VERIFICATION AND OPTIMIZATION OF NONLINEAR SHELL BUCKLING

FORMULA OF THIN-SHELL STRUCTURES

Jie Xu
Student Number: 4701732
2
 VERIFICATION AND OPTIMIZATION OF NONLINEAR SHELL BUCKLING
FORMULA OF THIN-SHELL STRUCTURES

by

Jie Xu
Student Number: 4701372

ADDITIONAL GRADUATION WORK

Submitted in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

in

Delft University of Technology

Supervisors:

Dr. ir. P. C. J. Hoogenboom

Dr. ir. C. van der Veen

3
4
 Preface
By taking shell buckling as the research subject I could be able to gain a better comprehension
of shell buckling behavior, at the same time, I could also gain 10 ECTs to take the report as the
assignment of CIE5050-09 (Additional Graduation Work, Research Project) in Delft University of
Technology.

This report is intended for engineers that are facing to designs of shell structures and students to
help them gain a better understanding of shell buckling behavior.

Appreciation to Dr. Hoogenboom and Dr. van der Veen for their patience and generous help.
Appreciation to Zhenheng Kong and Jijia Que for lending me their lap-tops.

5
6
 VERIFICATION AND OPTIMIZATION OF NONLINEAR SHELL BUCKLING
FORMULA OF THIN-SHELL STRUCTURES
Author: Jie Xu
Student Number: 4701372
E-mail Address: [email protected]

SUMMARY

Shell is a popular form of structures as it can withstand relative large load with small thickness,
which satisfies the aesthetic preference. Due to the curvatures of shell structures, shell can carry
the distributed load as membrane forces instead of bending moments. Due to the small thickness,
shell structures are highly sensitive to imperfections. As the loads are carried by membrane forces,
buckling failure often govern the design of shell structures. The buckling load could be predicted
accurately by performing a nonlinear finite element analysis. However, performing a nonlinear finite
element analysis is very time-consuming and expensive as a lot of computational efforts required.
The currently used design formula of shell structures can lower down the requirement of computation
in a great degree while the accuracy is not very satisfying.

Different from other types of structures, shell buckling often starts locally, therefore, an assumption
is made that shell buckling is a combination of membrane forces and curvatures. Thus, an improved
buckling formula is proposed. In this report, the proposed formula is verified and optimized. In
addition, the shell behavior before buckling has been illustrated by a series of figures that show the
deformation and stresses at several load steps.

Supervisors:

Dr. ir. P. C. J. Hoogenboom

Dr. ir. C. van der Veen

7
8
Contents
1 INTRODUCTION 5
1.1 Thesis Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Research Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Research Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Thesis Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 BACKGROUND 7
2.1 Shell Linear Elastic Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.1 Coordinate Systems and Curvatures . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Shell Buckling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.1 Static Instability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.2 Bifurcation Buckling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.3 Imperfection Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 FEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.1 Finite Element Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.2 FEM Solution Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4 Initial Imperfections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5 Proposed Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3 METHODS 16
3.1 Model Creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.1.1 Shape of Shell Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.1.2 Nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.1.3 Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.1.4 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2 Mechanical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2.1 Linear Buckling Analysis (LBA) . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2.2 Geometrically Nonlinear Analysis with Initial Geometric Imperfections (GNIA) 21

4 RESULTS 27
4.1 GNIA Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.2 Curvatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3 Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.4 Young’s Modulus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.5 Buckling Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.5.1 Displacements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.5.2 The Third Principal Stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

5 DISCUSSION AND CONCLUSION 35

5.1 Discussion of the Correctness of the Results . . . . . . . . . . . . . . . . . . . . . . . 35
5.2 Discussion of the Scatter Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.2.1 Curvatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.2.2 Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.2.3 Young’s Modulus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

1
 5.2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.3 Recommendations for Further Research . . . . . . . . . . . . . . . . . . . . . . . . . 36

A Command Lines of ANSYS Mechanical APDL (modified from [15]) 39

B Curvatures 45

C Thickness 58

D Young’s Modulus 59

2
List of Figures
1 Coordinate systems[9] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Normal section curvature[9] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3 Positive internal forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4 Load-deflection curves showing limit and bifurcation points: (a) General nonlinear
analysis, and (b) Asymptotic analysis[6, 15] . . . . . . . . . . . . . . . . . . . . . . . 10
5 Test results of axially loaded cylinders[18, 15] . . . . . . . . . . . . . . . . . . . . . . 11
6 (a) Postbuckling and imperfection sensitivity of externally pressurized cylinder, and
(b) Imperfection sensitivity of various shells (modified from [5])[15] . . . . . . . . . . 11
7 Flat shell element[9] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
8 Semi-loof element[9] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
9 Reduced solid element[9] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
10 Cylinder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
11 Thin-shell cylinder imperfection sensitivity[15] . . . . . . . . . . . . . . . . . . . . . . 17
12 Element type and material properties . . . . . . . . . . . . . . . . . . . . . . . . . . 17
13 Code of node creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
14 Element creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
15 Element creation code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
16 Shell281 Geometry[2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
17 Code of boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
18 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
19 Linear elastic analysis (LE) code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
20 Linear buckling analysis (LBA) code . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
21 First linear buckling modes of the example cylinder . . . . . . . . . . . . . . . . . . . 24
22 Get maximum deflection code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
23 Transformation Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
24 Update the geometry code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
25 Geometrically nonlinear analysis code . . . . . . . . . . . . . . . . . . . . . . . . . . 26
26 Contour plot of the displacements of the example cylinder . . . . . . . . . . . . . . . 27
27 Contour plot of the third principal stresses of the example cylinder . . . . . . . . . . 27
28 Membrane forces-displacement diagram of the example cylinder under GNIA . . . . 28
29 Nonlinear buckling membrane forces-sum of curvatures scatter diagram . . . . . . . 28
30 Nonlinear buckling membrane forces-radius scatter diagram . . . . . . . . . . . . . . 29
31 Nonlinear buckling membrane forces-thickness scatter diagram . . . . . . . . . . . . 29
32 Nonlinear buckling membrane forces-Young’s modulus diagram . . . . . . . . . . . . 30
33 Undeformed cylinder with initial geometric imperfections . . . . . . . . . . . . . . . 31
34 Deformation development of the example cylinder . . . . . . . . . . . . . . . . . . . . 31
35 Deformation of the example cylinder when buckling occurs . . . . . . . . . . . . . . . 32
36 The third principal stresses of the example cylinder (Phase 1) . . . . . . . . . . . . . 32
37 The third principal stresses of the example cylinder (Phase 2) . . . . . . . . . . . . . 33
38 The third principal stresses of the example cylinder (Phase 3) . . . . . . . . . . . . . 33
39 The third principal stresses of the example cylinder (Phase 4) . . . . . . . . . . . . . 34
40 The third principal stresses of the example of cylinder (Phase 5) . . . . . . . . . . . 34

3
List of Tables
1 Shell types and corresponding theory based on radius a and thickness t[9, 15] . . . . 8
2 Nonlinear buckling membrane forces-sum of curvatures . . . . . . . . . . . . . . . . . 45
3 Nonlinear buckling membran forces-thickness . . . . . . . . . . . . . . . . . . . . . . 58
4 Nonlinear buckling membrane forces-Young’s modulus . . . . . . . . . . . . . . . . . 59

4
1 INTRODUCTION
1.1 Thesis Statement
Shells tend to be thin because their curvature enables them to carry distributed load as membrane
forces[15]. Therefore, shell is a quite popular form of structures as it can withstand relative large
load with small thickness, which satisfies the aesthetic preference. By taking load as membrane
forces, buckling failure often govern the design of shell structures. At the same time, due to the
small thickness, shell structures are extremely sensitive to imperfections.

With nonlinear finite element analysis (NLFEA), the buckling load could be predicted accurately.
However, nonlinear finite element analysis is extremely time-consuming as there are a lot of com-
putational efforts are required, also very expensive. The currently used design formula could help
gain results easily, however, the accuracy is not very satisfying.

Different from other forms of structures, shell buckling often starts locally, therefore, an assumption
is made that shell buckling is a combination of membrane forces and curvatures. Thus, a new design
formula is proposed.

1.2 Research Purpose

The research purpose of this report is to verify and optimize the following formula for shell buckling.
kxx + kyy
nxx + nyy = −0.1Et2 (1)
2
where E is the Young’s modulus, t is the shell thickness, kxx and kyy are the curvatures of shell
structures in x and y direction and nxx and nyy are the membrane forces in the x and y direction.

An additional research purpose is to understand and explain the shell behavior before buckling.

1.3 Research Method

The program ANSYS Mechanical APDL has been used to generate cylinder shells of different size.
The shells were analyzed by a geometrical nonlinear procedure that increases an axial edge load
in small steps until buckling occurs. A database has been made of the buckling membrane forces.
The database was imported into MatLab and the proposed formula was verified and optimized.

1.4 Literature
The new design formula (Eq.1) is proposed by Dr. ir. P. C. J. Hoogenboom in meeting with the
author.

The APDL script used in this report is based on the APDL script used in the report of E. J. Giesen
Loo[15]. Loo also quantifies the influence of the initial geometric imperfections to the toroidal shell
segments. He also found that for toroidal shell segments, edge disturbance occurs at the locations
where Gaussian curvature equals to zero[15].

5
1.5 Thesis Overview
In this report, there are 5 chapters in total.

Chapter 2 introduces the basic theories that used in this report.

In Chapter 3, the methods used in this report are elaborated. Command batches are illustrated
and elaborated separately.

In Chapter 4, the results and the figures that illustrate shell behavior before buckling are listed.

In Chapter 5, the results are discussed and a conclusion of the formula is drawn.

6
2 BACKGROUND
2.1 Shell Linear Elastic Theory
2.1.1 Coordinate Systems and Curvatures
Shells can be described as plates with a curved middle surface. This curvature enables shells to
carry out-of-plane pressure loads as membrane forces instead of bending moments[3, 17, 15]. The
latter are restricted to so-called edge disturbances found at concentrated loads, edges and any other
discountinuities[17, 9, 15].

In order to describe the shell structures mathematically, the coordinate systems are used, thus,
the curvatures could be derived and Lamé parameters are introduced. In shell analysis three
coordinate systems are used: 1) a global coordinate system to describe the shape of the shell, 2) a
local coordinate system to define curvature, displacements, membrane forces, moments and loading,
3) a curvilinear coordinate system to derive and solve the shell equations[9].

Figure 1: Coordinate systems[9]

Curvature is also defined for surfaces. The z axis is a part of a local coordinate system. When the
normal plane includes the x direction vector the curvature is kxx . When the plane includes the y
direction vector the curvature is kyy , These curvatures can be calculated by[9]

∂2z
kxx = (2)
∂x2

∂2z
kyy = (3)
∂y 2
and

∂2z
kxy = (4)
∂x∂y

7
There are many possible normal planes of a point of a surface, and therefore, the principal curvatures
at this point could be calculated as Eq.5[9].
r
1 1
k1,2 = (kxx + kyy ) ± (kxx − kyy )2 + kxy
2 (5)
2 4
The mean curvature of a surface at a point is half the sum of the principal curvatures at this point
km = 21 (k1 + k2 ). It can be shown that also km = 12 (kxx + kyy ). The mean curvature is independent
of how we choose the local coordinate system except for the direction of z axis[9].

Figure 2: Normal section curvature[9]

Table 1: Shell types and corresponding theory based on radius a and thickness t[9, 15]
Type of Shell Slenderness Theory
Very thick shell a/t < 5 Solid elements (i.e., not a shell)
Thick shell 5 < a/t < 30 Mindlin-Reissner (includes shear deformation)
Thin shell 30 < a/t < 4000 Sanders-Koiter (membrane forces and moments)
Membrane 4000 < a/t Shell membrane (only membrane forces)

The geometry of a shell is described by its thickness and surface curvature. Thus, shell can be
categorized based on their radius-to-thickness ratio[3, 15]. Table 1 shows such a classification, with
suitable theories[15].

As seen in Table 1, the analysis of thin shells involves two distinct theories: the shell membrane
theory, which does not include bending and shear, and the Sanders-Koiter theory, which includes
bending deformation and shear stresses but not shear deformation[17, 15]. The positive internal
force resultants from the latter theory are shown in Fig.3[15].

8
 Figure 3: Positive internal forces

2.2 Shell Buckling

Shells tend to be thin because their curvature enables them to carry distributed load as membrane
forces. The property of thinness stems from shells’ capacity to store membrane strain energy
without much deformation. Yet, if this energy is converted into bending energy, shells may become
statically unstable and fail dramatically[6, 15].

2.2.1 Static Instability

Static instability, loosely termed buckling, is the condition when a structural member or system
exhibits a loss in its load-carrying capacity[19, 15]. Buckling may be divided into two categories: 1)
bifurcation of equilibrium (Fig.4, point B) and 2) collapse at the limit load without prior bifurcation
(point A). Bifurcation is exemplified by a sudden change in the load-carrying path, e.g., from
axial (or membrane) forces to bending moment, and corresponding deformations. Columns, plates
and cylindrical shells experience this type of instability. Shallow arches and spherical caps are
example of the second type of instability, also termed nonlinear buckling or ”snap-through”[19,
6, 15]. However, given initial geometric imperfections, even arches and spherical caps are prone
to fail in an asymmetric mode due to bifurcation prior to their limit load, i.e., curve 0-B-D in
Fig.4[19, 12, 6, 15].
The loads observed in Fig.4 are expressed as a multiplier λ to some reference load. λC is the
critical buckling load ratio at the bifurcation point. λL , or limit load ratio, is the maximum load
that can be achieved without prior bifurcation. λS is the maximum load that can be achieved by
a structure with initial geometric imperfections before static instability is reached[12, 15]. Chapter
1 refers to λS as the nonlinear buckling load because – just like for the limit load ratio – λS is
obtained by means of a geometrically nonlinear analysis (GNA). However, estimating λS requires
explicit modeling of the initial geometric imperfections in the finite element model. An analysis

9
 Figure 4: Load-deflection curves showing limit and bifurcation points: (a) General nonlinear anal-
ysis, and (b) Asymptotic analysis[6, 15]

that includes such imperfections is referred to as a geometrically nonlinear analysis with initial
geometric imperfections (GNIA)[15].

2.2.2 Bifurcation Buckling

The critical buckling load for discretized systems is the lowest eigenvalue from

Et2 ∼ Et2
nC = −[3(1 − ν 2 )]−1/2 = −0.6 (6)
a a
[3(1 − ν 2 )]−1/2 is approximated as 0.6 for realistic value of ν. Eq.6 is also valid for axially loaded
hyperboloids and for externally pressurized closed cylinders, spherical shells, domes, and hyper-
bolic paraboloids[9, 15]. The fact that Eq.6 makes no reference to the number of waves found
in the buckling pattern helps to explain this wide range of applicability[6, 15]. This quality has
further repercussions seen in subsection 2.2.3 which describes methods that estimate λS based on
asymptotic analyses that rest on the theoretical foundations established by Koiter and use the
postbuckling behavior as a starting point[12, 15]. 1

2.2.3 Imperfection Sensitivity

The classical example of imperfection sensitivity is the axially loaded thin-shell cylinder[9, 18, 15].
Fig.5 shows the test results of 172 axially loaded thin-shell cylinders compared to the critical load
predicted by Eq.6. nS has values as low as one sixth of nC . Kármán and Tsien attributed this
discrepancies to the highly unstable postbuckling regimes seen in both cylinderical and spherical
shells[13, 15].
Around the same time, Koiter found that asymptotically exact estimates λS can be obtained by
including the first-order effects of small initial geometric imperfections in the shape of the critical
1 According to Eq.6, when the Poisson’s ratio ν equals to 0.2721655270, the equation holds. Therefore, when the
Poisson’s ratio is larger than 0.2721655270, the results that derived from the equation are conservative. Therefore,
in this report, the Poisson’s ratio ν is assumed to be 0.3.

10
 Figure 5: Test results of axially loaded cylinders[18, 15]

buckling mode[12, 6]. If the magnitude of the initial geometric imperfection is denoted as δ̄, then
for as = 0 and bs < 0, λS can be estimated by[15]
−bs 1/3
λS ∼= λC [1 − 3( ) (ρδ̄)2/3 ] (7)
4
where ρ is a constant that depends on the imperfection shape. On the other hand, for a postbuckling
curve with as 6= 0 and bs = 0, λS is estimated using[15]

λS ∼
= λC [1 − 2(−ρas δ̄)1/2 ] (8)
In both cases, small values of δ̄ have a sizeable on λS [12, 15] which further substantiates the claim
by Kármán and Tsien[13, 15].

Figure 6: (a) Postbuckling and imperfection sensitivity of externally pressurized cylinder, and (b)
Imperfection sensitivity of various shells (modified from [5])[15]

11
2.2.3.1 Axially Loaded Thin-Shell Cylinder and Externally Pressurized Spherical Shell

Even with this limitation, it is possible to give a close estimate of PS for the axially loaded thin-
shell cylinder with the classical theory by using an imperfection in the shape of the axisymmetric
buckling mode[14, 12, 15]:

PS 2 3c δ̄ PS
(1 − ) = ( ) (9)
PC 2 t PC
or
r
PS 3c δ̄ 3c δ̄ 3c δ̄
=1+ − (2 + ) (10)
PC 4 t 4 t 4 t
p
where c = 3(1 − ν 2 . Koiter also found that the cylinder’s length and boundary conditions play
a negligible role[14, 15]. Similarly, Hutchinson developed an equation for a shallow section of
an externally pressurized spherical shell, taking in consideration the interaction between buckling
modes[11, 15]. The highest reduction in pressure was observed for two operative buckling modes
with one such mode having a zero wave-number associated with either the x or y coordinate[15]:
√
pS 2 27 3c δ̄ pS
(1 − ) = ( ) (11)
pC 32 t pC
or
√ r
pS 27 3c δ̄ 27c δ̄ √ 81c δ̄
=1+ − (32 3 + ) (12)
pC 64 t 1024 t 4 t
Equations 9, 10, 11 and 12 are plotted in Fig.6. Similar imperfection sensitivity studies were done
on axially compressed oval thin-shell cylinder[11, 15], externally pressurized thin0shell spheroids[8,
15] and externally pressurized thin-shell toroidal segments[10, 15]. It appears that imperfection
sensitivity disappears for toroidal segments of sufficiently large negative Gaussian curvature[5, 15].

2.3 FEM
2.3.1 Finite Element Implementation
The three most common shell finite elements are: flat shell elements, elements based on Sanders-
Koiter equations and reduced solid elements[9, 15]. Flat shell elements combine plane stress with
plate bending ad drilling degrees of freedom[9, 3, 15]. Their main disadvantage is that, by being
flat, a fine mesh is required to preserve the curvature of the shell model[7, 15].

Shell elements based on the Sanders-Koiter equations, such as semi-loof element (Fig.8), can be
very accurate but are often difficult to implement in finite element software[9, 15].

The most common element type is the reduced solid element (Fig.9) in which degrees of freedom are
combined the constitutive equations are simplified[9, 3, 15]. 8-noded quadrilaterals can be curved,
which reduces the need for fine meshes. The script provided in Appendix.A used this shell element
to generate shell models.

12
 Figure 7: Flat shell element[9]

Figure 8: Semi-loof element[9]

Figure 9: Reduced solid element[9]

2.3.2 FEM Solution Techniques

Two FEM procedures are described, each with its own strengths and pitfalls: linear buckling
analysis (LBA) and geometrically nonlinear analysis (GNA). LBA is an eigenvalue analysis based
on Eq.13[2, 16, 15]

(K + λi Kg )Ψi = 0 (13)
where K is the (linear elastic) stiffness matrix, Kg is the geometric stiffness matrix computed for
a reference load, λi is an eigenvalue (buckling load factor) and Ψi is a corresponding eigenvector
(buckling mode). LBA assumes negligible deflections prior to bifurcation of the loading path[16, 15].

13
The lowest eigenvalue is referred to as the critical buckling load, that is, λC .

The assumption of negligible displacements seldom holds: the transition to an alternate load path
is usually gradual due to deflections which may be enhances or even triggered by the presence of
initial geometric imperfections. GNA accounts for these deformation by updating the geometry
and satisfying equilibrium on the deformed geometry[16, 15].

GNA commonly tracks the equilibrium path via an incremental-iterative scheme: equilibrium is
established to prescribed tolerances by means of iterations at each load increment[16, 15]. The
reader is referred to[4, 15] for a comprehensive treatment of such schemes. The script provided
in App.A stipulates that analyses of this sort be executed using an arc-length controlled Newton-
Raphson method.

While GNA can yield a good approximation of λL , it fails to capture the effect of initial geometric
imperfections. GNA with explicitly modeled initial geometric imperfections is referred to as a
GNIA. Such analyses are typically required for cases in which the initial geometric imperfections
play a crucial role in triggering the nonlinear behavior and account for a significant reduction from
λC to λS [15].

2.4 Initial Imperfections

Chen proposes four approaches to adding imperfections[7, 15]. The first approach is to update the
geometry by rescaling the k th buckling load. This is achieved via

δ̄ k
δ Imp (u, v) = k
δ (u, v) (14)
δmax
where δ Imp (u, v) is the imperfection, δ̄ is the prescribed maginitude, δ k (u, v) is the deflection of the
k th buckling mode and δmaxk
is the absolute maximum of δ k .

The second approach that suggested by Chen is to apply a uniform combination of n buckling
modes based on[7, 15].
n
Imp δ̄ X
δ (u, v) = Pn δ k (u, v) (15)
max k=1 δ k (ui , vj )
k=1

The denominator equals the maximum deflection of the sum of the n modes for all possible (ui , vj ).

Chen further suggests using[7, 15]

δ̄
δ Imp (u, v) = rand
δ rand (u, v) (16)
δmax
instead of Eq.15, because the contribution of each buckling mode is randomized.

The buckling modes are combined to yield δ rand using[15].

n
X
rand
δ (u, v) = Ak δ k (u, v) (17)
k=1

14
where

Ak = rand(0, 1), k ∈ [1, n] (18)

The remaining two approaches suggested by Chen are using random noise imperfections, a draw-
back of which is its mesh dependency; and imperfection pattern based on sinusoidal waves[7, 15].
Only Eq.13 is implemented in the script in Appendix. However, using Eq.15 or Eq.16 could yield
remarkable results as the first mode may not govern as shown by the axially-loaded thin-shell cylin-
der and externally pressurized spherical shell[5, 12, 15]. Chen also observed that for structures
with closely spaces buckling loads the imperfection sensitivity tends to be significant and is often
controlled by a combination of buckling modes[7, 15].

2.5 Proposed Formula

Different from other forms of structures, buckling of shells often start locally. Therefore, an assump-
tion is made that the macro shape of shell structure only plays a negligible role in shell buckling.
Shell buckling start at a location where a combination of curvature and membrane forces is met[15].

Therefore, a formula is proposed for the relationship between the critical membrane forces and
curvatures for thin-shell structures.
kxx + kyy
nxx + nyy = −0.6Et2 (19)
2
1
Due to shell structures are highly sensitive to imperfections, a factor that equals to 6 is applied.
Therefore, the proposed formula with consideration of imperfections is
kxx + kyy
nxx + nyy = −0.1Et2 (20)
2
For the equations above, E is the Young’s modulus of the material applied in the structures. t
stands for the thickness of the shell.

15
3 METHODS
3.1 Model Creation
3.1.1 Shape of Shell Structures
In order to minimize the influence of the boundary conditions and edge disturbance due to the
application of loads, axially-loaded cylinders are chosen for this report.

Figure 10: Cylinder

The translational displacements at the bottom edges of cylinders are restrained. Vertically uniform
pressures are applied at both the top and bottom rings of cylinders to avoid the edge disturbance
at the bottom edges.
√
In case of cylinders buckle in bending, the length of the cylinder lv should equal to 10 aa · t where
aa is the radius of the cylinder, and t is the thickness of the cylinder.

At the same time, the radius-to-thickness ratio is limited between 30 and 1000 for this formula Eq.1

Due to the imperfection sensitivity of shell structures, the initial geometric imperfections should be
taken into consideration. As shown in Fig.11, where δ is the predefined scale of the initial geometric
imperfections, when the ratio δt increases, the λ decreases. And when the ratio δt goes infinite, the
λ tends to be a constant. Therefore, in this report, δ = 5t is chosen as the predefined scale of the
initial geometric imperfections.

As shown in Eq.1, the Poisson’s ratio of the material does not occur in this equation, therefore, for
all models used in this report, the Poisson’s ratio is assumed to be 0.3.

The cylinder is parameterized by equations that model the inner walls of a cylinder; hence it is easy
to change the mean curvatures of cylinders by changing the value of aa.

16
 Figure 11: Thin-shell cylinder imperfection sensitivity[15]

The model uses reduced solid 8-noded quadrilateral shell elements (SHELL281) with thickness
equals to t. The material is assumed to be linear-elastic with Young’s modulus equals to E and
Poisson’s ratio equals to ν. The element type is defined with ET, the thickness is defined with
R and the material properties are defined with MP. All commands (in bolds) are from ANSYS
Inc[1, 15].

Figure 12: Element type and material properties

The script in Fig.12 defines the element type and material properties. Note that the Poisson’s ratio
is represented by w to avoid confusion between the v parameter and the Greek letter ν.

Cylinders could be parameterized by

17
    
x̄(u, v) aa · cos(u)
r(u, v) = ȳ(u, v) = aa · sin(u) (21)
z̄(u, v) v
   
√
where v goes from 0 to lv = 10 aa · t and u goes from 0 to 2π. The length lv is the total height of
the cylinder.

Subsections 3.1.2 and 3.1.3 illustrate the creation of nodes (N) and elements (E), respectively.
Subsection 3.1.4 summarizes the enforcement of boundary conditions (D) and imposition of external
loads (SFE).

3.1.2 Nodes
Nodal counters, used exclusively for node creation, are denoted as ni and nj . Nodes are created
to accommodate nv rows by nu columns of 8-noded quadrilateral shell elements. 2nv + 1 nodal
rows and 2nu nodal columns are required to accommodate all elements. In reality, ’2nu + 1’ nodal
columns are omitted because the ’2nu + 1’ nodal column coincides (at u = 2π) with the first one
(at u = 0).

A nodal row is created in a loop with ni goes from 0 to 2nu − 1. This loop is nested inside another
loop with nj goes from 0 to 2nv to create nodal rows along the v−parameter. The loop counters
are chosen to fit the required number of nodal rows and columns and to ease the calculation of the
lv
u and v parameters. The nodes are equally spaces intervals of 2n v
along v and nπu along u. For
each node, u and v are obtained from ni and nj using
2π
u = ni (22)
2nu
and
lv
v = nj (23)
2nv
Eq.21 then yields x̄, ȳ and z̄, and the node is created via N. The variables tx and ty are introduced
that, when added together, act like a Boolean indicating when a node should or should not be
created. If and only if tx + ty is less than 1, a node is created. Thus, nodes are omitted if ni and
nj are simultaneously even. Fig.13 illustrates and elaborates the code that in charge of the node
creation.

3.1.3 Elements
Element rows and columns are denoted as j and i respectively. An element column is created using
a loop with i going from 1 to nu . This loop is nested within another loop with j going from 1 to
nv element rows along the u-parameter. An element is created using E whose arguments are the
numbers identifying the surrounding 8 nodes. Equations 24, 25 and 26 identify the numbers of the
left three nodes of an element, from bottom to top, based on i and j:

k1 = 1 + 2(i − 1) + 3nu (j − 1) (24)

k2 = i + 2nu + 3nu (j − 1) (25)

18
 Figure 13: Code of node creation

and

k3 = 1 + 2(i − 1) + 3nu j (26)

Figure15 shows the code that in charge of element creation.
Except the last column, the right three nodes could be expressed as equations 27, 28 and 29

k4 = k1 + 2 (27)

k5 = k2 + 1 (28)
and

k6 = k3 + 2 (29)
While for the last element column, k4 , k5 and k6 coincide with the first nodal column. Therefore,
the expression of k4 , k5 and k6 are listed below.

19
 Figure 14: Element creation

k4 = 1 + 3nu (j − 1) (30)

k5 = 1 + 2nu + 3nu (j − 1) (31)

and

k6 = 1 + 3nu · j (32)

3.1.4 Boundary Conditions

Fixities (Dirichlet boundary conditions) are enforced with D. The pressure load (Neumann bound-
ary condition) is applied on the element using SFE. The bottom ring nodes are restrained against
displacement in all directions, i.e., fully pinned. Therefore, the element number from 1 to 2nu − 1
are selected.

The uniform pressure is applied to the top ring elements and the elements close to the bottom ring
to avoid the edge disturbance at the bottom. As shown in Fig.16, for the top ring elements, the
external pressure is applied at the surface 4, and for the bottom ring elements, the external pressure
is applied at the surface 6. The code that in charge of the creation of boundary conditions is shown
in Fig.17.

The Fig.18 is a model with radius equals to 100mm, and the thickness equals to 1mm. The
Young’s modulus of this cylinder is 210000M P a and the Poisson’s ratio equals to 0.3. This cylinder
is referred as example cylinder in later.

20
 Figure 15: Element creation code

3.2 Mechanical Analysis

ANSYS Mechanical APDL’s solution procedure is preceded by /SOLU. The LE analysis is specified
with ANTYPE, STATIC[1, 15]. Additionally, PSTRES, ON (prestress effects) is required after
ANTYPE to save the state for the LBA[1, 15]. The code that in charge of the linear elastic analysis
is shown in Fig.19.

3.2.1 Linear Buckling Analysis (LBA)

The LBA is solved using the Block Lanczos methods. ANSYS Inc. recommends requesting a few
additional modes than needed to enhance the accuracy of the final solution[2, 15].

After the LBA, shell buckling results could be generated. The first buckling mode is plotted. Fig.21
shows the first linear buckling mode of the example cylinder.
Initial imperfections for the GNIA are based on Eq.12. The code uses a similar nomenclature: δ̄
k
(delta) is the prescribed imperfection amplitude and δmax (uz max) is the maximum deflection.
The latter is obtained by looping through all the nodes and storing the largest |uz | (Fig.22). The
deflections are transformed to local coordinates using Transformation Matrix (Fig.23).

3.2.2 Geometrically Nonlinear Analysis with Initial Geometric Imperfections (GNIA)

Geometrically nonlinear finite element analysis updates the geometry using a buckling mode and
k
δmax (uz max) from section 3.2.1. The geometry is updated with UPGEOM, whose arguments

21
 Figure 16: Shell281 Geometry[2]

Figure 17: Code of boundary conditions

are the factor (delta/uz max), load step, buckling mode, file and file extension is shown in Fig.24.

22
 Figure 18: Model

Figure 19: Linear elastic analysis (LE) code

Figure 20: Linear buckling analysis (LBA) code

The solution is then set up and executed as shown in Fig.25.

ANSYS Mechanical APDL uses TIME as the counter for both dynamic and nonlinear static anal-

23
 Figure 21: First linear buckling modes of the example cylinder

Figure 22: Get maximum deflection code

ysis. In a non-proportional analysis, time acts as a counter for indexing each load step. For single
load step analysis, time equals λ. Assuming λS does not exceed 1.

24
 Figure 23: Transformation Matrix

Since time defaults to 1 at the end of each analysis, the loop ensures that this value does not get
saved.

Figure 24: Update the geometry code

25
Figure 25: Geometrically nonlinear analysis code

26
4 RESULTS
4.1 GNIA Results
The contour plot of the example cylinder of displacements and the third principal stresses are shown
in Fig.26 and Fig.27 respectively. The load-displacement diagram (GNIA) of the example cylinder
is shown in Fig.28.

Figure 26: Contour plot of the displacements of the example cylinder

Figure 27: Contour plot of the third principal stresses of the example cylinder

27
 Figure 28: Membrane forces-displacement diagram of the example cylinder under GNIA

4.2 Curvatures
In order to check the relationship between the nonlinear buckling membrane forces and the curva-
tures of cylinders, a cylinder with thickness equals to 0.5mm, Young’s modulus equals to 21000M P a
and the Poisson’s ratio equals to 0.3 is applied with radius varies from 15mm to 500mm. The ra-
dius of the cylinder is generated randomly. In total there are 509 cylinders are analyzed. Then the
nonlinear buckling membrane forces generated with ANSYS Mechanical APDL and the nonlinear
buckling membrane forces derived from Eq.1 are imported into MatLab and scatter diagrams are
generated to illustrate the relation between the nonlinear buckling membrane forces and curvatures
of cylinders. The numerical results are listed in App.B.

Figure 29: Nonlinear buckling membrane forces-sum of curvatures scatter diagram

28
 Figure 30: Nonlinear buckling membrane forces-radius scatter diagram

4.3 Thickness
In order to check the relation between nonlinear buckling membrane forces and the thickness of
cylinders, a cylinder with radius equals to 100mm, Young’s modulus equals to 21000M P a is applied,
the thickness varies from 0.1mm to 2mm. The nonlinear buckling membrane forces generated with
ANSYS Mechanical APDL and the critical membrane forces derived from Eq.1 are imported into
MatLab and a scatter diagram is generated (Fig.31). The numerical results are listed in App.B

Figure 31: Nonlinear buckling membrane forces-thickness scatter diagram

29
4.4 Young’s Modulus
In order to check the relation between nonlinear buckling membrane forces and the Young’s modulus,
a cylinder with thickness equals to 1mm and the radius equals to 100mm is applied, the Poisson’s
ratio equals to 0.3 and the Young’s modulus varies from 1000M P a to 210000M P a. A scatter
diagram is generated to illustrate the relationship between Young’s modulus and nonlinear buckling
membrane forces (Fig.32). The numerical results are listed in App.D.

Figure 32: Nonlinear buckling membrane forces-Young’s modulus diagram

4.5 Buckling Procedures

In order to improve the design of shell structures, a better understanding of shell buckling behavior
is required. Therefore, a series of figures are generated for certain load steps to illustrate the shell
behavior before shell buckling. The representative figures are listed below.

4.5.1 Displacements
The representative figures of the deformed shape of the example cylinder are listed below.

Fig.33 illustrates the undeformed shape of the example cylinder with initial geometric imperfections.
The Gaussian curvature at the top part of the cylinder is not zero any more due to the initial
geometric imperfection.

Fig.34 illustrates the how the cylinder deform. The deformation first occurs at the locations with
negative Gaussian curvatures and the maximum deformation occurs at the top edge of the cylinder.

Fig.35 illustrates the deformed shape of the cylinder when buckling happens.

4.5.2 The Third Principal Stresses

The development procedures of the third principal stresses of the example cylinder are shown in
Fig.36, Fig.37, Fig.4.5.2, Fig.39 and Fig.40

30
Figure 33: Undeformed cylinder with initial geometric imperfections

Figure 34: Deformation development of the example cylinder

31
 Figure 35: Deformation of the example cylinder when buckling occurs

Figure 36: The third principal stresses of the example cylinder (Phase 1)

32
Figure 37: The third principal stresses of the example cylinder (Phase 2)

Figure 38: The third principal stresses of the example cylinder (Phase 3)

33
 Figure 39: The third principal stresses of the example cylinder (Phase 4)

Figure 40: The third principal stresses of the example of cylinder (Phase 5)

34
5 DISCUSSION AND CONCLUSION
5.1 Discussion of the Correctness of the Results
Due to the existence of the initial geometric imperfections, the Gaussian curvatures of the cylinder
in reality are not zero at every location. According to the Fig.33 and the Fig.34, Fig.35, it could be
observed that the maximum deformation of the cylinder occurs at the location where the Gaussian
curvature in reality is negative. It is reasonable as the shells with negative Gaussian curvatures
have smaller capacity compared to the shells with positive Gaussian curvature when it comes to
compressive membrane forces. Therefore, the deformation plots could be reliable.

According to the Fig.36, Fig.37, Fig.4.5.2, Fig.39 and Fig.40, the third principal stresses (also known
as the compressive stresses) at the locations that with positive Gaussian curvatures have higher
absolute value. The results are reliable as the location where the Gaussian curvature is positive is
stiffer and therefore, attracts larger stresses.

According to the Fig.28, the relationship between the membrane forces and the displacements is
almost linear. The diagram is reasonable as there is only geometric nonlinearity included in the
finite element analysis.

Therefore, the command batches could be applied for the nonlinear buckling analysis of cylinders
with ANSYS Mechanical APDL. And the database generated with ANSYS Mechanical APDL could
be reliable.

5.2 Discussion of the Scatter Diagrams

5.2.1 Curvatures
According to the Fig.29 and Fig.30, a linear relationship between the sum of membrane forces and
the sum of curvatures could be observed. Therefore, the relation between the nonlinear buckling
membrane forces and curvatures is verified. There are some horizontal parts of the Fig.29, this is
due to the variation of the curvatures is very small and the existence of the errors.

5.2.2 Thickness
As shown in Fig.31, the relation between the nonlinear buckling membrane forces of cylinders and
the thickness of the cylinders is quadratic. Therefore, the formula Eq.1 is correct in terms of
thickness.

5.2.3 Young’s Modulus

As shown in Fig.32, the nonlinear buckling membrane forces of the cylinder is linear to the Young’s
modulus of the material of the cylinder. Therefore, the formula is correct in terms of the Young’s
modulus. The reason of the horizontal parts in the diagram is due to the material with smaller
Young’s modulus is more ductile, therefore, the cylinders with those materials would have a more
stable performance as shell buckling in reality is a form of instability.

35
5.2.4 Conclusion
There are only 5 numerical results are lower than the theoretical results derived from the formula
out of 570 results in total, which is 0.8772%. Thus, the formula is safe and correct.

5.3 Recommendations for Further Research

In this report, as only the cylinder is analyzed, therefore, in reality, only the relation between the
nxx and kxx is verified, and the nyy and kyy have no influence in this report. Therefore, further
research on the formula should focus on the combination of nxx and nyy , kxx and kyy .

36
Reference
[1] ANSYS Inc. Command Reference for the Mechanical APDL, ANSYS Manual, 2009.
[2] ANSYS Inc. Theory Reference for the Mechanical APDL and Mechanical Applications, 2009.
[3] J. Blaauwendraad and J.H. Hoefakker. Structural Shell Analysis: Understanding and Appli-
cation (Solid Mechanics and Its Application, v.200). Dordrecht, the Netherlands: Springer,
2013.

[4] R. Borst and M. Crisfield. Nonlinear Finite Element Analysis of Solids and Structures. Wiley
series in computational mechanics. Chichester, West Sussex, United Kingdom: Wiley, 2nd ed.
edition, 2012.
[5] Bernad Budiansky and John.W. Hutchinson. A survey of some buckling problems. AAIA
Journal, 1505-1510, 1966.
[6] Bushnell. Buckling of shells-pitfall for designers. AAIA Journal, 19(9):1183–1226, 1981.
[7] T. Chen. On introduction imperfection in the non-linear analysis of buckling of thin-shell
structures. Master’s thesis, Delft University of Technology, 2014.

[8] D.A. Danielson. Buckling and initial postbuckling behavior of spheroidal shells under pressure.
AAIA Journal, 7(5):936–944, 1970.
[9] P.C.J. Hoogenboom. Cie4143 shell analysis, theory and application [lecture notes], 2018.
[10] John.W. Hutchinson. Initial postbuckling behavior of toroidal shell segments. International
Journal of Solids and Structures, 3:97–115, 1967.

[11] John.W. Hutchinson. Buckling and initial postbuckling behavior of oval cylindrical shells under
axial compression. Journal of Applied Mechanics, 35(1):66–72, 1968.
[12] John.W. Hutchinson and W.T. Koiter. Postbuckling theory. Applied Mechanics Reviews,
23(12):1353–1363, 1970.

[13] T. Kármán and H.S. Tsien. The buckling of thin cylindrical shells under axial compression.
Journal of the Aeronautical Science, 8(8):303–312, 1941.
[14] W.T. Koiter. A translation of the stability of elastic equilibrium. Technical report, STANFORD
UNIV CA DEPT OF AERONAUTICS AND ASTRONAUTICS, 1970.

[15] Erik Johannes Giesen Loo. Quantifying the influence of membrane forces, curvatures, and
imperfections on the nonlinear buckling load of thin-shells. Master’s thesis, Delft University
of Technology, 2017.
[16] W. McGuire, R.H. Gallagher, and R.D. Ziemian. Matrix Structural Analysis. 2nd ed. edition,
2000.

[17] G.Z. Voyiadjis and P. Woelke. Elasto-Plastic and Damage Analysis of Plates and Shells. Berlin:
Springer, 2008.

37
[18] V.I. Weingarten, E.J. Morgan, and P. Seide. Elastic stability of thin-walled cylindrical and
conical shells under combined internal pressure and axial compression. AAIA Journal, 3:500–
505, 1965.
[19] R.D. Ziemian. Guide to Stability Design Criteria for Metal Structures. Structural Research
Council, Hoboken, N.J., U.S.: John Wiley and Sons, 2010.

38
A Command Lines of ANSYS Mechanical APDL (modified
from [15])
! Preprocessing
!——————————————————————————————————————————–
/UIS, MSGPOP, 4 ! Sets pop-ups to YES
/UIS, ABORT, OFF ! No pop-ups about status of operation in progress

t=thickness
E=Young’s modulus
w=Poisson’s ratio
lu=2*pi
nu=number of elements along u-axis
nv=number of elements along v-axis
p=external pressure
delta=predefined scale of initial geometric imperfections
aa=radius
lv=10*SQRT(aa*t)

/PREP7
/VIEW,ALL,0,0,0 ! All windows: camera at point (0,0,0)

! Create element type: shell

!———————————————————————————————————————————

MP,EX,1,E ! Elastic modulus (linear elastic material)

MP,PRXY,1,w ! Poisson’s ratio (linear elastic material)

! Create shell nodes

!———————————————————————————————————————————

! ty and tx are dummy variables that, added together, act like a Boolean indicating when a node
! should be created. If ty+tx is less than 1, then a node is created.
ty=1
*DO,nj,0,2*nv ! Nodal rows going from nj=0 to 2*nv
ty=-ty
tx-1
v=nj*lv/nv/2
*DO, ni, 0, 2*nu-1 ! Nodal columns going from ni=0 to 2*nu-1
tx=-tx
*IF,tx+ty,LT,1,THEN ! If tx+ty=2 omit node creation
u=ni*lu/nu/2 ! Parameter u (U-V Plane)
x=aa*COS(u) ! x-coordinate x=x(u,v)
y=aa*SIN(u) ! y-coordinate y=y(u,v)
z=v ! z-coordinate z=z(v)
N,,x,y,z,,, ! Create node

39
*ENDIF
*ENDDO
*ENDDO

! Create shell elements

!———————————————————————————————————————————

SHPP,ON
*DO,j,1,nv ! j-th element row (along v-axis)
*DO,i,1,nu ! i-th element column (along u-axis)
k1=1+2*(i-1)+3*nu*(j-1)
k2=i+2*nu+3*nu*(j-1)
k3=1+2*(i-1)+j*3*nu
*IF,i,LT,nu,THEN ! i<nu
k4=k1+2
k5=k2+1
k6=k3+2
*ELSE ! i=nu
k4=1+3*nu*(j-1)
k5=1+2*nu+3*nu*(j-1)
k6=1+3*nu*k
*ENDIF
E,k4,k6,k3,k1,k5,k3+1,k2,k1+1
*ENDDO
*ENDD0

! Create Dirichlet boundary conditions

!———————————————————————————————————————————

*DO,j,1,2*nu
n bot=j ! bottom ring
D,n bot,UX,0,,,,UY,UZ
*ENDDO

! Create Neumann boundary conditions

!———————————————————————————————————————————

! Top and bottom rings

*DO,j,1,nu
i top=nv*nu-j+1
ESEL,S,ELEM,,i top,nu*nv
*ENDDO
SFE,ALL,4,PRES,0,p,,,
ALLSEL

*DO,j,nu+1,2*nu
i bot=j

40
ESEL,S,ELEM,,nu+1,i bot
*ENDDO
SFE,ALL,6,PRES,0,p,,,
ALLSEL

FINISH

! Linear Elastic Analysis

!———————————————————————————————————————————

/SOLU
ANTYPE,STATIC ! Linear elastic analysis
PSTRES,ON ! Prestress effects to be included in buckling analysis
SOLVE
FINISH

! Transformation Matrix
!———————————————————————————————————————————

*DMAT,Gamma,D,ALLOC,3,3

*VEC,i vector,D,ALLOC,3
i vector(1)=-aa*SIN(u)
i vector(2)=aa*COS(u)
i vector(3)=0
*NRM,i vector,NRM2,norm i,YES

*VEC,j vector,D,ALLOC,3
j vector(1)=0
j vector(2)=0
j vector(3)=1
*NRM,j vector,NRM2,norm j,YES

*VEC,m vector,D,ALLOC,3
m vector(1)=i vector(2)*j vector(3)-i vector(3)*j vector(2)
m vector(2)=i vector(1)*k vector(3)-i vector(3)*j vector(1)
m vector(3)=i vector(1)*j vector(2)-i vector(2)*j vector(1)
*NRM,m vector,NRM2,norm m,YES

*DO,k,1,3
Gamma(1,k)=i vector(k)
Gamma(2,k)=j vector(k)
Gamma(3,k)=m vector(k)
*ENDDO

! Linear Buckling Analysis (Find buckling mode)

41
!———————————————————————————————————————————

/SOLU
ANTYPE,BUCKLE ! Linear buckling analysis
BUCOPT,LANB,5,0,,CENTER ! Block Lanczos method, 5 buckling modes
SOLVE
FINISH

/POST1
buckling mode=1
/RGB,INDEX,100,100,100,0
/RGB,INDEX,80,80,80,13
/RGB,INDEX,60,60,60,14
/RGB,INDEX,0,0,0,15

/WINDOW,1,LEFT
/WINDOW,2,RIGHT

/PLOPT,INFO,0
/VIEW,ALL,0,0,0
/ANGLE,ALL,0
/DIST,ALL,AUTO

/WINDOW,2,OFF
SUBSET,1,buckling mode,FACT,,,,
/PLNSOL,U,SUM,0,1

/NOERASE
/WINDOW,1,OFF
/WINDOW,2,ON
SUBSET,1,buckling mode+1,FACT,,,,
PLBSOL,U,SUM,0,1

SUBSET,1,buckling mode,FACT,,,,
*GET,lambdaC,ACTIVE,0,SET,FREQ
*CFOPEN,ShellBuckling,csv,,APPEND
*VWRITE,lambdaC
(F11.5,’,’,$)
*CFCLOS

! Loop to read deflection of node next to i-th element

uz max=0
defl max=0
*DO,j,1,nv ! j-th element row (along v)
*DO,i,1,nu+1 ! i-th element column (along u)
nnode=1+2*(i-1)+(j-1)*(3*nu+2) ! nnode=k1(i,j)
*GET,u x,NODE,nnode,U,X ! Extract u x from nnode

42
*GET,u y,NODE,nnode,U,Y ! Extract u y from nnode
*GET,u z,NODE,nnode,U,Z ! Extract u z from nnode
defl=SQRT(u x*u x+u y*u y+u z*u z) ! Absolute deflection
u=(i-1)*(lu/nu)-lu/2 ! u-parameter
v=(j-1)*(lv/nv) ! v-parameter
*VEC,Defl global,D,ALLOC,3,,, ! Assemble Gamma matrix
*SET,Defl global(1),u x,u y,u z ! Allocate space for Defl global
*MULT,Gamma,,Defl global,,Defl local ! Defl local=Gamma*Defl global
u z=Defl local(3) ! local z displacement
*IF,ABS(u z),GT,uz max,THEN
uz max=ABS(u Z) ! uz max=max(u z)
defl max=ABS(defl) ! defl max=max(defl)
nnode=1+2*(i-1)+(j-1)*(3*nu+2) ! node with highest deflection
*ENDIF
*ENDDO
*ENDDO

defl diff=(defl max-uz max)/uz max*100

*CFOPEN,ShellLBA,csv,,
*VWRITE,E,t,w,p,aa,lv,nu,nv
(F16.4,’,’,F16.4,’,’,F16.4,’,’,F16.4,’,’,F16.4,’,’,F16.4,’,’,F16.4,’,’,F16.4,’,’,)
*CFCLOS

*DO,j,1,nv
*DO,i,1,nu+1
nnode=1+2*(i-1)+(j-1)*(3*nu+2)
*GET,u x,NODE,nnode,U,X
*GET.u y,NODE,nnode,U,Y
*GET,u z,NODE,nnode,U,Z
u=(i-1)*(lu/nu)-lu/2
v=(j-1)*(lv/nv)
*VEC,Defl global,D,ALLOC,3,,,
*SET,Defl global(1),u x,u y,u z
*MULT,Gamma,,Defl global,,Defl local
u x=Defl local(1)
u y=Defl local(2)
u z=Defl local(3)
defl=SQRT(u x*u x+u y*u y+u z*u z)
*CFOPEN,ShellLBA,csv,,APPEND
*VWRITE,u x,u y,u z,defl
(F16.8,’,’,F16.8,’,’,F16.8,’,’,F16.8,’,’,)
*CFCLOS
*ENDDO
*ENDDO

FINISH

43
! Update the geometry
!———————————————————————————————————————————

/PREP7
FACTOR=delta/uz max ! Factor for UPGEOM
UPGEOM,FACTOR,1,buckling mode,’file’,’rst’ ! Add imperfections
/RESET $ /ERASE $ /REPLOT ! Replot
FINISH

/SOLU
! Set analysis type: GNA
!———————————————————————————————————————————

NCNV,0 ! Do not terminate the program if not-converged

NERR,,,-1 ! Do not terminate the program if not-converged
ANTYPE,STATIC ! Static analysis
NLGEOM,ON ! Nonlinear geometry

! Set nonlinear controls / solution technique

!———————————————————————————————————————————

RESCONTROL,DEFINE,ALL,1 ! Write new files at every substep

OUTRES,NSOL,ALL,,,, ! Write nodal results at every substep
OUTRES,ESOL,ALL,,,, ! Write element results at every substep

! Solve
!———————————————————————————————————————————

SOLVE
FINISH

44
B Curvatures
Table 2: Nonlinear buckling membrane forces-sum of curvatures

nxx+nyy kxx E t time a analytical theoretical

20 0.00200012 21000 0.5 0.04999 499.970499 0.9998 0.52503098
20 0.00200167 21000 0.5 0.049968 499.583593 0.99936 0.52543759
20 0.00200782 21000 0.5 0.05016 498.053435 1.0032 0.52705188
20 0.002012 21000 0.5 0.049997 497.016978 0.99994 0.52815097
20 0.00201307 21000 0.5 0.050015 496.752779 1.0003 0.52843187
20 0.00201939 21000 0.5 0.049983 495.199373 0.99966 0.53008952
20 0.00202409 21000 0.5 0.049926 494.04967 0.99852 0.5313231
20 0.00202665 21000 0.5 0.049926 493.424142 0.99852 0.53199667
20 0.0020323 21000 0.5 0.049941 492.05308 0.99882 0.53347903
20 0.00203608 21000 0.5 0.049889 491.138663 0.99778 0.53447228
20 0.00203716 21000 0.5 0.049963 490.880604 0.99926 0.53475325
20 0.00204349 21000 0.5 0.050075 489.357724 1.0015 0.5364174
20 0.00205145 21000 0.5 0.050022 487.459915 1.00044 0.53850582
20 0.00205632 21000 0.5 0.050095 486.305051 1.0019 0.53978465
20 0.00205973 21000 0.5 0.050087 485.500771 1.00174 0.54067885
20 0.00206405 21000 0.5 0.050046 484.483756 1.00092 0.54181383
20 0.00207039 21000 0.5 0.05018 483.00105 1.0036 0.54347708
20 0.00207141 21000 0.5 0.05021 482.763298 1.0042 0.54374473
20 0.00207882 21000 0.5 0.050126 481.042525 1.00252 0.5456898
20 0.00208253 21000 0.5 0.050127 480.185401 1.00254 0.54666385
20 0.00208691 21000 0.5 0.050131 479.178376 1.00262 0.5478127
20 0.00209192 21000 0.5 0.050167 478.029265 1.00334 0.54912956
20 0.00209594 21000 0.5 0.05022 477.112982 1.0044 0.55018415
20 0.00210049 21000 0.5 0.05018 476.078387 1.0036 0.55137979
20 0.0021046 21000 0.5 0.050218 475.14916 1.00436 0.5524581
20 0.00210752 21000 0.5 0.050209 474.491813 1.00418 0.55322346
20 0.00211143 21000 0.5 0.050213 473.612006 1.00426 0.55425115
20 0.00211816 21000 0.5 0.050231 472.106924 1.00462 0.55601811
20 0.00212254 21000 0.5 0.050259 471.133788 1.00518 0.55716658
20 0.00212571 21000 0.5 0.050256 470.430271 1.00512 0.55799981
20 0.00212764 21000 0.5 0.050286 470.005258 1.00572 0.55850439
20 0.00212864 21000 0.5 0.050315 469.784033 1.0063 0.55876739
20 0.00213372 21000 0.5 0.050324 468.664617 1.00648 0.56010202
20 0.00214081 21000 0.5 0.050222 467.11367 1.00444 0.56196172
20 0.0021447 21000 0.5 0.050244 466.265602 1.00488 0.56298384
20 0.00215038 21000 0.5 0.050256 465.03453 1.00512 0.56447421
20 0.00215338 21000 0.5 0.050256 464.386025 1.00512 0.56526249
continued on next page

45
nxx+nyy kxx E t time a analytical theoretical
20 0.00215733 21000 0.5 0.050256 463.536979 1.00512 0.56629786
20 0.00216416 21000 0.5 0.050237 462.073615 1.00474 0.5680913
20 0.00216725 21000 0.5 0.050238 461.414986 1.00476 0.5689022
20 0.00217373 21000 0.5 0.05042 460.038919 1.0084 0.5706039
20 0.0021783 21000 0.5 0.050376 459.073251 1.00752 0.57180417
20 0.00218227 21000 0.5 0.050478 458.238024 1.00956 0.57284639
20 0.00218454 21000 0.5 0.050286 457.762107 1.00572 0.57344196
20 0.00219256 21000 0.5 0.050368 456.088057 1.00736 0.57554675
20 0.00219723 21000 0.5 0.050474 455.117607 1.00948 0.57677399
20 0.00219862 21000 0.5 0.05047 454.831529 1.0094 0.57713677
20 0.00220609 21000 0.5 0.050574 453.290879 1.01148 0.57909835
20 0.00221028 21000 0.5 0.050528 452.431187 1.01056 0.58019873
20 0.00221232 21000 0.5 0.05049 452.013518 1.0098 0.58073484
20 0.00221781 21000 0.5 0.050496 450.895931 1.00992 0.58217425
20 0.00222062 21000 0.5 0.050471 450.324698 1.00942 0.58291273
20 0.00222582 21000 0.5 0.050577 449.27322 1.01154 0.58427698
20 0.00223069 21000 0.5 0.050536 448.292224 1.01072 0.58555555
20 0.00223309 21000 0.5 0.050502 447.809199 1.01004 0.58618715
20 0.00223927 21000 0.5 0.050508 446.574427 1.01016 0.58780795
20 0.00224585 21000 0.5 0.050469 445.265742 1.00938 0.58953559
20 0.00224851 21000 0.5 0.050459 444.739546 1.00918 0.5902331
20 0.00225731 21000 0.5 0.050387 443.004253 1.00774 0.5925451
20 0.00226222 21000 0.5 0.050456 442.044548 1.00912 0.59383155
20 0.00226597 21000 0.5 0.050525 441.311972 1.0105 0.59481731
20 0.00227039 21000 0.5 0.050579 440.45386 1.01158 0.59597616
20 0.00227532 21000 0.5 0.05045 439.499091 1.009 0.59727086
20 0.002282 21000 0.5 0.050535 438.211504 1.0107 0.59902581
20 0.002287 21000 0.5 0.050623 437.254177 1.01246 0.60033732
20 0.00229391 21000 0.5 0.050542 435.93756 1.01084 0.60215045
20 0.00229433 21000 0.5 0.050491 435.857608 1.00982 0.60226091
20 0.00230344 21000 0.5 5.06E-02 434.133161 1.01160954 0.60465319
20 0.00230822 21000 0.5 0.050551 433.234733 1.01102 0.6059071
20 0.00231223 21000 0.5 0.050647 432.483119 1.01294 0.60696011
20 0.00231603 21000 0.5 0.050537 431.773306 1.01074 0.60795792
20 0.00232523 21000 0.5 0.050947 430.065049 1.01894 0.61037278
20 0.0023302 21000 0.5 0.050649 429.147516 1.01298 0.61167778
20 0.00233469 21000 0.5 0.050674 428.321563 1.01348 0.61285731
20 0.00234241 21000 0.5 0.05054 426.91052 1.0108 0.61488295
20 0.00234276 21000 0.5 0.050542 426.84646 1.01084 0.61497523
20 0.00235187 21000 0.5 0.050552 425.193642 1.01104 0.61736577
continued on next page

46
nxx+nyy kxx E t time a analytical theoretical
20 0.0023567 21000 0.5 0.050607 424.321948 1.01214 0.61863404
20 0.00235924 21000 0.5 0.05056 423.866129 1.0112 0.61929931
20 0.00236931 21000 0.5 0.050643 422.063597 1.01286 0.62194419
20 0.00237397 21000 0.5 0.050718 421.23591 1.01436 0.62316624
20 0.00237923 21000 0.5 0.050767 420.304744 1.01534 0.62454684
20 0.00238641 21000 0.5 0.050729 419.039537 1.01458 0.62643254
20 0.00239186 21000 0.5 0.050661 418.084633 1.01322 0.62786331
20 0.00239684 21000 0.5 0.050602 417.216811 1.01204 0.62916928
20 0.00240197 21000 0.5 0.050603 416.324796 1.01206 0.63051733
20 0.00240932 21000 0.5 0.050609 415.054149 1.01218 0.6324476
20 0.00241372 21000 0.5 0.050841 414.297782 1.01682 0.63360223
20 0.00241972 21000 0.5 0.078963 413.270279 1.57926 0.63517754
20 0.00242545 21000 0.5 0.078986 412.2948 1.57972 0.63668036
20 0.00243257 21000 0.5 0.079149 411.087248 1.58298 0.63855058
20 0.00243588 21000 0.5 0.078992 410.528929 1.57984 0.63941901
20 0.00244333 21000 0.5 0.079026 409.277851 1.58052 0.64137358
20 0.00244953 21000 0.5 0.079078 408.241365 1.58156 0.64300197
20 0.00245776 21000 0.5 0.079093 406.875186 1.58186 0.64516099
20 0.00245835 21000 0.5 0.079194 406.776181 1.58388 0.64531802
20 0.00246634 21000 0.5 0.079143 405.459355 1.58286 0.64741384
20 0.00247452 21000 0.5 0.079242 404.118656 1.58484 0.6495617
20 0.00248132 21000 0.5 0.079172 403.010747 1.58344 0.6513474
20 0.0024838 21000 0.5 0.079181 402.608677 1.58362 0.65199787
20 0.0024937 21000 0.5 0.07922 401.010433 1.5844 0.65459644
20 0.00249775 21000 0.5 0.079345 400.360094 1.5869 0.65565975
20 0.00250264 21000 0.5 0.07927 399.578782 1.5854 0.65694179
20 0.00250872 21000 0.5 0.079278 398.609263 1.58556 0.65853964
20 0.00251854 21000 0.5 0.0793 397.055632 1.586 0.66111642
20 0.00252357 21000 0.5 0.07939 396.263349 1.5878 0.66243825
20 0.0025293 21000 0.5 0.079457 395.365757 1.58914 0.66394217
20 0.00253801 21000 0.5 0.079588 394.008917 1.59176 0.66622858
20 0.00254431 21000 0.5 0.07939 393.03372 1.5878 0.66788163
20 0.00254864 21000 0.5 0.079462 392.366495 1.58924 0.66901737
20 0.0025558 21000 0.5 0.0797 391.266301 1.594 0.67089857
20 0.00255808 21000 0.5 0.079699 390.917877 1.59398 0.67149654
20 0.00256827 21000 0.5 0.080025 389.367905 1.6005 0.67416959
20 0.00257355 21000 0.5 0.079799 388.568206 1.59598 0.67555707
20 0.00258155 21000 0.5 0.080194 387.364418 1.60388 0.67765646
20 0.00258905 21000 0.5 0.079662 386.242547 1.59324 0.67962476
20 0.00259782 21000 0.5 7.97E-02 384.938077 1.59392 0.68192786
continued on next page

47
nxx+nyy kxx E t time a analytical theoretical
20 0.00260376 21000 0.5 7.97E-02 384.060038 1.5947 0.68348689
20 0.00260876 21000 0.5 0.079751 383.323199 1.59502 0.68480071
20 0.00261779 21000 0.5 0.079794 382.0013 1.59588 0.68717044
20 0.00262436 21000 0.5 0.079819 381.045122 1.59638 0.68889479
20 0.00262802 21000 0.5 0.07985 380.514385 1.597 0.68985565
20 0.0026325 21000 0.5 0.079859 379.867705 1.59718 0.69103005
20 0.00264166 21000 0.5 0.079867 378.549754 1.59734 0.69343593
20 0.00264807 21000 0.5 0.07989 377.632912 1.5978 0.6951195
20 0.00265637 21000 0.5 0.080056 376.45295 1.60112 0.69729829
20 0.00266394 21000 0.5 0.080068 375.384121 1.60136 0.69928371
20 0.00266689 21000 0.5 0.079943 374.967957 1.59886 0.70005982
20 0.00267333 21000 0.5 0.080076 374.065409 1.60152 0.70174893
20 0.0026805 21000 0.5 0.080161 373.064057 1.60322 0.70363251
20 0.00268632 21000 0.5 0.080267 372.256055 1.60534 0.70515979
20 0.00269381 21000 0.5 0.080147 371.220931 1.60294 0.70712607
20 0.00269719 21000 0.5 0.080324 370.755871 1.60648 0.70801306
20 0.00270279 21000 0.5 0.080391 369.987511 1.60782 0.70948341
20 0.0027156 21000 0.5 0.080708 368.242688 1.61416 0.71284511
20 0.00272363 21000 0.5 0.080461 367.156607 1.60922 0.71495377
20 0.00272844 21000 0.5 0.08077 366.509148 1.6154 0.71621678
20 0.00273655 21000 0.5 0.080301 365.423496 1.60602 0.71834461
20 0.00274403 21000 0.5 0.080654 364.427903 1.61308 0.72030708
20 0.002753 21000 0.5 0.080192 363.24009 1.60384 0.72266252
20 0.00276101 21000 0.5 0.08021 362.186759 1.6042 0.72476421
20 0.00276516 21000 0.5 0.080287 361.642642 1.60574 0.72585467
20 0.0027767 21000 0.5 0.08053 360.139481 1.6106 0.72888426
20 0.00278346 21000 0.5 0.08029 359.264867 1.6058 0.7306587
20 0.00279186 21000 0.5 0.080285 358.18468 1.6057 0.73286217
20 0.00279666 21000 0.5 0.080498 357.568778 1.60996 0.7341245
20 0.00280531 21000 0.5 0.08028 356.46695 1.6056 0.73639365
20 0.0028119 21000 0.5 0.080703 355.631821 1.61406 0.73812293
20 0.00282281 21000 0.5 0.080469 354.256302 1.60938 0.74098893
20 0.00283043 21000 0.5 0.080678 353.303361 1.61356 0.74298755
20 0.00283739 21000 0.5 0.080617 352.437117 1.61234 0.74481372
20 0.00284586 21000 0.5 0.080807 351.387703 1.61614 0.74703809
20 0.00285598 21000 0.5 0.080426 350.143104 1.60852 0.74969347
20 0.00286371 21000 0.5 0.080373 349.196809 1.60746 0.75172508
20 0.00286681 21000 0.5 0.080368 348.819547 1.60736 0.7525381
20 0.002878 21000 0.5 0.080386 347.463411 1.60772 0.75547523
20 0.00288889 21000 0.5 0.080448 346.154015 1.60896 0.75833296
continued on next page

48
nxx+nyy kxx E t time a analytical theoretical
20 0.00289584 21000 0.5 0.080474 345.3234 1.60948 0.760157
20 0.00290122 21000 0.5 0.080532 344.682832 1.61064 0.7615697
20 0.00290998 21000 0.5 0.080532 343.64529 1.61064 0.76386905
20 0.00291819 21000 0.5 0.080505 342.678379 1.6101 0.7660244
20 0.00293134 21000 0.5 0.08051 341.140926 1.6102 0.76947672
20 0.00293674 21000 0.5 0.080574 340.514059 1.61148 0.77089328
20 0.00294827 21000 0.5 0.080594 339.182415 1.61188 0.77391984
20 0.00295577 21000 0.5 0.080627 338.321404 1.61254 0.77588943
20 0.00296099 21000 0.5 0.0808 337.724686 1.616 0.77726033
20 0.00296961 21000 0.5 0.080401 336.744157 1.60802 0.77952355
20 0.00298243 21000 0.5 0.080431 335.29734 1.60862 0.78288721
20 0.00299122 21000 0.5 0.080747 334.312108 1.61494 0.78519441
20 0.00299818 21000 0.5 0.080581 333.535499 1.61162 0.78702267
20 0.00301052 21000 0.5 0.080599 332.167989 1.61198 0.79026278
20 0.00301343 21000 0.5 0.080629 331.847757 1.61258 0.79102538
20 0.00302854 21000 0.5 0.080629 330.191924 1.61258 0.79499219
20 0.00303666 21000 0.5 0.080484 329.309409 1.60968 0.79712268
20 0.00304405 21000 0.5 0.080487 328.509791 1.60974 0.79906294
20 0.0030559 21000 0.5 0.080662 327.235909 1.61324 0.80217358
20 0.00306718 21000 0.5 0.080746 326.032816 1.61492 0.80513368
20 0.0030691 21000 0.5 0.08093 325.828269 1.6186 0.80563912
20 0.00308174 21000 0.5 0.080551 324.491926 1.61102 0.80895695
20 0.00309511 21000 0.5 0.08069 323.090103 1.6138 0.81246686
20 0.00309996 21000 0.5 0.080559 322.585108 1.61118 0.81373874
20 0.0031112 21000 0.5 0.080734 321.419163 1.61468 0.81669057
20 0.00312215 21000 0.5 0.080598 320.291935 1.61196 0.81956481
20 0.00313064 21000 0.5 0.0806854 319.423607 1.613708 0.82179274
20 0.00313364 21000 0.5 0.080591 319.117818 1.61182 0.8225802
20 0.00314229 21000 0.5 0.080617 318.238806 1.61234 0.82485226
20 0.00314882 21000 0.5 0.080724 317.579702 1.61448 0.82656416
20 0.00315887 21000 0.5 0.080644 316.569273 1.61288 0.8292024
20 0.00317037 21000 0.5 0.080633 315.42057 1.61266 0.8322222
20 0.00317154 21000 0.5 0.08094 315.304307 1.6188 0.83252906
20 0.00318032 21000 0.5 0.080784 314.434 1.61568 0.83483338
20 0.00318995 21000 0.5 0.08083 313.484266 1.6166 0.8373626
20 0.00320049 21000 0.5 0.08088 312.451702 1.6176 0.84012985
20 0.00321249 21000 0.5 0.080677 311.284838 1.61354 0.84327911
20 0.0032232 21000 0.5 0.080623 310.2505 1.61246 0.8460905
20 0.00323263 21000 0.5 0.080768 309.345224 1.61536 0.84856652
20 0.00324063 21000 0.5 0.080886 308.581768 1.61772 0.85066594
continued on next page

49
nxx+nyy kxx E t time a analytical theoretical
20 0.00325526 21000 0.5 0.080651 307.195539 1.61302 0.8545046
20 0.00326452 21000 0.5 0.080626 306.323756 1.61252 0.85693648
20 0.00327527 21000 0.5 0.080992 305.318463 1.61984 0.85975803
20 0.00328775 21000 0.5 0.080987 304.159454 1.61974 0.86303416
20 0.00329456 21000 0.5 0.080905 303.531055 1.6181 0.8648209
20 0.003309 21000 0.5 0.080641 302.206534 1.61282 0.86861127
20 0.00331544 21000 0.5 0.080946 301.619536 1.61892 0.87030172
20 0.00331887 21000 0.5 0.08061 301.307422 1.6122 0.87120323
20 0.00332247 21000 0.5 0.080719 300.980583 1.61438 0.87214928
20 0.00334041 21000 0.5 0.080594 299.364226 1.61188 0.87685828
20 0.00334936 21000 0.5 0.080537 298.564556 1.61074 0.87920684
20 0.00336086 21000 0.5 0.08088 297.543166 1.6176 0.88222493
20 0.00337072 21000 0.5 0.080587 296.67222 1.61174 0.8848149
20 0.00338828 21000 0.5 0.080605 295.134717 1.6121 0.88942434
20 0.00339555 21000 0.5 0.080986 294.503409 1.61972 0.89133094
20 0.00340944 21000 0.5 0.080653 293.302928 1.61306 0.89497913
20 0.0034189 21000 0.5 0.080559 292.491659 1.61118 0.89746149
20 0.00342903 21000 0.5 0.08056 291.627797 1.6112 0.90011996
20 0.00344252 21000 0.5 0.080557 290.485129 1.61114 0.90366072
20 0.00345261 21000 0.5 0.080496 289.635668 1.60992 0.90631103
20 0.00346432 21000 0.5 0.08045 288.657107 1.609 0.90938346
20 0.0034806 21000 0.5 0.081512 287.306483 1.63024 0.91365847
20 0.00349638 21000 0.5 0.08147 286.009718 1.6294 0.91780098
20 0.00350234 21000 0.5 0.081476 285.523717 1.62952 0.91936321
20 0.0035125 21000 0.5 0.081462 284.697849 1.62924 0.92203015
20 0.00352797 21000 0.5 0.081448 283.449112 1.62896 0.92609216
20 0.00353735 21000 0.5 0.081467 282.697439 1.62934 0.92855457
20 0.00354031 21000 0.5 0.081408 282.461451 1.62816 0.92933035
20 0.00355434 21000 0.5 0.081407 281.34633 1.62814 0.93301377
20 0.00356459 21000 0.5 0.081393 280.537205 1.62786 0.93570477
20 0.00357667 21000 0.5 0.081375 279.589785 1.6275 0.9388755
20 0.00358882 21000 0.5 0.08118 278.642874 1.6236 0.94206608
20 0.00359132 21000 0.5 0.08139 278.449119 1.6278 0.9427216
20 0.00360383 21000 0.5 0.081375 277.482447 1.6275 0.94600578
20 0.00361002 21000 0.5 0.081365 277.006746 1.6273 0.94763035
20 0.00361937 21000 0.5 0.081373 276.291199 1.62746 0.95008455
20 0.00362956 21000 0.5 0.081392 275.515547 1.62784 0.9527593
20 0.00364002 21000 0.5 0.081323 274.723656 1.62646 0.95550563
20 0.00365397 21000 0.5 0.081302 273.675184 1.62604 0.95916625
20 0.00367363 21000 0.5 0.081228 272.210049 1.62456 0.96432884
continued on next page

50
nxx+nyy kxx E t time a analytical theoretical
20 0.00369942 21000 0.5 0.08107 270.312902 1.6214 0.97109682
20 0.00370759 21000 0.5 0.080784 269.716956 1.61568 0.97324248
20 0.00371315 21000 0.5 0.080927 269.313045 1.61854 0.97470213
20 0.00371625 21000 0.5 0.080795 269.088445 1.6159 0.97551569
20 0.00372208 21000 0.5 0.080722 268.666646 1.61444 0.97704722
20 0.00373983 21000 0.5 0.080747 267.391851 1.61494 0.98170531
20 0.00375282 21000 0.5 0.080671 266.466156 1.61342 0.98511572
20 0.00376758 21000 0.5 0.080648 265.422079 1.61296 0.98899082
20 0.00377869 21000 0.5 0.080681 264.64188 1.61362 0.9919065
20 0.00380051 21000 0.5 0.080604 263.122824 1.61208 0.99763295
20 0.00380977 21000 0.5 0.080563 262.482701 1.61126 1.00006591
20 0.00382801 21000 0.5 0.080583 261.232484 1.61166 1.00485206
20 0.00384239 21000 0.5 0.08054 260.254613 1.6108 1.00862765
20 0.00385436 21000 0.5 0.080451 259.446461 1.60902 1.01176944
20 0.00386676 21000 0.5 0.080536 258.614642 1.61072 1.01502373
20 0.00388191 21000 0.5 0.08045 257.604902 1.609 1.01900235
20 0.00389382 21000 0.5 0.080457 256.817494 1.60914 1.02212663
20 0.0039159 21000 0.5 0.080401 255.368814 1.60802 1.02792505
20 0.00392498 21000 0.5 0.080432 254.778423 1.60864 1.03030703
20 0.00393076 21000 0.5 0.080389 254.403929 1.60778 1.03182369
20 0.00393719 21000 0.5 0.080372 253.987932 1.60744 1.03351367
20 0.00394592 21000 0.5 0.080476 253.426193 1.60952 1.03580453
20 0.00395857 21000 0.5 0.080349 252.616476 1.60698 1.03912462
20 0.0039828 21000 0.5 0.080315 251.079599 1.6063 1.04548518
20 0.00399236 21000 0.5 0.080236 250.478704 1.60472 1.04799329
20 0.00400996 21000 0.5 0.080251 249.378868 1.60502 1.05261525
20 0.00402791 21000 0.5 0.080317 248.267598 1.60634 1.05732686
20 0.0040445 21000 0.5 0.080142 247.249254 1.60284 1.06168167
20 0.00405735 21000 0.5 0.080108 246.466387 1.60216 1.06505395
20 0.00407653 21000 0.5 0.080191 245.306386 1.60382 1.07009037
20 0.00408851 21000 0.5 0.080245 244.587714 1.6049 1.07323461
20 0.00410619 21000 0.5 0.080016 243.53483 1.60032 1.07787457
20 0.00412145 21000 0.5 0.080869 242.633156 1.61738 1.08188017
20 0.00413471 21000 0.5 0.080695 241.854853 1.6139 1.08536172
20 0.00416334 21000 0.5 0.080666 240.191696 1.61332 1.09287708
20 0.004196 21000 0.5 0.080728 238.321944 1.61456 1.10145124
20 0.0042096 21000 0.5 0.080442 237.552151 1.60884 1.10502052
20 0.00422459 21000 0.5 0.080409 236.709621 1.60818 1.10895366
20 0.0042509 21000 0.5 0.080769 235.244154 1.61538 1.11586195
20 0.00426458 21000 0.5 0.08106 234.489687 1.6212 1.11945222
continued on next page

51
nxx+nyy kxx E t time a analytical theoretical
20 0.00428042 21000 0.5 0.080549 233.622089 1.61098 1.12360951
20 0.00430482 21000 0.5 0.080395 232.297959 1.6079 1.13001423
20 0.00432468 21000 0.5 0.080397 231.231208 1.60794 1.13522739
20 0.00435077 21000 0.5 0.08063 229.84418 1.6126 1.14207808
20 0.0043783 21000 0.5 0.080703 228.399102 1.61406 1.14930399
20 0.00439085 21000 0.5 0.0806 227.746572 1.612 1.15259693
20 0.00441593 21000 0.5 0.08055 226.45262 1.611 1.15918288
20 0.00442802 21000 0.5 0.080539 225.834704 1.61078 1.16235457
20 0.00443303 21000 0.5 0.080556 225.579202 1.61112 1.16367111
20 0.00445188 21000 0.5 0.125068 224.624407 2.50136 1.16861744
20 0.00449497 21000 0.5 0.125123 222.470735 2.50246 1.17993047
20 0.00450745 21000 0.5 0.125216 221.855081 2.50432 1.18320482
20 0.00452999 21000 0.5 0.125047 220.751003 2.50094 1.18912257
20 0.00455186 21000 0.5 0.124886 219.690431 2.49772 1.19486315
20 0.00457078 21000 0.5 0.124836 218.781197 2.49672 1.19982889
20 0.00460337 21000 0.5 0.12488 217.232081 2.4976 1.20838506
20 0.00462498 21000 0.5 0.079757 216.217299 1.59514 1.21405642
20 0.00464791 21000 0.5 0.124784 215.15039 2.49568 1.2200768
20 0.0046702 21000 0.5 0.124735 214.123674 2.4947 1.22592703
20 0.00468975 21000 0.5 0.124762 213.231164 2.49524 1.23105833
20 0.00471302 21000 0.5 0.124743 212.177984 2.49486 1.23716889
20 0.00472854 21000 0.5 0.124766 211.481597 2.49532 1.24124275
20 0.00474956 21000 0.5 0.12462 210.545945 2.4924 1.24675876
20 0.00477434 21000 0.5 0.079339 209.453246 1.58678 1.25326298
20 0.00480058 21000 0.5 0.079276 208.308313 1.58552 1.26015134
20 0.00481686 21000 0.5 0.124445 207.604209 2.4889 1.26442523
20 0.00482962 21000 0.5 0.079182 207.055645 1.58364 1.26777514
20 0.00484824 21000 0.5 0.079112 206.260441 1.58224 1.27266284
20 0.00484895 21000 0.5 0.124159 206.230342 2.48318 1.27284859
20 0.00487012 21000 0.5 0.079051 205.33389 1.58102 1.27840562
20 0.00488331 21000 0.5 0.078905 204.779275 1.5781 1.281868
20 0.00492581 21000 0.5 0.078862 203.012279 1.57724 1.29302523
20 0.00494798 21000 0.5 0.078788 202.102571 1.57576 1.29884542
20 0.0049583 21000 0.5 0.12599 201.682134 2.5198 1.30155307
20 0.00496706 21000 0.5 0.080211 201.326258 1.60422 1.30385377
20 0.00499242 21000 0.5 0.125987 200.303641 2.51974 1.31051038
20 0.00502186 21000 0.5 0.125282 199.129261 2.50564 1.31823921
20 0.00503601 21000 0.5 0.125513 198.569802 2.51026 1.32195327
20 0.00506688 21000 0.5 0.08016 197.359996 1.6032 1.33005678
20 0.00509115 21000 0.5 0.079975 196.419395 1.5995 1.33642607
continued on next page

52
nxx+nyy kxx E t time a analytical theoretical
20 0.00510044 21000 0.5 0.080074 196.061586 1.60148 1.33886503
20 0.00512696 21000 0.5 0.079913 195.047436 1.59826 1.34582646
20 0.00514537 21000 0.5 0.079908 194.349296 1.59816 1.35066092
20 0.00517225 21000 0.5 0.1254 193.339421 2.508 1.35771587
20 0.00517794 21000 0.5 0.125476 193.126872 2.50952 1.35921012
20 0.00520445 21000 0.5 0.124495 192.143242 2.4899 1.36616827
20 0.00521719 21000 0.5 0.124948 191.6739 2.49896 1.36951353
20 0.00523893 21000 0.5 0.12483 190.878648 2.4966 1.37521929
20 0.00526974 21000 0.5 0.124699 189.762594 2.49398 1.3833074
20 0.0053188 21000 0.5 0.124513 188.012161 2.49026 1.39618628
20 0.00534079 21000 0.5 0.124452 187.23828 2.48904 1.40195691
20 0.00537175 21000 0.5 0.124364 186.159219 2.48728 1.41008327
20 0.00538266 21000 0.5 0.124326 185.781915 2.48652 1.412947
20 0.00541275 21000 0.5 0.124236 184.749138 2.48472 1.4208456
20 0.00546292 21000 0.5 0.124105 183.052364 2.4821 1.43401589
20 0.00548093 21000 0.5 0.078736 182.450946 1.57472 1.43874288
20 0.00549615 21000 0.5 0.123988 181.945528 2.47976 1.4427395
20 0.00555067 21000 0.5 0.123828 180.158404 2.47656 1.4570511
20 0.00557982 21000 0.5 0.123733 179.217163 2.47466 1.46470347
20 0.00562319 21000 0.5 0.123586 177.835011 2.47172 1.47608729
20 0.00564251 21000 0.5 0.123541 177.226163 2.47082 1.48115829
20 0.00565905 21000 0.5 0.123474 176.70821 2.46948 1.48549974
20 0.00566749 21000 0.5 0.123501 176.444942 2.47002 1.48771621
20 0.00569977 21000 0.5 0.123563 175.445738 2.47126 1.4961891
20 0.00573934 21000 0.5 0.123256 174.236209 2.46512 1.50657548
20 0.00577202 21000 0.5 0.123127 173.249578 2.46254 1.51515521
20 0.00578841 21000 0.5 0.123124 172.759049 2.46248 1.51945731
20 0.00583382 21000 0.5 0.122872 171.414165 2.45744 1.53137869
20 0.00586166 21000 0.5 0.122797 170.600053 2.45594 1.53868651
20 0.00590999 21000 0.5 0.122584 169.20513 2.45168 1.5513714
20 0.00591543 21000 0.5 0.122553 169.049448 2.45106 1.5528001
20 0.00593054 21000 0.5 0.122515 168.618684 2.4503 1.55676698
20 0.00597679 21000 0.5 0.122333 167.313813 2.44666 1.56890812
20 0.00598373 21000 0.5 0.122295 167.119935 2.4459 1.57072824
20 0.0060094 21000 0.5 0.122177 166.405933 2.44354 1.5774678
20 0.00604685 21000 0.5 0.124745 165.375495 2.4949 1.58729683
20 0.00608447 21000 0.5 0.124679 164.35288 2.49358 1.59717311
20 0.00611789 21000 0.5 0.124387 163.455038 2.48774 1.60594622
20 0.00617135 21000 0.5 0.124618 162.039173 2.49236 1.61997864
20 0.00620071 21000 0.5 0.123994 161.271727 2.47988 1.62768766
continued on next page

53
nxx+nyy kxx E t time a analytical theoretical
20 0.00623677 21000 0.5 0.123888 160.339414 2.47776 1.63715204
20 0.00628305 21000 0.5 0.123476 159.15838 2.46952 1.64930053
20 0.00632833 21000 0.5 0.123525 158.019485 2.4705 1.66118755
20 0.00636809 21000 0.5 0.12354 157.032968 2.4708 1.6716235
20 0.00644988 21000 0.5 0.123267 155.041606 2.46534 1.69309392
20 0.00648826 21000 0.5 0.123102 154.124514 2.46204 1.70316839
20 0.00652678 21000 0.5 0.122928 153.214896 2.45856 1.71327989
20 0.00657655 21000 0.5 0.122769 152.055479 2.45538 1.72634358
20 0.00662182 21000 0.5 0.12276 151.015907 2.4552 1.73822748
20 0.00663251 21000 0.5 0.122767 150.772443 2.45534 1.74103433
20 0.00669116 21000 0.5 0.122407 149.450865 2.44814 1.75643011
20 0.00674276 21000 0.5 0.122257 148.307247 2.44514 1.76997419
20 0.00679844 21000 0.5 0.122127 147.092519 2.44254 1.7845911
20 0.00683438 21000 0.5 0.121968 146.318946 2.43936 1.79402604
20 0.0068816 21000 0.5 0.121791 145.31512 2.43582 1.80641904
20 0.00692184 21000 0.5 0.121639 144.47036 2.43278 1.8169817
20 0.00698455 21000 0.5 0.121387 143.173047 2.42774 1.83344564
20 0.00701896 21000 0.5 0.121248 142.471279 2.42496 1.84247662
20 0.00707491 21000 0.5 0.121065 141.34448 2.4213 1.85716485
20 0.00711176 21000 0.5 0.120936 140.612136 2.41872 1.86683744
20 0.00716083 21000 0.5 0.120761 139.648703 2.41522 1.87971671
20 0.00721179 21000 0.5 0.187707 138.661851 3.75414 1.89309459
20 0.00729813 21000 0.5 0.120186 137.021336 2.40372 1.91576005
20 0.00734722 21000 0.5 0.122098 136.105972 2.44196 1.92864424
20 0.00735766 21000 0.5 0.187072 135.912845 3.74144 1.93138477
20 0.00745497 21000 0.5 0.18661 134.138615 3.7322 1.95693089
20 0.00751132 21000 0.5 0.19062 133.132387 3.8124 1.97172157
20 0.0075693 21000 0.5 0.19044866 132.112583 3.80897328 1.98694169
20 0.00762601 21000 0.5 0.19019 131.130237 3.8038 2.00182663
20 0.00768959 21000 0.5 0.190191 130.045994 3.80382 2.01851662
20 0.00775181 21000 0.5 0.189695 129.002193 3.7939 2.03484912
20 0.00780947 21000 0.5 0.189816 128.049654 3.79632 2.04998602
20 0.00785779 21000 0.5 0.189405 127.262213 3.7881 2.06267041
20 0.0079199 21000 0.5 0.189473 126.264257 3.78946 2.07897315
20 0.00798912 21000 0.5 0.189288 125.17027 3.78576 2.09714336
20 0.00806337 21000 0.5 0.188995 124.017626 3.7799 2.11663461
20 0.00811101 21000 0.5 0.188825 123.289181 3.7765 2.12914059
20 0.00819497 21000 0.5 0.18847 122.026131 3.7694 2.15117858
20 0.00824507 21000 0.5 0.187779 121.284563 3.75558 2.1643315
20 0.00831733 21000 0.5 0.188209 120.23093 3.76418 2.18329843
continued on next page

54
nxx+nyy kxx E t time a analytical theoretical
20 0.00838763 21000 0.5 0.187675 119.223183 3.7535 2.201753
20 0.00846221 21000 0.5 0.187393 118.172498 3.74786 2.22132902
20 0.00853508 21000 0.5 0.18714 117.163572 3.7428 2.24045747
20 0.00860509 21000 0.5 0.186854 116.210264 3.73708 2.25883663
20 0.0086905 21000 0.5 0.186075 115.068183 3.7215 2.28125615
20 0.00875869 21000 0.5 0.186152 114.172375 3.72304 2.29915512
20 0.00882844 21000 0.5 0.185963 113.270345 3.71926 2.31746447
20 0.00892784 21000 0.5 0.18545 112.009165 3.709 2.34355822
20 0.00899856 21000 0.5 0.185106 111.12884 3.70212 2.3621231
20 0.00907058 21000 0.5 0.184261 110.246569 3.68522 2.38102648
20 0.00912324 21000 0.5 0.184702 109.610206 3.69404 2.39484999
20 0.00925153 21000 0.5 0.184192 108.090251 3.68384 2.42852613
20 0.00934464 21000 0.5 0.183789 107.013217 3.67578 2.45296803
20 0.009384 21000 0.5 0.183632 106.56439 3.67264 2.46329942
20 0.00949874 21000 0.5 0.188803 105.27715 3.77606 2.49341855
20 0.00960234 21000 0.5 0.187724 104.141298 3.75448 2.52061387
20 0.00961795 21000 0.5 0.18782 103.97231 3.7564 2.52471066
20 0.00968472 21000 0.5 0.187534 103.255422 3.75068 2.54223938
20 0.00978704 21000 0.5 0.187323 102.175939 3.74646 2.56909801
20 0.00984838 21000 0.5 0.18693 101.53953 3.7386 2.58520006
20 0.00999343 21000 0.5 0.186563 100.065772 3.73126 2.62327462
20 0.01007988 21000 0.5 0.186257 99.2075288 3.72514 2.64596854
20 0.01015072 21000 0.5 0.185954 98.515197 3.71908 2.66456352
20 0.01030511 21000 0.5 0.18532 97.0392381 3.7064 2.70509131
20 0.01031587 21000 0.5 0.185261 96.9379798 3.70522 2.70791696
20 0.01051999 21000 0.5 0.18446 95.0571572 3.6892 2.76149643
20 0.01060731 21000 0.5 0.184043 94.2745705 3.68086 2.78442001
20 0.01074209 21000 0.5 0.183676 93.0917886 3.67352 2.81979758
20 0.01082993 21000 0.5 0.183363 92.3366981 3.66726 2.84285669
20 0.01096489 21000 0.5 0.182875 91.2001971 3.6575 2.87828325
20 0.01109652 21000 0.5 0.182426 90.1183151 3.64852 2.91283742
20 0.01119816 21000 0.5 0.281699 89.3004221 5.63398 2.93951578
20 0.01121277 21000 0.5 0.281628 89.1840365 5.63256 2.94335186
20 0.01121294 21000 0.5 0.281616 89.1826889 5.63232 2.94339634
20 0.01130758 21000 0.5 0.281166 88.4362342 5.62332 2.96824036
20 0.01134419 21000 0.5 0.280993 88.1508748 5.61986 2.97784906
20 0.01139435 21000 0.5 0.18128 87.7627623 3.6256 2.99101798
20 0.01140727 21000 0.5 0.280699 87.6633587 5.61398 2.99440957
20 0.01140908 21000 0.5 0.180983 87.6494955 3.61966 2.99488318
20 0.01145687 21000 0.5 0.18073 87.2838804 3.6146 3.00742816
continued on next page

55
nxx+nyy kxx E t time a analytical theoretical
20 0.01156951 21000 0.5 0.279916 86.4340803 5.59832 3.03699651
20 0.01175735 21000 0.5 0.278899 85.0531749 5.57798 3.08630454
20 0.01189636 21000 0.5 0.278037 84.0593119 5.56074 3.12279501
20 0.01200956 21000 0.5 0.285388 83.2670145 5.70776 3.15250885
20 0.01218304 21000 0.5 0.287066 82.0813182 5.74132 3.19804806
20 0.01228469 21000 0.5 0.286794 81.4021339 5.73588 3.22473119
20 0.01244581 21000 0.5 0.285769 80.3483218 5.71538 3.2670253
20 0.01265309 21000 0.5 0.284777 79.0320556 5.69554 3.32143708
20 0.01279354 21000 0.5 0.284155 78.1644486 5.6831 3.35830425
20 0.01282286 21000 0.5 0.28423 77.9857443 5.6846 3.3659998
20 0.01298164 21000 0.5 0.283474 77.0318784 5.66948 3.40768011
20 0.01306594 21000 0.5 0.283135 76.5348427 5.6627 3.42981041
20 0.01331729 21000 0.5 0.28204 75.0903425 5.6408 3.49578909
20 0.01351431 21000 0.5 0.28118 73.9956489 5.6236 3.54750589
20 0.01369556 21000 0.5 0.280253 73.0163765 5.60506 3.59508391
20 0.01387975 21000 0.5 0.279559 72.0474085 5.59118 3.64343431
20 0.01398386 21000 0.5 0.278909 71.5110009 5.57818 3.67076389
20 0.01425147 21000 0.5 0.277496 70.1681947 5.54992 3.74101117
20 0.01447467 21000 0.5 0.276896 69.0862155 5.53792 3.79960022
20 0.01461066 21000 0.5 0.27588 68.4431733 5.5176 3.8352985
20 0.01481222 21000 0.5 0.27545 67.5118299 5.509 3.88820745
20 0.01517543 21000 0.5 0.274287 65.8960063 5.48574 3.98354946
20 0.01540718 21000 0.5 0.273288 64.9047845 5.46576 4.04438598
20 0.01561208 21000 0.5 0.27252 64.0529672 5.4504 4.09817081
20 0.01575149 21000 0.5 0.28285 63.4860654 5.657 4.13476561
20 0.01604555 21000 0.5 0.281842 62.322568 5.63684 4.21195738
20 0.01616133 21000 0.5 0.281461 61.8760826 5.62922 4.24235002
20 0.0163615 21000 0.5 0.280666 61.1191072 5.61332 4.29489258
20 0.01665641 21000 0.5 0.279704 60.0369558 5.59408 4.37230697
20 0.01670272 21000 0.5 0.279486 59.870473 5.58972 4.38446511
20 0.01684356 21000 0.5 0.424664 59.3698649 8.49328 4.42143502
20 0.01698571 21000 0.5 0.278411 58.8730303 5.56822 4.4587479
20 0.01710244 21000 0.5 0.277989 58.4711835 5.55978 4.48939092
20 0.01751678 21000 0.5 0.422028 57.0881026 8.44056 4.59815597
20 0.01770538 21000 0.5 0.42122 56.4800002 8.4244 4.64766288
20 0.01843042 21000 0.5 0.4176 54.2581118 8.352 4.83798627
20 0.01904004 21000 0.5 0.414659 52.5209046 8.29318 4.99800988
20 0.02002609 21000 0.5 0.410184 49.9348501 8.20368 5.25684966
20 0.02105927 21000 0.5 0.542391 47.4850197 10.84782 5.5280592
20 0.02172547 21000 0.5 0.565788 46.0289263 11.31576 5.70293554
continued on next page

56
nxx+nyy kxx E t time a analytical theoretical
20 0.0220105 21000 0.5 0.56424 45.4328604 11.2848 5.77775641
20 0.02264587 21000 0.5 0.562806 44.1581658 11.25612 5.94454038
20 0.0230772 21000 0.5 0.56137 43.3328048 11.2274 6.05776619
20 0.02376944 21000 0.5 0.558798 42.070824 11.17596 6.23947845
20 0.02438112 21000 0.5 0.555937 41.0153449 11.11874 6.4000437
20 0.02495012 21000 0.5 0.553447 40.0799728 11.06894 6.54940564
20 0.02551941 21000 0.5 0.551174 39.1858596 11.02348 6.69884501
20 0.02625424 21000 0.5 0.548189 38.089091 10.96378 6.89173706
20 0.02668879 21000 0.5 0.546388 37.4689194 10.92776 7.00580653
20 0.02750609 21000 0.5 0.543103 36.3555888 10.86206 7.22034793
20 0.02835911 21000 0.5 0.539926 35.2620318 10.79852 7.44426758
20 0.02938637 21000 0.5 0.536279 34.0293769 10.72558 7.71392321
20 0.0298439 21000 0.5 0.534747 33.5076823 10.69494 7.83402437
20 0.03120274 21000 0.5 0.558439 32.0484705 11.16878 8.19071849
20 0.0321394 21000 0.5 0.55535 31.1144568 11.107 8.43659273
20 0.03286608 21000 0.5 0.552885 30.4265108 11.0577 8.6273448
20 0.03320454 21000 0.5 5.52E-01 30.1163622 11.03552 8.71619215
20 0.0343449 21000 0.5 0.547743 29.1164036 10.95486 9.01553653
20 0.03568094 21000 0.5 0.543344 28.0261713 10.86688 9.36624548
20 0.03664591 21000 0.5 0.678919 27.2881747 13.57838 9.61955142
20 0.03803401 21000 0.5 0.535698 26.2922566 10.71396 9.98392811
20 0.03953274 21000 0.5 0.668868 25.2954917 13.37736 10.3773432
20 0.04157688 21000 0.5 0.663273 24.0518314 13.26546 10.9139298
20 0.04305908 21000 0.5 0.657208 23.2239051 13.14416 11.3030087
20 0.04533648 21000 0.5 0.692611 22.0572932 13.85222 11.9008256
20 0.04743336 21000 0.5 0.687611 21.082208 13.75222 12.4512575
20 0.04978492 21000 0.5 0.679748 20.0864057 13.59496 13.0685402
20 0.05231091 21000 0.5 0.819762 19.1164718 16.39524 13.7316134
20 0.05464844 21000 0.5 0.666351 18.2987839 13.32702 14.3452156
20 0.05655518 21000 0.5 0.80575 17.6818472 16.115 14.845734
20 0.06231851 21000 0.5 0.786134 16.0465966 15.72268 16.358609
20 0.06631073 21000 0.5 0.776534 15.0805152 15.53068 17.4065671

57
C Thickness

Table 3: Nonlinear buckling membran forces-thickness

t a kxx+kyy theoretical lambda analytical nxx+nyy
0.1 100 0.01 0.105 0.031251 0.156255 5
0.2 100 0.01 0.42 0.189254 0.94627 5
0.3 100 0.01 0.945 0.280145 1.400725 5
0.4 100 0.01 1.68 0.565498 2.82749 5
0.5 100 0.01 2.625 0.691437 3.457185 5
0.6 100 0.01 3.78 0.807102 4.03551 5
0.7 100 0.01 5.145 0.68383 6.8383 10
0.8 100 0.01 6.72 0.687292 6.87292 10
0.9 100 0.01 8.505 0.837072 8.37072 10
1 100 0.01 10.5 0.674114 13.48228 20
1.1 100 0.01 12.705 0.688121 13.76242 20
1.2 100 0.01 15.12 0.817772 16.35544 20
1.3 100 0.01 17.745 0.548856 27.4428 50
1.4 100 0.01 20.58 0.539131 26.95655 50
1.5 100 0.01 23.625 0.700406 35.0203 50
1.6 100 0.01 26.88 0.691273 34.56365 50
1.7 100 0.01 30.345 0.68089 34.0445 50
1.8 100 0.01 34.02 0.815976 40.7988 50
1.9 100 0.01 37.905 0.80368 40.184 50
2 100 0.01 42 0.793236 39.6618 50

58
D Young’s Modulus

Table 4: Nonlinear buckling membrane forces-Young’s modulus

E t aa kxx+kyy lambda theoretical critical
1000 1 100 0.01 0.030313 0.5 0.60626
2000 1 100 0.01 0.07844 1 1.5688
3000 1 100 0.01 0.119985 1.5 2.3997
4000 1 100 0.01 0.11778 2 2.3556
5000 1 100 0.01 0.181058 2.5 3.62116
6000 1 100 0.01 0.277114 3 5.54228
7000 1 100 0.01 0.274915 3.5 5.4983
8000 1 100 0.01 0.272958 4 5.45916
9000 1 100 0.01 0.27104 4.5 5.4208
10000 1 100 0.01 0.411124 5 8.22248
11000 1 100 0.01 0.547676 5.5 10.95352
12000 1 100 0.01 0.547815 6 10.9563
13000 1 100 0.01 0.547246 6.5 10.94492
14000 1 100 0.01 0.545753 7 10.91506
15000 1 100 0.01 0.544407 7.5 10.88814
16000 1 100 0.01 0.54267 8 10.8534
17000 1 100 0.01 0.54116 8.5 10.8232
18000 1 100 0.01 0.539247 9 10.78494
19000 1 100 0.01 0.538002 9.5 10.76004
20000 1 100 0.01 0.535664 10 10.71328
21000 1 100 0.01 0.534597 10.5 10.69194
30000 1 100 0.01 0.11784 15 17.676
40000 1 100 0.01 0.180328 20 27.0492
50000 1 100 0.01 0.275621 25 41.34315
60000 1 100 0.01 0.273074 30 40.9611
70000 1 100 0.01 0.41234 35 61.851
80000 1 100 0.01 0.410074 40 61.5111
90000 1 100 0.01 0.547729 45 82.15935
100000 1 100 0.01 0.54648 50 81.972
110000 1 100 0.01 0.544834 55 81.7251
120000 1 100 0.01 0.542723 60 81.40845
130000 1 100 0.01 0.540421 65 81.06315
140000 1 100 0.01 0.538242 70 80.7363
150000 1 100 0.01 0.536254 75 80.4381
160000 1 100 0.01 0.672786 80 100.9179
170000 1 100 0.01 0.67273 85 100.9095
180000 1 100 0.01 0.66849 90 100.2735
190000 1 100 0.01 0.668828 95 100.3242
200000 1 100 0.01 0.665504 100 99.8256
210000 1 100 0.01 0.807841 105 121.17615

59