell at levels H5 and the top of cell.
Original cell
Case II0=180.
0=0°
At level H8
Final deformed cell(Case I + Case II)
Displacement scale
1 inch 0=0°
At level H6 (dredge line)
Figure 4.19. Horizontal displacements of backfilled cell at level H6 and H8.
Dredge linewANyx6N
1(1?
155 -19. 5Case II
Combined displacement(Case I + Case II)
1I//11'
I
,P - -39.0 48I
, A I
cp. ,a)
I
0-0°
\CI 58.5 0
0 =180°
-78. 0
/NNX/XNY
2 1 0 1 2 2 1 0 1 2
Displacement, inches
Figure 4.20. Radial displacement of backfilled cell at 8 = 0 and 180 degrees.
Dredge line
Case II
Combined displacement(Case I + Case II)
0=150° 0=30°
/1),N(//).Nx,
-78.0
2 1 0 1 2 2 1 0 1 2
Displacement, inches
Figure 4.21. Radial displacement of backfilled cell at 0 = 30 and 150 degrees.
Dredge line
1
------ Case II
0=120°
Combined displacement(Case I + Case II)
0=60°
II
II
19. 5
58. 5
78.0
2 1 0 1 2 2 1 0 1 2
Displacement, inches
Figure 4.22. Radial displacements of backfilled cell at 0 = 60 and 120 degrees.
2
/ANYANNe /"X/X\'V/>NN,t
1 0
_______ Case II
0=270°
1
Combined displacement(Case I + Case II)
2 2
Displacement, inches
0=90°
1
-19.5
-39.0 aa))
-58.50
-78.0
1
Figure 4.23. Radial displacement of cell at 0 = 90 and 270 degrees.
Dredge line
r11\Nxi./>\<
- - ________r--
L_
J
,,,,(Nriv/C
----- Deformed shape
Displacement scale
10.1 ft
-- _ _ _L, -- _ __ i- - -.i-I_ _ _ __------------------_ ___ ___ _ --- .,__
_I_ _ _ _ _ 1 -- - --- I
I
I-
_ ____ I(.... _ _ _ __ t_ -- - + - - -- _ _ _ _ _I_ _ - - - -
0=180° 0=0°
Figure 4.24. Deformed shape of circular cell bulkhead.
0
19.5
a)
039.0
4.,
a)
58.5
78.0
1, 1
outside surfacea- ZZ 0 inside surface
outside surfaceo-X00 Li inside surface
r
1"
cr
4o
Dredgeline level
.
o
Zi
Or.Al
Plillt
N.
-10 -5 5
Stress, kips per square inch-5 0
Figure 4.25. Vertical and hoop stresses of steel cell at 0 = 0 degree.
0
19.5
58.5
78.0
0 outside surfaceo-
aZZ o inside surface
o-00
outside surface
A inside surfaceC
0- 0-Zo
Dredgeline level-mg
0"0
A
Z i
icr
A i
-10 -5 5 -5Stress, kips per square inch
Figure 4.26. Vertical and hoop stresses of steel cell at A = 30 degrees.
5
0
19.5
a)
39.0
58.5
78.0
Ili
r0 outside surface
o- Ai
Z 0 inside surface1
outside surfaceo-
A inside surface
o-o
it
Zo
o-Zi
--Dredgeline level
11.11
-10 -5 0 5
Stress, kips per square inch-5 0
Figure 4.27. Vertical and hoop stresses of steel cell at 0 = 60 degrees.
5
0
19.5
a) 39.0
58.5
78.0
41. 0 outside surfaceo-ZZ o inside surface
outside surface6
00 A inside surface
o-ZoUzi
Dredgeline level
o-eiLeopr
1
111.
iir
-10 -5 5
Stress, kips per square inch-5 0
Figure 4. 28. Vertical and hoop stresses of steel cell at A = 90 degrees.
5
0
19. 5
39. 0
a)
58.5
78. 0
o- ZZ
X0 800
0 outside surface
inside surface
outside surface
A inside surface
I
0-
o-
Dredgeline level-
Cr
ooo
EliZ i
s,.
.
-10 -5 0 5
Stress, kips per square inch
Figure 4.29. Vertical and hoop stresses of steel cell at 0 = 120 degrees.
0
0
19. 5
(Do 39.0
58. 5
78. 0
o- ZZ
o-130
0 outside surface
0 inside surface
outside surface
A inside surface
I
2
Cro
Dredgeline level
cro
o-Zi
3
-10 5
Stress, kips per square inch-5
Figure 4.30. Vertical and hoop stresses of steel cell ate = 150 degrees.
0
19. 5
76.0
-
1
o-ZZ
I
0 outside surface
inside surfacea
outside surfacecr
A inside surface2
Dredgeline level
_____a 0-0
. o-o 0-
eiA
All).-10 -5 5
Stress, kips per square inch-5 0
Figure 4.31. Vertical and hoop stresses of steel cell at 0 = 180 degrees.
5
2
1
n
'CDCD
za)c.) -10
'4-4
17-14
00
- 2
- 3
Maximum hoop force due to cell ---7fill only at level H5 AN
I
fl II
E.
v 411101111111111111111
)' 1
I
vV
Level
H2H3H4
0 H5a H6V H80H9
1
0
1n An 9n 1 zn 150 11
Position 0, degrees
Figure 4.32. Hoop force in cell vs. position.
H5H4
H3
H2
H6Line ofsymmetry
H8
H90
H3
V H4
Level p H5A H6
0 H9
Line of anti-symmetryof shearing stress
30 60 90 120Position 0, degrees
150
Figure 4.33. Shearing stress in cell vs. position.
180 210
C)
- 1.0
- 1.5
- 2. 0
...-/ ....1 ---,... - .-/ -5. 0 ..... --__.... / /- -./ / JP \ \ .. , .... ...... ...... ......
, /,/
/
...... ,,,,,,..
//11 --- ........... ....-
.... ,,J .// / / ......--
....- --- ....- \ \/ / / .--I \ 5'/ . \I . \ \I
III I / I
/ \ \ \I
-1.0N./ N N./ \
// -1 5 \ or---Steel sheet pile/ ,.-. .
/ / . ./ / 1\,1
// \ II/ / . I. /. / -2.,0 \ \
// \ \/ / ....'\ \/ / \
\N,/...-
/ .. \ N
-3.0. \/ / \\
Stress contour in ksf.
- 3.
- 4.
0
0
/// ../
/< /. % -4. 0-3.0)
-4.0 -5.0 -6.0 -5.0CL of cell
Plane 0 = 180° IN. Plane 9 = 0°
Figure 4.34. Contours of vertical stress-in soil.
Steel sheet pile-0.5-0.5
Stress contour in ksf.
Plane 0 = /80°
Figure 4.35.
Plane 0 = 0°
Contours of radial stress in soil.
1.0
0.2
0.3
/., ..0 . //
/L 0._/ 5 / /
I
\\
1
1
1
1
0.5
1.3
1.0
Steel sheet pile
Stress contour in ksf.
\
_ 0.50.3
O. 5
10 5/.
/J,// -/ , ,
/ ,. /. /..II /
Plane 0 = 180° 41
0
Plane 0 = 0°
Figure 4.36. Contours of shear stress (TrZ)
in soil.
0.8
1.0
-0. 5
-1.0
-0 5
-1. 5
I
.-__ ----
. ,.'\\ =2. 2 0
i // [/ /\\
% /...
/ / - N.3. 0 . \\\ \/ ----
1
.__ \ \..- \
/ ---- \.., ,
N\ \. ---
\..- / /-- --- - 4 . 0 N . .\ N
/ / -- r- \ .I / ,.
I\ .
-2.0 q., -2.0
NY/7/(\\Y//AN
Steel sheet pile
-2.0
Stress contour in ksf.
Plane 0 = 180° 411 Plane 0 = 0°
Figure 4.37. Contours of circumferential stress in soil.
-1. 5
0
19.5
39.0
a)a)
58.5a)
78.0
97.5
117.0
Overburden
Dredge line level
Pile tip level
60 2 4 8Stress, ksf.
(a) Soil stress along section V1 at 0 = 0° (center)
0c
19. 5
39.0
58.5
78.0
97.5
117.002 4 6
Stress, ksf.(b) Soil stress along section V2 at 0 = 0°
Overburden
`Dredge line level
z z \ Pile tip level
Figure 4.38. Radial and vertical stresses in soil vs. depth of cell.
8
19.5
4)41)
39.0
58. 5
78.0
Position
Dredge line level
0.5K
1 0
Front
Position 1 -
Position 2 -
Position 3 -
Front position of the cell(section V3, 0 = 180°) andalso the front position of theisolated cell caseNear center position of thecell (section V1, 0 = 0°)Back position of the cell(section V3, 0 = 0°)
Figure 4. 40. Coefficients of lateral earth pressure inside the cell fill.
+8
cd
P,A ....
/i
..- . /.... .
..... . i." N.N. /
/A\,/Field measurement results
Finite element solution
after backfill (11)
60 120 180 240 300 360Position 0, degrees
Figure 4.41. Comparison of radial deformation of backfilled cell at level H5.
00. 4
0. 5
U
0. 6a)
0. 7
0. 8
Position 0, degrees60 120 180 240 300 360
Total settlement /5. 5
Total settlement/8,P., .
7.---, 7.,.., ./ ..., ki----A\ \. /--a----,
/ \ / // \\/ \\ \ // //\ \ ,,
Elastic settlement
-4\
/
0 just after backfill{Field measurement results (11)
A 2-1/2 months after backfill
O Finite element solutionsI I
Figure 4.42. Settlement of the top of the steel sheet piles vs. position.
10
20
4-3
a)
-30
40
50
60-1
\ \ \\ \\ \ \ =\ \ \\ \ \
\ \ \
120°
\\\\\.\ .\ \
1\\\
Finite element solutionsmeasurement results
of final fill (11)Fieldat stage
\ \\ \\
\
\\\
%
\
-Iiiii
\
,
\
= 90°
\ \I \ \1 \ \.I \ \,, .
\\\\_\0 180°
0 = 60°
\\
MILFr
\\
\\ \\\\ = 150°
0 2 0 1
Hoop tension force, kips per inch
Figure 4.43. Comparisons of hoop force in steel cell.
4
105
V. DISCUSSION OF RESULTS
5.1 Discussion of the Isolated Circular Cell
Figure 4.4 shows the radial displacement, hoop force and
stresses in steel sheets for the isolated circular cell subjected to the
gravity load of cell fill only. The radial displacement and hoop force
of the cell gradually increase as the depth of cell increases up to the
maximum bulging point and then drop off. The maximum bulging of
the cell is 0.57 inch which occurs at a point one-sixth (0. 17) of the
exposed height above the dredge line. The maximum bulging point
that was observed in the field (11) ranged from 0.18 to 0.28 of the
exposed height. The value of 0.18 corresponds to field measurements
on the front side of the cell where sheet pile length was the same as
in the present study.
From Figure 4.4 it can be seen that all vertical stresses in the
steel sheet are compressive with the maximum value of 5.3 ksi at
the bottom edge of the cell. This indicates that the steel sheet trans-
mits the downdrag due to the weight of soil adjacent to the pile, to the
soil foundation. The maximum hoop stress occurs on the outside
surface of the cell at a level one-sixth of the exposed height. There
are also some hoop stresses in the embedment zone.
The elastic settlement of soil fill inside the circular cell is
shown in Figure 4.5. It can be seen that the soil at the center of the
106
cell settles more than that away from the center. This may be
explained by recognizing that near the sheet piles there is friction
(actually no slip) between the soil fill and the wall. The stiffer steel
element deforms less. The maximum settlement of the soil element
at the center of the cell is 1.48 inches, while the top of the steel sheet
settles only 0. 66 inch.
The upward displacement along the vertical boundary line
outside the cell is due to the outward movement of the cell pushing
against the boundary. This is unreasonable for an actual installation.
In order to eliminate the upward displacement in the theoretical model
it is necessary to increase the distance of the vertical boundary from
the cell. By increasing this distance no significant change in the dis-
placement and stresses along the sheet pile node is expected. The
deflected shape and soil element stresses in the current boundary line,
however, may be changed considerably.
The horizontal boundary line at the base of the system as used
in the example is justified by observing the settlement of the nodal
points directly above which indicate the uniform settlement of the soil
element along the horizontal cross section. This indicates that the
horizontal boundary line is deep enough from the bottom edge of the
steel sheets.
Figures 4.6, 4.7 and 4.8 illustrate the contours of vertical,
radial, and circumferential stresses in the soil elements for the cell
107
filled, gravity loading case. In the body of the circular cell, all of
the stresses are compressive with the vertical stresses Cr greaterzz
than the radial stresses T and circumferential stressesrr creer
Outside the cell, just below the dredge line, the radial stress is con-
siderably higher than the vertical stress (Figures 4. 6 and 4. 7). This
distribution of normal stresses thus corresponds to development of a
passive pressure distribution where the sheet piles are being forced
into the soil.
The contours of shearing stress T in the soil are plotted inrz
Figure 4.9. It indicates that the maximum shearing stress occurs on
a plane just behind the sheet piles near mid-height above the dredge
line. Planes away from the sheet piles display a dramatic decrease
in shear stress near mid-height, while the top and bottom portions of
these planes show a slight decrease only.
Figure 4. 10 shows the principal stresses in the soil for the
circular cell under the gravity load of the cell fill. The stresses
inside the cell are all compressive stresses. The directions of the
maximum principal stress above the dredge line all incline toward the
sheet piles. This shows the tendency of the soil fill inside the cell to
bulge the sheet piles. Outside the cell the principal stresses tend to
resist such lateral movement.
108
5.2 Discussion of Circular Cell Bulkhead
As mentioned in Chapter 4 the loadings on circular bulkheads
are separated into two cases, therefore, the computed displacements
are presented separately and then combined.
The deformed shape of the circular cell due to connecting arc
tension loads (Case I) for various levels is shown in Figures 4.16 and
4.17. The radial displacement and tangential displacement data for
9 harmonic terms are plotted on the left side of the circular sections,
while the data for 11 harmonic terms are plotted on the right side of
the sections. It was found that going from 9 to 11 harmonics gave an
approximately 5% increase in the maximum displacement resultant.
The maximum movement occurs on the top level of the cell with 0.2
inch leaning toward the front side.
The radial displacement and tangential displacement due to
Case II loading are shown in Figures 4.18 and 4.19. The displace-
ment data for 6 and 8 harmonics are plotted on the opposite half of
the circular sections in the same manner as in Case I. The dis-
crepancy for displacement between the two harmonics was not greater
than 6% in any section.
The combined displacements are plotted in Figures 4.18 to
4.23. It can be seen that the connecting arc tension loading con-
tributes approximately 10 to 25% of the total horizontal displacements.
109
The deflected shape of the cell indicates that the cell was pushed
toward the water side with the front portion of the embedment sheet
piles forced into the front dredge line and the back portion moved
into the cell.
The maximum bulging point for the circular cell bulkhead occurs
on level H5 (0. 17 of the exposed height above the dredge line) on the
front side (0 = 180°) of the cell. This can be compared to the sug-
gested design value of one-fourth (0. 25) for TVA (22) and the field
measurement value (11) of 0.18 at the front sheet position. The maxi-
mum bulging on the back portion of the cell occurs at the mid-height
level (H3). This is in good agreement with the results from field
measurements (11).
Figure 4.24 shows the deformed shape of the circular cell
bulkhead on the front-back (180° -0° ) cross-section. It indicates that
the settlement of the fill in the center of the cell is consistently
greater than at the edges. The shape of settlement in the cell fill has
a pattern similar to that of field measurement results obtained by
White (26). From Figures 4.5 and 4.24, it can be seen that placing of
the backfill does not significantly change the settlement of the soil fill
inside the body of the cell.
Figures 4.25 to 4.31 show the steel sheet surface stress data
for various angles around the cell. The membrane stress is the
average of the outside surface and inside surface stresses, whereas
110
the bending stress is observed by noting one-half the difference
between these two stresses. It can be seen that the bending stress in
the vertical direction is more significant than that in the hoop direc-
tion. It also appears that both the vertical stress and the hoop stress
are maximum on the front sheet (0 = 180°) of the cell.
Figures 4.32 and 4.33 show how the hoop tension force and
shearing stress in the steel sheet vary around the cell at the levels
indicated. The hoop forces on the level above the dredge line appear
to increase from the back side (0 = 0°) of the cell to a maximum at
the arc connection (0 = 120°). They remain almost constant from this
point to the front sheet of the cell. From Figure 4.32, it can be
observed that in placing the backfill, the hoop forces in the back
portion decreases sharply since the back fill lateral pressures counter
the internal forces from the cell fill. At the front side of the cell,
the hoop forces increase to a very small degree. Moreover, it is
found that the maximum hoop force does coincide with the maximum
bulging point on level H5. Consequently, the TVA design rules (22)
in which maximum hoop tension occurs just inside the arc connection
at a point H/4 above the dredge line are appropriate.
The shearing stress TZ
in the steel sheet as shown in
Figure 4.33 appears to increase from zero at the back side of the cell
to a maximum at the arc connection. It then decreases gradually to
zero at the front sheet of the cell. The shearing stress diagrams are
111
anti-symmetric about the front to back cross-section. The maximum
shearing stress occurs on the dredge line level (H6).
The contours of vertical stress, radial stress, shearing stress
and circumferential or tangential stress in the soil on vertical cross-
sections of planes 0 = 0° and 0 = 180° are plotted in Figures
4.34 to 4.37, respectively. The stress contour data on other planes
of the bulkhead were also computed but they are not included here
because the variations from those shown here were small. The soil
stresses inside the cell appear to change very slightly due to the
effect of back fill (see Figures 4.6 to 4.9 and Figures 4.34 to 4. 37).
Outside the body of the cell, the application of backfill pressure
greatly increases the vertical and radial stresses around the cell.
The noticeable increase in radial stress in front of the cell results
from the passive resisting force due to the outward movement of steel
sheet piles in that area. The vertical stress is quite uniformly dis-
tributed across the base of the cell and approximately equal to the
overburden pressure.
Backfill lateral pressure slightly affects the shear stress in the
cell fill as shown in Figure 4.36. The shear stresses in the middle
are much lower than near the sheet piles. It therefore seems improb-
able that shear failure would start in the midplane and go towards the
steel sheet walls as suggested by Terzaghi (23). Instead it appears
that the failure should start near the dredge line of the sheet pile and
112
progress from there up to the top free surface and down through the
base of the sheets along a path of maximum shear stress. This type
of internal failure mode was also observed in photoelastic analysis (1).
Figures 4.38 and 4.39 show the vertical and radial stresses in
the soil elements inside the cell vs. the depth of the cell. Results for
radial stresses show that they increase more on the back side than the
front due to the lateral pressure of backfill trying to push the back
sheets inward. It also can be seen that levels above the dredge line
experience a small increase in stresses, while the lower portions of
the cell show a considerable increase. Specifically, there is an
abrupt increase in both vertical and radial stresses under the bottom
edge of the sheet piles. This is due to the weight of soil fill transmit-
ted by the sheet piles to the soil foundation. This observation may
help to explain larger than anticipated settlements which have been
experienced by structures of this type in Portland (11) and Long
Beach (26).
The coefficient of lateral earth pressure K shown in Figure
4.41 indicates that at the center plane of cell K decreases slightly
with depth. At the front plane, K is almost constant all the way
down the cell. At the back plane, it increases moderately due to
backfill lateral pressure. The mean values of K in the front posi-
tion (unloaded side), center position and back position (loaded side)
are 0.441, 0.412 and 0. 64, respectively. They indicate that the
113
coefficient of lateral earth pressure of the fill in the cell is higher
than the Rankine earth pressure coefficient. The maximum average
value of K in the field was 0.45 (11).
The results from the study and the field measurements obtained
by Khuayjarernpanishk (11) are compared with the one obtained by
using Cummings', TVA's and Terzaghi's recommended formulae as
shown in Table 5. 1. For the circular cell bulkhead case the maxi-
mum hoop tension force in the steel cell obtained in this study is about
0.75 of the values obtained from Cummings' and Terzaghi's formulae
and 0. 67 of the value obtained from TVA's formulae, for the isolated
cell case. At the same time, it is only 0. 61 of the field measurement
result for the bulkhead case. The computed vertical pressures at the
base of the cell are almost uniform and approximately equal to the
overburden pressure. This indicates that the bending effect due to
backfill results in a very small vertical pressure when compared to
the effect of the gravity load of the cell fill. Consequently, they are
not in the same trend as Terzaghi's concept which states that the
bending stress is linearly distributed across the horizontal section of
the cell by considering the cell as a rigid body. The calculated coef-
ficients of lateral earth pressure in the cell fill are in good agreement
with the values recommended by other investigators and those obtained
from field measurements.
Table 5. 1. Comparisons of results.
MaximurnHoop
Tensiono- ZZ at front base o- ZZ at back base
KSources kips /inch crZZ overburden crZZ overburden Front Center Back
Cummings (6) 2.53* -- -- - - -
TVA (22) 2.85** -- -- - 0.523***
Terzaghi (23) 2.53* 1.25 0.743 0.4 0.4-0.5 -
Kittisatra 1 1.90 1.04 1.05 0.441 0.412 0.64
Lacroix (13) -- -- -- 0.4 0.5 -0. 6 0.7-1.0
Khuayjarernpanishk (11) 3. 10+ -- 0.45+
"At dredge line level for isolated cell.At H/4 above dredge line just inside arc connection for isolated cell.
*** 2 2The Krynine factor, K = (cos 4)/(2-cos O.+Field measurement results for bulkhead1 Finite element results for bulkhead.
115
As mentioned in Chapter IV, the circular cell bulkhead used in
this study is not exactly the same configuration as the one instru-
meni-ed in the field (11). Nevertheless, the radial deformation at
level. H5, the settlement of the top of the cell and the hoop tension
force in the cell have been plotted in Figures 4. 41, 4. 42 and 4.43,
respectively, to compare and show some correlations.
Figure 4.41 shows that the computed deformations are in good
agreement with the general trend of observed values in the field
except for the positions near the left arc (0 = 1200). The exagger-
ated distortions at such points in the field during construction were
presumed to result from wave action, compaction of fill and tightening
of interlock slack (11). They are not associated with elastic
deformation.
Figure 4. 42 illustrates how the edge settlement varied around
the cell for the top level. It appears that the settlement of the back
portion of the cell is greater than the front. This is due to the back
sheet piles being subjected to a higher gravity effect of the backfill
than the front. The observed settlements in the field immediately
after backfilling and 2-1/2 months after backfilling are approximately
5.5 and 8 times the computed elastic settlement. This seems to
indicate that the total settlement in the field resulted from consolida-
tion of the fill in response to stresses due to fill weight in addition to
the immediate elastic distortion due to the same cause. The observed
116
settlements of the cell are in similar pattern to the computed results.
From Figure 4. 43, the computed hoop tension forces in the cell
have a similar pattern to the field measurement values but do
illustrate again significant differences in magnitude. All hoop forces
obtained from the field measurements are greater than in this study.
The main factors associated with the assumptions used in this
study that influence the results are that:
1. The elastic properties of a given soil in a cell depend to a
large extent on the method of filling the cell. There is no
laboratory procedure that would yield reliable advance
information on the state of density of the soil in the cell and
on the corresponding elastic properties of the soil. The
estimated moduli of elasticity E used in this study seemed
to be too high. They are approximately 7 times larger than
that used by Brown (2). If lower values of E had been
used, the nodal point displacements would have been signifi-
cantly greater. Since the steel sheet pile stress is a function
of displacements at the nodes, a corresponding increase in
sheet stress would also be expected.
2. The assumed interface condition (full friction) between the
soil and the steel sheets probably resulted in higher than
realistic axial stress in the steel sheets. The rougher the
steel sheet surface, the better the agreement will be between
117
analytic and field results. It is quite likely that a boundary
element could be developed to represent the actual state of
friction. Such an element is not now available. Additional
research and formulation would have to be undertaken to
develop such an element. It is believed that modification of
the analysis and the computer program can be made to
account for the interface friction force, once the appropriate
element is introduced and the experimental boundary inter-
action data are available.
3. The assumed cylindrical shell-like structure is much stiffer
than the actual interlocked sheet piling cellular bulkhead.
This assumption is a very essential factor, reducing the
horizontal displacements in the cell because no interlock
slack can exist in the shell. But on the other hand, interlock
friction in the actual cell probably reduces slippage to
negligible values for usual cell service conditions. The
interlocks would be expected to slip if the cell were caused
to distort a large amount (fail by bending).
118
VI. SUMMARY AND CONCLUSIONS
The finite element model and the corresponding computer
program developed in this investigation can be used to compute dis-
placements and stresses in a circular cell structure subjected to
gravity loads of cell fill, backfill and surcharge loads. The structure
can be founded on soil foundation or rock foundation. The loading
functions are expanded in Fourier harmonic series. For different
sets of Fourier force coefficients, the analyses must be carried out
separately. The components of displacement and stress for each
harmonic and each set of loadings are superimposed to form the total
solution.
The following conclusions are drawn from this study.
1. An almost uniform vertical pressure distribution on the base
of the circular cell bulkhead was obtained. The magnitude is
approximately equal to the overburden pressure.
2. Horizontal stresses in the soil within the cell are mobilized
as normal active and passive earth pressure.
3. The maximum elastic soil settlement occurred in the center
of the cell, with smaller settlements at the edges. It also
was found that the total elastic settlement at the back side of
the cell is greater than at the front.
4. The vertical shear stress in the cell fill for the bulkhead is
119
maximum near the front and back sheet piles and not on the
center plane. Therefore, the mid-plane shear failure pro-
posed by Terzaghi (23) seems unlikely.
5. Through shear transfer to and from the soil, the sheet piles
appear to transmit an appreciable amount of the load to the
foundation.
6. Placement of the backfill increases the hoop tension forces
in the front portion of the cell. A decrease in hoop tension
in the sides and back portion was found.
7. Both maximum hoop tension force and maximum bulging of
the cell occur on a level 1 /6 of the exposed height above the
dredge line. The maximum hoop tension occurs behind the
connections of the cell and arcs, whereas maximum bulging
occurs at the front sheet of the cell. This indicates that the
TVA (22) suggestion to compute the maximum hoop force in
the cell at the arc connecting sections is justified.
8. Both vertical and circumferential bending stresses do occur
in the steel sheet piles. These bending stresses are lower
than the membrane stresses. However, such stresses should
be taken into account in designing.
9. The coefficients of lateral earth pressure in the cell fill are
0.441 at the front, 0.412 at the center and 0.640 at the back
of the cell. These values agree closely with the values
120
recommended by other investigators.
10. The total eleven Fourier harmonics used in the analysis is
considered adequate for the loading conditions in this study.
121
BIBLIOGRAPHY
1. Burki, N.K. and Rowland Richards, Jr. Photoelastic analysisof a cofferdam. Proceedings of the American Society of CivilEngineers, Journal of the Geotechnical Division, pp. 129-144.February 1975.
2. Brown, P.P. Discussion of: Field study of cellular bulkhead,by A. White, J. A. Cheney and C.M. Duke. Transactions of theAmerican Society of Civil Engineers 128:503-506. 1963.
3. Burmister, D. M. Stress and displacement characteristics of atwo layer rigid base soil system.: Influence diagrams and practi-cal applications. Proceedings of the Highway Research Board35:773-814. 1956.
4. Carnahan, B. , H. A. Luther and J.O. Wilkes. Applied numericalmethods. New York, Wiley, 1969. 604 p.
5. Cook, R.D. Concepts and applications of finite element analysis.New York, Wiley, 1974. 402 p.
6. Cummings, E.M. Cellular cofferdams and docks. Proceedingsof the American Society of Civil Engineers, Journal of theWaterways and Harbors Division 83:13-45. September, 1957.(Also, Trans. ASCE, Vol. 125, 1960)
7. Dym, C. L. and I.H. Shames. Solid mechanics: VariationalApproach. New York, McGraw-Hill, 1973. 556 p.
8. Flugge, W. Stresses in shells. 2nd ed. New York, Springer-Verlag, 1973. 525 p.
9. Grafton, P.E. and D.R. Strome. Analysis of axisymmetricshells by the direct stiffness method. American Institute ofAeronautics and Astronautics Journal 1:2342-2347. October, 1963.
10. Herrmann, L.R. , R. L. Taylor and D.R. Green. Finite elementanalysis for solid rocket motor cases. Berkeley, 1967. (Uni-versity of California, Structural Engineering Laboratory. Reportno. 67-4)
122
11. Khuayjarernpanishk, T. Behavior of a circular cell bulkheadduring construction. Doctoral dissertation. Corvallis, OregonState University, 1975. 156 numb. leaves.
12. Kleins, S. A study of the matrix displacement method as appliedto shells of revolution. In: Proceedings of the First Conferenceon Matrix Methods in Structural Mechanics, Report no. AFFDL-TR-66-80, Wright-Patterson Air Force Base, 1965. p. 275-298.
13. Lacroix, Y,, M.I. Esrig and U. Luscher. Design, constructionand performance of cellular cofferdams. Proceedings of theSpecialty Conference, Lateral Stresses in the Ground and Designof Earth-Retaining Structures, The American Society of CivilEngineers, Soil Mechanics and Foundation Division. June, 1970.p. 271-328.
14. Laursen, H.I. Structural analysis. New York, McGraw-Hill,1969. 486 p.
15. Love, A.E.H. The mathematical theory of elasticity. 4th ed.Cambridge University Press, 1927. 643 p.
16. Milne, W. E. Numerical calculus Princeton, PrincetonUniversity Press, 1949. 393 p.
17. Novozhilov, V.V. Theory of thin shells (Translation).P. Noordhoff, Groningen, The Netherlands, 1959. 376 p.
18. Percy, J. H. , T. H. H. Pian, S. Kleins and D . R. Narvaratana.Application of the matrix displacement method to linear elasticanalysis of shells of revolution. American Institute of Aero-nautics and Astronautics Journal 3:2138-2145. November, 1965.
19. Popov, E.P., J. Penzien and Z. A. Lu. Finite element solutionfor axisymmetric shells. Proceedings of the American Society ofCivil Engineers, Journal of the Engineering Mechanics Division90:119-145. October, 1964.
20. Richart, F. E. , Jr. , J. R. Hall, Jr. and R . D. Woods. Vibrationsof soils and foundations. Englewood Cliffs, Prentice-Hall, 1970.414 p.
123
21. Speare, J. C. Discrete element static analysis of core-stiffenedshells of revolution for asymmetric mechanical loadings.Cambridge, 1968. 229 numb. leaves. (Massachusetts Instituteof Technology, Dept. of Aeronautics and Astronautics, Aero-elastic and Structures Research Laboratory. TR 146-5)
22. Tennessee Valley Authority. Steel sheet piling cellular coffer-dams on rock. TVA Technical Monograph no. 75, Vol. 1.
Knoxville, Tennessee, 1957.
23. Terzaghi, K. Stability and stiffness of cellular cofferdams.Transactions of the American Society of Civil Engineers,110:1083-1202. 1945.
24. Tezcan, S.S. Discussion of: Simplified formulation of stiffnessmatrices, by P.M. Wright. Journal of the Structural Division,American Society of Civil Engineers 89:445-449. December,1963.
25. Timoshenko, S. and K.S. Woinowsky. Theory of plates andshells. 2nd ed. New York, McGraw-Hill, 1959. 580 p.
26. White, A., J. A. Cheney and C. M. Duke. Field study of acellular bulkhead. Transactions of the American Society of CivilEngineers 128:463-508. 1963.
27. Wilson, E. L. Structural analysis of axisymmetric solids.American Institute of Aeronautics and Astronautics Journal3:2269-2274. December, 1965.
28. Zienkiewicz, 0. C. The finite element method in engineeringscience. London, McGraw-Hill, 1971. 521 p.
APPENDICES
APPENDIX A
Element Matrices
A. 1 Triangular Axisymmetric Ring Element
The [%] matrix of Equation (2. 16) is (also see Figure 2. 1)
1 rl Z1 0 0 0 0 0 0
0 0 0 1 rl Z1 0 0 0
0 0 0 0 0 0 1 r1 Z1
124
(A. 1)
125
The matrix
1.11L-0
[c3,1]
a1
b1
c1
0
0
0
0
0
0
of Equation (2. 17) is
0 0 a2
0 0
0 0 b2
0 0
0 0 c2
0 0
a1
0 0 a2 0
b1
0 0 b2
0
C10 0 c2 0
0 al 0 0 a2
0 b1
0 0 b2
0 0 01 c2
a3
b3
c3
0
0
0
0
0
0
0
0
0
a3
b3
c3
0
0
0
0
0
0
0
0
0
a3
b3
c3
(A. 2)2A
where A is the cross sectional area of the triangular element and
a, b, c are nodal coordinate functions as shown in Equation (2. 19b).
The matrix W(r, Z, 0)] of Equation (2.21) is shown below
where n is the Fourier harmonic number.
126
[I' (r, Z, 0)]
0
0
cos nO
cos nO
0
cos nO
0
n sin nO
0
0
cos nO
0
0
0
nZ sin ne
0 0
0 0
Z cos nO
0
0
0
cos nO
0
-n sin ne
0
nZ cos nO
(A. 3)
r
0
n sin nO
0r
cos nO 0
nZ sin nOr0
The matrix
0rn sin ne
0 - r
0 0
0 0
n cos nO n cos nOr0 0
sin ne
r0
Z sin nO0 _
r r0 0 sin nO
(2. 28a) can be found by rewriting
r
[Sn ] in Equation
the material constant C in Equation (2. 7) as
C11 C12 C13 0 0 0
C12 C22 C23 0 0 0
[C] =C13 C23 C33 0 0 0
(A. 4)0 0 0 C44 0 0
0 0 0 0 C55 0
0 0 0 0 0 C66
where
11C11
C1212
44C44
22
13
55
C33
C 23
C66
E(1-v)(l+v)(1-2v)
Ev(l+v)(1-2v)
E2(1+v)
127
Substituting Equations (A. 3) and (A.4) into Equation (2. 25c) and inte-
grating Equation (2. 28a), [Sn
] can be expressed as
[SUUn] [SUWn]
[SUVn]
[sn] = [SWU n] [SWW n] [SWVn] (A. 5)
[SVUn] [SVWn] [SVVn]
where
[SUU n]
(C33+n C55)I3 (C13+C33+n2C55)I1 (C
33+n2C
55)15
(C11+2C13+C33+n2 C55)I1 (C13+C33+n 2
C55)14
Symmetric
[SUWn [SWUn
=
7
(C33 )I +C I33 55 6 44 1
(A. 6)
(A.7)
[SUVn
] = [SVUn
IT
_n(C
33+C
55)13
= n(C 13+C 33 +C 55)12
n(C33+C55)1.5
nC33I2 n(C33
+C 55)15
n(C13
+C33)11
n(C13 +C 33+C55)I4
nC3314
2 2Cn C66I3 n 66I2
(C +n2C
[SWWn] = 44 66)I
1
[SVITn] =
Symmetric
_(n
2C33+C55 )13
Symmetric
[SWV n] = [SVWn]T =
0
0
n2C 33C3312
n2C33I1
0
0
n(C 33+C
55)16
n2C6615
n2C6614
C -,1 +n2C6616
2-. 1
(n 2C
33+C55 )15
n2C33I4
(n2C33 +C55)16 +C6611
-nC6612
- nC66I1
128
(A. 8)
(A. 9)
(A. 10)
(A. 11)
129
A. 2 Shell Element of Revolution
Matrix
o
1[co
0] of Equation (2.43c) is
1 0 0 0 0 0 0
0 - 1 /I 0 0 0 1// 0 0
0 0 1 0 0 0 0 0
0 0 -1// 0 0 0 1/1 0
(A. 12)1 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0
-3 /Q2 0 0 -2// 23// 0 0 -1/1
2// 30 0 1/12 -2 // 3
0 0 1/.e2
Matrix [X0] in Equation (2.44a) is
cos i) sin (I) 0 0 0 0 0 0
-sin 4 cos 1 0 0 0 0 0 0
0 0 1 0 0 0 0 0
0 0 0 1 0 0 0 0(A. 13)
0 0 0 0 cos 41) sin (I) 0 0
0 0 0 0 -sin g) cos g) 0 0
0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 1
The matrix [(6,00
1] of Equation (2. 44b) may be shown by
denoting
and
a = Z2 Z1 (see Figure 2.3)
b = r2
r1
cos (I) = a //
sin (I) = -b
where i is the length of the shell element.
130
(A. 14)
['PO
Substituting Equation (A. 14) into Equation (2.44c),
becomes
[(Poo- l]
b /1 a/1 0 0 0 0 0 0
-b // 2 -a // 20 0 b/i 2 a /i 2
0 0
0 0 1 0 0 0 0 0
0 0 - 1 /.12 0 0 0 1/i 0
a te -b /.2 0 0 0 0 0 0
0 0 0 1 0 0 0 0
-3a /23 313/13 0 -2/i 3a/f 3 -3b/13 0 -1/i
2a // 4- 2b/.24 0 1/12 - 2a/.24 2b if
40 1 /I
2
(A. 15)
Matrices [Nn(0)] and [Mn
(0)] in Equations (2. 47c) and
(2.47d ) are:
[N n(0)] =
and
E11
cn0 0 E12sOcn0 nE 12cn0 E
12c (1)cnO
E 44 sn0 -nE44 snO -E 44si)sn0 0
E 22 s2cknO+n
zE 44 snO nscHE22cnO+E44 snO) E 22 sOccOcnO
Symmetric nE cnO22 +E44sziOsne nE
22ccOcne
E22c2.0cnO
D11
cnO 0 D12 sc)cn0 -nD12
c)c nO -112D12cri°
D 44 c 2One nD 44cOn0 -D 44 c2cOscpsne -nD 44s (1)c On°
2 -n2 sci)(D22cnOD 22s2cl)cnO+n D44snO -nsizOcc1)(DzzcnO+D44snO)
(A. 16)
+D44sn0)
c2on2D
222crie+D44sOnO) nc(1)(nzDzzcnO+s ct)D44snO)Symmetric
n4D22
cn0 +n D 44c cl)snO
(A. 17)
132
where
cl) -= cos 4, c24) = cos 24)
s4) = sin (I), s24) = sin24)
cnO = cos2
nO, snO = sin2nO
Et EvtEll = E22E
121-v 2 1-v2
EtE 44 2(1+v)
Et3 Et 3
D11
= D2212(1-v2) 12(1 -v2))
Et3v Et3
D12 D12(1-v2)
44 12(1+v)
where E is Young's modulus, v is Poisson's ratio.
133
APPENDIX B
Fourier Harmonic Coefficients
There are two sets of Fourier force coefficients in this study;
the first set contains the coefficients due to the connecting arc tension
load as shown in Figure 4.13, the second set contains the coefficients
due to uniformly distributed lateral load and gravity load of backfill
as shown in Figure 4.14.
then
Let
p(0) = ancos nO
n=0
1a0
= p(0)d0-Tr
1Tr
an Tr= p(0) cos ned0, n = 1,2, ,N
-Tr
(B. 1)
(B. 2)
(B.3)
The resulting Fourier coefficients for each case of loadings are
shown in Table B. 1.
134
Table B. 1. Fourier force coefficients.
Coefficient
Case I Case II13Tr 11Tr
2.Tr 2Tr< 0 < -
18 18p(e) = p cos 90
11Tr 13Trp(0)=p, --3 <O<
2Tr 2Tr
< 0 <18 18
p(0) = 0, elsewhere<p(0) = 0, T < 13
3
a00.07073 0.66666
al -0.07053 0.55133
a2
-0.06992 -0.27566
a3
0.13783 0
a4
-0.06752 0.13783
a5
-0.06577 -0.11026
a6 0.12732 0
a7
-0.06124 0.07876
a8-0.05853 -0.06892
a9 0.11111 0
a10
-0.05236 0.05513
all -0 04899
135
APPENDIX C
Gaussian Quadrature Numerical Integration Procedure
Numerical integration by the Gaussian quadrature method is
chosen because high accuracy is obtained with relatively few nodal
stations and singularities along the boundary can easily be handled (21).
Reference (16) gives the following Gaussian quadrature formula:
where
and
1
f(X)dX = g(x)945dxa -1
i=1
66 (C. la)
g(x) = f(h(x)) (C. lb)
X = h(x)
w. is the Gaussian weighting functions given in Table C. 1.
x,1 , y. are the Gaussian integration stations given in Table C. 1.
N is the number of Gaussian stations.
is the differential transformation containing terms to change
the limits of integration to the interval [-1, 1] and terms
from the differential dX.
The extension to integrals involving more than one variable is
straightforward. For example, in the case of two variables, we
obtain
where
S
1
X, Y)dS = g(x, y),95dxdy-1 -1
N Nw.w.g(x.,
g(x, y) = f(hx(x, y), hy(x, y))
X = hX '
(x y)
Y = h (x, y)
136
(C. 2a)
(C.2b)
(C.2c)
(C.2(1)
Values of Gaussian weighting functions and stations are given
in Table C. 1.
137
Table C. 1. Gaussian weighting functions and
stations .
N i Station, xi.VV w.eight, .
I
2 1 -0.57735027 1.00000000
2 0.57735027 1.00000000
3 1 -0.77459667 0.55555555
2 0.00000000 0.88888889
3 0.77459667 0.55555555
4 1 -0.86113631 0.347854852 -0.33998104 0.652145153 0.33998104 0.652145154 0.86113631 0.34785485
5 1 -0.90617985 0.236926892 -0.53846931 0.47862867
3 0.00000000 0.568888894 0.53846931 0.478628675 0.90617985 0.23692689
6 1 -0.93246951 0.171324492 -0.66120939 0.36076157
3 -0.23861919 0.46791393
4 0.23861919 0.46791393
5 0.66120939 0.36076157
6 0.93246951 0.17132449
10 1 -0.97390653 0.066671342 -0.86506337 0.149451353 -0.67940957 0 219086364 -0.43339539 0.269266725 -0.14887434 0.29552422
6 0.14887434 0.295524227 0.43339539 0.269266728 0.67940957 0.21908636
9 0.86506337 0.14945135
10 0.97390653 0.06667134
138
APPENDIX D
Modulus of Elasticity of Soil
The modulus of elasticity of soil used in this study was obtained
by using Richert's formula (20). For an elastic soil as assumed in
Chapter I, the modulus of elasticity E can be expressed in terms of
shear modulus G:
E = 2(1+v)G (D. 1)
where v is Poisson's ratio of soil.
The shear modulus of angular grained materials can be
estimated from the empirical equation
1230(2.97-e) 2 ( 0.51+e
G cr )o
in which both G and T0
are expressed in psi, where
=1
(0- +0- +0- )o 3 xx yy zz
To is the average confining pressure of the soil element
(D.2)
(D. 3)
0- , 0- , T are normal stresses in the X, Y, and Z direction,xx yy zz
respectively.
e is void ratio which can be obtained from
e -GsYw
Yd1 (D.4)
where
where
139
Gs is specific gravity of soil
Yd is dry unit weight of soil.
The dry unit weight can be computed from the following equation:
Gs(Y--Vw)
Nd G -1
-y is total unit weight of soil
y is unit weight of water.
(D. 5)
Once the values of Gs, y, yw of 2.71, 117. 5 pcf. and 62.5 pcf.
respectively, are substituted into Equations (D.5) and (D.4), the
void ratio will be 0.94.
The vertical stress, 0 , in Equation (D. 3) was assumed tozz
be equal to the overburden pressure. Since the water level in this
study was at the top of the cell, o- can be written aszz
0- (y-yw)h
where h is depth of soil above point.
(D.6)
The horizontal stresses 6 and o- may be written asxx yy
V0- = 0-
XX yy -1) zz (D.7)
140
The moduli of elasticity of soil for the circular cells (Figures
4.1 and 4. 2) and the finite element model (Figure 4.3) are shown in
Table D. 1.
Table D. 1. Modulus of elasticity of soil.
Element Layer No. Depth, ft o-o, lb ift 2 E, kip/in 2
1 4.875 185 7.982 14.625 556 13.863 24.375 927 17.904 34.125 1298 21.185 43.875 1670 24.026 53.625 2041 26.557 63.375 2412 28.878 73.125 2783 31.019 82.875 3155 33.02
10 95.063 3619 35.3611 109.687 4176 38.00
141
APPENDIX E
User's Manual for Circular Cell Bulkhead Program
The program consists of a main program and 18 subroutines.
This appendix is included in order to explain the use of the computer
program developed for the analysis of circular cell bulkheads as well
as the capability of the computer program and restrictions concerning
the preparations of the finite element model and input data.
The computer program for the non-axisymmetrically loaded
axisymmetric solid and shell was written in FORTRAN IV for the CDC
3300 computer at Oregon State University.
E. 1 Program Capability
The computer program is used to analyze a circular cell
bulkhead which was filled with cohesionless soil. The cell can be
founded on rock or a sand foundation. The program described herein
does compute principal stresses on a vertical plane with 0 = 0,
stresses in soil elements on any desired vertical plane, shell forces
and deformations for a circular cell due to the gravitational effects of
the fills inside and/or outside the cell. The boundary axisymmetric
or non-axisymmetric surcharge loads can be added simultaneously.
The program can also analyze a shell of revolution or an
axisymmetric structure subjected to external forces. Present
142
limitations on the finite element model's size are as follows: 66
nodal points, 57 elements, 12 different material properties, 50
Fourier terms for the boundary force condition, a maximum band-
width of 28 and maximum degrees of freedom of 182. If the user
wishes to increase these limits for other computers the following
named common areas must accordingly be changed in the program:
COMMON A2, COMMON A3, COMMON A6, COMMON C7, COMMON
D2, COMMON D3 and COMMON F2. In addition, the indices on both
DO loops in subroutine INITL, the DIMENSION and EQUIVALENCE
statements in the main program and subroutines INPUT, SOISTR,
SHLSTR and TSTRES must also be changed.
E. 2 Preparation of the Finite Element Mesh
By making use of the axis symmetry of the circular cell system
and taking advantage of orthogonal properties of the harmonic func-
tions as mentioned in Chapter II, only one plane of revolution of the
structure needs to be discretized. Figure E. 1 is used as a model
structure to show how to prepare the finite element mesh. Two types
of elements are used to construct the model: quadrilateral elements
for the soil, and shell elements for the steel sheet pile. The quadri-
lateral element is sub-divided by the program into four triangles.
Therefore, a quadrilateral element will give the same results as four
triangular elements, but in the mean time a quadrilateral element has
A ZfAxis
63 64 65
59
55
51
47
43
36
29
22
15
8
1
of symmetry
60
56
52
48
44
37
30
23
16
9
61
57
53
49
45
38
31
66
143
@762
58
54
50
Steel sheet
46
39
Soil element number
37 Shell element number Roller40 41 42
Oa 032 33 34 35
24
17
10
O25g
26 27 28
18 19 20 21
11 12 13 14
O O O O O6 7
///////tifr//7///,77/////7//////,/,-//e/7/////7///7/7///7/.77///77/777///)>Fixed
2 3 4 5
Figure E. 1. Example of finite element mesh showing node andelement number scheme.
144
fewer sets of equilibrium equations in the final system than four
triangular elements.
Numbering of the nodal points should begin after the finite
element mesh has been established. Numbering of these nodes can
begin at any corner of the system and proceed either horizontally or
vertically. The primary objective of the numbering scheme is to
minimize the numerical difference between nodal point numbers
associated with the elements. The nodal numbering in this study
(Figure E. 1) was done horizontally.
After the numbering of the nodes has been accomplished, all
elements must be numbered sequentially.
E. 3 Preparation of Input Data
The following information describes the data cards which form
the necessary input data for the program.
E. 3. 1 Number of Problems Card (I5)
Columns
1-5 Number of sets of input (NPROBS)
145
E. 3.2 Control Card (815,3F10.4,215)
Columns
1-5 Total number of nodal points (NUMNPT, 66 max.)
6-10 Total number of elements (NUMELT, 57 max.)
11-15 Total number of different materials (NUMMAT, 12 max. )
16-20 Total number of pressure boundary conditions (NUMPC)
21-25 Total number of Fourier tersm (NUMFOU, 50 max.)
26-30 Total number of output print angles (NANGLE, 7 max.)
31-35 Total number of constrained boundary points (NCONP)
36-40 Problem type (NCUT)
Enter a 0 if no soil elements are present
Enter a 12 if both soil and shell elements are present
Enter a 13 if no shell elements are present
41-50 Radial acceleration (ACELR)
Enter a 0 or leave it blank
51-60 Axial acceleration (ACELZ)
Enter a -1.0 if own weight of soil elements is accounted
61-70 Angular velocity (ANGFQ)
Enter a 0 or leave it blank
71-73 Mesh check tag (ISTOP)
Enter a 0 on control card. The values of all essential
parameters are checked against their allowable values. If
146
ISTOP > 0, a mesh check is obtained then the program is
stopped and no further action is taken.
74-76 Total number of shell elements (NSHELT)
77 -80 Total number of nodal points with axisymmetric surcharge
loads (NAXISF)
E. 3. 3 Material Properties Cards (15, 3F15. 4)
Columns
1-5 Material identification number (MTYPE)
6-20 Material density (DENS, in kips per cu. ft. )
21 -35 Modulus of elasticity (EMOD, in ksi. )
36-50 Poisson's ratio (PR)
Number of material property cards must equal to value of
NUMMAT in Columns 11-15 of CONTROL CARD.
E. 3 . 4 Nodal Po int Card s (13,12, 6F 10.4 213, F9. 3)
Columns
1-3 Nodal point number (N)
4-5 Common node (NPCOM(N))
This entry indicates the number of elements being
connected at this node.
6-15 Radial coordinate (COOR(N, 1) > 0, in feet)
16 -25 Axial coordinate (COOR(N, 2) in feet)
147
26-35 Radial nodal force (FORCE(N, 1) in kips)
36-45 Axial nodal force (FORCE(N, 2) in kips)
46-55 Tangential nodal force (FORCE(N, 3) in kips)
56-65 Nodal moment (FORCE(N, 4) in kips-ft)
Columns are left blank if it is a soil nodal point.
66-68 Nodal point condition code (MCODE(N))
69-71 Type of node (NPTYPE(N))
This entry indicates the degree of freedom the program
will assign to the node. A3 is assigned all nodes except
the nodes on the shell elements which are assigned a 4.
72-80 Shell thickness (T in inch)
Columns are left blank if it is a soil nodal point.
In general every nodal point must be defined but since the
program has an automatic mesh generation feature, a minimum of two
nodal points per row need be input and the intervening points will be
assigned coordinates based on a linear interpolation procedure. For
example, if nodal point 1 is the first point in a row with coordinates
(0, 0) and nodal point 4 is the next point defined with coordinates
(33, 0), then the nodal point 2 will be located (11, 0) and etc.
The nodal point condition code (MCODE) will be set 0 unless
points 1 and 4 have the same MCODE, in which case all intervening
points will be assigned the same MCODE as the two end points.
148
The common node (NPCOM) and type of nodal point (NPTYPE)
of the omitted intervening points will be assigned the same NPCOM
and NPTYPE as the preceding point on that row.
The radial, axial, tangential forces and moments of all omitted
intervening points will be assigned 0 in all cases.
The shell thickness at each nodal point should be non-negative
and it is assumed constant over the shell length in each element.
All loads are total forces acting on a one-radian segment.
The nodal point condition code (MCODE in columns 66-68) is
interpreted in the following manner:
MCODE
Constrained Condition X
Radial (r) Axial (Z) Tangential (0)
0
1
2 X -3 X X X4 - X5 X X -6 X x7 - X X
The nodal point cards must be in order, starting with nodal
point number 1. The last nodal point card must be supplied.
E. 3.5 Connectivity Cards or Element Cards (615)
This data connects the element number with the nodal point
numbers on the element parameters. The nodal points for a
149
quadrilateral and triangular elements are listed counter-clockwise
sequentially around the element. The nodal points for a shell element
are arbitrary but all shell elements' nodal points must be in the
same pattern.
Columns
1-5 Element number (M)
6-10 Nodal point I (ICONN(M, 1))
11-15 Nodal point J (ICONN(M, 2))
16-20 Nodal point K (ICONN(M, 3)). Columns are left blank if it is
a shell element
21-25 Nodal point L (ICONN(M, 4)). Columns are left blank if it is
a shell element or a triangular element
26-30 Material identification (ICONN(M, 5))
In general, every element must be defined but with the
automatic mesh generation feature a minimum of 1 element per row
need be input. The program generates the omitted elements by
incrementing by 1 the preceding I, J, K and L. The material identifi-
cation code for the generated elements is the value specified on the
first generated card. For example if element 1 is read with values
I = 1, J = 2, K = 13, L = 12, material type = 11 and the next element
card read is element 11 with values I = 12, J = 13, K = 24, L = 23,
material type = 10 then element 2 would be assigned value 2, 3, 14,
150
13, 11, element 3 values 3, 4, 15, 14, 11 and etc. The last element
card must be supplied.
The nodal point array on the element cards must of course
correspond to nodal points on the nodal point cards and the material
identification must correspond to the materials in the material cards.
The element cards must be in order, starting with element
number 1.
E. 3. 6 Constrained Boundary Cards (515)
This data is applied primarily to nodal points on the boundaries
of the finite element model in the non-axisymmetric loading case.
A 1 is used to indicate the constraint of a nodal point.
Columns
1-5 Nodal point number of constrained node (KNOVA(N, 1))
6-10 Radial direction constraint (KNOVA(N, 2), 1 or blank)
11-15 Axial or vertical direction constraint (KNOVA(N, 3), 1 or
blank)
16-20 Tangential direction constraint (KNOVA(N, 4), 1 or blank)
21-25 Rotational constraint (KNOVA(N, 5), 1 or blank)
Enter 1 if the nodal point is not on the shell element
The number of cards must equal to NCONP, columns 31-35 of
control card. The program has an automatic constrained nodal point
151
generation i.e. , it will assign a 1 to tangential displacement
(KNOVA(N, 4)) of all nodal points of the system and radial displace-
ment (KNOVA(N, 2)) of all nodal points which are on Z-axis when
Fourier harmonic number equal to zero. The program will also
assign a 1 to rotation displacement (KNOVA(N, 5)) of all soil nodal
points.
E. 3.7 Boundary Pressure and /or Shear Cards (315,2F10.4)
If NUMPC in columns 16-20 of control card is zero or blank,
then omits these cards.
Columns
1-5 Element number (LBC)
6-10 Boundary nodal point I (IBC)
11-15 Boundary nodal point J (JBC)
16-25 Normal traction (PN in kips per sq. ft. )
26-35 Shear traction (PT in kips per sq. ft. )
Nodes I and J must be chosen such that the sequence of nodal
points on the traction cards are in the same manner as on the element
cards. The tractions are assumed constant over the length of the
element boundary. The positive senses of normal pressure and shear
are shown in Figure E. 2.
152
Figure E. 2. Boundary pressure sign convention.
The number of cards must correspond to the value of NUMPC
input in columns 16-20 of CONTROL CARD.
E. 3.8 Axisymmetric Nodal Surcharge Loading Cards (15, 4F15. 5)
If NAXISF in columns 77-80 of CONTROL CARD is zero, then
omits these cards.
Columns
1-5 Nodal point number (NPAX(N))
6-20 Radial axisymmetric nodal force (FORC(N, 1) in kips)
21-35 Axial axisymmetric nodal force (FORC(N, 2) in kips)
36-50 Tangential axisymmetric nodal force (FORC(N, 3) in kips)
51-65 Axisymmetric nodal moment (FORC(N, 4) in kips-ft)
153
If the surcharge loads are distributed over the boundary surface
of the elements, then the surface integral must be carried out
explicitly for the desired loading. The value of the nodal point forces
are total forces acting on a one-radian segment.
The number of these cards must correspond to the value of
NAXISF.
E. 3.9 Fourier Force Coefficient Cards (8F10. 5)
The Fourier force coefficients of the surface tractions and nodal
point forces of the non-zero displacement boundary conditions are as
follow:
Card 1 Columns 1-10 Fourier coefficient a0 (FORCOF(1))
11-20
71-80
Card 2 (if need) 1-10
71-80
Card 3 (if need) 1-10
rr It
r.
a1
(FORCOF(2))
a7
(FORCOF(8))
a8 (FORCOF(9))
a15
(FORCOF(16))
a16 (FORCOF(17))
aN-1 (FORCOF(N))
The value of N must correspond to the value of NUMFOU input
in columns 21-25 of CONTROL CARD.
154
E. 3.10 Angles Station of Displacements and Stresses PrintOut Card (7F10.5)
If value of NANGLE in columns 26-30 of CONTROL CARD less
than or equal zero, no angle station card is needed.
Columns
1-10 Angle station 1 (XANG(1) in radians)
11-20 Angle station 2 (XANG(2) in radians)
21-30 Angle station 3 (XANG(3) in radians)
31-40 Angle station 4 (XANG(4) in radians)
41-50 Angle station 5 (XANG(5) in radians)
51-60 Angle station 6 (XANG(6) in radians)
61-70 Angle station 7 (XANG(7) in radians)
E.3. 11 Output Control Card (215)
Columns
1 -5 IPRIN1. Enter a 1 if the output summation of total nodal
displacements are needed to be punched on cards, otherwise,
these columns are left blank.
6-10 IPRIN2. Enter a 1 if the output summation of total soil
element and shell nodal stresses are needed to be punched on
cards, otherwise, these columns are left blank.
155
E. 4 Printed Information and Data
The input data is printed first. This includes the control
numbers, the material properties, the nodal point coordinates, forces
and properties, the element connectivity, the constraint data, the
Fourier coefficients, and the angles of the results to be printed out.
For each non-zero Fourier coefficient harmonic, some computed
information is printed out next. First, the generated displacement or
rotation numbers for each nodal point in the model are printed. For
each soil nodal point there are two possible displacements for Fourier
numbers equal to zero and three for the succeeding Fourier terms.
For each shell nodal. point there is one more rotational displacement
in addition to the translational displacements. Each number repre-
sents one equation in the system. Secondly, the code numbers for
each element are printed. This information is generated from the
input connectivity and the displacement numbers and is used to
assemble the system of simultaneous equations. Finally, the maxi-
mum half-bandwidth for the system of simultaneous equations for each
Fourier harmonic is printed.
The output data is printed next. First, for every nodal point the
global displacement components and rotations (for shell node) are
printed. The units are feet for displacements and radians for rota-
tions.
156
The soil element stress coefficient on plane 0 equal zero
are printed next. A quadrilateral element will have four sets of com-
ponents, one for the center of each sub-triangle. A triangular ele-
ment will have one set of components. Listed for each element are
the radial stress Cr , the vertical stress Cr , the tangential orrr zz
circumferential stress SAO, , the shearing stresses: , TrO and
the maximum stress, the minimum stress, the maximumTZO'
shearing stress, and the direction of maximum stress from r-axis.
The unit of stresses is kips per square foot, the unit of angle is
degrees.
The next data printed is the shell nodal resultant coefficients
on plane 0 equal zero. Listed for each shell nodal point are the
vertical membrane force NzZ, the hoop membrane force Nzz , 00'
the shearing force Nz0, the vertical moment
moment M 00' and the cross moment Mz0.
M,zz the circular
Finally, the summation of nodal displacements, soil element
stresses, shell nodal stresses which were contributed from each
non-zero Fourier term for different desired angles are printed.
The shell stresses can be computed from the following equations:
N 12M dzz zzo- ,zz 12t
+
t3(E. 1)
where
0012M
00dTOO 12t t 3
Nz0 12MzOd
T= +z0 12t t3
o- ' 00, o- are vertical, hoop and shearing stresses,zz z0
157
(E. 2)
(E. 3)
respectively, in ksi.
Nzz,N00,Nz0 are shell resultant forces, in kips /ft.
Mzz' MOO' Mz0 are shell resultant moments, in kips - ft /ft.
d is the distance along the normal from the mid-plane to the
surface of shell, in inch.
t is the thickness of shell, in inch.
158
APPENDIX F
Description of Computer Program
Eighteen subroutines comprise the body of the computer
program. They are controlled by call statements. The main program
modifies the initial input boundary constrained nodal point data and
input nodal forces for each non-zero Fourier term. The calling
sequence of the subroutines is as follow:
1 INPUT 7 FVECTR
2 SUBCOD 8 BANWID
3 GENCOD 9 BANSOL
4 INITL 10 SOISTR
5 STIFF 11 SHLSTR
6 BIGK
Recycles subroutines 3-11 for each Fourier term
12 TSTRES
Recycles subroutines 1-12 for each problem.
The flow diagram for circular cell bulkhead analysis is
illustrated in Figure F. 1. The functions of the subroutines are as
follows:
Read input dataCALL INPUT
Generate quadrilateral element code numbersCALL SUBCOD
Store boundary conditions and boundary nodalforces on LUN 15
Begin Fourier loopNFOUR = 1
V
Read boundary conditions from LUN 15 and modify themfor each non-zero Fourier term
Yes
Is this the firstsucceeding Fourier term? Yes
159
Generate and print nodaldisplacement code numbersand element code numbersCALL GENCOD
Figure F. 1. Flow diagram for circular cell bulkhead analysis.
160
Initialize the structure to the initial conditionCALL INITL
Read boundary nodal forces from LUN 15 andmodify for each non-zero Fourier term
Is there anysurcharge load?
NAXISF > 0
Start DO-Loop (NN =1, NUMELT) to calculateand assemble element stiffness and load vectorCALL STIFF
Compute shell elementstiffnessCALL SHLSTFCALL SHLING
CALL SHLTRN
Compute triangular elementstiffness and body forcesCALL TRISTF, CALL SOILIN
Compute four sub-triangular elementstiffness and body forces andassemble into one quadrilateralelement CALL QUAD
Figure F. 1. Continued.
161
Assemble element stiffness matrices intosystem stiffness matrixCALL BILK
Assemble system load vectorsCALL FVECTR
NN=NN-I-1
Is DO-Loop satisfied?
Is this thefirst Fourier term or first
succeeding Fourier term?Calculate band widthCALL BANWID
Solve the system governing equationsCALL BANSOL
Compute and output the magnitude of nodaldisplacement and soil element stresscoefficients CALL SOISTR
Compute and output the magnitudeof shell nodal force coefficientsCALL SHLSTR
Figure F. 1. Continued.
162
Yes
Yes
Is there anynext problem?
NPROB <NPROBS
Compute and output the totaldisplacements and total stressesfor the required 0 -stationsCALL TSTRES
Figure F. 1. Continued.
163
1) Subroutine INPUT. The input data are read and written.
This subroutine generates a mesh of quadrilateral or shell
elements in the r-Z plane, checks the magnitude of all
essential variables against their maximum permissible sizes.
This subroutine also computes geometric properties of shell
elements and store on file LUN19. It calculates the nodal
point force from the boundary tractions.
2) Subroutine SUBCOD. This subroutine generates the 4 x 9
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 equations.
4) Subroutine INITL. This subroutine initializes a given
problem to the initial conditions.
5) Subroutine SOILIN. This subroutine calculates the 10
required volume integrals for a triangular element and stores
these data on file LUN16.
6) Subroutine TRISTF. This subroutine generates the 9 x 9
triangular element stiffness matrix and computes the ele-
ment body force vectors for the non-zero Fourier term.
Subroutine SOILIN is called for first non-zero Fourier term,
164
or read the corresponding volume integral data of triangular
element from LUN16 for succeeding Fourier term.
7) Subroutine QUAD. This subroutine computes the center nodal
point of each quadrilateral element using the mean value of
the 4 nodes of that element and then uses code numbers to
assemble four triangular elements into a quadrilateral ele-
ment. The 15 x 15 quadrilateral stiffness matrix and
associated load vector are obtained. Subroutine TRISTF is
called.
8) Subroutine SHLING. This subroutine computes shell volume
integrals using a 10 point Gaussian Quadrature Formula and
stores the results on file LUN17.
9) Subroutine SHLTRN. This subroutine generates 8 x 8
displacement transformation matrix for shell element.
10) Subroutine SHLSTF. This subroutine computes 8 x 8 shell
element stiffness matrix referred to the local coordinates
system, then transforms element stiffness matrix to global
coordinates by calling subroutine SHLTRN.
11) Subroutine STIFF. This subroutine performs the static
condensation process which reduces the 15 x 15 quadrilateral
stiffness matrix to a 12 x 12 element matrix. Subroutine
QUAD is called.
165
12) Subroutine BIGK. This subroutine uses the code number
technique to assemble the element matrices into a system
matrix.
13) Subroutine FVECTR. This subroutine uses the code number
technique to assemble the system load vector.
14) Subroutine BANWID. This subroutine computes the band
width of the system stiffness matrix for each Fourier term.
15) Subroutine BANSOL. This is a standard subroutine used to
solve the system of simultaneous equations by Gaussian
elimination.
16) Subroutine SOISTR. This subroutine outputs the magnitude of
the Fourier coefficients of the nodal point displacements,
stores the nodal point displacement coefficients of each
Fourier term on file LUNG. It computes and outputs the
magnitude of the Fourier coefficients of the element stresses
at the center node of each element, then stores the element
stress coefficients on file LUN10. This subroutine also
computes the principal stresses and their directions at the
plane 0 = 0 (plane of symmetric loading).
17) Subroutine SHLSTR. This subroutine computes and outputs
the magnitude of the Fourier coefficients of the generalized
shell stresses at the shell nodal points. The shell element
stress coefficients for each Fourier term are stored on
166
file LUN12.
18) Subroutine TSTRES. This subroutine computes and outputs
the total displacement for each of the 3 components at each
soil nodal point and the 4 components at each shell nodal
point for the required 0-stations by summing the contribution
of each Fourier term at each nodal point. It also computes
and outputs the total stress components for each element.
167
APPENDIX G
Program Listing
PROGRAM BULKHEADCC
COAmON /41/ WINPT,NUmELT. NCONP.NCOT,NSMEL'TCOMMON /12/ COOR(57,2),IS0NNI57,5),NPTYPE(66),NPCOM(66)
IF BOTH SOIL ELEMENT AND SHELL ELEMENT ARE PRESENTNCuT=12
COMMON /a3/ KNOVAC66,51.N3O0m(56,4) GO TO 13GoMmuN /91/ ACELR.ACELZ.ANGFQ,DENS(12)1 00 2 NP.1.NUNNPTCOMMOU /89/ MAX/SF,NPAX(1311FORC(13.4)
COMmr:N /C1/ FORCE(66.4) 00 3 NC=1.NCONDIF(NP.NE.4.NTEM(NC,1)) GO TO 3COMM's:: /0../ N3DO 5 L.1.5COMMON /F1/ NFOIIR,NFOF2,4UMFOUTNF.NFTIPMF1
5 KNOVA(NP,L)=KNTEM(NC.L):JIMON /F2/ FCRCOF(50),XAMG(7),NANGLEC . GO TO 6
DIMENSION KNTEM(6615) 3 CONTINUE
gg513:1=tPREAL NF, NF21 210C
KNOVA(NP,3)=0READ(60,1:0) NPROBS6 KN04A(NP,4)=1100 FORMATIT5) .
IF(NPTYPE(NP).E0.3) KNOVA(NP,5)4.1NPRO0 =1. IF (COOR(NP,1).NE.0.3) GO TO 210 CALL INPUTKNOVA(NF.2)=1CALL 5j9C0OCOORINR.1,=0.00001C
2 CONTINUEC "4' STSRE BOUNDARY CONDITION DATA ON FILE UNIT NO. 15 NCONR=NUMNPTCGO TO 45REMIND 19
7 00 B NP.1.NOMNPTkRITE115!NCONP,I(KNOVATAII),I21.5),N.1,NCONP)C 00 4 NC=1,NCONPIF(NP.NE.KNTEM(NC,1)) GO TO 9C STORE NODAL BOLINGARY LOAD:: ON FILE LON 15 DO 11 1=1.9C
11 KNOVA(NP.L)=KNTEM(NC,L)ARITE(15) IfFORCE(HP,I),I=1,4),NP=1,NUMNPT) G0 TO 12CC
Ti00AL DISPLACEMENT ANO ELEMENT STRESS-COEFFICIENTS 9 cOlTI.:DENNoVA(NP,i)=NPWILL 8E STORED ON FILES, LJN 6.7.8,10,12,20,21,22,23
C WAZ:i1:8REWI41 6KNOJAINP.4)=0VIN; 7 12 KNOV.INP,5)=1RE-41ND B
PEqINO 1) 8 CONTINUEPE-41;t1 1? talT8=a14PTg.,,.,:tip 20
C;'.Editt.) 21C ** BOTH SOIL ELEMENT AND SHELL ELEMENT ARE PRESENTAit':0 22 CE4r.r.) 23
13 DO 14 N=1,NUINPTDO 14 N=1.5
CC ".C
NF1=;i
BEGIN FOURIER LOOP
NFOUQ=1
14 KNO4A(N,M1=0NCN.GDO 16 NP.I.,NumNPTDO 17 NC=1,NCONPIFtNp.NE.KNTEm(NC,11) GO TO 1720 x--0-0,;.coFINF0uR1 NCN=NGN+1IF(X.FU.S.0) GO TO 310 00 15 L=1,5NF.rr;i?-1
hF2.',T.NF 18 KNOVA(NON,L)=KNTEM(NG,L)GO TO 16'iEMIN1 19
RE.i1":17 16 17 CONTINUEIF(NFTYPEINP).NE.3) GO TO 16RFHI,ris 17
kEwIN7i 11 NCN.NCN1KNOVaiNCN,1)=NP00 25 PL=1.NUMNPTKNOVA(NCN,2)=0GO 25 IG=1,51043vA(Nem,3)=66N)/1(IL.I)=0KN)vA(NON,41 .025 KliEm('IL,ID)=0KNO4A(Nr;N,5)=1ND14.=0
kEAD(15) NCW4PI(KSOVAIM,I/0.1,05),N31,NCONP) 16 CONTINUENDONP=NCNC
C MC017.1. E3',JNOARY CONOITIONS FOR EACH NON-ZERO CC FOLItIER Ti:Rm OF DIFFERENT SYSTEASC * GENERATE ELEMENT CODE NUMBERS AND INITIALIZE THECC00 25 NC=1.NCONP
45 CALL CEN000DO 26 K=1.5 91 CALL INITL26 P.NTEMOiC.K)=KNOYA(NC,K) CCC +" READ NODAL POINT BOUNDARY LOADS FROM FILE LUN 15IF(NrOUR.EC.1) Go TO 1 CIF(NFi.E0.11 GO TO 91 READC151 t( FORCE(NP,I1.I.1,4).NP.1.NUMNPT)C ,
8 ... IF NO SOIL ELEMENTS ARE PRESENT, NCuTwO cC P MODIFY SOLJNOARY NODAL FORCES FOR FOURIER TERN
IFINeuT.EO.0) GO TO 45 DO 109 N.1.NUMNPTDO 109 M=1,4i ". IF NO SHELL ELEMENTS ARE PRESENT, NCUT.13 109 FORCE(N,M)=FORCE(NOOX
IFINCUT.EO.13) GO TO 7 C IFIMAXISF.E0.0.0R.NFOUR.CT.13 GO TO 108
STRUCTURES
C ADD AxiSYmmETRIC SURCHARGE LOADS DATA (mAxNP.66).(mAxEL=57).(MAXmATs12).(NAxFOu$501u0 1'17 N=1,NJmNPTco 1,:6 J=1.NAX/SF REAL NF, NF2IFENFAx(JE.NE.N) GO TO 1EE1 8 READ AND PRINT OF CONTROL DATADO 1,5 K.1.4
C105 FORCE(N.K)=FORCE(N.K)+FOR:(.J.K)WPITE(61.1)GO TO 10'
1 F044AT(11 INPuT DATA t)106 :oNT:',JEREAD(60.100)NumNPT,NUmELT.NUmmAT.hUmPC.NUHFOU,NANGLE.107 :D4TINJE
1 NCO4P,NrJT,ACELROCELZ.AVSFO.ISTOP,NSHELT,NAXISF108 RERIND 19100 FORMAT(815.3E111.4.213,I4)C
IFEISTOP.E0.01 GO TO 5C ,GENERATE SYSTEM STIFFNESS MATRIX AND SYSTEM LOAD VECTOR WRITE(61.238)Cco llo sp4-1.NumELT 238 FO4mAT( /t MESH CHECK ONLY$,111
5 IF(NJMNPT.LE.MAXNP) GO TO 10CALL STIFEENN)wRITE(61.234) MAXNPCALL BIGK(NN)(STOP.1CALL Fs;:TE.ENN)
234 FORMAT( /) huHDER OF NODAL POINTS EXCEEDS THE MAXIMUM OF $.14)110 CCNTIN.,E10 IF(4omELT.LE.NAXEL) GO TO 20RE4I,.,' 15
RRIT-(51.235) MAXELRE4P,D 19235 FORmPT(/$ NUMBER OF FLEME4TS EXCEEDS THE MAXIMUM OF$,I4)IF1'(r1.E1.1) GO TO 111 ISTOP=iC20 IF(vJmrou.LE.mAxFoul GO T) 30C CALCULATE BAND WIDTH
wRITE(61,23.) MAXFOUC236 FOPHAT( /t NUMBER OF FOURIER TERMS EXCEEDS THE*,CALL BANKI0
1 $ MAXIMUM c1F-$,/A)111 RRITE(1.120) N9ISTOP=1120 FORMAT(//,t BAND WIDTH =$./5)
33 wRITE(u1,20c) NumNPT.NumE.T.NSNELT.N,immAT.NUmPC,NumF0o,NANGLE.C1 N00NP.NGuI,ACELR.A,EL!.AN:fuC 4P4". SOLVE THE STRUCTURE FOR EA:H NON-ZERO FOURIER TERM
200 FOR9AT(//$ NUMBER CF NODAL POINTS $$.14/C1$ NUMBER OF ELEMENTSX* NUMBER OF SHELL ELEMENIS2$ NUMBER OF DIFF. MATEIA.s---:ii
CALL EANSOL
i
CALL SoI:TRIF(N,)T.NE.13) CALL SH!STR 3t NUMBER OF PRESSURE CARDSIFENPCuP.ST.1.ArC. XNE.E0. 1.0) NF/s1 4$ NUM0FR OF FCUPIER TEPiSIF(NFOUR..SE.NUMFOU1 GO TO 400 5$ NU'^EER OF PRINT ANGLESC
6$ NumnfR OF CrNSTPAINEo NODES-AC SOLVE THE NEXT FOURIER TERM7$ SHELL CUT OFF NomiFRC8$ RA9IAL ACCELERATION
:i.M.t/9* AXIAL ACCELERATION300 NcoiR.---t.';oJR.1AC=L7.2.3 Xy ANGULAR VELOCITY st,F10.4/$4CtLa=J.0
CANGF0=0.0 CAN=1.0 8 PROPERTIES/F(Irovl.:T.NUmFOU) GO TO 400GO TO 20
00 50 M=1.NUHEIATCREA3(50.1a11 EiTyPE.DENS(RTYPF),EPODOSTYPE),PR(PTyPE)C ". SUM THE FOURIER COMPONENTSIE(HTYPE.LE mTyPE.GT.0) GO TO 40
C430 IF(NANILE.GE.0) CALL TSTRES
ITE(61.237) .lAx.ATRP/TF(61.250)237 PO4 IAT(/* NUMBER OF MATEPIALS EXCEEDS THE MAX. OF$.14)250 FORMAT( / //,t ENO OF THE PROBLEM$) IST0P=1IFiNPPoDS-NPROd) 30.90.80 40 wFITE(51.2a11 mTyPE,OENS(NTYpE).EN0D(NTyPE),PRImTYPE)C201 FORMAT(/* MATERIAL NuHLER - -4,15/C
.. SOLVE THE NEXT PROBLEM1$ MATERIAL DENSITY -$,FIC.3,t KIP.PER CU.FT.*/2$ m0OUtUS OF ELASTICITY--=$,F12.4,, KS1s$80 NPRoB=NPP00.13$ POISSON RATIOCO TO 10 C =*.F10.3./)
CC CALCULATE MATERIAL CONSTAmTS90 51 ±^C
=1=E).144.ENO
CC v
TE10...x(1.-y)/(11.tv)(1.-2.v))C
. Su3ROuTINE INPUTE(1.mTYPE)=TEmPEt2.mTYPEI=TEmPY/(1.-Y1COMMON /AI/ Num.:PT,NUMELT. NCONPOICUTINSHELTEl1.9TYPF)=E(201TYPE)COMMON /12/ Cno,,(67.2),ICONN(57.5).NPTYPE(66).NPCOM(66)E14.1TYPE)=E(1.mTYPE)COMMCN /AZ/ KNovA(66.51.NOOOm(66,4)E(5.mTYPFE=E(2.mTYPE)COMMON /A7/ OP .07 ,x4. ,S3 .13 ,XTt(5,MTYPE)=Ell.MTYP.)C144',4 191/ scELP,AGEL$014;$0.DENS(12)L(7.,ITYPI)=J.5TEmo(1.-2.Y)/(1.-Y)COMMON /9Z/ E(9.12)E(8,..TyPEI.E(7.11TY?E)COmoN /15/ NAXIS= NPAX(13),FORC(1394)E (4. EITYPF) =E I 7. mTyPE1COMMON $C1$ FCCE(5.4)
50 COvTiNuECcm114 /E1/ h$ou,R.NP.I.F2.mumFouoNF,NFT.NF1CDOmm.-..4 /F7/ Fo;cos(50).xAmG(7),NANGLEC READ AND PRINT OF NODAL POINT DATACOMMON $G1/ IPRINI,IPPIN2CC
wRITE(6E.2131DIMENSION MCOOE(661, T(65), roctis), J1C(15).LBC(15),213 FORmAT($1400Es.5X.$ R-oRn.s.5x,s z-oRn.:,:xy: NODE TYPEs.2X.I pNc15), pr(15). 11400112), PP(12) is NODE COOE$.2x,s SHL TNI:Ks0x.t R-LOAls,5x,s 2-LOAD:.EQUIVALENCE (mCODE(1).N000mt1.1)),(IOC(1).NODOm(1.2)1. 2 5X,* T-LOA0*.5X.$ MOMENT COMMON NODES. /)I ($8C(11.NODON(16.21).(L8:11).NODON(31.2)).(T(1)0000M(1.3)) L0
60 REA1(60.103) N.NRCOM( N),COOR(N.1),COOR(Na),(FORCE(N.J)sJwis4)01 M7.0 DE(N),NPTYPE(N),T(N)
103 FORMAT(I3,I2,6E10.4,213,F3.3)NN6=L+12X=N-LIF(L.LE.01 GO TO 70D=.(r-,C.7?(N.1)-COORIL,i))/IX02=(COD(A.2)..COOR(L,2))/ZX
70 L=L*1IF(9-1) 95.90.80
80 MC)]E(L)=5IF(mC0CE(NNL-1).EC.MCOOEI4)) MCODE(L)=MCODE(N)NPTYP=AL)=NPTYRE(L...1)IF(NPTYPE(NNL-1).ED.NPTYPE(N)) NPTYPE(L)=NPTYPE(N)NRCOm(L)=NPG0M(L...1)0)R(1.1)=CODRIL-1.1)+OR
CO-P(L.2)=000P(L-1,2)+02F0RCE(L.1)=0.0FORCE(L92)=0.0FORCE(L.3)=C.0FO,.OE(L.4)=0.0T(L)=T(L-1)CO TO 70
90 ,4,1ITE(61.203) (KI000R(K.1).COOR(K,2),NPTYPE(KI,MCOOE(K),TIK),1 (FOCE(K,J),J=1,4).NPCOMM.K=NNLIN)
203 FOR.AT(I5t2F12.4,2(7X.I5).5F12.4.I10)Ir(NUmPT-.N) 95,110.60
95 ii2ITE161.231) N239 FORMAT(/* NODAL POINT CARD ERROR N=$,I5)
ISTOP=1110 CONTINUE
C ""' REA] AN, PRINT ELEMENT PRDPERTIESC
WRITE(51.214)214 FORM:.T(ti ELEHENT*,4X,* Ity4X,* J*04X,* X$04Xs* LX.
1 ,.X,1 MATERIAL:/)N=0
130 ;4EAD(50.154) M. (ICONNIM.D.I.115)104 FORMAT(6I5)140 4.441
IF(M-N/ 185,170.150150 ICCIN(N,1)=I0ONN14-..1,1)+1
F;ONN(N.2)=ICONN(N-1,2)+1.AND. ICONN(N..1.4).NE. 0) GO TO 151ICONN(N,3).)
IXT1I(T0 ='151 1',0iN(N,T)=ICONN(N.-1,3)+1
ISOV(N,4)=ICCNN(N-1,4)41152 I.7,-.(N(N.5)=ICONN(N-1,5)170 KRiTE(151.234) N, (ICONN(NpI),I=1,5)204 F344AT(7/.515.19)
IFIN-N) 185.110.141150 IF(IUMELT -N) 190.193,130185 mRITE(51,231) N231 FORMAT(/* ELEMENT CARD EPR3R*.a5)
ISTOP=1190 CONTINUE
CC COMPUTE 54ELL ELEMENT RRORERTIESC
IFINCUT.E1.3) GO TO 199
C ". STORE SHELL ELEMENT PROPERTIES ON LUN 19PEmIND 10CO 111 N=1,NUMELTIF(ICON(9.5(.17.NCUT) GO TO 191I*ICr,NN(N.1)J*IGCNN(..,2)AT =;.5.(T(I)+T(J))/12.CR =';00Q(J.1)..LOOR(I11)CZ =CGO'(ls2)..COCP(I.2)XL =S1RTIDR 'OR 02 02 )SS =-02 /XLCS =02 /XLWRITE(19) N,XT,DROZ,XL.SS.CS
191 CONTINUE
8 READ ANO PRINT CONSTRAINED BOUNDARY NODAL POINTS
C199 WRITE(61,222)222 FORMAT(//t CONSTRAINED NODAL POINTS*,/
1 15X,* NODE:.GX,: R*,6X.* Z*,6X,* r*,2x,* ROTATION *)DO 223 M=1,NCONPREA0(60,224) (KNOVA(M,J),J=1,5)
224 FORMAT(5T5)223 WRITE(61.225) M, (KNOYA(M,J),J=1,5)225 FORMATUI5,7X15I8)
C * READ PRESSURE AND/OR SHEAR BOUNDARY STRESSES
IF(NUMPC.LE.0) GO TO 250WRITE(51.215)
215 FORMAT(//: PRESSURE AND -02 SMEAR BOUNDARY CONOTTIONS3//1 3 ELEMENT31143.3 Itg4Xs* .1314X,3 PRESSURE(KSF)3,4X,* SHERD/SF/Si)00 115 L=LINUMPCREA0(60,155) LBC(L), IBC(L),JBC(L),PN(L),PTIL)
105 FORMAT(3I5,2F10.4)195 WRITE(61.205) LBC(L), IOC(0,JBC(0,PN(L),PT(L)205 FORMAT(Iig216.2F12.4)
C * COMPUTE SURFACE INTEGRALSC
C
IF(NUMPC.LE.0) GO TO 250DO 253 N=1,NUMELTDO 260 M4.101DMPCIF(L6C(MM).NE.N) GO TO 26000 230 I=1,4J=I+1IF(ICCNN(4,4).NE.0) J4=4IF(ICONN(9.4).E0.0) J4=3IF(ICON9(4.3).EQ.0) J4=2IF(I.E1.J4) J=1K=ICONN(N,I)L=ICONN(N,J)
ICK=MCO5E(K)IF(K.NE.T3C(Mm).0R.L.NE.JBCOH)) GO TO 230
ICL=MOU0E(L)X=000R(1,11-.COOR(K,1)Y=COO4(1,2)*COOR(K,2)XX=X16.0YY=0.5'CO0R(K,1)+XXXX=1.7+XX/YYFRK=YYIPTIMM)X+PNIMM)Y)FR,=XX+FRKF2K=YY.(0T(MM),Y.-PN(MM)X)F21=XX.F7<
FORCFIK.1)=FORCE(K,t)+FRKFORCE(K,2)=FORCE(K.2)+F2KFORCE(L,1)=FORCF(L.1) +FRLFoRCE(L,2)=FoRCE(1,2)FZLIF(ICK.EJ.1.0R.ICK.EQ.6) ;0 TO 311IF(ICK.EO.2.OR.ICK.EQ.7) ;) TO 312IF(ICK.E"4.3.0R.ICK.E0.51 OD TO 313.GO TO 314
311 FORSE(K.1)=0.0GO TO 314
312 FORCE(K,2)=0.0CO TO 314
313 FORCE(,<,1)=0.0FORCE(K,21=0.0
314 IF(ICL.E1.1.0R.ICL.E0.5) 03 TO 315IF(ICL.E0.2.0R.ICL.E(1.7) GO TO 316IF(ICL.E0.3.0R.ICL.E0.5) GO TO 317GO TO 250
315 FoRCE(1,0=0.0Go TO 250
316 FORGE(1,2)=0.0GO TO 250
317 FORCEIL,11.0.0FORGE(1,2).0.0GO TO 250
'230 CONTINUE260 CONTINUE250 CONTINUE.. READ AX/SYMMETRIC SURCHAR;E LOADS
F.+
O
C COMMON /AI/ NumNPT,NUmELT, NCONP,NCUT,NSHELTIF(44.(ISF.P2.0) GO TO 410 COmmON /42/ CooR(67.2),ICO4N(57.5).NPTYPET661.NPCOm(66)NRITC((.1.411) COMMON /05/ KNOVA166.51.N330M(6C.4)
411 F04.1AT(///* AxiSrmmETRIC SURCHARGE LOAOS*// COMMON /44/ LADOMIt N07,Et.5X.: R-LOADIE.5x.A Z-L04C4,5X,t TLOAC4,5X, COMMON /46/ NC00E(57.12)2$ m3HENT//) COMMON /F1/ NFOUR,NFOF2,NDNFOU,XNE,NFT,NF1DO 412 N.i,NAXISF REAL NF,NF2RE40t2,413) NPAX(N),(FOR:TN,LT.L.1,4) C
413 FORmAT(T5.4F15.5) NFR= NFOUR -1412 HPITET61.4141 RAx(N),(FORC(N,L),L=1,4) C414 FORIAT( /i5.4F12.) C 4"Ps INITIALIZE
C . READ FOURIER COEFFICIENTS CC DO 23 1.1,NUmELT410 wRITE(61.216) 00 23 J.1.12216 FORMAT(t1 FOURIER NO.t,Sx,t FORCE COEFFICIENTS/) 23 NCCOE(I,J).0
READ(6C,106) (FOROOF(N),N.I.,NUmFOU) CO 30 I.2.NUMNPTG3 32C N.1,NumFOu DO 30 J=1,4M =N -1 30 NODOM(I.J)=0
379 wRITE( 51,206) M, FCRCOFTN1 K.0106 FG.!4Alt0r15.5) C
C *** TO GENERATE CODE NUMBERS IN TERM OF NODAL POINT NUMBERFORMAT(1110E2G.5)C
. READ ANGLES CF STRESS OUTPUT DO 12 I=/.NUMNPTC DO 2 N.1,NCONP
IFINANGLF:.GT.G) GO TO 350 IF(KNOVAtN.11.E0.I) GO TO 3311GO 11,, 3 00 4 m.2,5
350 ;FAN!,NIL.E.CT.7! NATILE.7 IF(KNOVA(140).E0.11 GO TO 4pFt.,,5,J.1).1 (XANG(N).N=IgNANGLE) K.K.I.1.,ITE151.5071 (XANG(NI.N.INANGLE) J=4-Im,207 FOR4fT(//A ANGLES OF STRESS PRINT OUT IN RADIANSt,///TF20.6/1)
C4 ENOOI0N(I
J)=K
nC * READ CuTRUT CONTROL CARDS V TOC 2 CONTINUE
IF(NRTYPE(I).GE.4) GO TO 3350 REA3(6C.J11) IPRINI.IPRIN2C111 FO?.44T(2I51
C C 4"14, THIS IS 4 SOIL NODAL POINTC CHECK FIRST FOURIER COEFFICIENT CC J=3
00 410 4=1,NUmFOU GO TO 11IFtFo4SOFIN).E0.0.0) CO TO 409 CN0 ITo
453C ." THIS IS A SHELL NODAL POINT
"400 CCNTINUE
C0 J=4
1,/TE(61,2331 11 00 31 1.1,J233 FOR4AT(//t BAD FOURIER COIFFICIENTSt) K=K+1
ISTOPTI450 IF(ISTOP.EO.0) RETURN 31
NODOM(I,L)=KCONTINUE
STOP 12 CONTINUEE NC, LADOm.K
Cc 8 .. TO GENERATE ELEMENT CODE NUMBERS
Su3RCUTINE SJBCOO CC 00 17 L=1,NUmELT
COMMON /A5/ ICODE(40) N.0C IF(ICONN(L.3).E0.0) GO TO 18
DO 111 0=1.4 IF (ICONN(L.4).E0.0) GO T3 19DO 199 J.1.9 C
CE(I,J1.0C199 NO S ,P,* THIS IS A SOIL ELEMENTI031 2c0 1.1.3CO 241 J.1.6 n10 20
201 1C30E(I.JI.J.N 19 4P.3200 N.'., 20 NOOF=3
GO TO 21CO 212 1=1.4103)1(1,7) .13 CI0 7:3F(I.9)=14 g ... THIS IS A SMELL ELEMENT
7)202 IC1Eli';)=151G3qE4,!1 .10
16ICi0E(4.2)=11 tHW.4. 1C3)C(4.31 .12
21 n414 LOU4=1.NP1C33C(4,4)ziICODE(4,,,)= NTE4P=ICONN(L.LDUM)2IC00.,:A4,61 3 UO 15 I=1, HOOFRETURN NOJI.I.K
15 NCODE(L,N3UM).NODOM(NTEMP,I)K=K+NDOF
En)
S 14 CONTINUE
17 52M711,50SUBROUTINE GENCOD
C / NFR
50 FORMAT(t1 OISPLACEMENT AND COOS NUMBERS FOR FOURIERY,30 TERmt.15,///It 103E1.18X,t DISPLACEMENTSt,/.22)(1* Rt,6X02$ 74,3x,t Portummt./1DO 51 I=IINUmNPT
51 WRITE(51.52) I. (N0)0M(I.J)1J=1.4)52 FOR4A1(15.14X.15.2(3X.I5).5X,15./)
wRITE(51.5J)53 FIRMAT(//,t ELEMENT*.28X,t CODE NUMBEPSW)
CO 5,- I=1.NUMELT54 WRI1E(51.55) I. (NC03E(I.AsJ.1,12)55 FORMAT(I5.5X.12I6,1)
RETURNEN)
8 10 P(CPIN)=7;(WI(NN.M)SU:ROUTINE INITL NP(5)=NUMNPT.1
DO 15 N=1,15CO4mON /Al/ NUMNAT.MUMELT,NCONP .NCUT 0(41=0.0COMMON /A4/ LA2OH DO )5 M4=1.15COMMON /C1/ FcRcE(66,4) 15 SYSK(MOM)=0.0COMMON /32/ 3CKitS2,28) J=NP(1)COMMON /01/ E(182) K=N0(2)00 3,1 I.,-1.tApol L=1P(3)F(L)=100
30 M=1.28 KK=NP(5)M=NP(4).0
30 iisvIL,H)z0.0 CO1RIKK,1).(COOR(J,1).0002(K,1)+COOR(L,1)+COOR(4,1))/44.0ETU CO3R(KK,2).(COOR(J,2)+COOR(K,2).COOR(L,2)+COOR(N,2))/4.0
END 00 25 N,..1A0=1.4C
j5:=Ang«ilIFINOUAD.EU.4) JJ=NP(1)SO N1B'OUT:',E STIFF()KK=NP(5)
C THIS SUBROUTINE PERFORMS THE STATIC CONDENSATION PROCESS CALL TRISTF(NN,II,JJ,KK)'C OF QUADRILATERAL ELEMENT
USE CODE NUMBER TECHNICUE TO ASSEMBLE QUADRILATERAL ELEMENTCOMMON /02/ COOR(57,2),ICONN(57,5),NPTYPE/66/0PCOM(56)COMMON /C2/ 2F(9).0(15).5TF(1) 00 17 M=1,9LGAMON /01/ SI9.51.STSK(15.15).SA(9.9) IF(TCOOC(q0000.M).E0.0) G) TO 17
C K=ICOCE(NQUAD,N)IF(ICONN(' :N.3).E:.°) GO TO 7 U(K)=0(K).0F(M)IF (ICONN(1,4).EQ.)) GO 10 8 01 17 N=1.9
IF(ICO)E(NOUAD.N).E0.0) CD TO 17". TMIS IS 4 ''.:UORILATERAL ELEMENT L.I0o0E(N2uAD,N)
Al(!4N)CAL!. SYSK(K.LI.SYSK(K,L).S(M,N)IT CONTINUE4 11,14
CO=STS<(1II.15)/SYSK(15.15) 25 CONTINUE..i(II)=O(II)-CEQt151 RETURN00 4 JJ=1.14 END
4 SYS<( II*JJ)=SYSK(II,JAC:SYSK(15,JJ)00 5 II=I.13CG.2YSK(II.141/SYSK(14.14) SUBROUTINE TRISTF(NN,II,JJ.KK)Q(II1,-;(r:)-cnc(14)
8CC
DO 5 J.1=1,13 THIS SOBROJTINE FORMS THE TRIANGULAR ELEMENT STIFFNESS MATRIX5 SYSK(II7jJ)=SYSK(II,JJ).-CC.SYSK(14,JJ)
8AND BODY FOPCE VECTOR FOR THE NON -ZERO FCURIER TERM
03 11 11=1.12YSvA II.13)/SYSKI13,13/ COMMON /A2/ COOR(57,2).1C11N157.51,NaTYPE(66)0PCOM(66)
COMMON /II/ ACELR,ACELZ,ANGFQ,DENS(12). COMMON /a2/ E(9,12)L JJ=1.12It STO S<1C1 II.J.H.SYSK(II,JJ1..C:SYSK(131JJ) COMMON /53/ C11.C12,C13.C22,C23,C33,C44,C55466
00 TO 15 COMMON /C2/ 5F(9).0(15).BTF(9)8 CONTINUE COMMON /01/ S(9.0).SYSK(15.15),SA(9.9)
C COMMON /El/ XI(10),A(9.91C ". THIS IS A TRIANGULAR ELEMENT COMMON /F1/ NFOUROF,NF2,NUMFOU,ANF,NFT,NFA
C '0**REAL NE, NF2
KK=ICO110.1.3) C NEOUR.GT.NFT, READ SOIL ELEMENT INTEGRAL INFORMATIONSLOLL TPI;TFINN.II.JJ.KK) C FROM FILE LUN 1633 15 4,1,4L.(1)=0.0
15 GI3 N)TO =3(M)BF(M) 8 INITIALIZE GEOMETRY ANO MATERIAL PROPERTIESG 10
7 CONTINUE MTYRE=ICONN(NN,5)C C11.E(1,MTYPEI
THIS IS A SMELL ELEMENT C12=E(2,MTYPE1CALL SHLSTF(NN) C13=C1203 i M.1,6 C22=C1t
9 Q(MI=0.0
ECC23.C1233.10 kNO ETURN44CmE(7cti04TYPEA
C
SUBROUTINE QUAO(NN)
C '" THIS SUBROUTINE USES THE "DOE NUMBER TECHNIQUE TO ASSEMBLEC FOUR SUB-TRIANGULAR ELEMENTS INTO A QUADRILATERAL ELEMENT
COMMON /Al/ NUMNPT.NUMELT. NCONP,NCUT,NSHELTCOMMON /02/ COOR(67,2),ICONN(57,5),NPTYPE(66),NPCON(65)COMMON /A5/ ICOOE(4.9)COMMON /C2/ OF(9),Q(15),BTF(9)COMMON /01/ S(919),SYSK(15,15)ISA(9,9)DIMENSION NP(5)
C55=C44c6c,.n44IF(N,:oiR.GT.NFT1 GO TO 15
CC FORM STIFFNESS MATRICES IN GENERALIZED CCORDINATES
CALL SOILIN(NN,II.J..1000GO TO 17
C15 REA3(1,) NN,II,JJ,KK, (XI(4),N=1,10),TIA(ItOtI*119)1,9)17 X=::33NF2C55
s(1.1)=,.xx(3)St1.2)=1C13,x)xI(2)S(1.3)=x°xI(5)S(2,2)=IC112.0C13.X1XI(1)(2.3)=C-.13.).X1(4)
5(3,3) =xxI(61.C44XI(1)Y.,1.72.C5Sf4.4)=Yxi(3)5o4.)=Y.x1(217.(,6)=Yx0(5)sc,;.5)=(C44.Y1XI(11:;(5.6)=Yxl(41S(5,1',)=Y"(I(6).C22XI(1)x=7.5s.Nr?o33Y=C33NF2S(7.71=xxI(3)S(7.A)=Y.XIC21517.9)=KxI(5):.,(4,,,)=Y.x:(1)519,=r.sx1(4)Stl ))=1,,x1(614.C66XII11x.NFC=55*NFV.K.C134=S(2,!4)=T*X1(1)V=Y+CC=44CS(1,7)=C*XII3)sf1,81=Xx112)S(1.,)7C.XI(5)5(2,7) =rX1121512.31=vx1(4)5(3,7)=C'xI(5)511,=)=X.xI(4)st3,1)=C.xI(6)511.41=0.05(I.5)=0.0S(1,4,1=C23.XI(2)S(2,4)=0.0512,51=3.0St2.61 =(C12.023)*XI411sts.41 =2.3St!.:,i:44.X1(1)S(3.:)=C22.XI141x,":654,4FY=:23.iFS(4.7)=3.0S(4.91=0.05(4,4).-xx1(2)S(i,7)=0.3515,51=6.L5(5,11=-rxI(115(0,7) =1.X1(2)S(6.8)=1.x1(1)SI6.91=tr-xlXI(4)03 10 1=2.330 13 J=1.I
10 S(/..1)=S(J.1)
FIRM BODY FORCE VECTORi
MTYDE=ICON4(NN,5)AFC=ACELR'OENS(MTTPE)OF2=-ACELZOENS(MTTPE)BFR=3ENSt4TYPE)ANSFQANGCDIF(NFOU-C.GT.11 GO TO 60caFt1)=BFRXI(7)BTF(2)=9FP'XIlliDTF(1)=3,:.XI(11BTF(4)z8FRX1(9)
C
C
CC
CCCCC
C
CCC
60
20
4,,0*
70
40
50
'6"
10
4
20
BTE(51=9FVXI(7)OTF(6)=BFC.XIl7)BTF(7)=8FRX1110)OTF18)=HFZ.x1181BTF(9)=gFC.xI(8)GO TO 70CONTINUEDO 20 1=1,9BTFII) =0.0IF(NFOUR.GT.2) GO TO 70BTF(1) =BTF(.31=BFC.XI(1)BTF141=3T=I6)=OFCXI(71UTF(7)=BTF(91=DFCXI(8)
TRANSFORM THE STIFFNESS MATRIX AND 800Y FORCEVECTOR TO GLOBAL COOROINATES
DO 40 K=1.9CO 40 L=1,9SA(K,L)=0.0DO 40 M =1.9SA(KIL)=SAIK,L)S(K,M)A01,L)DO 50 K=1.9BF(K1=0.0DO 50 L=1,9S(K,L)=3.0HF(K)=0F(<14A(L,BTF(L)DO 50 m=1.9SIK.L)=SOC,L)+A(M,K)SA(M,L)RETURNEND
SUBROUTINE SOILININN,II.JJ,KK)
THIS SURROUTINE CALCULATES THE TRIANGULAR VOLUME INTEGRALSANO FORMS THE DISPLACEMENT TRANSFORMATION MATRIX FOR THESTIFFNESS MATRIX AND BODY FORCE VECTOR
COmmoN /A2/ CooP(67.21,ICONN(57,51,NPTyPE(661,NPCOmf661CORMON /E1/ x1(10),A(9.9)COMMON /F1/ NFOUR,NF,NF2.4UMFOU,XNFOFT,NF1DIMENSION XM(6),R(6),Z(6)REAL NF,NF2
00 10 J=3.10xI(J)=0.0
FORM THE AREA AND VOLUME INTEGRALS USING THEEXACT EXPRESSIONS
R11)=COOR(II.1)2(1)=Cw3R(I1.2)R(?)=CO(J.J.1)Z(2)=CocR(JJ,2)R(31=CcO(!K.1)1(3)=CO3R(KK,2)xI(2)=5.5cR(1)C2(2)-2(3))+R(2)(2(31-2(1))
1 2(3)(2(1)-2(2)11C.xI(2)/3.0XI(I)=C.(R11113(214.4(311X1(4)=C(2(111.2(21.2(3))C=0.25CR(4)=0.544R(114.R(2))R(5)=0.5sOR(2).R(3)1R(5)=0.5(R(3).R(11)Z(4)=0.5.(/(1).2(2)12(5)=0.5(2(2)+2(311Z(6)=0.5(2(3).Z(1)10=3.,::.0DO 20 1=1,3J=143XM(I)C.RtI)XMIJI=0*RIJ)DO 30 1=1,6A=XMII)B=R(I)C=Z(I)D=ABxim=xt(7).0
1 .
...4
to
XI(8)=XI(8).A°C JJ=ICONN(NN.2)/I(1)=X1(9).0.3 HTYPE=ICONN(NN.5)XI(10)=AI(10)0C RFA011.)) N.xToR,OZ,xL,SS,CSIF(9.E7.4.).L) 8=0.00001 IFINFOUR.GT.NFT) GO TO 33U.1/03.31 IFICOOR(II.1).NE.0008(J).11.28.A113)=x14314.n i C00-2(11.2).NE.COOR(JJ.2)) GO TO 1021(5)=xE(5)+0CxI(6)=x/(6)4,0CIT 10
ISToP=1IF(GOW:(II,11.LT.0.0) IST30=1
30 CONTINUE IF(COOR(i..1.1).LT.0.1) ISTOD.1. 1F(XT .LE.O.G) IST3P=1
i '09" FORM THE DISPLACEMENT TRANSFORMATION MATRIX IF(ISTOP.EC.0) Go TO 30WRITE(61.20) NN
C=0.5/x/(2) 20 FORMAT(il BAO SHELL ELEmENT.14/1CO 4C 1.1,9
30 V1TA/215. mTYPE)00 4C J=1,940 A(I. ;)=S.0 E11=XT u(1,HTypE)-E(30TyPE)*E(30TyPE)gx)
A1.R(2).7(3)-P(3).2(2) E12=XT "(E(2.MTYPE)-E(301TYPE).E(5,MTYPE)")e)A2.1(3).2(1)-R(1)Z(3) E22=YT "(E(4.MTYPE)-E(501TYPE)*E(5.MTYPE).X)A3=8(1).?(2)-R(2)"2(1) 144=XT E(701TYPE)b1=Z(2)-Z(3)
:::::::::44X=XT "Xi /12.82=2(3)-7111
83=Z(1)-7(2) Di1=X"Et1C1 rR(3)-8(2)C2=R(1)-2(3) 022=X*E22C3=8121-'<(1)4(1.11=It4,21 =A(7.31 =11"O 39
044=X"E44DO 40 1=1.8
A(2,11=A(5,21=A(5.31=11"C 0)1)=0.3113.1)=1(6,21.1(9,3)=01"C DO 40 J=1.84(1,4)=A (4.5)=A(7.,3)=A2*C 40 0(1,J)=0.0A(2,41./(i.:1=118,01.3,2"C CAry,4)=A(6,,),AC).E0=C2.0 8.. FCP.M STIFFNESS MATRIX IN SENERALIZEO COORDINATESA(1,7)=1(4.C)=4(7,9)=13CA(2.7)=1(5.81 =1(8.91=93C 81(1,1)=E11A(3.7)=A(6.8)=1(9,9)=C3*C gilt:in°"E12IF(NrOUR.ST.NFT) G) TO 50WRITE(16) NN.II,JJ.KK, (XI(N),N=1,10),((A(I,J) 131(1,4)=11,-.E12
1 ,I=1,9),J=1,9) 61(1,5)=CS *E1250 RETURN 131(2.21=E44
ENO 01(2.3)=-NFE44C 8112,4)=-SS "E44C
I 81(2,5)=0.0SU3ROUTINE SHLSTFOIN) al(3,3)=0S 'SS *E22NF24144
01(3,41=NF"SS "(E22.E44)C... THIS SUBROJTINE CALCULATES SHELL STIFFNESS MATRIX 81(3.5)=CS "SS "E22
C. 81(4.4)=N2'E22+SS 'SS E44COMMON /12/ CC07(67.2).1C)4N(57.5),NPTYPE(66)0000M(661 131(4.5)=Nr"CS "E22COMMON /A7/ OR ,07. .XL .SS .CS ,XT 61(5,5)=GS 'CS "122COMMON /82/ E(9.12) CC04404 /341 Eti,E12.E22.E44.011.012,022.044 02(1.1)=011COmMON /C2/ OF(9).Q(15)0TF(8) 82(1,2)=0.0COMmON /S1/ S(90),SYSK(15.151,S1(9.9) 02(1,3)=!;S 112cc44:-,m /E2/ xrs(10.15) 02(1,4)=-NFcS '012COi4on /F3/ A'..:(807) 82(1.5)=-4F2.012COMMON /F1/ NFOLP.OF0F2.4UMFOU.XNF.NFT.NF1 de(2.2),xs ."-,S "044
C 82(2,3)=4F"CS "044DIMENSION IG(10), J5110),KG(101.IH(12),JH(12). 8212.4)=-CS 'CS 'SS "044
1 01(12).1(12).01)5.5).32(5,5) U2(2,5).-NFCS "SS "044MEAL NF. NF2 62(3,3).ss 'OS 022fNr2044
C 02(3,4)=-4Fcs 'OS 0122+744)OATA (1G(1)=1),(IG( 2)=2),(IG(3)=3),(IG(4)=s),(IG(5)=4), b2(3.5).-4F2.SS '(022,044)1 (IG(6).41.(VA7)=5).(IG(i)=5),(IG(91=5),(IG(101 =5), B2(4.4)=CS 'CS "INF20224SS *SS '044)2 IJC(1)=11.1.1G(2).1),IJG(3)=2),(JG(4)=3),(JG15)=2),IJG(6)=3), 02(4,51=NFcs INF2.022.SS 'SS '044)3 (J;(7)=2),(JGtm)=3),(JG(3)=4),(JG(10)=5). (KC(1)=2) 82(5 X 6)=4F2*(NF2022fSS 'SS '044)OATH (KG(2)=4),(KG(3)=1),(KG(4)=2).(KG(5)=3), U0 4., I=2.5
5 (G(6)=4).(<6(7).5),(G1I)=6),(KG(0)=7),(KG(1C)=8). CO 10 J=1.I6 ) I1 (1)=11.(IM(2)=1).(I8(3)=21(TH(4) .3).(!m(5)=3). B1(I.J)=n1(J.I)7 (1,4(6)31.(I4'(7)=4).(1M(4)=410I1(9)=5).(IH(10)=5).(IH(11)=61 90 62(I,J)=112(..),I)DATA (In(12)=5), (JH(1).11,(JH(2)=6).(.1H(3)=2), C
9 (JH(4)=2),(Jm(5)=31.(JH(5)=4),(JH(7)=7),(JH(8)=4), X=XT.XT/12.1 (Ji1(9)=7),(JH(10)=1),(im(11) =9),(JH(12)=10),2 (01(1)=7),(KH(2)=8).(xH(3)=4),(HH(4)=6),44H(5)=7),(KH(6)=8),
E44=1.44044=X.E41,
3 (<4(7)=3),(104(8)=4),(x)4(91=51,(KH(10)=6),(xH(11)=7),(KH(12)*!) IF(NFOUR.ST.NFT) GO TO 91DATA (H(11z2.0),IM(2)=6.0),(M(3)=1.0),(H(4)=1.0), CALL SHLINGINN,IIIJA
5 (M(5)=2.3),(M(6)=3.0Ig(M(7)21.0),(M(8)41.0),(M(9).1.0)41 GO TO 926 (m(10)=1.0).(m(11)=1.0).(H(12)=1.0) 91 READ(17) NM.II,JJ,((XIS(IgAvIa1,10)oJ=1,10)
C 92 DO 100 NGz1,10
CINITIALIZATION GEOMETRY AND MATERIAL PROPERTIES I.KG(NG)
X=JG(NG)/STOP=0 N=IGINGI
DIIgICONMNN,1) O 100 MGz1,10
J=tG(mG)L=.11-.(mS1N=IG(HS) .
70DO 70 .1.1.1xISII.J)=XIS(J,I)IF(NFOUR.GT.NET) GO TO 90
100 S(I.J).31I.J1+81(M,H)XISCK.L1 wRITE(17) NN.II.JJ.C1XISII.A.I=1.101.J=1.1C)CO 110 NS=1.12I .KH(NS)
80 RETURNEND
0,.....,H(N3 )
m=IH(N3) CDO 111 MO=1,12 SUBROUTINE SHLTRN(NN)J.KHIm3) CL=.11(43) FORM DISPLACEMENT TRANSFORMATION MATRIXFCR SHELL.
(Mil . C110 SII,J)=S(I,J)+HING102(MoN)H(MG)XISIK,L) COMMON /A7/ OR gOZ ,XL ,SS .CS ,XT
C COMMON /E3/ AS(8.8)C TRANSFORM STIFFNESS MATRIX TO GLOBAL COORDINATES COMMON /F1/ NFOUROIFOF2.NUmFOU.XNFOFTOF1.C REAL NF.NF2
IF(NrCjR.ST.NFT) GO TO 111CALL =HLT;?N(NN)
CxLI.I.C/XL
,,D TO 112 OCS=xLIo2ill READ(1E) Mo((AS(MoN),M=108),N=1,8) OSS=xLI1R112 OD 11: 1..1.8 X=XL1xLI
G3 130 i=1.4 DO 120 1=1.8SA(I.J)=S.G 00 120 J=118DO 133 K.1.5 120 AS(I.J)=0.2
130 SA(I.J)=SA(I.J)+S(I.K)AS(K.J) AS(1,1)=OSSDO 140 1=1.8 AS(1.2)=1CSDO 140 J=1.8 As(a.t).-DssnrSYS<II,J).0.3 AS(2.2)=-OCSxLI'IX) 141 K=1.8 AS(2.5)=-AS(2,1)
140 SYSK,:.Ji=SYSK(I.J)+AS(K.1)SA(K,J1 AS(2,G)=-AS(2.2)9 =ET.iN AS(3.3)=1.0
END AS(4.3)=-x1I
SAS(4,7)-.xLIAS(5.1)=AS(1,21
SUBROUTINE SHLING(NN,II.JJ) AS(5,2)=-AS(1.1)C AS(6,4)=1.0C THIS SUBROUTINE CALCULATES THE SHELL INTEGRALS AS(7,1)=-3.00CSXC AS17,21= :..3.0SSX
,30m-ION /A2/ COOR(57,2).13344(57.510PTYRE(6610PCOM(86) AS(7.4)=-2.0xLI,com01 /qz/ op .oz .XL .SS .O .XT AS(7,5)=-AS(7.1)
com..^N ,Z2/ xIS(10.10) AS(7.6)=-AS(7.2)ce,,:::4 /F1/ FOUR,NF.NF2.NUmFOU,XNF,NFT,NF1 AS17,91=AS14.3IGImE;4SION xS(13). XN(10). Y(10) AS(3.1)=2.0.)CS.X.XLI
C AS(5.21=-2.5USSxXLIEOJIVALEN3E (r(2).RS).(Y(1.S) AS(3.4)=X;;ATA (YS(1)=-C.,17393E53),(xs(2)=-0.95505337),(XS(3). AS18.5)=-AS(8,1)I -0.N7940,57),(xS(4)=-:,.433395391,(xS15).-0.14987434).(XS16). AS(5,6)=-AS(8,2)2 0.1 494743'.), (XS(7)=0.43334535), tXS(81=3(7940957), (XS(N I= AS(6,81=3 Ogib5V,i171,(AS(/0).0.97390653), ( %M(1)= 0.066571 4),1 %M12)= IF1NFOJR.GT.NFT) GO TO 104 1.1494;135).(XN(3)=0.21918636).(XW(4)=0.26926672). WRITC(18) NNol(AS(ItJ),I=1,8),J=1.0)5 W14(51.1.2955242Z/o(XW(61.0.29552422),(XM(7)=0.26926672),6 114.4011.2.21906636),(XW191=0.14945125)11 M(10)=0.06667134),
10 EUURN7 (Y(1).1.6) CREA.L. NroF2 C
C SUBROUTINE BIGK(NN)00 50 I=1.10 C0.1 AO J.1110 8 THIS SUBROUTINE USES CODE NUMBER TO ASSEMBLE
50 Ais(I.J)=0.0 STIFFNESS MATRIX OF THE SYSTEMA=0.5xL Cx9=3.5OR COMMON /82/ COOR(6712),ICONN(57,5)OPTYRE(66),NPCOM(66)kI=C0OR(I1o1) COMMON /85/ NCODE(57,12)
COMMON /nl/ S(9,9),SYSK(15,15),SA(9,91S 4 `HELL INTEGRALS COMMON /02/ BGK(182.28)
DO 51 K=1.10 Cxicri.Ov,;(K) IF(ICONN(VNg3).E0.0) GO T3 3S.Xxx IFAICONN(NN,41.E12.01 GO T3 4RS.41.x8XX LN.12RS=1.0/RS GO TO 5Y(3)=S47: 3 LN.8Y441=SY(31 GO TO 5Yt5)=SY(4) 4 LN=9T(7)=ASRS 5 00 1 M.I.LNY(5).SY(7) DO 1 N.MILNV(91=5Y(51 X=NCODE(NNIM)Y(11)=ST(9) L.NCODE(NNoN)XXsX/RS IF(K.EQ.0.0R.L.E12.0) GO TO I00 60 I=1,10 IF(K.LE.L1 GO TO 6DO 63 JIc10 TENP=K
60 XIS(I,J).XISII.J)0WL
IXNIK)*Y(I)*Y(J) R=LDO 2U 1=2,10 =TEMP
6 HN34=L-K#1IF(10oRN(NN.4).NE.0) CO TO 7IF(IcoN4(4,4,3).E0.0) Go TO 7
IF(NCOOE(NIT).E0.0) GO TO 20DO 30 L=I.12IF(N000E(N.1).E0.0) GO TO 30
HSK(K03ANP=BGK(K.maAN)+S(401 KK=IA4s(N000E(N,I)-NCOOE(4,01GO To 1 IFIKK.GT.J) J=KK
7 !s)(K,111,11=BGK(X,M3AN)+SYSK(4,41 30 CONTINUE1 G1,)T1uuE 20 COAT INUE
.2E-TORN 10 CONTINUEEND NB=J#1
C RETURNC END
SUBROUTINE FVECTR(NN) CC CC ** THIS SUORouTINE USES CODE NUMBERS TO ASSEMBLE SUBROUTINE BANSOL
THE SYSTEM LOAD VECTCR CCOMMON /A4/ LADOM
CommoN /A2/ CooP(67,2),ICONN(57.5),NPTYPE166),NPCON(66) COMMON /02/ BGK(182,281COm1cN /A3/ NNovA(66.5)04300m(66,4)commN /56/ N000E(57,12)
CGmm0N /03/ F(182)COMMON /D4/ NB
C0mmN /DS/ NAKISF,NPAK(13).FORC(13.4)commoi /01/ F0RCE(66.41 C
DIMENSION T(28)
00,.1111 /02/ NF(9)0(15).BTF(9) NO.LAOom00m1ON /03/ F(182) N=0CO,m3N /F1/ NFouP,NFOF2,NUMFOU,KNF,NFT,NFL 500 N.N+1CO,.10N /F2/ FORCOF(50),XAVG(7),NANGLEREAL NF,NF2 C 'ogy DIVIaE RIGHT SIDE BY DIAGONAL ELEMENT
C F(N)=F(N)/OGK(N.1)
!AllgMl%;1; C " CHECK FOR LAST EQUATIONIF(N-ND) 950.700,700
IF(KK.EO.2) GO TO 5 C '"" DIVIDE N TH EQUATION 9Y DIAGONAL ELEMENTiFtLL.ED.0) GO TO 6 550 00 6J0 K=204L4.1? T(K)=EGK(v.K)CO TO 7 600 BGK(vo()=BGK(N,K)/BGK(N,1)LN=P.GO TO 7 C REDUCE REMAINING EQUATIONS
6 04.9 DO 660 L=2097 U0 1 N././NK=N;N':E(NN.m) IF(NO-I) 660,640.640IF (( 3) GO TO 1 640 J.0CO 2 LP=1,4 DO 650 K=1.001L II , LP) J.J4.1IFtL.E0.01 GO TO 3 650 6GKII.J)=BGK(I,J)-T(L)BCK(N,K)IF( KL) 3,9,3 F(I)=F(I)-T(L)+F(N)
9 .;-,Tcm.IT 660 CONTINUETO 15 GO TO 500
3 L=Nc::-;m(JJ,LP)IF(L.E3.u) GO TO 4 C BACK SUBSTITUTION1F(K-L) 4,8,4 700 N.N1
8 NPTEM=JJGO TO 10 C CHECK FOR FIRST EQUATION
4 IF(K.K.E0.0) GO TO 2 IF(N) 900000.7501.0004(KK,LP) CIF(L.E74.0) GO TO 11 C ** CALCULATE DISPLACEMENTSIF(K -L) 11,12.11 150 Do 800 K=2.N3
12 NPTEM=KK L=N.K-1GO TO 13 IF(NO-L) 800.770.770
C
C
C
C
it IFIL.1.13.0) GO TO 2
IF( -1.) 2.14.214 (.PTEM.LL10 T=PCOm(NDTEm)
FIK)=F(K).0(m),FORCE(NPTE1,LP)/YGO TO
2 CONTINUECONTINUE
RETJRNEN)
SUBROUTINE 0ANNIO
COMMON /AL/ NumNpT,NumELT,COMMON /86/ NCOOE457,12)COMMON /04/ N8
J000 10 N.1,NUMELTBB 20 181,12
.770 F(N)=F(N)-BGK(N.K)*F(L)800 CONTINUE
GO TO 700500 RETURN
ENO
C.CC
Su3ROUTIN2 SOISTR
COMMON /Ai/ NumNPT.NumFLT, NCONP,NCUT,NSHELTCOMMON /A2/ COOR(G7,2),I.70NN(67,5),NPTYRE(66),NRCOM(66)COMMON /A3/ NNovA(60.5).NO]om(66.4)COMMON /AL/ N030E(57,12)COMMON /12/ F(9.12)COMMON /93/ C11,C12.C13.C22,C23,C33,C44455,C66
NCONP,NCUT,NGMELT COMMON /Cl/ FoRGE(660)COMMON /C2/ 3F(9).0(15)0TF(9)COMMON /01/ 5(9,9),SYSK(15,15).S8(9,9)COMMON /03/ F(182)COMMON /FL/ NFOUR,NF,NF2.NUMFOU,XNF,NFTOF1COMMON /F2/ FORCOF(50)0(A4G(7)0NANGLE
DIMENSION TOISP(151.ST(12) ,DISP(66,41 C 4"s THIS IS A QUADRILATERAL ELEMENTEQUIVALENCE IFORCE(1.11.0ISP(1,1)) CREAL NF. Nr2 CALL OUAOINN)
C 00 12 N=1,12C "P PRINT Or NODAL POINT VARIABLES IF(NC00E(NN.N).E0.0) GO TO 105C ,
K=NSOOE(NN.N)DO 41 I=1,NUMNPT TOISF(N)=F(K)CO 41 J=1,4 GO TO 12
41 DISP(I.J)=0.0 105 TOIsR(N)=0.0KFP.FcuR*1MPRINT=3 12 CONTINUEDo 50 I=1.NUMNIRT ST(K)=Q(0.121IF(CooR(I.1).LT.0) GO TO 50 DO 14 L=1.1200 4 J=1,4 14 ST(K)=ST(K)-SYSK(Kf12,L)*TDISP(L)K=N1001(I,J) Al1=SY3K(14,14)*SYSKI15.15)-SYSK(15,14)*SYSK(14.15)IF(K.E7.01 GO TO 4 A21=SY1K(13,14).SYSK(15,15)-SYSK(15,14)*SYSK(12,151DISR(I,J)=DISP(I,J)+FIK) A31=SYSK(13,14) SYSKI14.15)..SYSK(13,15).SYSK(14,14)
4 CONTINUE '422=SYSK(11,13)*SYSK(15.15)..SYSK(12,15)*SYSK(15,13)IFIHRRINT.GT.0) GO TO 40 A32=Sr1K112,13).SYSK(14,15).-SYSK(14,13)*SYSK(13,15)rIppt%T=5c, A33=SYSK(13,13).SYSK(14,14)-..SYSK(14.13)*SYSK(13.14)NR1TE(b1,2C0) NFR DETA=SYSK(13,13).411-SYSK(14,13)*A21+SYSK(15,131431
200 FoRmr.T(s1 CISFLAC7HcNTS FOR THE FOURIER TFRm, NumBER*,/3// DET1=ST(1),,Ali-ST(2),A21+3T(3)*A3iIs 432Eg,3X,* R-OISPLACEmE4Tx.3x.* Z-DISPLACENENTA,3Y. DET2=-ST(1).421,ST(2).A22-ST(3)A32Zs T-)ISPLACEmENTs.9X,* OTATIONs./) OET3=ST(1)4A3L-ST(2),A324.3Tt3)*A33
40 oRITEA.2,,,I) I, (DISPtI.J),J=1.4) TOISP(13)=DET1 /DETA201 FOl1Arli5.4E18.5/1 TDISP(14)=DET2/DETA50 miARI4T=sPRINT-1 TOISP(15)=DET3/DETA
AR4I,cf=0 CC
1,',7123nM:11C . IF LOADINGS ARE AxISY..mETRIC, NANGLE LESS THAN ZEROC LL=ICONN(NM,3)
IF(NAN;LE.LT.C) GO TO 51 HM=Ii:CNUINN,4)C KK=NUPT4.1C .0.* STORE NOOAL DISPLACEHENT COEFFICIENTS ON FILE, LUN 6,7,0 CO3', (IOC, 1.) = (COOR My 1) +COOR WO.) +COOR (LL, 1)+COOP (Mt1.1) )/ 4.C COOR(KK,2)=KOOR(II.2).+COOR(JJ,2)*COORILL.2)+
IFCNFOJR.LT.Bi GO TO I.I SOOR(M4,2))/4.
IF(NFouR.LT.15) GO TO 2iIRITEI81 claisP(K,L).L=1.4)K=1NUmNPT)
DO 16 NuILJAO=1,4II=ICONN(NN,NQUAC)
CO 1,-) ii JJ=ICONN(NN.NQuA041)1 ARITE(b) C(OISP(K,L),L=10),K=1,NumNPT) IF(Nou40.EQ.4) JJ=ICoNN(NN,L)
CO TO 51 AJ=C00R(JJ.1)-CoOR(II.1)2 wRITF(s) l(DISP(K.L),L=1.:).<=1.NUmNPT) AK=CO1Rtp(K.11-COoR(TI,1)
0J=cuoR(JJ,2)-COOR(II,2)C ... IF NO SOIL ELEMENT IS PRESENT ,NCUT=0 BK=CooR(KK.2)-COOR(II.2)
51 IF(NSUT.E.C) GO TO 300 A.AJ-AKNR1TF(61,36) NFR
36 FORHAT(g1 ELEMENT STRESSES FOR FOURIER TERM NUMBER*,U=3J-HKD=AJK-..AK*F5,1
1 13,//i.119Y,, OIR.OF mAx.s/s EL NOnE RR-STRESS*, RI.,COOP(II,1)2t 2/-STRESS TT-STRESS RZ-STRESS RI-STRESS*, 2I=C00R(II.213s 2Y-STRESS MAX-STRESS MIN-STRESS MAX-SHEAR$1 RJ.CooR(JJ,1)4$ PRIN:.sTRESSsis NO NO NO*,106X, 2J=CODRAJJ.2)5$ FROM R-AXISCOEGW) RK=COORCK<,1)
C 2K=CoOklY,<.2)C *4'0 CALCULATION OF CENTER NODE VARIABLES RC=CRI+RJ+RK)/3.C
DO 300 N4=1.NUMELT2C=17I+7J.2i0/3.I=NluAl
mTyRF=ICoNN(NN,5) J=NOUGOtiIF(ITYPE.SE.NCUT) GO TO 300 IF(NQUAO.EQ.4) J=1C11=FlI,HTYPE) K=5C12=E(2.ITYPEI CC13=n12 IZT=-A TOISP(31)-AKT015(3.J1+AJTOTSP(3,01!1C22.C11 EPRRr(2 TIISP(3.1-2)r0KT)ISP(3*J-2)-RJ+TnIS0(3,0K-211/0C?1=C12 ERZZ=I-A TOI;R( 3' I1)-AK.TOISP(3J.1).AJTCISPI3K1))/0C33=C11 ERRZ=l-A"TDISIR(3 1..214.9T)ISP(3.1...1).-AK.Tn/SF(2.J2)C44=E(704TYPE) 1 ,0,0,T0Isp(3 J-1).AJ.TOISP(3,1(-2)-BJ.TOISp(3K-1))/0C55=C44 ER7T=TZT/0C65=C44 EPRT=-((RJ2K-RK.2J)*TOISD(3.I1s(RK.ZI-RI,21414TOISP(3J)
C 1 +IRI 2.1-RJZI)*TOISP(3.K)+20TZT)/(RC0)C ". INITIALIZE STRAINS u=(TOIsR(3.1-2)+TOISR(3J-2).ToISP(31<-2))/3.C W=( TDISP13.1-1).TDISP(JJ-114,TOISR(3.4(-1))/3.
EP?R=0.0 V=( TiIS2(3°I).TOISR13.J)44OISP(3K))/3.EP22=0.0 EPTI=(u+Nrv)/RCEPTT=0.0 EPRT=EPRT-NF.usRCERR2=0.0 EP2T=EPZT-NFm/RCEPRT=0.0 8 . CALCULATE STRESSESEP2T=G.000 15 N=1,9 ST(1)=C11EPRR*Cl2'EPZZ.C13EPTT
1S ST(N)=0.0 ST(2)=C12.EPRR4,C2EP224,CE3EPTT ,-
IFIICONN(NN.4).E0.0) GO T3 104 ST430=C13EPRR.C23EPEZ*C33EPTT -.4
C ST(4)=C44EPRZ -.1
ST(5)=C55EPRT EPRR=EPRP+ERRST(6)=C66*EPZTIF(NFOUR.GT.1) GO TO 26
EPZZ=EP17+E77EPTT=ITOISR(10)+NFTOISR(1211/RK+EPTT
ST(5)=0.0 LPRZ=FPR7+ERZST(6)=6.0 EPRT=EPRT+CERT-NFTOISP(10)1/RK
26 IF(NAIiSLE.LT.0) GO TO 24 EPZT=EPZI+EZT-NFTOISP(111/RKIFINFOj2....T.5) GO TO 5IF(NOUR.LT.9) GO TO 6 E g" CALCULATE STRESSESIF(NOUR.LT.131 GO TO 7 CIr(4FOUR.LT.17) GO TO 8 ST(1)=C11.EPRIR+C12*EPZZ+C13EPTTwRITE(23) NN,II,JJ.(ST(N)04=1.6) ST(2)=C12EPRR+C22EPZ74:23,EPTIGO TO 24 ST(3)=C13+EPRP+C23EPZZ+C33EPTT
5 WRITE(101 NN.II.JJ.(ST(111,14=1.6) ST(4)=C44.EPRZGO TO 24 ST(5)=G55+ERRT
6 NRITE(23) NN.II.JJ.(ST(N).4=1.6 ST(b)=066+EP2TGO TO 24 IF(N0 '!GLE.LT.0) GO TO 20
7 WRITE(211 NN.II,JJ. (ST(N).N=1,61 IF(NFUR.LT.5) GO TO 42GO TO 24 IF(NF)OR.LI.91 GO TO 43
8 wRITE(22) NN.II,JJ. (ST(N).N=1.6) IF(NFOUR.LT.131 GO TO 44C IF(NFOUR.LT.171 GO TO 45C ". CALCULAT7 PRINCIPAL STRESSES AND DIRECTIONS WRITE(23) NN.II,JJ.(ST(N),N=1.6)
24 C414'4.(ST(1)+ST(211/2. CO TO 23C4O1R=(3T(11-ST(2))/2. 42 WRITE(101 NN,II.JJ,(ST(N).N=1,61Rml4QT(0T141'..2+0m0HP'.2) GO TO 20ST(7)=OMOIRPm0HR 43 WRITE(2C) NN.II.JJ.(ST(N).N=1.6)sT(8)=c4nAR-RmomR GO TO 23ST13)=RYCHR WRTE21AN.-,Li:=CAIINFIST(4)/ST(1)157.2957795)/2.
44GO
ITO (20
) NN.II.JJ.(ST(N1.4=116)
IF(GT(11.LT.S1(2)) GO TO 22 45 wRITE(22) NN,II.JJ.(ST(N),4=1.6)G7 TO 23
22 1,.(ST(4).LT.O.C) GO TO 25C8 ... CALCULATE PRINCIPAL STRESSES ANO DIRECTIONS
ANGLE=17.7-A8SF(ANSLE)GO TO 23 20 CMOHR=AST(1)+ST(21)/2.0
25 ANGLE=-90.C+ATISF(ANGLE) OMOHR=(ST(11-ST(2))/2.023 W?ITE(51.37) NN,II JJ, (ST(L),L=1.9).ANGLE RMOHR=SORT(ST(4)**2+0MOMR"2)37 F1=-441t 513 .9E1Z .4.F1,33/1 ST(71=0M0HcaRm0H2lE C'4TINVE ST(A)=CMO1R-RMOHR
GO TO 370 ST(9)=RMOH2NF(ST(41/T(91
C " THIS IS A TRIANGULAR ELEMENT 1ANGLE=( ASI *57.295745S)/.10
C IF(ST(1).LT.ST(2)1 GO TO 32104 CONTINUE GO TO 33
00 10 N=1.9 32 IFIST(4).LT. 0.0) GO TO 35W=NCOr2E(U4,N) ANGLE= 90.47-AOSF(ANGLE)IF(K.E0.01 GO TO 110 GO TO 33T:ISR(N)=F(K) 35 ANGLE= -90.0+ABSF(ANGLE)GO TO 10 C
110 T3IS('(N)=0.0 C '4', PRINT STRESSES FOR EACH F)URIER COEFFICIENT10 COATINJE C
00 11 L=1,3 33 WRITE(61.37) NN,II,JJ, (ST(1.),L=1.91.ANG;-EI=3+LJ=3+L
300RETURN
K=O+L ENDIt TO/SP(I)=(TOISP(L)(1TOISP(J)+TOISF(K)1/3.0 C
II=ICONN(NN.11 CJJ=IO7f41(44.2) SU1ROUTINE SHLSTR10,..ICON(NN,3)RT=100
CA(II.1) COMMON /01/ NUMNPT.NUmCLT, NCONP,NCUT NSHFLT
zr-4,2^;'(I(.2) COMMON /A2/ COOR(67,21.IGO4N(57.510PTYPE(6610PCOM(661RJ....)7:'(JJ,1) COMMON /05/ ICODE(4,9)2J=C774(JJ,2)R<=01)4(K<.11 Eglg Wi; W5P;;L2:55. ,CS ,XT2.(=C00 4(K.2) COMMON /R4/ E11,E12,E2Z,E44,011,012022.044AJ=RJ-RI COMMON /02/ OGK(182,28)AK=RK-RI COMMON /03/ F(18216.1=2J-ZI COMMON /ES/ AS(8.8)BK=ZK-zI COMMON /F1/ NFOUR,NFOF2.4UMF0J.XNE.NFT.NF/A=AJ-AK COMMON /F2/ FORCCF(50),XANG (7) ,NANGLE.8=3J-9K0. A J `,3<- l< 3..1
CDIMENSION TOISP(15), PN(8). ST(12),STR(66.6)
T7T=-A.TOISP(3I-4K,TOTSP(5)+AJTDISP(9) (UI.PN(5)). (WI,PN(1)). (VTIPN(3)) , (T10214(6)3EP.2.(5.T) isP(1)35K.TOIsP(4)-9JTOTSp(71)/13 (EUstliw(2)),(ETT,PNC4Ii,(STRI1,11,0Gx(1,11)EZ2=(-4T:ISP(2)-AKTOISP(51+AJTOISP(811/0LR2=(-AT3I5P11)+8 TOISP(2)-AK.TOISP(4).0KTDISP(5)
1 .11J+TDISP(7)-8JTDIS2(811/0 C '4'4' INITIALIZATIONEZT=TZT/0ERTs-c(RJ lx-RKZJ).TDISP(3).(RK,o2I-RI.ZK)*TOISR(6) 00 Li NR=1,NUmNPT
1 .(RI.FZJ-RJZI)*TOISR(11.2KTZT)/0 00 11 Im=1,6,RK=(RI.R.DRK)/3.0 11 STR(NR,Im) =0,8.2KsIZI42.1.210/3.0 NFR4NFOUR.1
ML=0 1* NODE 8,6X,8 LONG STRESS 8,3X,8 CIRC. STRESS 822X,00 500 NN=1,NUMELT 2$ SHEAR-STRESS 8,3X.8 LONG MOMENT *,3X,* CIRC MOMENT *.mTyPE.ICCNN(NN,5) 3 2X,8 CROSS MOMENT 81IFINTYPE.LI.NCUT1 GO TO 500 00 1.40 Np=1,NumNFT4L=m1.41 IF(NPTYPE(NP1.NE.4) GO TO 140II=ICON4(NN,1) NRITE(61.201) NP,(STR(NP,N),N=1.E)JJ=ICCN4(NN,2) 201 FORMAT(/3X,I5,6E18.5)RSI.1.0/DOORIII,1) 140 CONTINUERSJ=1.0/CCOP(JJ.1) CREAD(111 I,XT,OR,OZ,XL,SS.CS C IF LOADINGS ARE AXISYMMETRIC , NANGLE LESS THAN ZERODO 6 N=1.9 CK=NO3OEINN,N) IFINANGLE.LT.0) GO TO 310IF(K.E0.01 GO TO 103 CTDISP(N).F(K) C STORE SHELL NODAL STRESS COEFFICIENTS ON FILE, LUN 12GO TO 6 C
103CONTINUE 310
3 TDISP(q)=0.0 WRETURRITE(121 ((STRINP,N),N=1,6).NP=1,NUMNPT)
CN
C ." COMPUTE GENERALIZED DISPLACEMENT COORDINATES, PN8
END
CREA0(181 NN, c(AS(I,J),I=1,8),J=1,61 SUBROUTINE TSTRES00 100 K=1,8 CPN(4)=0.: COMMON /01/ NUMNPT.NUMELT, NCONP,NCUT,NSHELTDO 100 L=1.8 COMMON /32/ COOP(67,2),IC1VN(57,5),NPTYPE(66),NPCOM(66)
100 PN(K)=0N(C)+ASIK,L)*TOrSP(L) COMMON /43/ KNrjVA(66,3).N030M(6°,4/C COMMON /C1/ FORCE(b6.4)C COMMON /02/ OCK(182.24)C "4. COMPUTE 7,TPAINS COMMON /F1/ NFOUR,NP,NF2.4UMFOU,XNE,NFT,NF1
UJ=UIXL'lTI.XL0IPN(7),XLPN(8))/ COMMON /F2/ *014.DoF(50),*A4G(7),NANGLEwJ=WI,xL:253I COMMON /01/ IPRIN1,IPRIN2 .VJ=1I.XLETT CTJ=TI,XL(2.0.PN(7).3.0*XLPN(8)) DIMENSION 8ANGLEC7), DIS(66,4),ST(6),0/SP(66,4).EDIS.I.EP;SI 1 STR(66.6).STRES(66,241EoTTI=R31*(NF*V19.CS.UI.SSWI) EQUIVALENCE (FORCE41.DOISP(1,1/),ISTR(1.11,0GK(1,11).EPTTJ=F,SJ*(N,"'VJ.CS.UJI.SS.WJ) 1 (01.5(1,1),OGK(33,3)),(STRES(1,1),OGK(115,41)
DATA (RAD=57.2957795)GAMIETT-6I.ISS.VI.NP.WI)G14J=ETT-7JolSS*VJ.NF.WJ) REAL NF, NF2X<S7,I,-2.3.RN(7) Cki',;CJ,..8KCS1-6.0*PN(9)XL NF3=NUMF0J-3X<TTI.S.I.(74SI*(NF 2fOI.CS.NF.vI)-SS*TI))(KTTJ=R;J.(.--zsd(Nr2..uJ,csN*4,c))-SSTJ)
C 04. SUM THE DISPLACEMENTSC
x*STI.2.0*PSI*(NF0T/+0SI*NF.SSUI+CS(ETT-RSI.SSVI)) DO 200 NA=1,NAUGLEXK3TJ=2.0.RSJ.INF.TJ.RSJ*14F*SSUJ.CS*(ETT-RSJ4SS.VJ)) D3 19 NP=1,NUMNPT
C 00 19 J=1,4C 'A" COMPUTE STESSES 19 DIS(NP,J1=0.0C XANGLE(NA)=XANGINA)RAO
03 110 N=1,12 NRITE(61.15) XANGLE(NA1110 STIN1=0. IF(IPRINI.E0.01 GO TO 52:
ST(11=E11EPCSI.E12'EPTTI PUNCH 17, XANGLEINA)5T17) =E11.FPCSJtE12.ERTTJ 52 DO 50 N=1,NUMNPTST(2)=E124EPSSI.E22'EPTTI CO 53 L=1,4ST(3)=E12°F_PCSJ.E22EPTTJ 50 OIS(N.L) =0.0ST(3)=E4'.'54mI REWINO 6ST111=E44oC,AmJ REWINO 7ST(4).011*KSS1+012**TTT REWIND 8ST(10)=011**SSJf.D12**TTJ DO 100 NFOUR=1,NUMFOUST(5)=312*XKCSI.022XXTTI 1*(*DRDortNFouR).E0.0.0) GO TO 100ST(11)=012YWSSJ.022.XKITJ NFR=NFOUR-1ST(6)=C4'.XKSTI IFIIFRIN1.E0.0) GO TO 55$1*(121=0XR;TJ PUNCH 54,NFR03 12,3 N=1.6 54 FO;MATUA SUM OF NODAL DISPLACEMENTS UPTO FOURIERf,I5,H.4.6 18 HARmONICSx./1IF(mL.NE.11 GO TO 121 55 IF(NFOUR.LE.NF3) GO TO 57STF'CII,N)=STRIII,N1.ST(N) NRITF(61.56) NFRGO TO 121 56 FORMAT(///8 SUM OF NODAL DISPLACEMENTS UPTO FOURIER*,
121 ST(II,t()=5TR(II,N)+ST(N)/2. 1 15,8 HARmoNICS*,/)123 IF(mL.NE.mSHELT) GO TO 122 C
C C . READ DISPLACEMENT COEFFICIENTS FROM LUN 6,701STR1JJ01/=STR(JJ,N).ST(11)GO TO 120 57 IFINFOUR.LT.I/ GO TO 1
122 STR(JJ.N1=STR(JJ.N).STIM)/2. IF(NFOUR.LT.151 GO TO 2120 CoNTLNUF READ(*) ((DISP(N,I),I=1,4),N=1,NumNPT)
IF(mL.LT.NSHELT) GO TO 503 GO TO 3
500GGONTO To
INUE125 1 REA71(6) 11DISP(N,I),I.1.41,N.I.NUMNPT)
2CGO TO 3RE40171 ((01SP(N,I),I.1.41.H141NUOINPT)
C + PRINT SMELL STRESSE 3 NF=NFOUR-1S FOR EACH FOURIER COEFFICIENT1....'
C CO,,COS(NFXANG(NA))125 WRITE161,200/ NFR $I.SIN(NFxANG(NA)) -4200 FORMAT(A1 SHELL STRESS FOR FOURIER TERM NUMBER*,I3//0 00 90 N1914UNNPT .0
00 83 L=1,4 250 STRES(N.M)=STRES(N.4)+ST(4)00K=41';04(N.L) DO 26G M=5.6IF(K.E0.)) GO TO 70 260 STRE7(N,M)=STRES(N.M)*ST(4).S1IF(L.E0.3) GO TO 60 IF(IPRIN2.EQ.0) GO TO 411DISL)=DIS(N.L).0ISP(N.L)*C0 PUNCH 23. N.(STRES(N.m).m=1.6)GO TO 50 411 F(NFOUR.LE.NF3) GO TO 270
60 DIS'N.L)=01S(N.L)+OISP(N.L)SIr.3 T7 eo
WRITE(61.410) N,II.JJ.(STRES(N01)021,6)
70 OIG(N.L)=0.0 CGO TO 270
80 COVITY),- C THIS IS A QUADRILATERAL ELEMENTIF(N5nR.LE.NF31 GO TO 58 6 DO 7 N)UAD=1.4WRITE(51.25) N.(0IS(N.L).L.1.4)
58 ;F(IRRINI.E0.0) GO TO SOPUNCH 15. N. tO/Stu,L).L=1.4)
90 CONTINDE100 CONTINUE READ(23) 4.II.JJ.IST(J).J=1.6)17 FCRMAT(t TOTAL FOURIER OISPLACEMENTS ATY,F7.3.1 DEGREES*, GO TO 361//A NGOEA.6Y.g R-OISPL.X,iX.$ ZDISPL./.6)(si T-OISPL.F, 31 REA0(10) M.II.JJ.(ST(J).J=1.6)2 61,.1 ',70TATIGt) GO TO 36
15 FORMAT(A1 TOTAL FOURIER DISPLACEMENTS AT :.117.3,* OEGREESS. 32 REA0(20) M,II.JJ.(ST(J).J=1.6)1/1/A 'IO0Es.67,8 R-DISPi.8.6X,$ 2-DISPL.$6)(,* T-DISPL.*, 60 TO 352 68,: R77ATIONA/) 33 READ(21) M.II,JJ.(ST(J).J.1.6)
25 F0,4aT(I0,,E15.5/) 00 TO 3610 F7 ?4LT(Iti,4E15.5) 34 REA0(22) M.II.JJ.(ST(J).J=1.6)200 004TINUE 36 NO1= NOUAO -1
C =1C ", IF NO SOIL ELEMENT IS PRESENT , NCUT=0
DO
CJ=NO4 1.I6.I 6IF(I.GE.5) GO TO 12
IF(NCUT.F.O.G) GO TO 131 STRES(N.J)=STRES(N,J)+ST(I).00
C ... SU4 SOIL ELEMENT STRESSESGO TO 9
12 STFES(NO)=STRES(N.J)+ST(I)SIC 9 CONTINUc
IF(NFOUR.LT.5) GO TO 31IF(NFOUR.LT.9) GO TO 32IF(mFouR.LT.13) GO TO 33IF(NFOUR.LT.17) GO TO 34
29
61
6362
641
CCC
65
CC
23
20
21
2224
00 5)C NA=1.NANGLEXAq0LE( N.)=Y1NG(NA)*RAO00 2R N.1,NDmELT3C 2'? H=1,24STREct.,,m),,c.0rf,(IPL,I.Eo.o) GO TO ElPLir:h 27, XANGLE(NAI%1TEIS1.35) XANGLE(NA)
REwl'i 1]REAINO 2.?REdINO 21RE4I.1l 22REM:NO 2300 TYL ric7u,?=1,NumFOUIFIFO,'C?5(NF)UR).EO.G.0) ;0 TO 3004FR-..:J-;-1IF(I0,IN2.EC.G) GO TO 62PJ.U.:H 53. NFRFO,MAT(z SUM IF ELEMENT STRESSES UPTO FOURIERS,I4,sHARMONICStlIF(NFOUR.LE.NFS) GO TO 65m2ITE1;1,:,4) NFRFORMAT( / //t SUM OF SOIL ELEMENT STRESSES UPTO FOURIER:,IS,i HimONICSss)
READ SOIL ELEMENT STRESS :3EFFICIENTS FROM LUN 13,20,21,22,23
hi= m.:F0;J-100=CCS)Nr7ANG)N41)SI=3:uW"x.:04G(NA)l00 27G N=1,NlmELTLF(/GT.N(N,5).;E.NCUT) GO TO 270IFAICONN)N,41.NE.C) GO TO 5
THIS IS A TRIANGULAR ELEMENTIF(N=OJR.1T.51 GO TO 23IF(F0jR.LT.91 GO TO 20IF(NFOjR.LT.131 GO TO 21IF(NGiR.LT.17) GO TO 22REA11231 H.II.JJ,(ST(J).J=1.6)GO Ti Z4T,Ei0(1,;) 4.II,JJOST(A.J.1.6)GO 73 24READCZC) M.II,JJ.(ST(J).J=1,6)GO TO 24REA0(211 M.II.JJ.(ST(J),J.1,61GO TO 24RE40(22) 4,II.J.1,(ST(J),J=1.6)00 250 M=1,4
CCc
CCC
CCC
J1=J-5IFiNrOUR.LE.NF3) GO TO 66hRITE(61,410) N.II.JJ.(STRES(N.K).K=JI.J)
66 IFLIPRIN7.E0.01 GO TO 7PUNCH 28, N.(STRES(N.K1.K=J1,J)
7 CONTINUE270 C3N7INJE300 CONTINUE35 FORMAT(g1 TOTAL FOURIER SOIL ELEMENT STRESSFS AT A,
1 F7.3.8 DEGREESt,///A ELEM NODES:OX,: RP-STRE.'33:,57.: 72-ST RFS8,2 07.8 TT-STRESSS,SX's RZ-STRESSx.5X,t RT-STFFSSi,5X.X 2TSTRESST,3 1)
27 FORMAT): TOTAL FOURIER SOIL ELEMENT STRESSES ATs,F7.3.IS DEGRrrSt//t ELEM RR-STRESS 77-STRESS$,2A TI-STRtS., R7-STRESS RI-STRESS ZT-STRESS:)
410 FORAT(315,61:15.5/)28 FORmaT(15,6E12.41500 CONTINUE
"' IF NO SHELL ELEMENT IS PRESENT , NCUT3
131 IF(NCUT.E0.13) GO TO 10
s" SUM SHELL ELEMENT STRESSES
DO 600 NA.I,NANGLEXANGLF(NA=xANG(NA).RAOIF(IPRIN?.EQ.0) Go TO 601PUNCH 43, XANGLE(NA)
601 WRITE(61,45) XANGLE(NA)00 600 N=1,NUMNPTCO 610 m=1,6
610 sTREs(N.4).0.0REHINO 12DO 70.0 NFOUR=1.NUMFOUIF(FORCOP(NFOUR).EQ. 0.0) ;0 TO 700NER=NFOUP-1IF(IPRIN?.E0.0) GO TO 682PUNCH 603, NFP
603 FORMAT): SUM OF SHELL N000L STRESSES UPT0*.I5,is FOURIER HARmONICSt)
602 IF(NFOUR.LE.NF3IGO TO 67WRITE(51.604) NFR
604 FORMAT( / //t SUM OF SHELL NODAL STRESSES UPTO..I5.1$ FOURIER HARMONICS: /) 1....
CO. READ SHELL STRESS COEFFICIENTS FROM LUN 12 C)