Plastic Buckling of Mindlin Plates

PLASTIC BUCKLING OF MINDLIN PLATES
TUN MYINT AUNG
NATIONAL UNIVERSITY OF SINGAPORE 2006
PLASTIC BUCKLING OF MINDLIN PLATES
TUN MYINT AUNG B. Eng, (Yangon Institute of Technology)
A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF CIVIL ENGINEERING NATIONAL UNIVERSITY OF SINGAPORE
2006
ACKNOWLEDGMENT
The author wishes to express his sincere gratitude to Professor Wang Chien Ming, for his guidance, patience and invaluable suggestions throughout the course of study. His extensive knowledge, serious research attitude and enthusiasm have been extremely valuable to the author. Also special thanks go to Professor J. Charabarty for his valuable discussions and help in the research work. The author is grateful to the National University of Singapore for providing the research scholarship during the four-year study. The author also would like to express his gratitude to his girlfriend, Ms Chuang Hui Ming, and to his friends; Mr. Kyaw Moe, Mr. Sithu Htun, Ms. Thida Kyaw, Ms. Khine Khine Oo, Mr. Kyaw Myint Lay, Mr. Vo Khoi Khoa, Mr. Tran Chi Trung, Mr. Vu Khac Kien for their kind help and encouragement. Finally, the author wishes to express his deep gratitude to his family, for their love, understanding, encouragement and continuous support.
TABLE OF CONTENTS
ACKNOWLEDGEMENT ................................................................................... TABLE OF CONTENTS SUMMARY .. NOMENCLATURE . LIST OF FIGURES .
i ii vi ix xii
LIST OF TABLES xviii CHAPTER 1. INTRODUCTION ....................................................................... 1.1 Literature Review ... 1.1.1 Plastic Buckling Theories of Plates 1.1.2 Inclusion of Effect of Transverse Shear Deformation ... 1.2 1.3 Objectives and Scope of Study ..... Organization of Thesis .... 1 3 3 7 8 11
CHAPTER 2. GOVERNING EQUATIONS FOR PLASTIC BUCKLING ANALYSIS .. 2.1 Mindlin Plate Theory .. 2.1.1 Assumptions 2.1.2 Displacement Components . 2.1.3 Strain-Displacement Relations .... 2.2 Stress-Strain Relations in Plastic Range ................................................. 2.2.1 Derivation of Stress-Strain Relations Based on Prandtl-Reuss Equation .. 18 13 13 14 15 16 17
ii
Table of Contents
2.2.2
Derivation of Stress-Strain Relations Based on Henckys Deformation Theory ... 23 25 29 33 34 35 37 43 49 59 60 61 61 63 88 97
2.2.3 2.3
Material Modeling ......
Derivation of Energy Functional and Governing Equations ..
CHAPTER 3. RITZ METHOD FOR PLASTIC BUCKLING ANALYSIS .. 3.1 3.2 3.3 3.4 3.5 3.6 3.7 Introduction . Boundary Conditions .. Ritz Formulation in Cartesian Coordinate System . Ritz Formulation in Skew Coordinate System ... Ritz Formulation in Polar Coordinate System Solution Method for Solving Eigenvalue Problem . Computer Programs
CHAPTER 4. RECTANGULAR PLATES . 4.1 4.2 4.3 4.4 4.5 Introduction ......... Uniaxial and Biaxial Compression . Pure Shear Loading . Combined Shear and Uniaxial Compression .....
Concluding Remarks 103
CHAPTER 5. TRIANGULAR AND ELLIPTICAL PLATES 104 5.1 5.2 5.3 Introduction . 104 Triangular plates Elliptical Plates ... 104 112
iii
Table of Contents
5.4
Concluding Remarks ..
117
CHAPTER 6. SKEW PLATES .. 118 6.1 6.2 6.3 6.4 Introduction . 118 Uniaxial and Biaxial Loading . 119 Shear Loading . 133 Concluding Remarks .. 145 146
CHAPTER 7. CIRCULAR AND ANNULAR PLATES .. 7.1 7.2 7.3 7.4 7.5 7.6
Introduction ..... 146 Analytical Method .. 148
Circular Plates . 153 Annular Plates Circular and annular plates with intermediate ring supports .. 157 168
Concluding Remarks ... 178 180
CHAPTER 8. PLATES WITH COMPLICATING EFFECTS .. 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8
Introduction . 180 Internal line/curved/loop supports .. 181 Point supports . Elastically restrained edges and mixed boundary conditions . Elastic foundation ... 185 191 194
Internal line Hinges . 199 Intermediate in-plane loads . 203 Concluding Remarks ... 205
iv
Table of Contents
CHAPTER 9. CONCLUSIONS AND RECOMMENDATIONS 207 9.1 9.2 Conclusions.. 207 Recommendations for Future Studies . 210
REFERENCES ..... 212 Appendix A: Ritz Program Codes ... 222 Appendix B: Elements of Matrices .. 249 List of Authors Publications .. 256
SUMMARY
This thesis is concerned with the plastic bifurcation buckling of Mindlin plates. So far, most of the limited studies on plastic buckling of plates have been based on the classical thin plate theory (which neglects the effect of transverse shear deformation). In view of this, the first-order shear deformation theory proposed by Mindlin is adopted in this study to model the plates so as to take into consideration the effect of transverse shear deformation that becomes significant when the plate thickness-towidth ratio exceed 0.05. Moreover, earlier studies have been confined to simple rectangular and circular plate shapes. The present study considers, for the first time, the plastic buckling of arbitrary shaped plates that include skew, triangular, elliptical and annular shapes in addition to the commonly treated rectangular and circular shapes. The plates may also be subjected to in-plane normal and/or shear stresses and the edges may take any combination of boundary conditions. In order to capture the plastic behaviour of the plates, two widely used plasticity theories are considered. These two theories are the incremental theory (IT) of plasticity and deformation theory (DT) of plasticity. The explicit expressions of stress-strain relations in elastic/plastic range are derived for the first time based on the adopted plasticity theories for arbitrary in-plane loading. On the basis of these stress-strain relations, the total potential energy functional is formulated and the Ritz method is used for the plastic buckling analysis of plates. It is worth noting that the Ritz method is automated for the first time for such plastic buckling plate analysis. This automation is made possible by employing Ritz functions comprising the product of mathematically complete two-dimensional polynomials and boundary equations raised to appropriate powers that ensure the satisfaction of the vi
Summary
geometric boundary conditions a priori. The Ritz formulations are coded for use in MATHEMATICA for three types of coordinate systems, namely Cartesian, skew and polar coordinate systems so as to better suit the plate shapes considered. The employment of MATHEMATICA enables the differentiation, integration and algebraic manipulations to be done in a symbolic mode and permits executions of the mathematical operations in an exact manner. It should be remarked that the presented Ritz program codes can be used to study not only the plastic buckling of plates but also the elastic buckling of plates. One can calculate the elastic buckling stress parameter based on the classical thin plate theory by setting the thickness-to-width ratio to a small value. If one wish to calculate the elastic buckling stress parameter based on Mindlin plate theory, it can be easily done by setting the tangent modulus and secant modulus equal to Youngs modulus. The results obtained using the developed Ritz method were compared with some limited plastic buckling solutions found in the open literature. The good agreement of the results provides verification of the formulations and program codes. The effects of transverse shear deformation, geometrical parameters such as aspect ratios, thicknessto-width ratios and material parameters on the plastic buckling stress parameter are investigated for various plate shapes, boundary and loading conditions. Moreover, the vast plastic buckling data presented in this thesis should serve as a useful reference source for researchers and engineers who are working on analysis and design of plated structures. Based on comparison and convergence studies, the Ritz method is found to be an efficient and accurate numerical method for plastic buckling of plates. It can be used to study not only the plates with traditional boundary conditions and loading conditions but also the plates with complicating effects such as the presence of internal
vii
Summary
line/curved supports, elastically restrained boundary conditions, elastic foundations, internal line hinges, intermediate in-plane loads. The treatments of these complicating effects in the formulations and solutions technique are also presented. In addition to the Ritz method, a new analytical method is featured for handling the asymmetric buckling problems of circular and annular Mindlin plates. This method is based on the Mindlin (1951) approach where the governing equations of the plates were transformed into standard forms of partial differential equations by using three potential functions. The general solutions of the governing equations are then substituted into the boundary equations and the resulting homogeneous equations are solved for the exact plastic buckling stress parameters. The exact solutions should be useful to researchers for verifying their numerical solutions.
viii
NOMENCLATURE
a b c, k D DT E G h IT N n M nn M ns p Qn R
length of plates width of plates Ramberg-Osgood parameters flexural rigidity of plates deformation theory of plasticity Youngs modulus effective shear modulus thickness of plates incremental theory of plasticity number of polynomial terms number of nodal diameters bending moment normal to the plate edge twisting moment at the plate edge degree of polynomial transverse shear force Radius of circular or annular plate stress deviator tensor displacement components
s ij
u , v, w U V X
strain energy potential energy contact function

parameters used in stress-strain relations
, , , , , ,
ix
loading ratio y / x
shear strain components Kronecker delta normal strain components total effective strain effective elastic strain effective plastic strain aspect ratio of plates (a/b or b/R ) ratio of the radius of first internal ring support to the radius of circular or annular plate (b1/R)
xy , xz , yz
ij x , y , z
e

1
p
Mindlins shear correction factor buckling stress parameter elastic buckling stress parameter plastic buckling stress parameter for R shear plastic buckling stress parameter for S shear nominal yield stress 2% offset yield stress secant yield stresses buckling stress hydrostatic stress ( x + y + z ) / 3 normal stresses effective stress
e
R
S 0 0.2 0.7 , 0.85 c m x , y , z
Nomenclature
xy , xz , yz
shear stresses Poissons ratio thickness-to-width ratio h/b or thickness-to-radius ratio h/R rotations about y-, x-, - and r- axes skew angle of skew plate
x , y , r ,
xi
LIST OF FIGURES
Fig. 2.1 Fig. 2.2 Fig. 2.3 Fig. 2.4 Fig. 3.1 Fig. 3.2 Fig. 3.3 Fig. 3.4 Fig. 4.1 Fig. 4.2
Deformation of cross-section of Mindlin plates Ramberg-Osgood stress-strain relation . Illustration of 0.7 and 0.85 on the stress-strain curve . Offset yield stress .. Forces and moments on a plate element Rectangular plate under in-plane compressive stresses . (a) Skew coordinates and plate dimensions (b) Directions of stresses in the infinitesimal element .. Definitions of and FSCS rectangular plate under in-plane compressive stresses Normalized plastic buckling stress c / 0.2 versus b/h ratio for SSSS square plate .. Normalized plastic buckling stress c / 0.2 versus b/h ratio for CCCC square plate Normalized plastic buckling stress c / 0.2 versus b/h ratio for SCSC square plate . Normalized plastic buckling stress c / 0.2 versus b/h ratio for SSSF square plate .. Effect of transverse shear deformation on plastic buckling stress (IT) of a square plate . Effect of transverse shear deformation on plastic buckling stress (DT) of a square plate Plastic buckling stress parameter versus aspect ratio a/b for SSSS plate based on IT .. Plastic buckling stress parameter versus aspect ratio a/b for SSSS plate based on DT Plastic buckling stress parameter versus aspect ratio a/b for CCCC plate based on IT
16 26 27 29 37 38 44 51 62
71
Fig. 4.3
72
Fig. 4.4
72
Fig. 4.5
73 73 74 76 76 77 xii
Fig. 4.6 Fig. 4.7 Fig. 4.8 Fig. 4.9 Fig. 4.10
Fig. 4.11 Fig. 4.12 Fig. 4.13 Fig. 4.14 Fig. 4.15 Fig. 4.16
Plastic buckling stress parameter versus aspect ratio a/b for CCCC plate based on DT .. Plastic buckling stress parameter versus aspect ratio a/b for SCSC plate based on IT . Plastic buckling stress parameter versus aspect ratio a/b for SCSC plate based on DT ... Plastic buckling stress parameter versus aspect ratio a/b for SSSF plate based on IT . Plastic buckling stress parameter versus aspect ratio a/b for SSSF plate based on DT Contour and 3D plots of simply supported rectangular plate under uniaxial loading ( = 0.05 , a / b = 1.1 ) for (a) IT, (b) Elastic and (c) DT .................................................................................... Plastic buckling stress parameter versus stress ratio for SSSS square plate Plastic buckling stress parameter versus stress ratio for CCCC square plate Plastic buckling stress parameter versus stress ratio for SCSC square plate Plastic buckling stress parameter versus stress ratio for SSSF square plate Mode switching with respect to stress ratio for SSSS square plate ... Effect of material properties on normalized plastic buckling stress c / 0.2 for SSSS square plate (IT) .. Effect of material properties on normalized plastic buckling stress c / 0.2 for SSSS square plate (DT) . Normalized plastic buckling stress c / 0.2 versus modified slenderness ratio for SSSS square plate . Comparison study of buckling stresses obtained by DT and IT with test results (Al-7075-T6) Comparison study of buckling stresses obtained by DT and IT with test results (Al-2014-T6)
77 78 78 79 79
80
Fig. 4.17
82
Fig. 4.18
82
Fig. 4.19
83
Fig. 4.20
83
Fig. 4.21
84 85
Fig. 4.22
Fig. 4.23
85
Fig. 4.24
86 87 87
Fig. 4.25 Fig. 4.26
xiii
List of Figures
Fig. 4.27
Normalized plastic buckling stress xy / 0.2 versus b/h ratio for
SSSS square plate .. Fig. 4.28 Normalized plastic buckling stress xy / 0.2 versus b/h ratio for CCCC square plate
Normalized plastic buckling stress xy / 0.2 versus b/h ratio for SCSC square plate . Fig. 4.30 Normalized plastic buckling stress xy / 0.2 versus b/h ratio for SSSF square plate .. Fig. 4.31 Plastic buckling stress parameter xy versus aspect ratio a/b for SSSS rectangular plate . Fig. 4.32 Plastic buckling stress parameter xy versus aspect ratio a/b for CCCC rectangular plate ... Fig. 4.33 Plastic buckling stress parameter xy versus aspect ratio a/b for SSSF rectangular plate ..... Fig. 4.34 Fig. 4.35 Fig. 4.36
92
93
Fig. 4.29
93
94
95
95
96
Interaction curves for SSSS rectangular plates under combined uniaxial compression and shear stresses ( = 0.05) 101 Normalized interaction curves for simply supported rectangular plate (a/b = 4) 102 Normalized interaction curves for clamped rectangular plate (a / b = 4) ... 102 CS*F isosceles triangular plate and coordinate system . 105 CS*F right-angled triangular plate and coordinate system ... 105 Variation of with respect to a/b ratio for S*S*S* isosceles triangular plates . 111 Variation of with respect to a/b ratio for S*S*S* right angled triangular plates . 111 Elliptical plate under uniform compression and coordinate system .. 112 Variation of with respect to a/b ratio for simply supported elliptical plates ... 116 Variation of with respect to a/b ratio for clamped elliptical plates ... 117
Fig. 5.1 Fig. 5.2 Fig. 5.3 Fig. 5.4 Fig. 5.5 Fig. 5.6 Fig. 5.7
xiv
List of Figures
Fig. 6.1 Fig. 6.2
Skew plate under biaxial compression stresses .. Normalized plastic buckling stress c / 0.2 versus b/h ratio for simply supported skew plate ( = 30) ... Normalized plastic buckling stress c / 0.2 versus b/h ratio for clamped skew plate ( = 30) .. Plastic buckling stress parameter versus aspect ratio a/b for simply supported skew plate ( = 30) ... Plastic buckling stress parameter versus aspect ratio a/b for clamped skew plate ( = 30) . Plastic buckling stress parameter versus skew angle for simply supported skew plate (a/b = 1) Plastic buckling stress parameter versus skew angle for clamped skew plate (a/b = 1) .. Two kinds of shear loading conditions on skew plates (a) R shear loading and (b) S shear loading ... Normalized buckling stress xy / 0.2 versus b/h ratio for simply supported skew plate ( a / b = 1 ) ..
120
130
Fig. 6.3
130 131 131 132 132 133
141
Fig. 6.10
Normalized buckling stress xy / 0.2 versus b/h ratio for clamped skew plate ( a / b = 1 ) ... 142 142 143 143 144 144 145 153 158
Fig. 6.11 Fig. 6.12 Fig. 6.13 Fig. 6.14 Fig. 6.15 Fig. 6.16 Fig. 7.1 Fig. 7.2
Variation of plastic buckling stress parameters with respect to aspect ratio for SSSS skew plate with = 30 Variation of plastic shear buckling stress parameters with respect to aspect ratio for CCCC skew plate with = 30 .. Variation of plastic (R shear) buckling stress parameters with respect to skew angle of SSSS skew plate .. Variation of plastic (S shear) buckling stress parameters with respect to skew angle of SSSS skew plate .. Variation of plastic (R shear) buckling stress parameters with respect to skew angle of CCCC skew plate Variation of plastic (S shear) buckling stress parameters with respect to skew angle of CCCC skew plate Clamped circular plate under in-plane compressive stress . C-S annular plate under in-plane loading
xv
List of Figures
Fig. 7.3
Variation of buckling stress parameter with respect to ratio for C-C annular plate ... Variation of buckling stress parameter with respect to ratio for C-S annular plate ... Variation of buckling stress parameter with respect to ratio for C-F annular plate ... Variation of buckling stress parameter with respect to ratio for S-S annular plate ... Variation of buckling stress parameter with respect to ratio for S-C annular plate ... Variation of buckling stress parameter with respect to ratio for S-F annular plate ... Variation of buckling stress parameter with respect to ratio for F-C annular plate ... Variation of buckling stress parameter with respect to ratio for F-S annular plate ... Critical buckling mode shapes for C-C annular plates of various ratios (DT, = 0.05) .. Simply supported (a) annular plate (b) circular plate with multiple concentric ring supports .. Variation of buckling stress parameter with respect to ring support location 1 for annular plate ( = 0.2, IT) .
164
Fig. 7.4
164
Fig. 7.5
165
Fig. 7.6
165
Fig. 7.7
166
Fig. 7.8
166
Fig. 7.9
167
Fig. 7.10
167 168
Fig. 7.11
Fig. 7.12 Fig. 7.13
169
175
Fig. 7.14
Variation of buckling stress parameter with respect to ring support location 1 for annular plates ( = 0.2, DT) . Variation of buckling stress parameter with respect to ring support location 1 for circular plates Critical Buckling mode shapes for C-C annular plates with various ring support locations (b/R = 0.2,DT, h/R = 0.05) .. Plates with internal (a) line, (b) curved and (c) loop supports Simulation of internal supports using point supports Rectangular plate with an internal line support ..
175
Fig. 7.15
176
Fig. 7.16 Fig. 8.1 Fig. 8.2 Fig. 8.3
177 182 182 183
xvi
List of Figures
Critical buckling mode shapes of square plates with internal line support . Four point supported square plate ... Square plates with mixed boundary conditions under uniaxial compression . Typical buckling mode shapes for bilateral and unilateral buckling of clamped rectangular plates ... A rectangular plate with an internal hinge under in-plane compressive stress ... Rectangular plate under edge and intermediate compressive stress ..
185 190 193 199 201 204
xvii
LIST OF TABLES
Table. 2.1
Values of 0.7 , 0.85 and n for various aluminum alloys under room temperature (From Bruhn, 1965) ... Sample basic functions for rectangular plates with various boundary conditions Comparison and convergence study of for square plates under uniform uniaxial loading with various boundary conditions .. Comparison and convergence study of for rectangular plates under uniform uniaxial loading with various boundary conditions (a/b = 2) ... Convergence and comparison studies of plastic buckling stress parameters for square plates under uniform biaxial loading .. Convergence and comparison studies of plastic buckling stress parameters for rectangular plates under uniform biaxial loading (a/b = 2) ... Parametric study on the value of transverse shear correction factor for simply supported square plate under biaxial loading .... Convergence and comparison studies of plastic buckling stress parameters for SSSS rectangular plates under shear . Convergence and comparison studies of plastic buckling stress parameters for CCCC rectangular plates under shear ... Buckling mode shapes for Rectangular plate under pure shear ( = 0.03)
28 63 66
Table. 4.1 Table. 4.2 Table. 4.3
67 68
Table. 4.4 Table. 4.5
69 71 91 92
Table. 4.6 Table. 4.7 Table. 4.8 Table. 4.9
96
Table. 4.10 Convergence study of plastic buckling shear stress parameters xy for simply supported plates under combined loadings ( = 0.03) ... Table. 4.11 Convergence study of plastic shear buckling stress parameters xy for clamped square plates under combined loadings. ( = 0.03) . Table. 4.12 Comparison study of elastic shear buckling stress parameters xy for simply supported plates under combined loadings. ( = 0.001, = 0.3) .. 101 100 100
xviii
List of Tables
Table. 5.1
Convergence and comparison studies of plastic buckling stress factors for equilateral triangular plates (a / b = 3 / 2) under uniform compression ...
108
Table. 5.2
Convergence and comparison studies of plastic buckling stress factors for right angled triangular plates under uniform compression (a/b = 1) .. Effect of transverse shear deformation on for S*S*S* isosceles triangular plates ... Effect of transverse shear deformation on for CCC isosceles triangular plates ... Convergence and comparison studies of plastic buckling stress factors for circular plates under uniform compression .... Convergence and comparison studies of plastic buckling stress factors for elliptical plates under uniform compression (a/b=2) .. Effect of transverse shear deformation on for simply supported circular (a/b = 1) and elliptical plates (a/b = 3) ... Effect of transverse shear deformation on for clamped circular (a/b = 1) and elliptical plates (a/b = 3) ... Convergence study of plastic buckling stress parameters for simply supported skew plate under uniaxial compression ( = 0 , a/b=1) .. Convergence study of plastic buckling stress parameters for simply supported skew plate under biaxial compression x = y = ( = 1 , a/b=1) . Convergence study of plastic buckling stress parameters for clamped skew plate under uniaxial compression ( = 0, a / b = 1 ) Convergence study of plastic buckling stress parameters for clamped skew plate under biaxial compression x = y = ( = 1 , a/b = 1) . Comparison study against existing elastic solutions for simply supported skew plates ( = 0.001) ... Comparison study against existing elastic solutions for clamped skew plates ( = 0.001)
108 110 110 114
Table. 5.3 Table. 5.4 Table. 5.5 Table. 5.6
114 115 115
121
Table. 6.2
122
Table. 6.3
123
Table. 6.4
124
Table. 6.5 Table. 6.6
125 126
xix
List of Tables
Effect of transverse shear deformation on for simply supported skew plates under uniaxial compression (a/b = 1) .. Effect of transverse shear deformation on for clamped skew plates under uniaxial compression (a/b = 1) ... Convergence study of plastic buckling parameters for simply supported skew plates under shear loading ( = 45) .
127 128 135 135 136 136 137
Table. 6.10 Convergence study of plastic buckling parameters for simply supported skew plates under shear loading ( = -45) Table. 6.11 Convergence study of plastic buckling parameters for clamped skew plates under shear loading ( = 45) .. Table. 6.12 Convergence study of plastic buckling parameters for clamped skew plates under shear loading ( = -45) . Table. 6.13 Comparison study of elastic shear buckling stress parameter of skew plates under shear (a/b = 1) Table. 6.14 Effect of transverse shear deformation on R for simply supported skew plates (a/b = 1) .. Table. 6.15 Effect of transverse shear deformation on S for simply supported skew plates (a/b = 1) .. Table. 6.16 Effect of transverse shear deformation on R for clamped skew plates (a/b = 1) Table. 6.17 Effect of transverse shear deformation on S for clamped skew plates (a/b = 1) Table. 7.1 Convergence and comparison study of plastic buckling stress parameter for circular plates under uniform radial compression .. Convergence study of plastic buckling stress parameter for annular plates ( = b / R = 0.4 ) .. Comparison study of plastic buckling stress parameter for annular plates ( = b / R = 0.4 ) .. Convergence study of plastic buckling stress parameter for annular plates with one concentric ring support ( = 0.2, 1 = 0.6 ) Convergence study of plastic buckling stress parameter for circular plates with one concentric ring support ( 1 = 0.6 )
138
138
139
139
157 160
Table. 7.2
Table. 7.3
161
Table. 7.4
172
Table. 7.5
173
xx
List of Tables
Table. 7.6 Table. 8.1 Table. 8.2 Table. 8.3 Table. 8.4 Table. 8.5 Table. 8.6 Table. 8.7 Table. 8.8
Comparison study between the optimal ring support radius and radius of nodal circle of second mode shape Convergence and comparison studies of plastic buckling stress parameter for rectangular plate with internal line supports . Convergence and comparison study of plastic buckling stress parameter for four points supported square plate . Comparison studies of buckling stress parameter square plates with mixed boundary conditions Comparison study of unilateral buckling stress parameter for rectangular plates under shear ... Unilateral plastic buckling stress parameters for rectangular plates under shear (IT) .. Unilateral plastic buckling stress parameters for rectangular plates under shear (DT) Buckling stress parameter for square plate with an internal line hinge under uniaxial loading . Convergence and comparison of plastic buckling stress parameters 1 of rectangular plates under edge and intermediate loads. ( 2 = 50, = 0.04 ) ...
178 184 190 194 197 198 198 203
205
xxi
CHAPTER 1
INTRODUCTION
Plates and plated structures are widely used in many engineering structures such as aircrafts, ships, buildings, and offshore structures. Owing to their relatively small thickness when compared to the length dimensions, the design strength of these structures is commonly governed by their buckling capacities. Buckling is a phenomenon in which a structure undergoes visibly large transverse deflection in one of the possible instability modes. Buckling of a structural component may affect the strength or stiffness of the whole structure and even triggers unexpected global failure of the structure. Therefore, it is important to know buckling capacities of the structures in order to avoid premature failure. The first event that highlighted the importance of plate buckling was related to a series of tests carried out in the laboratories of University of London on box beams for Britannia Bridge in 1845 (Walker, 1984). A number of failure cases were due to local instability, which is now referred to as local buckling. However, the first theoretical
Introduction
examination of buckling of plates was attributed to Bryan (1891). He presented the buckling analysis for rectangular plates under uniform uniaxial compression using the energy criterion of stability. It should be noted that he was not only the first to analyze the stability of plates, but also the first to apply the energy criterion of stability to the buckling problem of plate. The application of the energy criterion was later proven to be a powerful tool for the analyses of buckling problems which could not be solved by conventional analytical methods because of the difficulty in the mathematical treatment. Since then, many researchers have studied the buckling of plates of various shapes, loading and boundary conditions using a variety of methods for analysis. Research on the buckling of plates can be generally divided into two groups: elastic buckling and plastic buckling. In the case of elastic buckling, it is assumed that the buckling stress is below the proportional limit of the plate material and the linear stress-strain relation is adopted. In many practical problems, however, the plates may be stressed beyond the proportional limit before the buckling take place i.e. buckling observed in the plastic range. This phenomenon may be likely to occur in the cases of plates whose materials posses a low proportional limit when compared to the nominal yield stress, for example aluminium alloy and stainless steel. In the plastic range, the stress-strain relation is no longer linear and thus the actual buckling load is smaller than the value determined by the elastic buckling analysis. Therefore, plastic buckling analysis needs to be carried out in order to obtained more accurate buckling loads when the aforementioned conditions hold. In the next section, a literature review on the application of plasticity theories in plastic buckling problems of plates and the success of these theories is presented.
Introduction
1.1
Literature Review
In this section, the development of plastic buckling theories and the paradox in the plastic buckling of plates will be presented.
1.1.1
Plastic Buckling Theories of Plates
Various plasticity theories have been proposed in the literature for the plastic buckling analysis of plates. The most commonly used theories are the deformation theory of plasticity (DT) and the incremental theory of plasticity (IT). The deformation theory of plasticity (DT) was pioneered by Ilyushin (1946) and later modified by Stowell (1948) based on the assumption that no unloading take place during buckling (Shanley, 1947). On the other hand, the incremental theory (IT) was first developed by Handelman and Prager (1948). They pointed out that certain postulate of the theory of plasticity which cannot be satisfied by using the stress-strain relation developed by Ilyushin (1946). The main difference between the two theories is that IT relates plastic strain increment to the stress while DT relates the total strain to the state of stress. Pride and Heimerl (1949) performed experiments on simply supported plates in order to assess the various plastic buckling theories in predicting plastic buckling stress. They tested square tubes of 14S-T6 aluminum alloy, for which each of the four plates may be considered simply supported. The results obtained were in good agreement with Stowells unified theory for the plastic buckling of columns and plates (1948). It was found that the results obtained from deformation theory of plasticity (Ilyushin 1948) gave a close correlation with the test results while the incremental 3
Introduction
theory of plasticity (Handelman and Prager 1948) yielded results that do not agree with experimental results. Another experimental investigation on the plastic buckling of plate was studied by Dietrich et al. (1978). Their experimental results are slightly lower than the critical buckling loads as predicted by Stowells theory (DT). Despite the fact that the deformation theory is mathematically less correct especially in nonproportional loading (Hill, 1950), many researchers have continued to use it for solving plastic buckling problems since the results show good agreement with experimental results. For example, Bijlaard (1949) assumed that plastic deformation is governed only by the amount of elastic shearing energy. He derived the stress-strain relations by writing the infinitely small excess strains as total differentials and computing the partial derivatives of the strains with respect to the stresses. ElGhazaly and Sherbourne (1986) proved that the deformation theory can be used for the elastic-plastic buckling analysis of plates under nonproportional external loading and nonproportional stresses. Loading, unloading, and reloading situations have been conveniently considered via conducting multistage analysis after casting the constitutive relations of the deformation theory in an incremental form. On the other hand, many researchers have tried to justify the theoretically more acceptable incremental theory of plasticity, even though it does not yield results that are in agreement with experimental results. Generally, the incremental theory proposed by Handleman and Prager (1948) yield the results that are higher than experimental results. Pearson (1950) modified the Handelman and Pragers calculation by the assumption of continuous loading. He solved the plastic buckling problem of rectangular plate under compressive edge thrust and showed that results were in fair agreement with experimental work by Pride and Heimerl (1949).
Introduction
As another attempt, Sewell (1964) showed that the bifurcation buckling stress is rather sensitive to the direction of normal to the local yield surface. This cast doubt on the validity of any experiment that does not give proper attention to the yield surface. Dubey and Pindera (1977) also studied the effect of rotation of principal axes on the shear modulus in elastic-plastic solids. They concluded that, by taking proper account of the rotation of principal axes, the effective shear modulus of an elastic-plastic solid can be considerably reduced from the elastic-shear modulus which, in turn, lower the plastic buckling load obtained from the incremental theory. However, Hutchinson (1974) pointed out that basic experiments on the multiaxial stress-strain behaviour of typical metals failed to show the occurrence of corners on yield surfaces. Another drawback for the corner theory is that although one can fit the appropriate corner to observe experimental results in order to obtain good agreement, one cannot estimate the appropriate corner a priori. Moreover, based on imperfection sensitivity, Onat and Drucker (1953) tried to justify the inaccuracy of incremental theory in the plastic buckling of plates. They considered the case of an axially loaded cruciform and showed that unavoidable initial imperfections may significantly lower the plastic buckling load predicted by the incremental theory of plasticity. In addition, using the general von Karman non-linear plate equations, Hutchinson and Budiansky (1974) confirmed analytically OnatDruckers conclusion concerning imperfection sensitivity. Neale (1975) also studied the effect of imperfections on the plastic buckling of simply supported rectangular plates. He showed that plastic buckling load obtained from incremental theory can be reduced considerably in the presence of small initial imperfections in geometry. He concluded that a proper consideration of imperfections in geometry provides an explanation for the apparent plastic buckling paradox for axially compressed
Introduction
rectangular plates. Although this justification seems to be logical, it leaves the value of the buckling load dependent on the amount and distribution of imperfections. This lends credibility to the use of the deformation theory in the prediction of bifurcation buckling load for engineering purpose. Similarly, as another justification, Gjelsvik and Lin (1987) investigated the plastic buckling of plate including the effect of friction acting on the loaded edges during tests. The stresses due to friction, which die out away from the edges, are shown to effectively reduce the incremental theory buckling loads to those predicted by the deformation theory and the experimental results. It was argued that the amount of friction required for such a reduction is less than normally expected at the loaded edges of a real test. Recently, Tucu (1991) also showed that incremental theory predictions were significantly reduced when small amounts of biaxial and in-plane shear stresses are included in the prebuckling state as these stresses can be produced by the experimental set-up. Unlike the analysis of Gjelsvik and Lin (1987), where the edges effects were assumed to die out away from the edges, Tucu (1991) assumed that these non-zero stress components exist uniformly in the prebuckling state. In spite of the continuous efforts and progress made in the last few decades, plastic buckling of plates continues to remain a paradox for researchers in the field. Accordingly, some researchers studied elastic/plastic buckling problems of plates based on both the deformation theory and the incremental theory (for e.g. Shrivastava, 1979; Shrivastava, 1995; Durban and Zuckerman, 1999; Wang et al. 2001). Durban and Zuckerman (1999) carried out a detailed parametric study on the elasto-plastic buckling of rectangular plates under biaxial compression by an analytical method. Apart from confirming that the deformation theory furnishes lower buckling stresses than those computed using incremental theory, they reported the existence of an
Introduction
optimal loading part for the deformation theory model. Shrivasta (1979) and Wang et al. (2001) studied the plastic buckling problems of rectangular plates based on Mindlin plate theory. In this study, both plasticity theories (IT and DT) are adopted for plastic buckling analysis of plates in order to compare the results from these two theories.
1.1.2
Inclusions of Effect of Transverse Shear Deformation
Most of the aforementioned studies on plastic buckling analysis of plates adopted the classical thin plate theory. It is well known that the classical thin plate theory neglects the effect of transverse shear deformation. This effect becomes significant when the plate thickness increases. Since plastic buckling is most likely to occur in relatively thick plates, the effect of transverse shear deformation should be taken into account in the plastic buckling analysis. One of the earlier studies on plastic buckling of thick plate was done by Shrivastava (1979). Three cases of plastic buckling of rectangular plates were considered based on the Mindlin plate theory. The thick plate results were compared with thin plate results based on both incremental (IT) and deformation theory (DT) of plasticity. It was found that the reduction due to the effect of transverse shear deformation in plastic buckling stress parameter was generally larger for IT than that for DT. However, this difference in the reduction was not large enough to compensate for the differences between the results from two theories. Another study on plastic buckling of thick plate was investigated by Wang et al. (2001). They used an analytical method to solve for the plastic buckling of (a) uniformly loaded rectangular plates with two opposite edges simply supported while
Introduction
the other edges may take any boundary conditions (b) uniformly loaded circular plate with simply supported or clamped boundary condition. Recently, Wang et al. (2004) developed an analytical relationship between the plastic buckling loads of thick plate and its corresponding thin plate elastic buckling loads. Since elastic buckling stresses of classical thin plates are well-known, plastic buckling stresses of thick plates can be readily obtained without the tedious nonlinear analysis to solve more complicated plastic buckling equations. However, the relation is only valid for simply supported plates of polygonal shapes. All the aforementioned studies on the plastic buckling of thick plates were solved by analytical methods for simple plate shapes and boundary conditions. These analytical methods can only be employed for only a limited number of buckling problems of practical importance due to the complexity arising from material nonlinearity and the difficulty in getting a function to satisfy the geometric and static boundary conditions of non-simple plate shapes. Therefore, for plastic buckling problems of plates with general shapes and boundary conditions, numerical methods such as the finite element method, the finite difference method, the finite strip method and the Ritz method have to be used.
1.2
Objectives and Scope of Study
There are few analytical expressions of plasticity correction factors (Stowell, 1948; Bijlaard, 1949) in order to account for the effect of material nonlinearity. Plastic buckling solutions can be calculated from corresponding elastic buckling solutions through these plasticity correction factors. However, due to the material nonlinearity, these calculations still result in transcendental equations which need iterative method
Introduction
for determining the plastic buckling solutions. Moreover, the analytical expressions of plasticity correction factors are only available for limited loading and boundary conditions. On the other hand, the numerical plastic buckling solutions exist for several loading and boundary conditions for rectangular plates (El-Ghazaly et al., 1984; Bradford and Azhari, 1994, 1995; Smith et al., 2003). It would be useful for the design profession if a simple computing algorithm was available, that is readily adaptable to specific problems in determining the plastic bifurcation buckling loads. In this thesis, an algorithm is developed for the plastic buckling analysis of Mindlin plates. The analysis makes the use of the Ritz method. The major advantage of the Ritz method is that much less computer memory is needed when compared to the finite element method (Leonard and Moon, 1990; Surendran, 2001). So far, the Ritz method has been extensively used in the studies of elastic buckling and vibration problems of plates (Liew et al., 1998; Wang et al., 1994; Liew and Wang, 1995). The scope of this thesis is to extend the Ritz method for the plastic buckling analysis of Mindlin plates and to examine the plastic behaviour of Mindlin plates with various shapes, boundary conditions and loading conditions. In addition to developing a simple algorithm, an analytical method is featured for tackling the asymmetric plastic buckling problems of circular and annular Mindlin plates. Hitherto, no analytical asymmetric buckling solutions are available in the literature for circular and annular Mindlin plates, even for the elastic buckling case. Therefore, this analytical method and the exact solutions will be useful for researchers who wish to verify their numerical solutions and methods. It should be noted that this thesis mainly focuses on the plastic bifurcation buckling analysis of Mindlin plates. It has been assumed that all the plates considered are initially perfectly flat. Moreover, the assumption of small deflection (effect of
Introduction
geometrical nonlinearity is neglected) needs to be adopted in order to obtained plastic bifurcation buckling load. Therefore, the postbuckling analysis (which accounts the effect of geometrical nonlinearity and initial imperfections) is beyond the scope of this thesis. In summary, the main contributions of this thesis include: (1) Developing a Ritz formulation and program codes to tackle plastic buckling problems of Mindlin plates for various shapes (sides are defined by polynomial functions) and boundary conditions. The program was coded in
MATHEMATICA programming language which is a high level programming language that allows readers not only to understand the program easily but also be able to perform mathematical operations in a symbolic and exact manner. The Ritz program code should be useful for engineers and designers who wish to perform plastic buckling analysis of plates since the procedure is simple to understand and easy to use. (2) An analytical approach is developed for the first time in order to solve asymmetric plastic buckling problems as well as elastic buckling problems of circular and annular Mindlin plates. (3) So far plastic buckling of thick plates has been studied only for rectangular plates with two simply supported opposite edges. This study treats general plate shape and investigates the plastic buckling behaviour of various shapes such as rectangular, skew, triangular and elliptical, circular and annular plates with all possible combinations of boundary conditions. Extensive plastic buckling results are presented for the aforementioned plate shapes.
10
Introduction
(4) Modifications of the Ritz method for complicating effects such as the presence of internal line/curved/point supports, internal line hinges, elastic restrained and mixed boundary conditions, elastic foundation and intermediate in-plane loads. In the next section, a brief overview of the organization of thesis is given.
1.3
Organization of Thesis In this chapter, the background information of plate buckling and literature review
on the plastic buckling of plates is presented. Moreover, the objectives, scope of study and organization of thesis are also presented. In Chapter 2, the explicit expressions of stress-strain relations in the plastic range are derived for the first time based on the Prandtl-Ruess and the Hencky equations. Finally, adopting the Mindlin plate theory, the energy functional and the governing equations are presented for plastic buckling of plates under arbitrary loading. In Chapter 3, the detailed formulations of the Ritz method for plastic buckling of Mindlin plate based on three coordinate systems, namely Cartesian, skew and polar coordinate systems are presented. In Chapters 4 and 5, the plastic buckling analyses are carried out by using the Ritz method with the Cartesian coordinate system. Chapter 4 presents the parametric studies on the plastic buckling of rectangular Mindlin plate under in-plane compressive stresses, shear stress and the combinations of these stresses. In Chapter 5, the applicability of the Ritz method to various shapes is illustrated by considering the plastic buckling problems of isosceles triangular plates, right angled triangular plates and elliptical plates.
11
Introduction
In Chapter 6, plastic buckling of skew Mindlin plates is studied by using the Ritz method with the skew coordinate system. The effects of aspect ratio, skew angle and thickness-to-width ratio on the plastic buckling stress parameter are investigated. In Chapter 7, plastic buckling of circular and annular plates is studied by using the Ritz method with the polar coordinate system. In addition to the Ritz method, the development of an analytical method is presented for the first time for asymmetric buckling of circular and annular plates. The plastic buckling analyses of circular and annular plates with/without an internal ring support are performed by both methods. The independent check of the results by both methods confirms the correctness of the methods. In Chapter 8, modifications to the Ritz formulation are presented for handling the plastic buckling of Mindlin plates with complicating effects such as internal line/curved/point supports, elastic foundation, internal line hinges and intermediate inplane loads. The results are verified using existing elastic buckling solutions. Chapter 9 presents conclusions of the research and recommendations for future studies.
12
CHAPTER 2
GOVERNING EQUATIONS FOR PLASTIC BUCKLING OF PLATES
This chapter presents the derivation of the energy functional and the governing equations for the bifurcation plastic buckling of thick plates under in-plane loadings. The significant effect of transverse shear deformation on the buckling stress for thick plates is taken into consideration by adopting the Mindlin plate theory. In order to capture the plastic buckling behaviour of the plate, the two commonly used plasticity theories, namely the incremental theory of plasticity (IT) and the deformation theory of plasticity (DT) were considered. Based on these plasticity theories, the derivation of the stress-strain relations in the plastic range is presented in detail.
2.1. Mindlin Plate Theory Generally, plate analyses are based on the classical thin plate theory (also called Kirchhoff plate theory since it is first proposed by Kirchhoff, 1850) in which the effect
13
Governing Equations for Plastic Buckling Analysis of Plates
of transverse shear is normally neglected. However, when dealing with thick plates, the classical thin plate theory underpredicts the deflections and overpredicts the buckling loads and natural frequencies due to the effect of transverse shear deformation. As we are dealing with thick plates, it is necessary to adopt a more refined plate theory such as the Mindlin plate theory that will allow for the effect of transverse shear deformation.
2.1.1
Assumptions
The assumptions of the Mindlin plate theory are (Mindlin 1951): The plate thickness remains unchanged during deformation, i.e. transverse inextensibility of the plate is maintained. No in-plane deformation in the middle surface of the plate after deformation. The normals during bending undergo constant rotations about the middle surface while maintaining the straightness and thereby admitting a constant shear strain through the plate thickness (Fig. 2.1). These constant rotations of the normals to the middle surface about y- and x- axes now become unknown independent variables and are denoted by x and y , respectively. It should be noted that the first two assumptions are the same as their classical thin plate counterparts. The third assumption that allows the constant rotation of normal is the main difference between the Mindlin plate theory and the classical thin plate theory. The allowance of constant rotation implies that transverse shear strain is constant through the thickness of the plate. This, however, contradicts the fact that the actual transverse shear strain distribution is parabolic through the thickness. As the
14
constant strain (stress) violates the statical requirement of vanishing shear stress at the surface of the plate, a shear correction factor 2 was proposed by Mindlin (1951) to compensate for the error. He pointed out that for an isotropic plate, the shear correction factor 2 depends on Poissons ratio v and it may vary from 2 = 0.76 for v = 0.3 to 2 = 0.91 for v = 0.5. On the other hand, by comparing the constitutive
Mindlin shear force with the one proposed by Reissner (1945), who assumed a parabolic shear stress distribution at the outset of his plate theory formulation, the implicit shear correction factor of Reissner (1945) becomes 2 = 5 / 6 . This value of shear correction factor has been commonly used for the analyses of Mindlin plates, for example, Raju and Rao (1989), Liew et al. (2004) and Hou et al. (2005). Therefore, the shear correction factor 2 = 5 / 6 will be used in this study.
2.1.2
Displacement Components
Based on the Mindlin assumptions, the displacement fields are given by,
u ( x , y , z ) = z x ( x , y )
(2.1a)
(2.1b) (2.1c)
v ( x , y , z ) = z y ( x , y )
w( x, y, z ) = w( x, y )
in which u, v are in-plane displacements in the x and y directions, respectively, w
the transverse displacement in the z direction and x , y the rotations about the y- and x- axes, respectively as shown in Fig. 2.1. Note that by setting x = w / x and
y = w / y , one recovers the displacement fields of the classical thin plate theory.
15
x
z w x
Deformed shape
h x
Undeformed shape
Fig. 2.1 Deformation of cross-section of Mindlin plates
2.1.3
Strain-Displacement Relations
The normal strains x , y and shear strains xy , xz , yz are given by (Mindlin, 1951).
u x v y u v + y x
u w + z x
x = y = xy =
xz =
(2.2a)
(2.2b)
(2.2c)
(2.2d)
16
yz =
v w + z y
(2.2e)
In view of Eqs. (2.1a-c), the strain-displacement relations can be expressed as

x x
x = z
(2.3a)
y = z
y y
(2.3b)
y xy = z x + y x
(2.3c)
xz = x + yz = y +
w x w y
(2.3d)
(2.3e)
2.2
Stress-Strain Relations in Plastic Range
While the strains are linearly related to the stresses by Hookes law in the elastic range, the relations between stress-strain are nonlinear in the plastic range. It has been pointed out in Chapter 1 that many plasticity theories have been proposed in order to capture the nonlinear stress-strain relations. In this study, the two commonly used plasticity theories (IT and DT) are considered. In the sequel, the detailed derivation of stress-strain relations for commonly used plasticity theories of IT and DT are presented.
17
2.2.1
Derivation of Stress-Strain Relations Based on Prandtl-Reuss Equation
Prandtl (1925) and Reuss (1930) assumed that the plastic strain increment is, at any instant of loading, proportional to the instantaneous stress deviation (Mendelson, 1968); i.e.
d ijp = s ij d
i, j = 1,2,3
(2.3)
where d ijp is the plastic strain increment, d is a positive scalar which may vary throughout the loading history and s ij is the stress deviator tensor which is given by
s ij = ij m ij
(2.4)
where ij is the Kronecker delta and m denotes the hydrostatic stress which is equal to ( x + y + z ) / 3 . Equation (2.3) states that the increments of plastic strain depend on the current values of the deviatoric stress state, not on the stress increment required to reach this state. According to the assumption that the material remains isotropic throughout the deformation, the state of hardening at any stage is therefore specified by the current uniaxial yield stress denoted by . The quantity is known as the equivalent stress or effective stress, which increases with increasing plastic strain, and is given by von Mises yield criterion as follows
3 (sij sij ) 2
(2.5)
or 1 2
{(
2 2 2 y ) + ( y z ) + ( z x ) + 6 xy + yz + zx 2 2 2
)}
1/ 2
(2.6)
18
The corresponding equivalent or effective plastic strain increment d

2 d ijp d ijp 3
is given by
(2.7)
or
2 d xp d yp 3
p +6 d xy
{(
) + (d
2
p y
d zp
) + (d
2
p z
d xp
p + 6 d yz
p + 6 d zx
)}
2 1/ 2
(2.8)
In view of Eqs. (2.3) and (2.5), the Eq. (2.7) can be written for a work-hardening Prandtl-Reuss material as
2 = d 3
(2.9)
By using the value of d from Eq. (2.9), the Prandtl-Reuss flow rule, i.e. Eq. (2.3), may be written as (Chakrabarty, 2000)
3d p 3d 3d sij = sij = 2 2 H 2
d ijp =
1 1 sij E t E
(2.10)
since
1 d p d d e d d e 1 1 = = = = H d d d d Et E
(2.11)
where Et is the tangent modulus that is equal to d / d , d is the total effective strain increment which is equal to d e + d p and d e is the effective elastic strain increment. It can be seen that Eq. (2.10) relates the increments of plastic strain to the stress and thus such theory is called the incremental theory of plasticity. According to the generalized Hookes Law, the elastic strain increment is given by:
19
1 2 1 + e d ij = ij d kk dsij + 3E E
(2.12)
Therefore, the rate form of complete Prandtl-Reuss equation relating the stress rate to the strain rate is given by adding Eqs. (2.10) and (2.12)
& 3 E 1 2 & 1 sij & & E ij = (1 + ) sij + kk ij + E 2 t 3
(2.13)
during the continued loading of a plastically stressed element. Let x = p, y = q and xy = s at the point of bifurcation. According to the usual assumption that the plate is in a state of plane stress i.e. z = yz = zx = 0 , the effective stress in Eq. (2.3) can be expressed as
2 2 2 = x2 x y + y + 3 xy
(2.14)
A straightforward differentiation of Eq. (2.14) gives
(2 p q)d x + (2q p)d y + 6sd xy 2 2
(2.15)
on setting x = p , y = q , xy = s at the point of bifurcation. Using the expression for d / from Eq. (2.15) and the fact that 2 = p 2 pq + q 2 + 3s 2 at the onset of buckling, the constitutive Eq. (2.13) therefore furnishes a system of homogeneous equations which can be expressed in the following matrix form:
20
c11 c12 d x c 22 d y Et d xy = d xz d yz sym where
c13 c 23 c33
0 0 0 c 44
d x d y d xy 0 d xz c55 d yz 2 0 0 0
(2.16)
s2 Et q 2 c11 = 1 31 + E 4 2 2
(2.17a)
E 1 s 2 E pq c12 = 1 (1 2 ) t 31 t + 2 2 E E 2 2 c13 = 3 E t 2 p q s 1 2 E
(2.17b)
(2.17c)
s2 E p 2 c 22 = 1 31 t + 2 E 4 2
(2.17d)
c 23 =
3 Et 2q p s 1 2 E
(2.17e)
Et Et s 2 c33 = 2(1 + ) + 91 2 E E
c 44 = c55 =
(2.17f)
2(1 + ) Et E
(2.17g)
and 2 is the Mindlin shear correction factor. By using the process of matrix inversion, we finally obtain the following constitutive equations:
21
& x &y & xy = E & xz & yz sym

where
0 0 0 2G E
& x & y & xy 0 & xz 2 G & yz E 0 0 0
(2.18)
= = = = = =
(c
1 1
22 33
2 c c 23 )
(2.19a)
(c13 c 23 c12 c33 )

11 33 2 c c13 )
(2.19b)
(c
1
(2.19c)
(c12 c23 c13 c 22 ) (c12 c13 c11c23 )

11 22 2 c c12 )
(2.19d)
(2.19e)
(c
(2.19f)
G=
Et E = c 44 2(1 + )
(2.19g)
and
c11 sym c12 c 22 c13 c 23 c33
E = Et
(2.20)
22
2.2.2
Derivation of Stress-Strain Relations Based on Henckys Deformation Theory
Unlike the Prandtl-Reuss constitutive relations, Hencky (1924) proposed the total stress-strain relation which relates the total strain component to the current stress. Therefore, instead of Eq. (2.10), one has 3 p sij 2
ijp =
(2.21)
Since the strain rate vector in this case is not along the normal to the Mises yield surface in the stress space, the yield surface must be supposed to have locally changed in shape so that the normality rule still holds. The possibility of the formation of a corner on the yield surface may also be included. The incremental form of the Hencky equation Eq. (2.21) is easily found as:
3d 2 d p p 3 p sij + dsij d 2
d ijp =
(2.22)
where p / can be written as
p e 1 1 = = Es E
(2.23)
By combining Eq. (2.12) and Eq. (2.22), and using Eqs. (2.11) and (2.23), one obtains the following rate form of the complete stress-strain relation for DT:
3E 1 2 & E ij = 2E 2 s & E 1 2 3 E s ij sij + & & ij kk + 3 2 Et E s
(2.24)
23
In view of Eq. (2.15), the constitutive equation (Eq. 2.24) furnishes c11 c12 d x c 22 d y Et d xy = d xz d yz sym where
E q 2 s2 c11 = 1 31 t E 4 2 + 2 s
c13 c 23 c33
0 0 0 c 44
d x d y xy 0 xz c55 yz 2 0 0 0
(2.25)
(2.26a)
Et Et pq 1 s 2 c12 = 1 (1 2 ) 31 E 2 2 + 2 2 E s
c13 = E 2p q s 3 1 t E 2 s
(2.26b)
(2.26c)
E p 2 s2 c 22 = 1 31 t + 2 E 4 2 s c 23 = E 2q p s 3 1 t E 2 s Et E E s2 (1 2 ) t + 91 t 2 E Es E s
(2.26d)
(2.26e)
c33 = 3
(2.26f)
E E c 44 = c55 = 3 t (1 2 ) t E E s
(2.26g)
It can be seen that Eq. (2.25) has the same form as Eq. (2.16). Therefore, by the process of matrix inversion, the stress-strain relation for DT can be simply expressed as in the matrix form as that given in Eq. (2.18). The corresponding parameters
24
, , , , , , G and for DT can be obtained by substituting c11 ,....., c 44 of Eq.

(2.26a-g) into Eqs. (2.19) and (2.20).
2.2.3
Material Modeling
The strain hardening plastic behavior of most structural materials used in aerospace and off-shore applications can be adequately described by the strain-stress relationship proposed by Ramberg and Osgood (1943) as follow
= +k 0 E E 0
(2.27)
where 0 is a nominal yield stress depending on the value of k, c is a dimensionless constant that describes the shape of the stress-strain relationship with c = for elastic-perfectly plastic response and k is the horizontal distance between the knee of
c = curve and the intersection of the c curve with the / 0 = 1 line as shown in
Fig. 2.2. By differentiating both sides of Eq. (2.27), and noting that the tangent modulus
Et = d / d , one obtains
c 1
E = 1 + kc Et 0
c >1
(2.28a)
while after some manipulations and by noting that E s = / , one obtains

c 1
E = 1 + k Es 0
c >1
(2.28b)
25
1.5
c=
c=5
c = 10
c = 20
c=
= +k 0 E E 0
c
0.5
1+k
0 0 1 2 3 4
0
Fig. 2.2 Ramberg-Osgood stress-strain relation
Based on the materials in the experiments of Aitchison and Miller (1942), Ramberg and Osgood assumed the expressions of k and c as: 3 7
ln 17 7
k=
(2.29a)
c = 1+
ln 0.7 0.85
(2.29b)
0 = 0.7
(2.29c)
26
where 0.7 and 0.85 are the secant yield stresses corresponding to the intersections of the stress-strain curve and the lines of slope 0.7E and 0.85E respectively from the origin as shown in Fig. 2.3.
0.85E 0.7E
0.7 0.85
Fig. 2.3 Illustration of 0.7 and 0.85 on the stress-strain curve
With the above assumption of k = 3/7, only three parameters ( E , 0.7 , c ) are needed to specify the Ramberg-Osgood stress-strain model. These parameters are given by Bruhn (1965) for a wide range of materials and some of them are shown in Table 2.1. The modified version of Eq. (2.27) was proposed by Hill (1944) by using a 2% offset yield stress, 0.2 , for 0 and is given by
c
= + 0.002 E 0.2
(2.30)
c=
0.301 log 0.2 0.1
(2.31)
27
in which 0.1 is the 1% offset stress as shown in Fig. 2.4. Table 2.1. Values of 0.7 , 0.85 and c for various aluminum alloys under room temperature (Bruhn, 1965) Material Stainless Steel AISI 301 Hard Sheet Transverse Compression Longitudinal Compression AISI 301 Hard Sheet Transverse Compression Longitudinal Compression Low Carbon & Alloy Steels AISI 4130, 4140, 4340 Heat Treated Aluminum Alloys 2014-T6 Forgings 29.0 125 113 111 102 10.9 27.0 26.0 150 150 118 58 116.5 48 105 37 9.2 4.4 27.0 26.0 125 125 80 43 73 28.2 63 23 6.9 5.2
E 106 psi
tu
ksi
0.2
ksi
0.7
ksi
0.85
ksi
h 4 in
2024-T3 Sheet & Plate Heat treated, h 0.25 in 2024-T4 Sheet & Plate Heat treated, h 0.5 in 6061-T6, Heated & aged
10.7 10.7 10.7 10.1 10.5
62 65 65 42 76
52 40 38 35 67
52.3 39 36.7 35 70
50 36 34.5 34 63
20 11.5 15.6 31 9.2
h < 0.25
7075-T6, Bare Sheet & Plate
h 0.5
Magnesium Alloys AZ61A Extrusions
h 0.249
6.3
38
14
12.9
12.3
19
28
0.2 0.1
0 0.001 0.002 Fig. 2.4. Offset yield stress
In the literature, both Ramberg-Osgood models, i.e. Eq. (2.27) and Eq. (2.30), have been used. It can be seen that Eq. (2.27) is the more general form of Ramberg-Osgood model since Eq. (2.30) can be deduced from it by equating
0 = 0.2 and k = 0.002
0.2
. Therefore, the more general Ramberg-Osgood formula
Eq. (2.27) will be used herein.
2.3
Derivation of Energy Functional and Governing Equations
The strain energy functional of the Mindlin plate is given by (Chakrabarty, 2000)
1 {& x & x + & y & y + &xy & xy + &xz & xz + & yz & yz }dV 2V
U=
(2.32)
29
In view of Eqs. (2.2a-e) and the stress-strain relations given by either of Eq. (2.18) or Eq. (2.25) and after the integration over the plate thickness, the strain energy
functional can be expressed as 1 Eh 3 x Eh 3 y U= + 2 A 12 x 12 y

2
Eh 3 x + 6 x
y y
Eh 3 x y + 12 y + x
Eh 3 y y
y Eh 3 x x x y + + + x 6 x y 6 x y x x y y
2 2 w w dA + Gh x + + y + x y 2
(2.33)
where A is the plate area and dA = dxdy The potential energy V for the plate subjected to uniform in-plane compressive stresses is given by
2 2 w 1 w w w V = x h + y h + 2 xy h dA y 2 A x x y
(2.34)
The total potential energy functional can be expressed as

= U +V
(2.35)
which can be expressed in the form:

y y y w w dxdy = F x, y, w, , , x , x , , y , , x y x y x y A
(2.36)
30
Using calculus of variations, the Euler-Lagrange differential equations associated with the minimization of the total potential energy functional with respect to arbitrary variation of w, x and y are respectively given by (Forray, 1968): F y w, y =0 =0 =0
F F w x w, x
(2.37a)
F F x x x , x F F y x y , x
F y x, y F y y, y
(2.37b)
(2.37c)
where () , x and () , y refer to the differentiation with respect to x and y respectively. In view of Eqs. (2.33), (2.34) and (2.35), the governing equations are given by
x y w w + 2 w x h yh + x y x y y x
2 Gh
w w xy h xy h =0 y x y x
(2.38a)
Eh 3 x Eh 3 y Eh 3 x y w 2 Gh x + + + + x 12 x x x 12 y 12 y + Eh 3 y 12 x y Eh 3 y Eh 3 x y + x + 12 y + 12 x = 0 x y y + x w 2 Gh y + y (2.38c) (2.38b)
Eh 3 y Eh 3 x Eh 3 + + y 12 y 12 x 12 Eh 3 + x 12
x y Eh 3 y Eh 3 x y + x + 12 y + 12 x = 0
31
in which the parameters , , , , , are given by Eqs. (2.19a-g). By using the boundary conditions at the plate edges, the aforementioned governing equations can be solved for plastic buckling stress.
32
CHAPTER 3
RITZ METHOD FOR PLASTIC BUCKLING ANALYSIS
This chapter presents the formulations of the Ritz method for plastic buckling of Mindlin plates under in-plane compressive loadings. In this version of Ritz method, it is proposed that the displacement functions be approximated by the product of mathematically complete, two-dimensional polynomial functions and basic functions comprising boundary equations that are raised to appropriate powers so as to ensure the satisfaction of the geometric boundary conditions at the outset. By minimizing the total potential energy functional with respect to the parameterized displacement functions, one obtains a system of homogeneous equations which forms the governing eigenvalue equation. The Ritz formulations will be presented for three different types of coordinate systems, namely the Cartesian, the skew and the polar coordinate systems for use in handling plates of various shapes.
33
Ritz Method for Plastic Buckling Analysis
3.1
Introduction
In 1909, Walter Ritz published a paper that demonstrates his famous method for minimizing a functional, determining frequencies and mode shapes. Since then, the Ritz method has been widely used because of its simplicity in implementation. Two years after Ritzs paper (1909), Rayleigh (1911) published a book where he complained that Ritz had not recognized his similar work (Rayleigh, 1877). Therefore it is sometime referred to as the Rayleigh-Ritz method. However, Leissa (2005) investigated carefully the historical works of Rayleigh and Ritz and came to conclusion that Rayleighs name should not be attached to the Ritz method. In the Ritz method, the displacement function, ( x, y ) is approximated by a finite
linear combination of the trial functions in the form

m
( x, y ) cii ( x, y )
i =1
(3.1)
in which i ( x, y ) are the approximate functions which individually satisfies at least the geometric boundary conditions to ensure convergence to the correct solutions. The static boundary conditions need not be satisfied by these approximate functions. By minimizing the energy functional with respect to each of the unknown coefficients
ci, a set of homogeneous equations is obtained as follows
= 0; ci
i = 1,2,......, m
(3.2)
For buckling and vibration problems, the above set of homogeneous equations is reduced to eigenvalue and eigenvector problems.
34
The exact solution is obtained if infinite terms are adopted in Eq. (3.1). However, it is impractical to use infinite number of terms and so the number of terms is usually truncated to N terms in applications. The choice of the approximate functions is very important in order to simplify the calculations and to guarantee convergence to the exact solution. Some of the commonly used trial functions in the Ritz method for plate analysis are the products of eigenfunctions of vibrating beams (Leissa, 1973; Dickinson and Li, 1982; Narita and Leissa, 1990), the generated beam functions (Bassily and Dickinson, 1975, 1978), the spline functions (Mizusawa, 1986), the beam characteristic orthogonal polynomials (Bhat, 1985) and two dimensional polynomial functions with appropriate basic functions (Bhat, 1987; Laura et al., 1989; Liew, 1990; Liew and Wang, 1992, 1993; Geannakakes, 1995). Among them, the latter trial functions can be used for the analysis of plates of general shape and boundary conditions while the others may not be so convenient. Moreover, the computational accuracy may also be increased since the polynomial functions admit exact calculations of differentiation and integration of the functions (Liew et al., 1998). Therefore, in this study, mathematically complete, two-dimensional polynomial functions are adopted together with basic functions comprising boundary equations that are raised to appropriate powers in order to ensure the satisfaction of the geometric boundary conditions.
3.2
Boundary Conditions
The following boundary conditions are considered for the plastic buckling problems of Mindlin plates; (1) Simply Supported Edge (S and S*)
35
There are two kinds of simply supported edges for Mindlin plates. The first kind (S), which is commonly referred to as the hard type simply supported edge, requires M nn = 0, s = 0, w = 0 (3.3)
The second kind (S*), commonly referred to as the soft type simply supported edge, requires M nn = 0, M ns = 0, w = 0 (3.4)
where M nn is the bending moment normal to the edge, M ns the twisting moment at the edge as shown in Fig. 3.1, and s the rotation of the mid-plane normal in the tangent plane sz to the plate edge. (2) Clamped Edge (C) The boundary conditions for a clamped edge are
n = 0, s = 0, w = 0
in which n is the rotation of the mid-plane to the edge. (3) Free Edge (F) For a free edge, the boundary conditions are
(3.5)
M nn = 0, M ns = 0, Qn N nn
w =0 n
(3.6)
where Qn is the transverse shear force in the normal direction n to the plate edge and N nn is the inplane compressive force in the normal direction of the edge.
36
Fig. 3.1 Forces and moments on a plate element
3.3
Ritz Formulation in Cartesian Coordinate System
Consider an isotropic, rectangular plate of uniform thickness h, length a, width b and subjected to in- plane uniform compressive stresses of magnitudes x , y and uniform shear stress xy as shown in Fig. 3.2. The edges of the plate may take on any combination of simply supported, clamped or free edges. For generality and convenience, the coordinates are normalized with respect to the maximum length dimensions (a and b as shown in Fig. 3.2) and the following nondimensional terms are introduced:
2x 2y h 2w a ; = ; = ; w = ; = ; A = dd a b b b b
(3.7)
37
xy hb 2 y hb 2 x hb 2 Eh 3 x = 2 ; y = 2 ; xy = 2 ; D = D 12(1 2 ) D D
(3.8)
y,
y xy
x, x xy a y b x
Fig. 3.2 Rectangular plate under in-plane compressive stresses
Then, in view of Eqs (3.2) and (3.3), the strain energy and potential energy functionals in Eqs. (2.33) and (2.34) become respectively,
Eh 3 2 Eh 3 y 2 Eh 3 y x x + 2 + 6 12 12 x Eh 3 x y Eh 3 y y + + + 6 + 12
+
1 U = 2 A
Eh 3 x 6
x y +
2 2 w 1 w + y + dd x +
2 Ghb 2 2
4
(3.9)
2 w 2 w w w 1 2 Eh 3 + y + 2 xy V = x dd 8 12(1 2 ) A
(3.10)
38
Therefore, the total potential energy functional is given by

= U +V
(3.11)
The adopted admissible p-Ritz functions for the deflection and rotations of the plate are given by (Liew et al. 1998)
p q
w w ( , ) = c m m ( , ) q =0 i =0
(3.12a)
x x ( , ) = d m m ( , ) q =0 i =0
(3.12b)
y y ( , ) = em m ( , ) q =0 i =0
(3.12c)
where p is the degree of the complete polynomial space, ci, di, ei the unknown coefficients to be varied with the subscript m given by
m = ( q + 1)( q + 2) / 2 i
and the total numbers of c m , d m , em are given by ( p + 1)( p + 2) 2
(3.13)
N=
(3.14)
w x y The functions m , m , m are given as follow:
w m = bw i q i
x m = bx i q i y m = by i q i
(3.15a) (3.15b) (3.15c)
39
In view of Eqs. (3.15a-c), the Eqs. (3.12a-c) give the mathematically complete two dimensional polynomials which follow the Pascal-type of expansion, for example: For p = 1, w ( , ) = bw (c0 + c1 + c3 )
(3.16a)
For p = 2
w ( , ) = bw c1 + c 2 + c3 + c 4 2 + c5 + c6 2
(3.16b)
The functions bw , bx , and by are the basic functions which are the product the boundary equations raised to appropriate power so the admissible Ritz functions satisfy the geometric boundary conditions as follow.
bw = [ j ( , )]
ne j =1
w j
(3.17a)
bx = [ j ( , )]
ne j =1
x j
(3.17b)
by = [ j ( , )]
ne j =1
y j
(3.17c)
where ne is the number of edges, j the boundary equation of the jth edge and
w , xj , y are given by j j
0 free( F ) w = j 1 simply supported ( S and S ) or clamped (C ) 0 free ( F ) or simply supported ( S *) or ( S ) in the y - direction xj = 1 simply supported ( S ) in the x - direction or clamped (C ) 0 free ( F ) or simply supported ( S *) or ( S ) in the x - direction y = j 1 simply supported ( S ) in the y - direction or clamped (C )
(3.18a)
(3.18b)
(3.18c)
On the basis of foregoing energy functional and p-Ritz functions, the application of the Ritz method yields:
40
, , = 0,0,0 cm d m em
(3.19)
where m = 1,2,.,N. The substitution of Eqs. (3.11) and (3.12a-c) into Eq. (3.19) yields the following homogeneous set of equations
c [K ]d = {0} e
(3.20)
where vectors c, d and e are given by

c1 c c = 2; : c m d1 d d = 2; : d m e1 e e= 2 : em
(3.21a-c)
The matrix [K] has a structure of
K cc K cd K dd symmetric
K ce K de K ee
(3.22)
and its elements are given by

w w 2 G 1 w jw iw j x 1 iw j i = 2 + 48(1 2 ) A 4 E w iw jw iw jw xy iw j dd + 48(1 2 ) 48(1 2 )
cc ij
(3.23a)
41
cd K ij =
2 G iw x j dd 4 2 E A 2 G iw y j dd 4 2 E A
(3.23b)
ce K ij =
(3.23c)
dd ij
1 x jx x jx x jx x jx i i dd = + + i + i 12 12 12 A
2 G x x + i j dd 4 E 2
K
de ij
(3.23d)
ix jy ix jy = + 12 12 A
ix iy ix iy + + dd 12 12
y y y y jy iy j iy j iy j ee i + + + K ij = 12 12 12 A
(3.23e)
2 G y y i j dd 4 E 2
(3.23f)
where i, j = 1, 2,.N and , , , , , , G are parameters depending on the material properties and stress level as shown in Chapter 2. The plastic buckling stress parameter can be obtained by setting the determinant of the [K ] matrix in Eq. (3.20) to zero and solving the characteristic equation for the roots (eigenvalues). The lowest eigenvalue corresponds to the critical plastic bifurcation buckling stress.
42
3.4
Ritz Formulation in Skew Coordinate System
For the analysis of skew or parallelogram plates, it is more expedient to adopt the skew coordinate system. For example, the hard type simply supported boundary condition of oblique edge can be readily implemented in a skew coordinate system while the Ritz formulation with the Cartesian coordinate system has to be modified in order to implement that this kind of boundary condition. Consider a skew plate of uniform thickness h with length a, oblique width b and skew angle as shown in Fig. 3.3a. The expression of energy functional in skew coordinate system can be readily obtained by transforming of Cartesian coordinate system into skew coordinate system. It can be seen from Fig. 3.3a that these coordinates are related by (Liew et al. 1998)
x = x y tan y = y sec
(3.24a) (3.24b)
The rotations x , y can be expressed in the skew coordinate system as
x ( x, y ) = x ( x , y ) cos y ( x, y ) = x ( x , y ) sin + y ( x , y )
(3.25a) (3.25b)
The relationship between the partial derivatives of rectangular and skew coordinates are given by
() () = x x () () () = tan + sec y x y
(3.26a)
(3.26b)
43
For generality and convenience, the following nondimensional terms are adopted:
h a 2x 2y 2w ; = ; = ; w = ; = ; A = dd a b b b b
(3.27)
x
y
x
y ,
y ,y
b/2 b/2 x, x ,
a/2
a/2
(a)
xy x y
(b) Fig. 3.3. (a) Skew coordinates and plate dimensions (b) Directions of stresses in the infinitesimal element
In view of Eqs. (3.24) to (3.27), the strain energy functional in Eq. (2.33) become
2
U =
A
1 2
3 Eh 3 x cos 12
Eh 3 + cos 6
2 x sin
sin x
sin
x x y Eh 3 + x + sec 12
y 2 x sin sin
44
y y x Eh 3 x sin x + 2 sin + cos 12
Eh 3
6
y x y x y + + sin
+ sin
y 2 sin 2 x 3 sin x 3 sin y + 2
y x Eh 3 x cos x 2 sin 6
2
+ cos
2 Ghb 2
4
2
w w + y sin x tan cos x +
w + sec
dd
(3.28)
The potential energy functional, V, of the in-plane uniform loads is given by

2 2 w 2 w w w 1 2 Eh 3 V = x cos + y sec + 2 sin 2 8 12(1 ) A 2 w w w sin + 2 xy dd
w sin
2
(3.29)
The total potential energy functional, , in the skew coordinate system is obtained by summing the strain energy functional and the potential energy functional. It should be noted that the plastic buckling stress parameters, x , y , xy used in Eq. (3.29) are based on the stress resultants on an infinitesimal rectangular element as shown in Fig. 3.3b.
45
The admissible Ritz displacement functions for the skew coordinate system have the same form as those for the rectangular coordinate system (Eqs. 3.12 to 3.18) except for the notation of the coordinate system. However, for clarity and convenience, these admissible functions for skew coordinate system are repeated as follows:
p q
w w ( , ) = c m m ( , ) q =0 i =0
(3.30a)
x x ( , ) = d m m ( , ) q =0 i =0
(3.30b)
y y ( , ) = em m ( , ) q =0 i =0
(3.30c)
w x y where the functions m , m , m are given by:
w m = bw i q i x m = bx i q i
(3.31a) (3.31b) (3.31c)
y m = by i q i
The functions bw , bx , and by are the basic functions which are the product the boundary equations raised to appropriate power so the admissible Ritz functions satisfy the geometric boundary conditions as follows:
bw = [ j ( , )]
ne j =1
w j
(3.32a)
bx = [ j ( , )]
ne j =1
x j
(3.32b)
by = [ j ( , )]
ne j =1
y j
(3.32c)
46
w , xj , y are given by j j
0 free( F ); w = j 1 simply supported ( S and S ) or clamped (C ) 0 free ( F ) or simply supported ( S *) or ( S ) in the y - direction; xj = 1 simply supported ( S ) in the x - direction or clamped (C ) 0 free ( F ) or simply supported ( S *) or ( S ) in the x - direction; y = j 1 simply supported ( S ) in the y - direction or clamped (C )
(3.33a)
(3.33b)
(3.33c)
By substituting the displacement functions given by Eqs. (3.30a-c) into Eq. (3.19), the resulting eigenvalue equation has the same form as Eq. (3.20) and the elements of the matrix [K ] are given by
w 1 w jw iw jw iw jw iw j i sec + sin + A w w iw jw iw jw i j sin +
cc ij
2G = 4 E 2
w 2 x 2 cos iw j y sec 48 (1 2 ) 48(1 2 )
w xy 2 iw jw iw jw sin 2 iw j + + 2 48(1 ) w 2 sin iw j dd
(3.34a)
cd K ij =
2G 4 E 2 2G 4 E 2
iw x w j sin i jx dd A iw y w j sin i jy dd A
(3.34b)
ce K ij =
(3.34c)
47
dd ij
x x sin 2 ix jx ix j 1 ix j 3 = cos + tan sin sin 12 A 12 x x x 2 ix j ix j ix j sin + + sin cos sin 12 x x x x 4 ix j ix j ix j ix j 2 + + cos sin 12
ix jx ix jx x jx ix jx + cos 2 i sin + + 12
x x x j 3 sin 4 sin ix j + sin i + 2 6 x 2 sin 2 ix j 2 G cos ix jx dd + 4 E 2
ix jx ix jx +
(3.34d)
y x x jy ix jy jy sin 2 ix j de i + sin i K ij = tan + A 12
ix jy 2 sin ix jy ix jy sin ix jy + cos + cos 12 12
y y y x j x j 3 sin 2 ix j + i + cos 2 i 12 12
ix jy ix jy 2 G 2 sin sin cos ix jy dd + 4 E 2
(3.34e)
y y y sin 2 iy jy y j y j y j ee sin i sin i + i K ij = sec A 12 y y y y cos iy j 2 sin iy j iy j iy j + + + + dd 12 12
48
2G cos iy jy dd 2 4 E
(3.34f)
where i, j = 1, 2,.N.
3.5
Ritz Formulation in Polar Coordinate System
The energy functionals in polar coordinates can be obtained from the corresponding Cartesian coordinates equations by using the relations x = r cos and y = r sin . However, the transformation procedure involves some lengthy mathematical manipulations. Therefore, the procedure would be simpler if the energy functional is derived directly from the strain-displacement relations in the polar coordinates. Based on the assumptions of the Mindlin plate theory, the displacement components in the polar coordinates can be expressed as
u r ( r , , z ) = z r ( r , )
(3.35a) (3.35b) (3.35c)
u (r , , z ) = z (r , )
w(r , , z ) = w(r , )
where u r , u are in-plane displacements in the r and directions, respectively and
r , are rotation about and r axes respectively. The corresponding straindisplacement relations are given by
r r
rr = z
(3.36a)
z + r r
(3.36b)
49
r = z
1 r + r r r w r 1 w r
(3.36c)
rz = r + z = +
(3.36d)
(3.36e)
In view of Eq. (3.36a-e), the strain energy functional and potential energy functional can be expressed in polar coordinates as:
Eh 3 1 Eh 3 r Eh 3 r U = r + + r + + 2 12r 2 A 12 r 6r r
2 2
Eh 3
r Eh 3 + r 6 12r 2
2 r
1 r
1 r + r
1 1 + r 2 r + r r +
r +
Eh 3 r 1 r + 6 r
2 2 1 w w + 2 Gh r + + + rdrd r r
(3.37)
V =
2 2 1 1 w 2 w w w rr h + h 2 + r h rdrd r r 2 r r A
(3.38)
where rr , are the inplane stresses in the radial and tangential directions, respectively and r is the shear stress. The polar coordinates (r, ) can be transformed into a set of nondimensional coordinates (, ) by the following relations: R [(1 + ) + (1 ) ] 2
r=
(3.39a)
50
1 2
(3.39b)
where the parameters (, ) are defined in order to get the limits of integration from -1 to 1 as shown in Fig. 3.4. Note that, in order to avoid singularity problems in the case of a circular plate, needs to be set to a very small number (e.g. 10-7) and the inner edge is made free.
R R
(a) Circular plate = 0, = 2
(b) Annular plate 0, = 2
(c) Sectorial plate = 0, 0
(d) Annular sectorial plate 0, 0
Fig. 3.4 Definitions of and
51
For generality and convenience, the following nondimensional terms are adopted r h w ; = ; w = ; A = dd R R R
r=
(3.40a)
r =
hR 2 hR 2 rr hR 2 ; = 2 ; r = r 2 2D D D
(3.40b)
In view of aforementioned nondimensional terms, the strain energy functional can be expressed as
(1 ) Eh 3 U= 12 8 A
2
2 r (1 )
2
r 2 2 Eh 3 + r + 6 (1 )r
2 + + 2 r 12r
Eh 3
+ 12r 2
2r 2 r (1 )
Eh 3 2 2 2 r (1 )r + r 2 6
2 2 2 2 r + r (1 )r
2 r + 2 r r 2
Eh 3
6
2 r (1 )
2 r 2 r + (1 )
(3.41)
2 2 2 w 2 w + + r dd + GhR r + r (1 ) 2 2
The potential energy functional, V, becomes

(1 ) 1 w 2 Eh 3 w V = r + r 2 24(1 ) A (1 )
2 2
+ 2 r
1 w w r dd r
(3.42)
52
The Ritz displacement functions can be written as follows:

p q
w w ( , ) = c m m ( , ) q =0 i =0
(3.43a)
r r ( , ) = d m m ( , ) q =0 i =0
(3.43b)
( , ) = em m ( , )
q =0 i =0
(3.43c)
where p is the degree of the complete polynomial space, cm, dm, em the unknown coefficients and the subscript m is given by Eq. (3.8).
w r The functions m , m , m are given as follows:
w m = bw i q i r m = br i q i
(3.44a) (3.44b) (3.44c)
m = b i q i
in which the basic functions bw , br , and b are the product of boundary equations with
appropriate power as follows:
bw = [ j ( , )]
ne j =1
w j
(3.45a)
br = [ j ( , )]
ne j =1 ne
r j
(3.45b)
b = [ j ( , )]
j =1
(3.45c)
53
w , rj , j are given by j
0 free( F ); w = j 1 simply supported ( S and S ) or clamped (C ) 0 free ( F ) or simply supported ( S *) or ( S ); rj = 1 clamped (C ) 0 free ( F ) or simply supported ( S *); j = 1 simply supported ( S ) or clamped (C ) The elements of matrix [K ] are given by
w w 2 G r iw j (1 ) iw j cc K ij = 2 + r E (1 ) A
(3.46a)
(3.46b)
(3.46c)
r r iw jw (1 ) iw jw 2 + 2 r 12(1 ) (1 )
iw jw iw jw dd + r +
(3.47a)
cd ij
2 G r iw r = j dd E 2 2 A
2 G (1 ) iw j dd 2 E 2
(3.47b)
ce K ij = A
(3.47c)
r r ir j (1 ) r r 2 Gr r r dd + i i K ij = i j + 4 12r E 2 12(1 ) A r r r j ir r (1 ) ir j + + j + i 24 12r
54
r jr ir jr ir jr (1 ) i r r + dd + + j + i 12 24r
(3.47d)
ir (1 ) r ir G (1 ) ir j j j de K ij = j + + i 12 24r 24 Er 12 A
(1 ) ir j r j (1 ) r i j + + i 12 r 2 4r 2r ir j ir j dd 24 (1 )
(3.47e)
ee ij
(1 ) i 2 Gr (1 ) G r i j j = + i j + 2 12 E (1 ) 4 E A 12r
(1 ) 1 i j + + i i j j 4r 2
i j i j (1 ) i j dd j + i + 12 2r
(3.47f)
where i, j = 1, 2,.N . For the special cases of circular and annular plates, the deflection and rotations of the plate may be written as (Irie et al., 1982)
w ( , ) = w ( ) cos(n )
(3.48a) (3.48b) (3.48c)
r ( , ) = r ( ) cos(n )
( , ) = ( ) sin(n )
where n is the arbitrary integer number which represents the number of nodal diameter.
55
It should be noted that the trigonometric functions which appear in the energy integration can be calculated as follows:
1 1 when n = 0 S1 = cos 2 (n )d = 2 when n > 0 1
(3.49a)
0 S 2 = sin 2 (n )d = 1 1
1
when n = 0 when n > 0 for n = 0, 1, 2, ......
(3.49b)
cos(n ) sin(n )d = 0
(3.49c)
In views of Eqs. (3.48a-c) and (3.49a-c), the strain energy and potential energy functionals can be expressed as
(1 ) Eh 3 2 r U= 12 S1 (1 ) 8 A
2
2 Eh 3 + S1 r ( r + n ) 6 (1 )r
2
Eh 3
12r 2
S1 ( r + n )
2r S 2 + + n r 2 (1 ) 12r
2 2 2 w 1 + S 2 nw r d + GhR S1 r + r (1 ) 2 2 2 2 2 w (1 ) 2 Eh 3 1 + S 2 nw r dd V = S 2 1 r 8 12(1 ) A r (1 )
(3.50)
(3.51)
In this special case, the Ritz displacement functions may be assumed as:
N
w ( ) = ciiw
i =1 N
(3.52a)
r ( ) = d iir
i =1
(3.52b)
56
( ) = eii
i =1
(3.52c)
where N is the degree of the one-dimensional polynomial and ci , d i , ei are the unknown coefficients. The functions iw , ir , i are the product of the one-dimensional polynomial and basic functions as follows
iw = bw i ir = br i i = b i
where the functions bw , br , and b are the basic functions given by
(3.53a) (3.53b) (3.53c)
= [ j ( )]
w b ne j =1
w j
(3.54a)
br = [ j ( )]
ne j =1 ne
r j
(3.54b)
b = [ j ( )]
j =1
(3.54c)
w , rj , j are given by Eqs. (3.46a-c). j
The elements of matrix [K ] in Eq. (3.20) become

w 2 G r w j (1 ) 2 cc + K ij = 2 S1 i n S 2iw jw 4r E (1 ) A
57
w r r w j (1 ) 2 2 S1 i n S 2iw jw dd + 2 4r 12(1 ) (1 )
(3.55a)
cd K ij = A
2 G r iw r S1 j dd E 2 2
(3.55b)
ce ij
2 G (1 ) nS 2iw dd = j 2 4 E A
r r ir j (1 ) 2 Gr = + S1 S1ir jr + S1ir ir 2 12(1 ) 4 E 12r A
(3.55c)
dd ij
r jr ir r (1 ) 2 + + S1 i j + n S 2 ir jr dd 24 48r
(3.55d)
de ij
ir j G (1 ) ir ir j (1 ) j r nS1 nS1i nS 2 j + + = 48r 24 24 Er A 24
(1 ) ir j r j (1 ) r i j + + i 12 r 2 4r 2r ir j ir j dd 24 (1 )
(3.55e)
ee ij
(1 ) i 2 Gr (1 ) G r i j j + = i j + 2 12 E (1 ) 4 E A 12r
(1 ) 1 i j + i j + i j 4r 2
i j i j (1 ) i j dd j + i + 12 2r
(3.55f)
where i, j = 1, 2,.N .
58
3.6 Solution Method for Solving Eigenvalue Problem
For elastic buckling, the aforementioned eigenvalue problems can be expressed in the standard form of generalized eigenvalue problem and a standard eigenvalue routine, for e.g. EISPAC (Smith et al., 1974), can be used to solve the problem. However, for plastic buckling, the matrix [K ] depends nonlinearly on the stresses since the parameters , , , , , , G contains the buckling stress term. In order to solve this nonlinear eigenvalue problem, the Sturm sequence iteration method with bisection strategy was adopted. Normally, the Sturm sequence procedure is done by the Gaussian elimination which reduces the matrix in to an upper triangular matrix form (Gupta, 1978). The number of sign changes in the diagonal elements of the triangular matrix indicates the number of eigenvalues of the matrix that are smaller than the trial stress level. It was found that the built-in function Eigenvalues in MATHEMATICA (Wolfram, 1999) can be used instead of Gaussian elimination. The use of built-in function Eigenvalues is found to be faster than using the Gaussian elimination in the determination of number of eigenvalues that are smaller than the trial stress. Therefore, the following procedure was adopted for the calculations of plastic buckling stresses. A trial stress value of is first assumed and the eigenvalues of matrix [K (trial )] are calculated. Since the number of eigenvalues of [K ], which are less than trial is equal to the number of negative eigenvalues of [K ( trial )] , the value of trial which gives one negative eigenvalue is sought. Upon identifying the value of
trial that gives one negative eigenvalue, a simple bisection algorithm is used to solve
59
the characteristic equation of the matrix [K ] until the desired accuracy of the solution is achieved. In this way, the lowest eigenvalue of the matrix [K ] will not be missed.
3.7 Computer Programs
Three computer programs have been written in MATHEMATICA language for the foregoing Ritz formulations. One program, called BUCKRITZRE, is for determining the plastic buckling stresses of rectangular, triangular, elliptical and trapezoidal plates and it is based on the aforementioned rectangular coordinate formulations (Eqs. 3.23af). Another program, called BUCKRITZSKEW, is written based on the skew coordinate formulations (Eqs. 3.33a-f) and it is to compute the plastic buckling stress of skew plates. The last program, called BUCKRITZPOLAR, is used to solve the plastic buckling problems of circular and annular plates and it is based on the polar coordinate system. The program codes and detailed explanations of the codes are given in Appendix A.
60
CHAPTER 4
RECTANGULAR PLATES
4.1
Introduction
This chapter is concerned with the plastic buckling analysis of rectangular Mindlin plates under uniaxial compressive stress ( x ), biaxial compressive stress ( x , y ),
shear stress ( xy ) and combinations of these stresses as shown in Fig. 4.1. The boundary conditions of the plate edges may take on any combination of simply supported, clamped and free edges. The Ritz method associated the rectangular Cartesian coordinate system is employed for the plastic bifurcation buckling analyses of such plates. The Ritz program code is called BUCKRITZRE and it is given in Appendix A. The program code is designed to handle Mindlin plates subjected to any combinations of in-plane and shear stresses and arbitrary boundary conditions. The preparation of the input file for the program code is so simple that only the basic functions and desired degree of polynomial need to be defined, apart from the usual information on the material and geometric properties of the plate such as aspect
61
Rectangular Plates
ratio and thickness-to-width ratio. The sample basic functions for rectangular Mindlin plates with various boundary conditions are given in Table 4.1. For brevity and convenience, the boundary conditions of the plate will be represented by a four-letter designation. For example, a FSCS plate will have a free boundary condition at x = a / 2 , a simply supported boundary condition at y = b / 2 ,
a clamped boundary condition at x = b / 2 and a simply supported boundary condition at y = b / 2 as shown in Fig. 4.1.
y,
xy
S
x, x
F C S
x a y
Fig. 4.1 FSCS rectangular plate under in-plane compressive stresses
In computing the buckling solutions, a typical aerospace material of Al 7075 T6 is adopted. Its material properties are: Poissons ratio = 0.33, E = 10.5 10 6 psi,
0.7 = 70 ksi, 0.2 = 67 ksi and the Ramberg-Osgood parameters k = 3/7 and c = 9.2 .
In order to verify the results, comparison studies are carried out against existing plastic buckling solutions wherever available. The effect of thickness-to-width ratio, aspect ratio, loading ratio and material properties on the plastic buckling behaviour of rectangular plates are investigated.
62
Rectangular Plates
Table 4.1 Sample basic functions for rectangular plates with various boundary conditions Boundary conditions Basic functions
bw
SSSS
( + 1)1 ( + 1)1 ( 1)1 ( 1)1
bx by bw
( + 1) 0 ( + 1)1 ( 1) 0 ( 1)1 ( + 1)1 ( + 1) 0 ( 1)1 ( 1) 0 ( + 1)1 ( + 1)1 ( 1)1 ( 1)1

( + 1) 1 ( + 1) 1 ( 1) 1 ( 1) 1
CCCC
bx by bw
( + 1) 1 ( + 1) 1 ( 1) 1 ( 1) 1
( + 1) 1 ( + 1) 1 ( 1) 0 ( 1) 1
CSFS
bx by
( + 1) 1 ( + 1) 1 ( 1) 0 ( 1) 1 ( + 1) 1 ( + 1) 0 ( 1) 0 ( 1) 0
4.2
Uniaxial and Biaxial Compression
For generality, y is expressed in terms of x as y = x . Therefore, the loading ratio = 0 represents an applied uniaxial stress while = 1 indicate an applied uniform biaxial stress. For a given value of loading ratio , the plastic buckling stress parameter = c hb 2 /( 2 D) can be determined by solving the eigenvalue equation given in Eq. (3.20). Since xy = 0 , the parameters , , , , G, , used in stressstrain relations become simpler and can now be expressed in an explicit form given by
for the deformation theory of plasticity (DT)

E E 2 3E + (1 2 )2 (1 2 ) t 31 t 2x E Es E s
(4.1a)
63
Rectangular Plates
E 2 1 4 31 t x2 E s
2 Et Et x 1 2 2 2(1 2 ) 31 E E s 2 E 2 x 1 4 31 t E 2 s
(4.1b)
(4.1c)
(4.1d)
G = E
1 E 2 + 3 1 E s
(4.1e)
= =0
= x 1 + 2
for the incremental theory of plasticity (IT)
2 Et E 3(1 2 )1 t 2x E E
(4.1f)
(4.1g)
= (5 4 ) (1 2 ) 2
(4.2a)
2 1 E 4 31 t x2 E 2 E 1 E 2 2(1 2 ) t 31 t 2x E E
2 2 1 E x 4 31 t E 2
(4.2b)
(4.2c)
(4.2d)
G 1 = E 2(1 + )
(4.2e)
Noted that by setting E s = E , the expressions of , , , , G, , based on DT reduced to those corresponding to IT.
64
Rectangular Plates
Rectangular plates with two opposite sides simply supported while the other two sides take on any combination of boundary conditions are first considered. Such supported rectangular plates allow us to verify the convergence of the Ritz solutions because of the availability of analytical solutions (Wang et al. 2001a). It is worth noting that, for the special case of uniformly stressed simply supported rectangular plate (i.e. = 1 ), the analysis given in Wang (2001a) can be further simplified by noting that ( 2G / E ) = 0. After some simplifications, the exact expression of plastic buckling stress parameter can be expressed in a neat transcendental equation given by 12 2 (1 v 2 )( m 2 + n 2 2 ) E 12 2 2 + 2 2 ( m 2 + n 2 2 ) G
(4.3)
where m and n are the integers depending on the buckling mode shape of the plate. For a square plate ( = 1 ), the integers m and n become unity for the lowest plastic buckling stress parameter c . Therefore Eq. (4.3) furnishes 12 2 (1 v 2 ) c = E 6 2 + 2 2 G
(4.4)
The analytical solutions, which were derived by Handleman and Prager (1948), Shrivastava (1979), Durban and Zuckerman (1999) and Wang et al. (2001a), are presented in Tables 4.2 to 4.5 for comparison purposes. Convergence and comparison studies for square (a/b = 1) and rectangular (a/b = 2) plates under uniaxial and biaxial compression are shown in Tables 4.2 to 4.5. It can be seen that the solutions for SSSS plates converge very fast while a higher polynomial
65
Rectangular Plates
degree is needed for plates with free edges. Generally, the solutions from the Ritz method converge well when the degree of polynomial p = 12 is used. Hence, the degree of polynomial p = 12 will be used for all subsequent calculations. It can be seen that the converged results are found to be within 0.08% of analytical solutions given by Wang et al. (2001). This shows that the accuracy of plastic buckling results determined by the Ritz method is excellent. However, the results from Shrivasta (1979) are slightly higher than the present results. This is attributed by the fact that Shrivastas results shown in Tables 4.2 are recalculated from the truncated expressions presented in his paper. It was found that the present results are slightly lower than the corresponding results of Handleman and Prager (1948) and Durban and Zuckerman (1999). This is due to the neglect of the effect of transverse shear deformation in their analysis. Table 4.2 Comparison and convergence study of for square plates under uniform uniaxial loading with various boundary conditions. BC
p
2 4 6 8 Wang et al. (2001a) Shrivastava (1979) Handleman and Prager (1948) 8 10 12 14 Wang et al. (2001a)
= 0.001 IT DT 4.4581 4.4581 4.0021 4.0021 4.0000 4.0000 4.0000 4.0000

4.0000 4.0000 4.0000 1.9632 1.9632 1.9631 1.9631 1.9616 4.0000 4.0000 __ 1.9632 1.9632 1.9631 1.9631 1.9616
= 0.050 IT DT 3.7718 2.8735 3.4969 2.7958 3.4955 2.7954 3.4955 2.7954

3.4955 3.5278 3.5740 1.8552 1.8455 1.8419 1.8409 1.8407 2.7954 2.8058 __ 1.8388 1.8299 1.8265 1.8256 1.8254
= 0.1 IT DT 3.5950 0.9250 3.1990 0.9144 3.1908 0.9144 3.1908 0.9144

3.1908 3.4636 3.5193 1.0380 1.0177 1.0151 1.0149 1.0149 0.9144 0.9205 __ 0.7627 0.7619 0.7619 0.7619 0.7700
SSSS
FSFS
66
Rectangular Plates
BC
p
8 10 12 14 Wang et al. (2001a) 8 10 12 14 Wang et al. (2001a)
SSFS
Table 4.2 Continued = 0.001 = 0.050 IT DT IT DT 2.2829 2.2829 2.0479 2.0222 2.2828 2.2828 2.0404 2.0155 2.2828 2.2828 2.0378 2.0131 2.2828 2.2828 2.0371 2.0125 2.2811 2.3038 2.3037 2.3037 2.3037 2.3019 2.2811 2.3038 2.3037 2.3037 2.3037 2.3019 2.0369 2.0565 2.0489 2.0462 2.0455 2.0454 2.0124 2.0303 2.0235 2.0212 2.0206 2.0204
= 0.1
IT 1.3933 1.0339 1.0334 1.0334 1.0334 1.0404 1.0353 1.0349 1.0348 1.0348 DT 0.7701 0.7672 0.7667 0.7667 0.7667 0.7672 0.7667 0.7667 0.7667 0.7667
CSFS
Table 4.3 Comparison and convergence study of for rectangular plates under uniform uniaxial loading with various boundary conditions (a/b = 2) BC
p
4 6 8 10 Wang et al. (2001a) Handleman and Prager (1948) 8 10 12 14 Wang et al. (2001a) 8 10 12 14 Wang et al. (2001a)
= 0.001
IT 4.0114 4.0001 4.0000 4.0000 4.0000 4.0000 2.1854 2.1852 2.1852 2.1852 2.1835 2.2311 2.2310 2.2310 2.2310 2.2293 DT 4.0114 4.0001 4.0000 4.0000 4.0000 __ 2.1854 2.1852 2.1852 2.1852 2.1835 2.2310 2.2310 2.2310 2.2310 2.2293
= 0.05
IT 3.5001 3.3397 3.3225 3.3222 3.3222 3.4759 2.0344 2.0212 2.0101 2.0021 1.9944 2.0461 2.0341 2.0251 2.0198 2.0155 DT 2.7972 2.7954 2.7954 2.7954 2.7954 __ 2.0064 1.9948 1.9849 1.9778 1.9708 2.0186 2.0080 2.0000 1.9952 1.9914
= 0.1
IT 3.3423 3.0655 3.0506 3.0503 3.0503 3.2839 1.1059 1.0569 1.0389 1.0341 1.0318 1.0914 1.0507 1.0366 1.0333 1.0326 DT 0.9146 0.9078 0.9069 0.9069 0.9068 __ 0.7729 0.7687 0.7671 0.7667 0.7479 0.7717 0.7682 0.7670 0.7667 0.7666
SSSS
FSFS
SSFS
67
Rectangular Plates
BC
p
8 10 12 14 Wang et al. (2001a)
CSFS
Table 4.3 Continued = 0.001 = 0.05 IT DT IT DT 2.2314 2.2314 2.0462 2.0187 2.2314 2.2314 2.0342 2.0081 2.2314 2.2314 2.0253 2.0001 2.2314 2.2314 2.0199 1.9953 2.2297 2.2297 2.0157 1.9915
= 0.075 IT DT 1.0907 0.7716 1.0503 0.7682 1.0365 0.7670 1.0332 0.7667

1.0326 0.7666
Table 4.4. Convergence and comparison studies of plastic buckling stress parameters for square plates under uniform biaxial loading. BC
p
2 4 6 8 Wang et al. (2001a) Durban and Zuckerman (1999) 8 10 12 14 Wang et al. (2001a) 8 10 12 14 Wang et al. (2001a) 8 10 12 14 Wang et al. (2001a)
= 0.001
IT 2.2291 2.0010 2.0000 2.0000 2.0000 2.0000 0.9159 0.9159 0.9159 0.9159 0.9158 1.0287 1.0287 1.0287 1.0287 1.0284 1.1104 1.1104 1.1104 1.1104 1.1099 DT 2.2291 2.0010 2.0000 2.0000 2.0000 2.0000 0.9159 0.9159 0.9159 0.9159 0.9158 1.0287 1.0287 1.0287 1.0287 1.0284 1.1104 1.1104 1.1104 1.1104 1.1099
= 0.050
IT 2.0073 1.8720 1.8713 1.8713 1.8713 1.8713 0.9054 0.9041 0.9035 0.9033 0.9033 1.0073 1.0054 1.0047 1.0046 1.0045 1.0816 1.0790 1.0781 1.0779 1.0777 DT 1.9969 1.8656 1.8649 1.8649 1.8649 1.8649 0.9054 0.9041 0.9035 0.9033 0.9032 1.0072 1.0053 1.0047 1.0045 1.0044 1.0815 1.0789 1.0780 1.0778 1.0777
= 0.1
IT 0.8484 0.8086 0.8084 0.8084 0.8084 0.6661 0.6401 0.6401 0.6401 0.6401 0.6401 0.6865 0.6865 0.6865 0.6865 0.6865 0.7236 0.7236 0.7236 0.7236 0.7236 DT 0.7445 0.7300 0.7300 0.7300 0.7300 0.7346 0.6249 0.6249 0.6249 0.6249 0.6250 0.6567 0.6567 0.6567 0.6567 0.6567 0.6775 0.6775 0.6775 0.6775 0.6775
SSSS
FSFS
SSFS
CSFS
68
Rectangular Plates
Table 4.5. Convergence and comparison studies of plastic buckling stress parameters for rectangular plates under uniform biaxial loading (a/b = 2). BC
p
2 4 6 8
= 0.001
IT 1.4388 1.2508 1.2500 1.2500 DT 1.4388 1.2508 1.2500 1.2500 1.2500 0.9390 0.9390 0.9390 0.9390 0.9389 0.9753 0.9753 0.9753 0.9753 0.9752 0.9833 0.9833 0.9833 0.9833 0.9831
= 0.050
IT 1.4158 1.2371 1.2363 1.2363 1.2363 0.9297 0.9286 0.9274 0.9266 0.9257 0.9632 0.9615 0.9601 0.9593 0.9587 0.9704 0.9685 0.9670 0.9661 0.9654 DT 1.4151 1.2369 1.2361 1.2361 1.2361 0.9297 0.9285 0.9274 0.9265 0.9257 0.9632 0.9615 0.9601 0.9593 0.9586 0.9704 0.9685 0.9670 0.9661 0.9654
= 0.1
IT 0.7180 0.6869 0.6868 0.6868 0.6868 0.6402 0.6402 0.6402 0.6402 0.6402 0.6527 0.6527 0.6527 0.6527 0.6527 0.6574 0.6574 0.6574 0.6574 0.6574 DT 0.6835 0.6623 0.6622 0.6622 0.6622 0.6254 0.6254 0.6254 0.6254 0.6257 0.6351 0.6351 0.6351 0.6351 0.6351 0.6385 0.6384 0.6384 0.6384 0.6384
SSSS
Wang et al. 1.2500 (2001a) 8 10 FSFS 12 14 0.9390 0.9390 0.9390 0.9390
Wang et al. 0.9389 (2001a) 8 10 SSFS 12 14 0.9753 0.9753 0.9753 0.9753
Wang et al. 0.9752 (2001a) 8 10 CSFS 12 14 0.9833 0.9833 0.9833 0.9833
Wang et al. 0.9831 (2001a)
Effect of thickness-to-width ratio and transverse shear deformation

The variations of normalized plastic buckling stress with respect to b/h ratios are shown in Figs. 4.2 to 4.5 for square plates under unaxial ( = 0 ) and biaxial compressive stresses ( = 1 ). It can be seen that the buckling stresses associated with
69
Rectangular Plates
DT are consistently lower than their IT counterparts. The difference between the buckling stresses obtained based on IT and DT increases with the b/h ratio decreases, i.e. plate thickness increases. Moreover, the difference is more significant for more restrained boundary conditions for the same b/h ratio. On the other hand, all the plastic buckling curves (i.e. IT and DT) converge to the elastic buckling curve as b/h ratio increases since buckling takes place in the elastic range for large b/h ratios. In Figs. 4.2 to 4.5, the buckling curves based on the classical thin plate theory are also presented with the view to examine the effect of transverse shear deformation. The buckling stress, which is equivalent to that of the classical thin plate theory (CPT), can be readily obtained by setting a large number 2 = 1000 for the shear correction factor (see Table 4.6). Note that, in Figs. 4.2 to 4.5, the dashed lines refer to the classical thin plate theory and the solid lines represent the Mindlin plate theory. It can be seen that the effect of transverse shear deformation lowers the buckling stress and the effect is more significant when the plate is relatively thick. The effect is not significant for plastic buckling when compared to that of the elastic buckling counterparts. For example, the reductions for SSSS square plate under biaxial compression (b/h = 6) are about 13%, 4.8% and 1% in the cases of elastic theory, IT and DT, respectively. Generally, the reductions are larger for the buckling stress associated with IT when compared with the corresponding reductions in their DT counterparts. This is because the effect of transverse shear deformation depends on the value of plastic buckling stress as shown in Figs. 4.6 and 4.7. It can be seen that the influence of the effect of transverse shear deformation increases when the plastic buckling stress increases.
70
Rectangular Plates
Table 4.6 Parametric study on the value of shear correction factor for simply supported square plate under biaxial loading.
2
Mindlin plate theory Classical thin plate theory 5/6 1000 -
= 0.001
2.0000 2.0000 2.0000
0.05 1.8713 1.8896 1.8896
0.10 0.8084 0.8168 0.8168
0.15 0.6429 0.6710 0.6711
0.20 0.6210 0.6700 0.6700
Incremental theory of plasticity
Deformation theory of plasticity Mindlin plate theory Classical thin plate theory 5/6 1000 2.0000 2.0000 2.0000 1.8649 1.8830 1.8830 0.7300 0.7346 0.7346 0.3694 0.3729 0.3729 0.2239 0.2270 0.2270
1.5
1.25
=0
a b
c/0.2
0.75
=1
a
Elastic IT DT
0.5
0.25
0 5 10 15 20 25 30 35 40
b/h
Fig. 4.2 Normalized plastic buckling stress c/0.2 versus b/h ratio for SSSS square plate
71
Rectangular Plates
1.5
=0
1.25
a b
c/0.2
0.75
a b
=1
Elastic IT DT
20 25 30 35 40
0.5
0.25 5 10 15
b/h
Fig. 4.3 Normalized plastic buckling stress c/0.2 versus b/h ratio for CCCC square plate
1.5
=0
1.25
a b
c/0.2
0.75
=1
a b
Elastic IT DT
15 20 25 30 35 40
0.5
0.25 5 10
b/h
Fig. 4.4 Normalized plastic buckling stress c/0.2 versus b/h ratio for SCSC square plate
72
Rectangular Plates
1.5
1.25
c/0.2
0.75
a b
=0
0.5
a b
=1
Elastic IT DT
0.25
0 5 10 15 20 25 30 35 40
b/h
Fig. 4.5 Normalized plastic buckling stress c/0.2 versus b/h ratio for SSSF square plate
0.15
0.12
CPT
SSSS plate (Mindlin)
c h3 / D
(Mindlin)
0.09
0.06
CCCC plate (Mindlin)
0.03
0.00 0 0.03 0.06 0.09 0.12 0.15
c h 3 / D (CPT)
Fig. 4.6 Effect of transverse shear deformation on plastic buckling stress (IT) of a square plate
73
Rectangular Plates
0.015
CPT
0.012
SSSS plate (Mindlin)
c h3 / D
(Mindlin)
0.009
CCCC plate (Mindlin)
0.006
0.003
0 0 0.003 0.006 0.009 0.012 0.015
c h 3 / D (CPT)
Fig. 4.7 Effect of transverse shear deformation on plastic buckling stress (DT) of a square plate
Effect of aspect ratio a/b

The variations of plastic buckling stress parameters with respect to aspect ratios a/b are presented in Figs. 4.6 to 4.13 for various boundary conditions. Aspect ratios between 0.5 and 4.0 are considered since the buckling capacity has the most significant variation over this range of aspect ratios. Generally, it is found that the plastic buckling stress parameters are almost constant for aspect ratios greater than 4. The thickness-towidth ratios in these figures are chosen in order to capture the plastic buckling behaviour when the normalized buckling stress ( c / 0.2 ) is in the vicinity of unity. Therefore, according to Figs. 4.2 to 4.5, = 0.05 is used for SSSS, SCSC, CCCC plates while = 0.1 is used for SSSF plates. In order to observe the effect of loading ratio, we consider = 0, 0.25, 0.5, 0.75 and 1.
74
Rectangular Plates
For the cases of SSSS, CCCC, SCSC, it can be seen that some of the buckling curves are composed of individual curve segments that correspond to different buckling modes. This is due to the fact that the number of half-sine waves in the buckling mode shape of the plates increases as the aspect ratio increases. The locations of cusps, at the aspect ratios of coincident buckling modes, are not the same for different plate thicknesses because of the effect of transverse shear deformation. It was also found that these cusps locations depend on the plasticity theories used. For example, the buckling mode comprises two half sine waves for IT while the plate buckles in the shape of one half sine wave in the cases of elastic buckling and plastic buckling based on DT for h/b = 0.05 and a/b = 1.1 (see Fig 4.14). This information would be useful when one needs to assume the distribution pattern of initial imperfection that yield the lowest buckling load in a postbuckling analysis. The phenomena of mode switching disappear at a certain value of . For example, a smooth buckling curve is obtained when = 0.5 in the case of SSSS plates. In contrast to the aforementioned cases, the plastic buckling stress parameters decrease monotonically with increasing aspect ratios in the case of SSSF plates, regardless of the value of , i.e. no mode switching phenomenon is observed. When the buckling curves of different values of are compared, the buckling stress parameters are observed to decrease as the value of increases except for some cases of a / b < 1 . For a/b <1, an unexpected behaviour of buckling stress parameter with respect to value is observed. This behaviour is due to mode switching which will be proven later.
75
Rectangular Plates
3.5
=0 = 0.25 = 0.5 = 0.75 = 1.0
= xhb2/(2D)
2.5
1.5
1
0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 4.8 Plastic buckling stress parameter versus aspect ratio a/b for SSSS based on IT
3.5
=0 = 0.25 = 0.5
= xhb2/(2D)
2.5
= 0.75
1.5
= 1.0
1 0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 4.9 Plastic buckling stress parameter versus aspect ratio a/b for SSSS based on DT
76
Rectangular Plates
= xhb2/(2D)
=0 = 0.25 = 0.5 = 0.75
= 1.0
2
0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 4.10 Plastic buckling stress parameter versus aspect ratio a/b
3.75
3.5
= 0.25
= 0.5
=0
3.25
= xhb2/(2D)
= 0.75
2.75
= 1.0
2.5
2.25 0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 4.11 Plastic buckling stress parameter versus aspect ratio a/b
77
Rectangular Plates
5.5
=0
4.5
= xhb2/(2D)
= 0.25 = 0.5 = 0.75 = 1.0
3.5
2.5
2
0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 4.12 Plastic buckling stress parameter versus aspect ratio a/b for SCSC based on IT
3.5
= 0.25
3.25
= 0.5
=0
= xhb2/(2D)
= 0.75
2.75
2.5
= 1.0
2.25 0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 4.13 Plastic buckling stress parameter versus aspect ratio a/b for SCSC based on DT
78
Rectangular Plates
1.25
=0 = 0.25 = 0.5 = 0.75
= xhb2/(2D)
0.75
0.5
0.25
=1
0
0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 4.14 Plastic buckling stress parameter versus aspect ratio a/b for SSSF based on IT
1.25
=0
= xhb2/(2D)
= 0.25 = 0.5 = 0.75
0.75
0.5
0.25
=0
0 0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 4.15 Plastic buckling stress parameter versus aspect ratio a/b for SSSF based on IT
79
Rectangular Plates
(a)
(b)
(c)
Fig. 4.16. Contour and 3D plots of simply supported rectangular plate under uniaxial loading ( = 0.05, a/b = 1.1) for (a) IT, (b) Elastic and (c) DT
Effect of loading ratio

Figures 4.15 to 4.18 show the variations of plastic buckling stress parameter over a range of loading ratio for square plates with various boundary conditions. Since the thickness-to-width ratio has a strong influence on the plastic buckling behaviour, four different thickness-to-width ratios, namely = 0.001, 0.05, 0.075, and 0.1 are considered. In the case of = 0.001 , it can be seen that the IT curves coincide with their corresponding DT curves for all boundary conditions. Moreover, the plastic buckling stress parameter decreases steadily with the increasing value of loading ratio . For large thickness-to-width ratios, the plastic buckling stress parameter may decrease or increase with increasing . The optimal value of for the maximum buckling capacity in DT curves is observed. For example, the maximum buckling strength (DT) is obtained when the value of 0.5 in CCCC and SCSC plates for
= 0.075 and 0.1 . In contrast to DT buckling curves, the concave buckling curves of
80
Rectangular Plates
IT are observed except for SSSF plate. In the case of SSSF, the IT and DT buckling curves are close to each other and the optimal value of is found to be about 0.4. These unexpected behaviours were also observed in Figs. 4.7 to 4.9 and Fig. 4.11 for small a/b ratios. Similar observations were also reported in Durban and Zuckerman (1999). It is believed that these unexpected buckling behaviours are observed due to the mode switching. In order to verify this hypothesis, we perform an analytical analysis where the numbers of half wave of buckling shapes were predetermined. The analysis is carried out for SSSS square plate and the results are shown in Fig. 4.19. In this figure, the solid lines refer to the buckling mode shape with m = 1 and n = 1 while the dashed lines represent the buckling mode shape with m = 2 and n = 1 (m and n are numbers of half wave in the x and y direction, respectively). It can be seen that the plate always buckles with a mode shape associated with m = 1 and n = 1 for small thickness-towidth ratios, regardless of the loading ratio. However, for large thickness-to-width ratios, the plate first buckles in a m = 2, n = 1 mode shape up to certain value of loading ratio and then the critical mode shape changes to a m = 1, n = 1 mode shape as the loading ratio increases. This m = 2 mode shape is not evident in the case of elastic buckling for a square plate. Only in the plastic buckling, such a two half sine waves buckling mode shape is possible depending on the loading ratio for a square plate.
81
Rectangular Plates
IT
4
= xhb2/(2D)
= 0.001 = 0.05 = 0.075
DT
= 0.1
0
0 0.2 0.4 0.6 0.8 1
Stress ratio Fig. 4.17 Plastic buckling stress parameter versus stress ratio for SSSS square plate
12
10
= 0.001
IT DT
= xhb2/(2D)
= 0.05
4
= 0.075
= 0.1
0
0 0.2 0.4 0.6 0.8 1
Stress ratio Fig. 4.18 Plastic buckling stress parameter versus stress ratio for CCCC square plate
82
Rectangular Plates
IT
= 0.001
6
DT
= xhb2/(2D)
= 0.05
4
= 0.001
= 0.1
0
0 0.2 0.4 0.6 0.8 1
Stress ratio Fig. 4.19 Plastic buckling stress parameter versus stress ratio for SCSC square plate
IT
1.5
= 0.001 = 0.05 = 0.075
DT
= xhb2/(2D)
= 0.1
0.5
0
0 0.2 0.4 0.6 0.8 1
Stress ratio Fig. 4.20 Plastic buckling stress parameter versus stress ratio for SSSF square plate
83
Rectangular Plates
= xhb2/(2D)
= 0.001 = 0.025 = 0.05

0.25 0.5
= 0.001
0.75
= 0.1
1
0 0
Stress ratio
Fig. 4.21 Mode switching with respect to stress ratio for SSSS square plate
Effect of material properties

The aforementioned discussions are based on the analysis with Al 7075-T6 and the results may vary if a different material is used. Therefore, a parametric study is carried out for SSSS square plate under uniaxial compression in order to examine the effect of material properties on the buckling capacity. The normalized buckling curves are presented in Figs 4.20 and 4.21 for various Ramberg-Osgood parameters c and E / 0.7 ratios. It can be seen that the buckling capacity increases before a certain stress level and after that it decreases as the value of c increases. This behaviour can be explained by the trends of the stress-strain curves with respect to the c value which were shown in Fig. 2.2. The buckling curves are found to be shifted to the right if the ratio E / 0 increases. This effect can be incorporated in the slenderness ratio (b/h) as shown in Fig. 4.22.
84
Rectangular Plates
1.5
1.25
c=2 c=5 c = 10
c=2 c=5 c = 10
c = 15
c/0.2
0.75
c = 15
0.5
0.25
E/0 = 150 E/0 = 500

15 20 25 30 35 40 45 50
b/h
Fig. 4.22 Effect of material properties on normalized plastic buckling stress c / 0.2 for SSSS square plate (IT)
c=2 c=5
1.5
c = 10
c/0.2
c = 15
0.5
E/0 = 150 E/0 = 500

0 10 15 20 25 30 35 40 45 50
b/h
Fig. 4.23 Effect of material properties on normalized plastic buckling stress c / 0.2 for SSSS square plate (DT)
85
Rectangular Plates
1.5
1.25
c/0.2
0.75
0.5
0.25
IT DT
0 10 15 20 25 30 35 40 45 50
b h E /(100 0.7 )
Fig. 4.24 Normalized plastic buckling stress c / 0.2 versus modified slenderness ratio for SSSS square plate
Comparison with experiments

It has been shown in Fig. 4.22 that the effect of the variation of E / 0.7 can be absorbed by multiplying
0.7 / E in the value of slenderness ratio. Therefore the

b 12(1 2 ) (where e = e hb 2 / D is the elastic h e
commonly used slenderness ratio
buckling parameter) becomes = 0.7 / e , where e refers to the elastic buckling stress. Since the plastic buckling stresses obtained by IT and DT differ greatly when the thickness to width ratio is large, the comparison studies against test data for rectangular plates under uniaxial compression are presented in Figs. 4.23 and 4.24. It can be seen that the test results are in favour with the results obtained by DT which violates the fundamental rule of plastic flow of metal. On the other hand the results
86
Rectangular Plates
given by analysis based on IT, which is regarded as most satisfactory theory for treating plasticity problems, are significantly higher than test data for small slenderness ratios.
1.5
IT
1.25
Elastic
a b
DT
c 0.2
0.75
0.5
0.25
Anderson (1956)
0 0 0.5 1
1.5
2.5
Fig. 4.25 Comparison study of buckling stresses obtained by DT and IT with test results (Al-7075-T6)
1.5
IT
1.25
Elastic
c
a b
c 0.2
DT
0.75
0.5
Anderson (1956)
0.25
Pride (1949)
0 0 0.5 1
1.5
2.5
Fig. 4.26 Comparison study of buckling stresses obtained by DT and IT with test results (Al-2014-T6)
87
Rectangular Plates
4.3
Pure Shear Loading
In the case of pure shear, the in-plane stresses x and y are set to zero in Eqs. (2.19a-g) and Eq. (2.20). Hence the effective stress and the parameters
, , , , G, , become as follows
For deformation theory of plasticity (DT)
2 = 3 xy
(4.5a)
E 3E + (1 2 )2 (1 2 ) s Es E
4
(4.5b)
= =
(4.5c)
Es 1 2 2(1 2 ) E
1
E 2 + 3 1 E t E E 2 + 3 E 1 s
(4.5d)
(4.5e)
G=
(4.5f)
= =0
For incremental theory of plasticity (IT)
1 1 2
(4.5g)
= =
(4.6a)
1 2
(4.6b)
88
Rectangular Plates
1
E 2 + 3 1 E t
(4.6c)
G=
E 2(1 + )
(4.6d)
where xy is the shear stress. The convergence studies in Tables 4.7 and 4.8 show the convergence of the results with respect to the degree of polynomial. It can be seen that p = 12 is sufficient for reasonably converged results and therefore, p = 12 will be used for subsequent calculations. For comparison purposes, the plastic buckling stress parameters are computed from their corresponding elastic thin plate results by multiplying with the plasticity reduction factors which were derived by Gerard (1948), Bijlaard (1949) and Gerard and Becker (1957). Gerard (1948) modified the secant modulus method for plates under shear by replacing the secant modulus Es with the shear secant modulus Gs. The latter modulus is equal to E in the present study. In view of modified secant method, the plasticity reduction factor becomes Gs/G. It was found that the results obtained by using the secant modulus method are consistently lower than the present results based on DT and IT. Bijlaard (1949), it was found that the results, which are obtained by Bijlaards (1949) formula, are consistently higher than the present Ritz results for both thicknessto-width ratios ( = 0.001 and = 0.05). This is due to an error in the parameters used in the plastic reduction factor formula. An erratum was reported by Peters (1954) who pointed out that the symbols, A, B, and F appearing in the case 6 of Table 5 are not the same symbols as A, B, and F as given in Eq. 21 of Bijlaard (1949). Instead, they are
89
Rectangular Plates
the same symbols as those given in Eq. 36 of Bijlaard (1947). Unfortunately, the paper by Bijlaard (1947) is not accessible to the author and hence the corrected results cannot be computed and verified. On the other hand, the present results agree well with the solutions obtained from plasticity reduction factor reported in Table 2 of Gerard and Becker (1957). When the thickness-to-width ratio = 0.025, the present solutions are however slightly lower than Gerards solutions due to the effect the transverse shear deformation.
Effect of transverse shear deformation and thickness-to-width ratio
The variations of normalized buckling stress with respect to b/h ratio are shown in Figs. 4.23 to 4.26. The effect of transverse shear deformation can be observed by comparing the solid lines (Mindlin plate theory) and dashed lines (classical thin plate theory). In general, the effect of shear deformation is more pronounced in the case of IT than DT. It is interesting to note that the IT buckling curves are parallel with the elastic buckling curves for small b/h ratios.
90
Rectangular Plates
Table 4.7 Convergence and comparison studies of plastic buckling stress parameters for SSSS rectangular plates under shear
a/b
1 IT DT IT 6.5460
2 DT IT 5.8403
3 DT IT 5.6246
4 DT
Elastic thin 9.3246 plate results 0.001 8 9.3387 9.3387 10 9.3252 9.3252 12 9.3245 9.3245 14 9.3244 9.3244 Gerard (1946) Bijlaard (1949) Gerard and Becker (1957) 0.025 __ __ 9.3246 9.5327
6.5544 6.5544 6.5463 6.5463 6.5460 6.5460 6.5460 6.5460 __ __ 6.5460 6.6923
5.8464 5.8464 5.8404 5.8404 5.8402 5.8402 5.8401 5.8401 __ __ 5.8403 5.9706
5.7437 5.7437 5.6326 5.6326 5.6249 5.6249 5.6246 5.6246 __ __ 5.6246 5.7503
__
9.3246
6.5460
__
5.8403
__
5.6246
8 6.7903 6.2047 10 6.7820 6.2010 12 6.7815 6.2008 14 6.7814 6.2008
5.5982 5.4344 5.5951 5.4321 5.5949 5.4319 5.5949 5.4319 __ __ 4.8793 5.8606
5.2718 5.1550 5.2673 5.1514 5.2671 5.1513 5.2670 5.1513 __ __ 4.6875 5.4975
5.2176 5.1091 5.1385 5.0427 5.1327 5.0379 5.1325 5.0377 __ __ 4.6187 5.3673
Gerard (1948) Bijlaard (1949) Gerard and Becker (1957)
__ __
5.3610 6.7260
__
6.4507
__
5.5858
__
5.2503
__
5.1330
91
Rectangular Plates
Table 4.8 Convergence and comparison studies of plastic buckling stress parameters for CCCC rectangular plates under shear
a/b
1 IT DT IT
2 DT IT 9.5343
3 DT IT 9.2952
4 DT
Elastic thin 14.6421 plate results 0.001 8 14.659 14.659 10 14.643 14.643 12 14.642 14.642 14 14.642 14.642 0.025 8 11.1994 7.1910 10 11.1781 7.1890 12 11.1770 7.1889 14 11.1769 7.1889 Gerard (1948) Gerard and Becker (1957)
1
10.2480 10.578 10.578 10.262 10.262 10.248 10.248 10.248 10.248 8.3101 6.6143 8.0692 6.5635 8.0591 6.5615 8.0587 6.5615 __ 5.4709
9.8601 9.8601 9.5575 9.5575 9.5353 9.5353 9.5341 9.5341 7.9557 6.5070 7.6356 6.4280 7.6044 6.4211 7.6029 6.4208 __ 5.3875
10.325 10.325 9.4936 9.4936 9.3406 9.3405 9.2984 9.2984 8.6245 6.6516 7.6610 6.4269 7.4493 6.3723 7.4231 6.3658 __ 5.3573
__
5.8463
__
7.2690
__
6.6432
__
6.4973
__
6.4440
xy
0.75
a b
xy/0.2
0.5
0.25
Elastic IT DT
0 20 30 40 50 60 70
b/h
Fig. 4.27 Normalized plastic buckling stress xy / 0.2 versus b/h ratio for SSSS square plate
92
Rectangular Plates
0.75
xy/0.2
0.5
Elastic IT DT
0.25 10 20 30
xy
a b
40
50
60
70
b/h
Fig. 4.28 Normalized plastic buckling stress xy / 0.2 versus b/h ratio for CCCC square plate
0.75
xy/0.2
0.5
Elastic IT DT
0.25 10 20
xy
a b
30
40
50
60
70
b/h
Fig. 4.29 Normalized plastic buckling stress xy / 0.2 versus b/h ratio for SCSC square plate
93
Rectangular Plates
0.75
xy
a b
xy/0.2
0.5
0.25
Elastic IT DT
10 20 30 40 50 60 70
b/h
Fig. 4.30 Normalized plastic buckling stress xy / 0.2 versus b/h ratio for SSSF square plate
Effect of aspect ratio a/b
The variations of plastic buckling stress parameters with respect to aspect ratio a/b are shown in Figs. 4.27, 4.28 and 4.29 for SSSS, CCCC and SSSF rectangular plates respectively. It is observed that the buckling mode switches in the cases of SSSS and CCCC but the locations of cusps cannot be seen clearly in those figures. On the other hand, no mode switching phenomenon is observed in the case of SSSF. Typical buckling mode shapes of rectangular plates under shear are shown in Table 4.9 for various aspect ratios by using DT.
94
Rectangular Plates
25
20
= 0.025 = 0.05
Elastic
xy=xyhb2/(2D)
15
xy
a b
IT
10
DT
0 0.5 1 1.5 2 2.5 3 3.5 4
Aspect ratio a/b Fig.4.31 Plastic buckling stress parameter, xy, versus aspect ratio, a/b, for SSSS rectangular plate
40
Elastic
30
= 0.025 = 0.05
xy
IT
a b
xy=xyhb2/(2D)
20
DT
10
0 0.5 1 1.5 2 2.5 3 3.5 4
Aspect ratio a/b Fig.4.32 Plastic buckling stress parameter, xy, versus aspect ratio, a/b, for CCCC rectangular plate under shear stress
95
Rectangular Plates
20
= 0.025 = 0.05
xy=xyhb2/(2D)
xy
10
a b
0 0.5 1 1.5 2 2.5 3 3.5 4
Aspect ratio a/b Fig.4.33 Plastic buckling stress parameter, xy, versus aspect ratio, a/b, for SSSF rectangular plate under shear stress
Table 4.9 Buckling mode shapes for rectangular plates under pure shear. ( = 0.03) BC
a/b
3D Plot
Contour Plot
SSSS 2
96
Rectangular Plates
Table 4.10 Continued.
CCCC
1 SSSF
4.4
Combined Shear and Uniaxial Compression
In practice, structural plate elements are often subjected to both inplane compressive and shear loads simultaneously. Elastic buckling of plates under such combined inplane loads has been studied by many researchers and well documented in text books such as Timoshenko and Gere (1961). In the case of plastic buckling, Peters (1954) conducted experiments on long square tubes and compared the results with plastic buckling stresses predicted by the flat plate theory. The results were found to agree
97
Rectangular Plates
well with the theoretical results obtained by Stowell (1949) and an interaction formula was proposed. Another theoretical study was performed by Tugcu (1996) based on the incremental theory of plasticity and the deformation theory of plasticity with the view to understand the effect of axial force on the shear buckling of long plate strips. It was found that the shear buckling results based on IT were highly sensitive to the presence of axial loads while those based on DT were less affected. Recently, Smith et al. (2003) studied the plastic buckling of rectangular plate under various loading conditions. However, their study was confined to the classical thin plate theory and to the case where the external shear stress is small. Hitherto, there are apparently no study being done on the plastic buckling of rectangular Mindlin plates under uniaxial compression and shear loads. The convergence studies of buckling stress parameters are shown in Tables 4.10 and 4.11 for rectangular plates under uniaxial (in the x-direction) and shear stresses. The converged results are obtained within reasonable accuracy when the degree of polynomial p =12 is used. Owing to the lack of plastic buckling solutions, the comparison study is performed against existing elastic buckling solutions and shown in Table 4.12. The buckling stress parameters obtained in this study are slightly lower than the corresponding existing elastic solutions obtained by Batdorf and Stein (1947). This is due to the fact that the existing solutions were obtained by using 10 terms of the Fourier series that represent the deflected shape and it is believed that these solutions have not converged yet. The interaction curves for the plastic buckling shear and uniaxial stresses are shown in Fig. 4.30 for simply supported, rectangular plates with a/b = 1 and a/b = 4. It can be seen that the differences between the plastic buckling stress parameters obtained by IT and DT are more pronounced when the plate is under pure shear (i.e. x = 0 ). With the
98
Rectangular Plates
presence of a small value of uniaxial compression, the reductions in shear plastic buckling stress parameters given by IT are larger than their corresponding DT counterparts. This is consistent with Tugcus (1996) observation that the plastic buckling shear stress parameters based on IT are more sensitive to the presence of uniaxial compression than their DT counterparts. Moreover, the interaction curves based on IT are affected significantly by the aspect ratios whereas the corresponding DT interaction curves are negligibly affected by the aspect ratios for the thickness-towidth ratio = 0.05 . In order to examine the effect of transverse shear deformation, the interaction curves based on the classical thin plate theory are presented in Fig. 4.30 as well. As expected, the classical thin plate theory overpredicts the plastic buckling load parameters as it normality assumption stiffens the plate by not permitting any transverse shear deformation. It is found that the reductions due to the effect of shear deformation in the results obtained by IT are generally larger than those reductions in their DT counterparts. Peters (1954) found that the experimentally obtained plastic buckling loads of simply supported plates are in good agreement with the theoretical plastic buckling results computed by Stowell (1948). He went on to propose a simple interaction
2 formula, Rxy + Rx2 = 1 , where R xy is the ratio of shear stress to the critical shear stress
due to pure shear and R x is the ratio of uniaxial stress to the critical buckling stress due to pure uniaxial compression. In order to investigate the validity of Peters formula for different thickness-to-width ratios , the normalized theoretical interaction curves based on DT for various thickness to width ratios are determined and compared with the interaction curves of Peters formula (see Figs. 4). It can be seen from these figures
99
Rectangular Plates
that the shapes of the interaction buckling curves depend on the thickness-to-width ratios and that the theoretical curves are bounded by the curves obtained by Peters
2 2 formula R xy + R x2 = 1 and the elastic interaction formula, Rxy + Rx = 1 (Batdorf and
Stein, 1947). Therefore it can be concluded that Peters formula provides an upper bound solution and it overestimates the buckling capacity of the plates with small to moderate thickness-to-width ratios.
Table 4.10 Convergence study of plastic buckling shear stress parameters xy for simply supported plates under combined loadings ( = 0.03).
p
x
1
a/b = 1
a/b = 2
a/b = 4
8 10 12 14 8 10 12 14
IT 5.0118 5.0067 5.0062 5.0060 4.0189 4.0173 4.0170 4.0169
DT 4.3807 4.3793 4.3972 4.3792 3.8152 3.8144 3.8142 3.8142
IT 4.2216 4.2201 4.2199 4.2198 3.5760 3.5753 3.5752 3.5752
DT 3.9907 3.9898 3.9897 3.9897 3.4740 3.4735 3.4735 3.4734
IT 4.0267 3.9846 3.9825 3.9824 3.4271 3.3953 3.3940 3.3939
DT 3.8502 3.8230 3.8216 3.8215 3.3440 3.3178 3.3167 3.3166
Table 4.11 Convergence study of plastic buckling shear stress parameters xy for clamped square plates under combined loadings. ( = 0.03)
p
x
1
a/b = 1
a/b = 2
a/b = 4
8 10 12 14 8 10 12 14
IT 10.0704 10.0527 10.0518 10.0517 8.9635 8.9478 8.9469 8.9468
DT 5.1586 5.1578 5.1577 5.1577 4.8768 4.8761 4.8760 4.8760
IT 7.3376 7.1084 7.0984 7.0981 6.4573 6.2306 6.2197 6.2193
DT 4.8265 4.7996 4.7985 4.7985 4.5381 4.5082 4.5069 4.5069
IT 7.8656 6.7100 6.4496 6.4132 7.1827 5.8685 5.6013 5.5577
DT 4.8904 4.7229 4.6833 4.6780 4.6471 4.4233 4.3744 4.3666
100
Rectangular Plates
Table 4.12 Comparison study of elastic buckling shear stress parameters xy for simply supported plates under combined loadings. ( = 0.001, = 0.3)
a/b
x
Present study Batdorf and Stein (1947) Present study Batdorf and Stein (1947) Present study Batdorf and Stein (1947)
1 7.92 8.11 5.60 5.71 4.85 5.00
2 6.46 6.62 4.56 4.66 4.02 4.10
3 4.55 4.68 3.21 3.29 2.85 2.94
1 2 4
Mindlin plate theory

6
IT, a/b = 1 IT, a/b = 2 IT, a/b = 4 DT, a/b = 1
xy = xyhb2/(2D)
Classical thin plate theory

xy x a b x
DT, a/b = 4
0 0.5 1 1.5 2 2.5 3 3.5 4
x = xhb2/(2D)
Fig. 4.34 Interaction curves for SSSS rectangular plates under combined uniaxial compression and shear stresses ( = 0.05)
101
Rectangular Plates
1.2
= 0.001 = 0.025 = 0.05
0.8
0.6
Rxy
0.4
= 0.1
0.2
2 Rxy + Rx = 1 2 2 Rxy + Rx = 1
0 0 0.2 0.4 0.6 0.8 1 1.2
Rx
Fig. 4.35 Normalized interaction curves for simply supported rectangular plate (a/b = 4)
1.2
= 0.001 = 0.025 = 0.05 = 0.1
0.8
Rxy
0.6
0.4
2 Rxy + Rx = 1 2 2 Rxy + Rx = 1
0.2
0 0 0.2 0.4 0.6 0.8 1 1.2
Rx
Fig. 4.36 Normalized interaction curves for clamped rectangular plates (a/b = 4)
102
Rectangular Plates
4.5 Concluding Remarks
The plastic buckling behaviour of rectangular Mindlin plates under uniaxial, biaxial and shear stresses are examined based on the extensive buckling results generated by the Ritz computer code. The effect of transverse shear deformation is taken into account in the analysis by adopting the Mindlin plate theory. As in the case of elastic buckling, the effect of transverse shear deformation lowers the plastic buckling stress parameter and the effect is more pronounced for the results obtained by IT when compared with their DT counterparts. The effects of aspect ratios a/b, thickness-to-width ratios , loading ratio on the plastic buckling stress parameters were investigated. It was found that the buckling mode switches at particular aspect ratios for SSSS and CCCC plates. The locations of cusps, at the aspect ratios of coincident buckling modes, are different from those locations in the case of elastic buckling. Moreover, the cusps locations also depend on the type of plasticity theory used. This indicates that the plastic buckling mode shape may be different from its elastic counterpart for a given aspect ratio. For small thickness-to-width ratios, the plastic buckling solutions are close to their elastic counterparts while, for large thickness ratios, the plastic buckling solutions are substantially lower than their elastic counterparts. Generally, the plastic buckling stress parameters obtained by IT are larger than their DT counterparts. The difference increases with increasing thickness-to-width ratios. With the presence of a small value of uniaxial compression, the reductions in shear plastic buckling stress parameters given by IT are larger than their corresponding DT counterparts. 103
CHAPTER 5
TRIANGULAR AND ELLIPTICAL PLATES
5.1
Introduction
In the previous chapter, plastic buckling problems of rectangular plates were investigated by using the Ritz method based on the Cartesian coordinate system. In fact, the Ritz method can be employed to solve arbitrary plate shapes which are expressed by polynomial functions. All that is needed is to modify the integration limits and the basic functions depending on the plate shapes and boundary conditions. This will be illustrated in this chapter by considering the plastic buckling problems of isosceles triangular plates, right angled triangular plates and elliptical plates.
5.2
Triangular Plates
Consider isosceles and right-angled triangular plates under uniform compressive stress as shown in Figs 5.1 and 5.2, respectively. The boundary conditions for the triangular plates are represented by a three-letter designation which is arranged in the 104
Triangular and Elliptical Plates
order starting from the left end edge (x = 0) followed by the other edges in an anticlockwise direction. For example, see Fig. 5.1 for CS*F isosceles triangular plate and Fig. 5.2 for CS*F right-angled triangular plate.
y, F C S* a Fig. 5.1 CS*F isosceles triangular plate and coordinate system x,
y, b x, C F S* a Fig. 5.2 CS*F right-angled triangular plate and coordinate system
105
In order to illustrate the compositions of the basic functions, we use the CS*F isosceles and CS*F right angled triangular plates examples. The basic functions for these plates are given by for CS*F isosceles triangular plate
bw = ( + 1)1 ( 0.5 + 0.5)1 ( + 0.5 0.5) 0 = ( + 1)( 0.5 + 0.5)

bx = ( + 1)1 ( 0.5 + 0.5) 0 ( + 0.5 0.5) 0 = ( + 1) by = ( + 1)1 ( 0.5 + 0.5) 0 ( + 0.5 0.5) 0 = ( + 1)
for CS*F right angled triangular plate
(5.1)
(5.2) (5.3)
bw = ( + 1)1 ( + 1)1 ( + ) 0 = ( + 1)( + 1) bx = ( + 1)1 ( + 1) 0 ( + ) 0 = ( + 1) by = ( + 1)1 ( + 1) 0 ( + ) 0 = ( + 1)
(5.4) (5.5) (5.6)
Note that a relationship between the plastic buckling solutions of polygonal Mindlin plates and their corresponding elastic buckling solutions based on classical thin plate theory was presented by Wang (2004). The relationship is given by
p
2 p 2 E 1 2 12 2 G
= e
(5.7)
where the subscript p refers to the plastic buckling solutions, the subscript e refers to the elastic buckling solutions and the parameters , G are given by Eqs. (4.1b) and
(4.1e) or Eqs. (4.2b) and (4.2e) depending on the type of plasticity theory used.
106
Based on the classical thin plate theory, Han (1960) gave the elastic buckling solution e = 16 / 3 for simply supported equilateral triangular plates ( a / b = 3 / 2 ). In view of this result, Eq. (5.7) can be expressed as
48 2 (1 v 2 ) E 9 2 + 4 2 2 G
p =
(5.8)
Note that the above equation is valid only for simply supported edge of hard type (S). However, the adopted Ritz functions in Ritz method fail to satisfy the geometric boundary condition of this hard type of simply supported edge when the plate edges are inclined (as in triangular plates) or curved (as in elliptical plates) edges. In order to handle the hard type of simply supported edge, a modification in the formulations is needed and it will be discussed in Chapter 8. Table 5.1 presents the comparison and convergence studies on the plastic buckling stress parameters of equilateral triangular plates with various boundary conditions; S*S*S*, CCC and S*CC. The convergence study for right-angled triangular plate with a/b = 1 is presented in Table 5.2. From Tables 5.1 and 5.2, it can be seen that the results associated with the degree of polynomial p = 10 show satisfactory convergence. Therefore, p = 10 will be adopted for the subsequent calculations of plastic buckling solutions of triangular plates. For simply supported equilateral triangular plates, it can be seen that present results are slightly less than the corresponding results obtained by Eq. (5.8). This is due to the different implementation of simply support boundary condition as discussed in the previous paragraph. However, the buckling results for the simply supported edge of soft type become closer to their hard type counterparts when the plate thickness is thin (for example = 0.001). For other boundary conditions, no
107
plastic buckling results are available in the literature and thus comparison studies are made only with existing elastic buckling results. It can be seen that results agree well with existing elastic solutions when the thickness-to-width ratio = 0.001 .
Table 5.1. Convergence and comparison studies of plastic buckling stress factors for equilateral triangular plates (a/b = 3 / 2 ) under uniform compression. BC p = 0.001 IT DT 5.3331 5.3331 5.3331 5.3331 5.3330 5.3330 5.3332 5.3332 = 0.050 IT DT 2.7492 2.6572 2.7441 2.6542 2.7428 2.6533 2.8033 4.8175 4.8173 __ 3.6207 3.6171 3.6164 2.6885 3.2301 3.2301 __ 3.0497 3.0491 3.0490 = 0.075 IT DT 1.6106 1.3670 1.6046 1.3662 1.6038 1.3661 1.8038 4.5489 4.5487 __ 3.0713 3.0684 3.0681 1.3912 1.5973 1.5973 __ 1.5219 1.5218 1.5218
8 10 S*S*S* 12 Wang et al. (2004) 8 10 CCC Xiang et al. (1994) 8 S*CC 10 12
14.8351 14.8351 14.8340 14.8340 14.8340 (Elastic) 10.6051 10.6051 10.6050 10.6050 10.6050 10.6050
Table 5.2 Convergence and comparison studies of plastic buckling stress factors for right angled triangular plates under uniform compression (a/b = 1). BC p = 0.001 IT DT 4.9998 4.9998 4.9998 4.9998 4.9997 4.9997 4.9997 (Elastic) 14.1625 14.1625 14.1417 14.1417 14.1412 14.1412 14.1413 (Elastic) 10.4139 10.4139 10.4130 10.4130 = 0.050 IT DT 2.6971 2.6180 2.6920 2.6147 2.6905 2.6138 __ 4.6217 4.6193 4.6193 __ 3.5857 3.5801 __ 3.2074 3.2072 3.2072 __ 3.0405 3.0396 = 0.075 IT DT 1.5605 1.3530 1.5548 1.3520 1.5538 1.3518 __ 4.3551 4.3537 4.3537 __ 3.0192 3.0133 __ 1.5883 1.5883 1.5883 __ 1.5183 1.5163
8 10 S*S*S* 12 Xiang et al. (1994) 8 10 CCC 12 Xiang et al. (1994) 8 S*CC 10
108
Effect of transverse shear deformation Tables 5.3 and 5.4 show the effect of transverse shear deformation on the plastic buckling stress parameters of simply supported and clamped isosceles triangular plates. It can be seen that the inclusion of the effect of transverse shear deformation in the analysis lowers buckling stress parameters. The reductions are shown as the percentages of the classical thin plate solutions. It can be seen that the reduction increases with increasing thickness-to-width ratios. Generally, those reductions associated with IT are larger than their DT counterparts.
Effect of aspect ratio The variations of plastic buckling stress parameter with respect to the aspect ratio are shown Figs. 5.3 and 5.4 for simply supported, isosceles and right angled triangular plates. In order to show the effect of thickness-to-width ratios, three thickness-to-width ratios are considered (i.e. = 0.001,0.05,0.1 ). It should be noted that the buckling curve of = 0.001 may be regarded as an elastic buckling curve. Hence the deviation of plastic buckling stress parameters from the corresponding elastic parameters can be seen for various aspect ratios. Generally, the plastic buckling stress parameter decreases with increasing aspect ratios.
109
Table 5.3 Effect of transverse shear deformation on for S*S*S* isosceles triangular plates Plate theory a/b = 0.5 IT 10.0000 9.9990 0% 8.6119 8.4259 2.16% 3.3500 2.0271 39.49% a/b = 2 IT 2.7262 2.7262 0% 2.7261 2.6893 1.53% 0.9676 0.8677 10.32%
Kirchhoff 0.001 Mindlin Difference Kirchhoff 0.025 Mindlin Difference Kirchhoff 0.1 Mindlin Difference
Elastic 10.0000 9.9990 0% 10.0000 9.5888 4.11% 10.0000 7.1522 28.48%
DT 10.0000 9.9990 0% 8.5474 8.3714 2.06% 0.9343 0.8929 4.43%
Elastic 2.7262 2.7262 0% 2.7262 2.6893 1.53% 2.7262 2.4068 11.72%
DT 2.7262 2.7262 0% 2.7261 2.6893 1.53% 0.7760 0.7577 2.36%
Table 5.4 Effect of transverse shear deformation on for CCC isosceles triangular plates Plate theory a/b = 0.5 IT 28.2837 28.2820 0% 12.4332 12.3117 0.97% 9.4750 7.2632 23.34% a/b = 2 IT 7.8271 7.8280 0% 7.4427 7.3587 1.13% 2.6221 2.4129 7.98%
Elastic 28.2837 28.2820 0% 28.2837 26.6956 5.61% 28.2837 14.8924 47.34%
DT 28.2837 28.2820 0% 11.4777 11.3982 0.69% 1.0584 1.0166 3.95%
Elastic 7.8271 7.8280 0% 7.8271 7.7001 1.62% 7.8271 6.2244 20.48%
DT 7.8271 7.8280 0% 7.4190 7.3361 1.12% 0.9054 0.8917 1.51%
110
8 7 6
b a
= 0.001 = 0.05 = 0.10 IT DT
= hb2/(2D)
5 4 3 2 1 0 0.25
0.5
0.75
1.25
1.5
1.75
Aspect ratio a/b Fig. 5.3 Variation of with respect to a/b ratio for S*S*S* isosceles triangular plates
b
4
= 0.001 = 0.05 = 0.10 IT DT
= hb2/(2D)
0 1 1.5 2 2.5 3 3.5 4
Aspect ratio a/b Fig. 5.4 Variation of with respect to a/b ratio for S*S*S* right angled triangular plates
111
5.3
Elliptical Plates
Consider an elliptical plate of constant thickness h and aspect ratio a/b as shown in Fig. 5.5. The plate is subjected to in-plane uniform compressive stress . The problem is to determine the plastic buckling stress of such loaded elliptical plates. y, x, b
a Fig. 5.5 Elliptical plate under uniform compression and coordinate system The basic functions for the Ritz formulations are given by: for simply supported elliptical plate
bw = ( 2 + 2 1)1 = ( 2 + 2 1) bx = ( 2 + 2 1) 0 = 1 by = ( 2 + 2 1) 0 = 1
for clamped elliptical plate
(5.9) (5.10) (5.11)
bw = ( 2 + 2 1)1 = ( 2 + 2 1) bx = ( 2 + 2 1)1 = ( 2 + 2 1) by = ( 2 + 2 1)1 = ( 2 + 2 1)
(5.12) (5.13) (5.14)
112
There are hitherto, no studies been done on plastic buckling of elliptical plates despite a few studies (Voinovsky-Krieger, 1937; Shibaoka, 1956; Wang and Liew, 1993b; Wang et al., 1994a) been conducted on the elastic buckling of such plates. Therefore, comparison studies can only be carried out against existing elastic buckling solutions. In order to obtain elastic buckling stress parameters, very small thickness-towidth ratio ( = 0.001 ) is adopted in the Ritz plastic buckling analysis. Note that, as a special case, the elliptical plate becomes a circular plate when a/b = 1 and the analytical plastic buckling solutions for circular plates are available (Wang et al., 2001a). The exact expression of plastic buckling parameter for circular plates under uniform compression can be expressed in the transcendental form of
J 0 ( ) = 1 J 1 ( )
for simply supported plates
(5.15)
2 (1 v 2 )
0.6722 + 0.8225
2
for clamped plates

2
(5.16)
where J n () is Bessel function of the first kind of order n and 12 2 2 =b (12 2 (1 v 2 ) 2 2 E / 0
(5.17)
The convergence and comparison studies are shown in Tables 5.5 and 5.6 for simply supported and clamped circular and elliptical plates for various aspect ratios (a/b = 1, 2). It can be seen that satisfactory convergence is achieved when the degree of polynomial p = 10 is used. The converged solutions are found to agree well with the existing solutions.
113
Table 5.5 Convergence and comparison studies of plastic buckling stress factors hb 2 /( 2 D ) for circular plates under uniform compression (a/b=1). BC p 8 Simply 10 supported Wang et al. (2001a) 8 Clamped 10 Wang et al. (2001a) = 0.001 IT 1.7322 1.7322 1.7322 5.9503 5.9503 5.9503 DT 1.7322 1.7322 1.7322 5.9503 5.9503 5.9503 = 0.050 IT 1.6623 1.6623 1.6622 2.9035 2.9035 2.9035 DT 1.6617 1.6617 1.6617 2.7536 2.7536 2.7536 = 0.075 IT 1.0690 1.0690 1.0690 1.9673 1.9673 1.9673 DT 1.0663 1.0663 1.0663 1.4152 1.4152 1.4152
Table 5.6 Convergence and comparison studies of plastic buckling stress factors hb 2 /( 2 D ) for elliptical plates under uniform compression (a/b=2). BC p 8 Simply supported 10 12 Wang et al. (1994a) 8 Clamped 10 Wang et al. (1994a) = 0.001 IT 1.2467 1.2467 1.2467 DT 1.2467 1.2467 1.2467 = 0.050 IT 1.2281 1.2271 1.2266 __ 2.6054 2.6054 __ DT 1.2280 1.2270 1.2265 __ 2.5403 2.5403 __ = 0.075 IT 0.9814 0.9811 0.9810 __ 1.5564 1.5564 __ DT 0.9762 0.9759 0.9758 __ 1.3379 1.3379 __
1.236 (Elastic) 4.2286 4.2286 4.2286 4.2286
4.231 (Elastic)
Effect of transverse shear deformation Tables 5.7 and 5.8 show the effect of transverse shear deformation on the plastic buckling stress parameters of simply supported and clamped circular and elliptical plates. It can be seen that the effect of transverse shear deformation lower the plastic buckling stress parameters. The effect on the plastic buckling stress parameters
114
obtained by using both plasticity theories is not as significant as that on corresponding elastic buckling stress parameters. Generally, the reductions are larger for the buckling stress associated with IT when compared with the reduction in DT counterparts. Table 5.7 Effect of transverse shear deformation on for simply supported circular (a/b = 1) and elliptical plates (a/b = 3) h/b Plate theory Elastic 0.001 Thin plate Mindlin Difference Thin plate Mindlin Difference Thin plate Mindlin Difference 1.7322 1.7322 0% 1.7322 1.7267 0.32% 1.7322 1.6481 4.86% a/b = 1 IT 1.7322 1.7322 0% 1.7322 1.7267 0.32% 0.6788 0.6762 0.38% DT 1.7322 1.7322 0% 1.7322 1.7267 0.32% 0.6761 0.6735 0.38% Elastic a/b = 3 IT DT 1.1834 1.1834 1.1834 1.1834 1.1834 1.1834 0% 0% 0% 1.1834 1.1834 1.1834 1.1787 1.1787 1.1787 0.40% 0.40% 0.40% 1.1834 0.6686 0.6522 1.1299 0.6611 0.6459 4.52% 1.12% 0.97%
0.025
0.100
Table 5.8 Effect of transverse shear deformation on for clamped circular (a/b = 1) and elliptical plates (a/b = 3) h/b Plate theory Elastic 0.001 Thin plate Mindlin Difference Thin plate Mindlin Difference Thin plate Mindlin Difference 5.9504 5.9503 0% 5.9504 5.8859 1.08% a/b = 1 IT 5.9504 5.9503 0% 5.9037 5.8439 1.01% DT 5.9504 5.9503 0% 5.9001 5.8405 1.01% 0.8728 0.8631 1.11% Elastic a/b = 3 IT DT 4.0392 4.0392 4.0392 4.0392 4.0392 4.0392 0% 0% 0% 4.0392 4.0378 4.0377 4.0063 4.0050 4.0049 0.81% 0.81% 0.81% 4.0392 1.3583 0.8258 3.5721 1.3030 0.8179 11.56% 4.07% 0.96%
0.025
0.100
5.9504 1.9937 5.0628 1.8833 14.91% 5.54%
115
Effect of aspect ratio The variations of plastic buckling stress parameter with respect to the aspect ratio are shown in Figs. 5.6 and 5.7 for simply supported and clamped elliptical plates. Generally, plastic buckling stress parameters decreases with increasing aspect ratios. For large thickness-to-width ratios, the variation of plastic buckling stress parameters with respect to aspect ratio is negligible. Interestingly, the plastic buckling stress parameters obtained by using IT and DT are close to each other for simply supported elliptical plates.
b a
1.75
= 0.001 = 0.050 = 0.100 IT DT
= hb2/(2D)
1.5
1.25
0.75
0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 5.6 Variation of with respect to a/b ratio for simply supported elliptical plates
116
7 6 5
b a
= 0.001 = 0.050 = 0.100 IT DT
= hb2/(2D)
4 3 2 1 0 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 5.7 Variation of with respect to a/b ratio for clamped elliptical plates
5.4
Concluding Remarks
The applicability of the Ritz method for various plate shapes is demonstrated by solving the plastic buckling problems of isosceles triangular plates, right angled triangular plates and elliptical plates. The results are verified by comparing with the limited existing elastic and plastic solutions. Based on the triangular and elliptical plate examples, it can be concluded that the developed Ritz method can be readily applied for the plastic buckling analyses of arbitrarily shaped plates with sides are defined by polynomial functions.
117
CHAPTER 6
SKEW PLATES
6.1
Introduction
Buckling behaviour of skew plates is especially important to engineers who are involved in designing aircraft wings, missile fins and skew bridges. The elastic buckling of skew plates has been studied extensively by many researchers who employed various numerical methods for solutions. Anderson (1951), Wittrick (1953), Mizusawa et al. (1980) and Wang et al. (1992) used the Ritz method, York and Williams (1995) adopted the Lagrangian multiplier method, Huyton and York (2001) made used of the commercial finite element software ABAQUS and Wang et al. (2003) employed the differential quadrature method. When the skew plates are thick, the effect of transverse shear deformation must be considered. Otherwise the buckling capacity will be overestimated. The buckling problem of thick skew plates was
118
Skew Plates
investigated by Kitipornchai et al. 1993, Wang et al. 1993a and Xiang et al. 1995. However, there are hitherto no plastic buckling solutions for thick skew plates published in the literature. This chapter presents for the first time the plastic buckling solutions of thick skew plates. The plastic bifurcation buckling problem of skew Mindlin plates is solved by using the Ritz method. The skew coordinate system is adopted because it is the most natural coordinate system for this plate shape. The detailed formulation of the Ritz method using the skew coordinate system was presented in Section 3.4 and the computer code called BUCKRITZSKEW was developed (see in Appendix A). In the calculations of plastic buckling stress, both the incremental theory (IT) and the deformation theory (DT) of plasticity were adopted. The effects of transverse shear deformation, thickness-to-width ratios, aspect ratios and skew angles on the plastic buckling stress parameters are investigated. For convenience and clarity, the letters, IT and DT, will be stated in the parentheses to indicate the plasticity theories used. For example, (IT) refers to the plastic buckling stress parameter based on IT. In generating the numerical results, a typical aerospace material Al 7075 T6 is adopted. The material has the following properties: Poissons ratio = 0.33, E = 10.5 10 6 psi, 0 = 0.7 = 70 ksi, 0.2 = 67 ksi and the
Ramberg-Osgood parameters k = 3/7, c = 9.2 .
6.2
Uniaxial and Biaxial Loading
Consider a skew plate of uniform thickness h, length a, oblique width b and skew angle as shown in Fig. 6.1. The plate is subjected to in-plane uniaxial or biaxial compressive stresses and the boundary conditions of the plate edges may take any
119
Skew Plates
combination of simply supported, clamped and free edge conditions. The problem at hand is to determine the plastic bifurcation stress of such a loaded plate.
y = x
a/2 a/2
b/2
y = x
Fig. 6.1. Skew plate under biaxial compression stresses
The plastic buckling stress parameters x , y used in the program code are defined
by:
x hb 2 x = 2 = D
y = y hb 2 2D
=
(6.1)
(6.2)
where is the key parameter to be evaluated. Before generating the buckling results, convergence studies of the plastic buckling stress parameter = x hb 2 /( 2 D) with respect to the degree of polynomial functions were conducted. Tables 6.1 to 6.4 present typical convergence studies for simply supported and clamped skew plates. Monotonic convergence is observed with an increasing degree of polynomials for the Ritz approximate functions. The convergence rate is found to be slower when the skew angle increases. Generally, a polynomial
120
Skew Plates
degree set of p = 14 is found to furnish converged solutions. This degree of polynomial will be used to calculate all subsequent buckling results. Table 6.1. Convergence study of plastic buckling stress parameters for simply supported skew plate under uniaxial compression ( = 0 , a/b=1)
h/b
p IT 8 10
= 0o
DT 4.0000 4.0000 __ __ __ 3.9848 3.9848 __ __ __ 2.7954 2.7954 2.7954 __ __
= 30 o
IT 6.0105 5.9637 5.9361 5.9184 5.9062 5.9621 5.9163 5.8893 5.8719 5.8599 4.6798 4.6069 4.5547 4.5192 4.4937 DT 6.0105 5.9637 5.9361 5.9184 5.9062 5.9544 5.9092 5.8825 5.8652 5.8534 3.0723 3.0632 3.0574 3.0535 3.0508
= 45 o
IT 10.8428 10.5315 10.3651 10.2528 10.1710 9.6919 9.4476 9.3321 9.2549 9.1977 7.8187 7.3994 7.2705 7.1860 7.1015 DT 10.8428 10.5315 10.3651 10.2528 10.171 9.3467 9.1711 9.0815 9.0202 8.9736 3.4149 3.3831 3.3696 3.3614 3.3551
4.0000 4.0000 __ __ __ 3.9850 3.9850 __ __ __ 3.4955 3.4955 3.4955 __ __
0.001
12 14 16 8 10
0.025
12 14 16 8 10
0.05
12 14 16
In order to check the correctness of the program, comparison studies are shown in Tables 6.5 and 6.6 against existing elastic buckling solutions because there are no plastic buckling solutions available in the open literature. The elastic buckling solutions can readily be obtained from the plastic buckling analysis program by setting Et = E s = E . Alternatively, elastic buckling solutions can also be obtained by setting the thickness-to-width ratio to a small value (for eg. = 0.001) in the plastic buckling analysis. This is because the plastic buckling solutions approach the corresponding
121
Skew Plates
elastic buckling solutions for relatively thin plates. It can be seen from Tables 6.5 and 6.6 that thin plate ( = 0.001) plastic buckling results agree well with all existing elastic buckling solutions for clamped skew plates. For simply supported skew plates, the present results agree with some existing solutions (Fried and Schmitt, 1972; Kitipornchai et al, 1993; Saadatpour et al., 2002) but are slightly higher than the existing solutions obtained by Saadatpour et al. (1998), Huyton and York (2001) and Wang et al. (2003). This is due to the treatment of the corner condition as discussed in the paper by Wang et al. (1992).
Table 6.2. Convergence study of plastic buckling stress parameters for simply supported skew plate under biaxial compression x = y = ( = 1 , a/b=1)
h/b
p IT 8 10
= 0o
DT 2.0000 2.0000 __ __ __ 1.9927 1.9927 __ __ __ 1.8649 1.8649 __ __ __
= 30 o
IT 2.5757 2.5597 2.5500 2.5437 2.5393 2.5634 2.5474 2.5378 2.5315 2.5271 2.1672 2.1607 2.1567 2.1541 2.1523 DT 2.5757 2.5597 2.5500 2.5437 2.5393 2.5634 2.5474 2.5378 2.5315 2.5271 2.1497 2.1436 2.1398 2.1373 2.1356
= 45 o
IT 3.8364 3.7640 3.7167 3.6836 3.6592 3.8068 3.7348 3.6874 3.6541 3.6293 2.5255 2.5093 2.4985 2.4909 2.4852 DT 3.8364 3.7640 3.7167 3.6836 3.6592 3.8068 3.7347 3.6874 3.6540 3.6293 2.4742 2.4605 2.4512 2.4446 2.4397
2.0000 2.0000 __ __ __ 1.9927 1.9927 __ __ __ 1.8713 1.8713 __ __ __
0.001
12 14 16 8 10
0.025
12 14 16 8 10
0.05
12 14 16
122
Skew Plates
Table 6.3. Convergence study of plastic buckling stress parameters for clamped skew plate under uniaxial compression x ( = 0 , a/b = 1)
h/b
p IT 8 10
= 0o
DT 10.0764 10.0739 10.0737 __ __ 8.7840 8.7833 8.7832 8.7831 __ 3.2398 3.2392 3.2392 3.2392 __
= 30 o
IT DT
= 45 o
IT 22.6083 20.617 20.2042 20.1346 20.1172 15.6945 13.6173 12.9930 12.8328 12.8009 13.1786 11.6271 11.0101 10.8042 10.7524 DT 22.6083 20.617 20.2042 20.1346 20.1172 11.7879 11.4179 11.3007 11.2756 11.2716 3.6760 3.6272 3.6098 3.6054 3.6046
10.0764 10.0739 10.0737 __ __ 8.9361 8.9352 8.9351 8.9350 __ 5.3547 5.3451 5.3441 5.3439 __
13.6543 13.6543 13.5476 13.5476 13.5399 13.5399 13.5382 13.5382 13.5378 13.5378 10.4743 10.0177 10.3926 10.3885 10.3881 10.3880 7.8350 7.3715 7.2784 7.2690 7.2680 9.9772 9.9753 9.9751 9.9750 3.4300 3.4146 3.4115 3.4113 3.4112
0.001
12 14 16 8 10
0.025
12 14 16 8 10
0.05
12 14 16
123
Skew Plates
Table 6.4. Convergence study of plastic buckling stress parameters for clamped skew plate under biaxial compression x = y = ( = 1 , a/b = 1)
h/b
p IT 8 10
= 0o
DT 5.3044 5.3037 5.3036 __ __ 5.2285 5.2282 5.2281 5.2280 __ 2.6829 2.6829 2.6828 __ __
= 30 o
IT 6.8618 6.8561 6.8550 6.8550 6.8549 6.6278 6.6254 6.6251 6.6250 6.6249 3.0443 3.0439 3.0439 3.0438 DT 6.8618 6.8561 6.8550 6.8550 6.8549 6.6178 6.6155 6.6151 6.6150 6.6150 2.8326 2.8324 2.8324 2.8324 -
= 45 o
IT 10.0546 9.9203 9.8953 9.8883 9.8854 8.5286 8.4919 8.4866 8.4855 8.4851 3.5891 3.5746 3.5726 3.5722 3.5721 DT 10.0546 9.9203 9.8953 9.8883 9.8854 8.4654 8.4308 8.4259 8.4248 8.4245 3.0328 3.0297 3.0293 3.0292 3.0292
5.3044 5.3037 5.3036 __ __ 5.2298 5.2295 5.2294 5.2293 __ 2.7957 2.7957 2.7957 __ __
0.001
12 14 16 8 10
0.025
12 14 16 8 10
0.05
12 14 16
124
Skew Plates
Table 6.5. Comparison study against existing elastic solutions for simply supported skew plates ( = 0.001) a/b
References Uniaxial compression ( = 0) Durvasula (1971) Fried and Schmitt (1972) Kitipornchai et al. (1993)
Method
= 0o
= 30 o = 45 o
Ritz Method1 F.E.M Ritz Method2
4.000 __ 4.000 4.000 __ 4.000
6.41 5.91 5.897 5.868 5.85 5.904 5.83 5.918 6.03 5.621 5.59 5.625
12.3 10.2 10.103 9.756 9.67 10.061 9.39 10.253 10.3 8.905 8.80 8.927
Saadatpour et. al. (1998) Galerkin Method Huyton and York. (2001) ABAQUS
Saadatpour et al. (2002) Ritz Method3 Wang et al. (2003) Present Study Durvasula (1971) Kitipornchai et al. (1993) Huyton and York (2001) Present Study Biaxial Compression ( = 1) 1 2
1 3
Differential 4.000 Quadrature Method Ritz Method3 Ritz Method1 Ritz Method2 ABAQUS Ritz Method3 4.000 4.000 4.000 __ 4.000
Wang et al. (1993) Present Study Wang et al. (1993) Present Study
Ritz Method3 Ritz Method Ritz method Ritz Method

3 3 3
2.000 2.000 1.250 1.250
2.544 2.544 1.616 1.616
3.684 3.684 2.356 2.356
double Fourier series; 2 orthogonal polynomial functions; simple polynomial functions
125
Skew Plates
Table 6.6. Comparison study against existing elastic solutions for clamped skew plates ( = 0.001) a/b Uniaxial compression ( = 0) Wittrick (1953) Durvasula (1970) Kitipornchai et al. (1993) Huyton and York. (2001) 1 Saadatpour et al. (2002) Wang et al. (2003) Present Study Kitipornchai et al. (1993) 2 Huyton and York (2001) Present Study Biaxial compression ( = 1) Ashton(1969) 1 Wang et al. (1993) Present Study 2
1 3
Method
= 0
= 30 = 45
Ritz Method1 Galerkin Method Ritz Method ABAQUS Ritz Method3 Differential Quadrature Method Ritz Method3 Ritz Method ABAQUS Ritz Method
3 2 2
__ 10.07 10.074 __ 10.107 10.07 10.074 7.867 __ 7.867
13.636 13.58 13.538 13.5 13.538 13.54 13.538 10.283 10.30 10.284
21.64 20.44 20.112 20.1 20.106 20.10 20.135 15.197 15.1 15.172
Ritz Method4 Ritz Method3 Ritz Method3 Ritz Method3 Ritz Method3
5.39 5.304 5.304 3.923 3.923
6.981 6.854 6.854 5.208 5.208
10.092 9.888 9.888 7.788 7.788
Wang et al. (1993) Present Study
trigonometric series; 2 double Fourier series; simple polynomial functions; 4 beam characteristic functions
Effect of transverse shear deformation and thickness-to-width ratio Tables 6.7 and 6.8 show the effect of transverse shear deformation on the plastic buckling stress parameter for simply supported and clamped skew plates under uniaxial compressive stress. The Kirchhoff plate (classical thin plate) solutions are obtained by setting the shear correction factor 2 = 1000 as discussed in Chapter 4. It
126
Skew Plates
can be seen that the effect of transverse shear deformation lowers the plastic buckling stress parameter . The effect is more pronounced as the thickness-to-width ratio and the skew angle increases. Moreover, the effect of transverse shear deformation is more significant in the case of clamped boundary condition than in the case of simply supported boundary condition. Generally, the reductions due to the effect of transverse shear deformation for the plastic buckling stress parameter x (IT) are larger than those for x (DT).
Table 6.7 Effect of transverse shear deformation on for simply supported skew plates under uniaxial compression (a/b = 1) h/b Plate theory Elastic 0.001 Kirchhoff Mindlin Difference Kirchhoff Mindlin Difference Kirchhoff Mindlin Difference Kirchhoff Mindlin Difference 5.9184 5.9184 0% 5.9184 5.8876 0.52% 5.9184 5.7978 2.04% 5.9184 5.4651 7.66%
= 30 o
IT 5.9184 5.9184 0% 5.9019 5.8719 0.51% 4.6296 4.5192 2.38% 4.6199 4.1714 9.71% DT 5.9184 5.9184 0% 5.8950 5.8652 0.50% 3.0700 3.0536 0.53% 0.9638 0.9513 1.30% Elastic
= 45 o
IT DT 10.2530 10.2530 10.2530 10.2528 10.2528 10.2528 0% 0% 0% 10.2530 10.1496 1.01% 10.2530 9.8671 3.76% 10.2530 8.8833 13.35% 9.3182 9.2549 0.68% 7.5279 7.1861 4.54% 7.5265 6.2560 16.88% 9.0724 9.0202 0.58% 3.3895 3.3614 0.83% 1.0275 1.0034 2.35%
0.025
0.05
0.1
127
Skew Plates
Table 6.8 Effect of transverse shear deformation on for clamped skew plates under uniaxial compression (a/b = 1) h/b Plate theory Elastic 0.001 Kirchhoff Mindlin Difference Kirchhoff Mindlin Difference Kirchhoff Mindlin Difference Kirchhoff Mindlin Difference
= 30 o
IT DT Elastic 13.5387 13.5387 13.5387 13.5382 13.5382 13.5382 0% 0% 0% 13.5387 10.4965 10.0531 13.2683 10.3881 9.9751 2.00% 1.03% 0.78% 13.5387 7.9399 12.5310 7.2690 7.44% 8.45% 3.4633 3.4113 1.5%
= 45 o
IT DT 20.1366 20.1368 20.1368 20.1346 20.1346 20.1346 0% 0% 0% 20.1366 13.1523 11.3825 19.4957 12.8328 11.2756 3.18% 2.43% 0.94% 20.1366 12.3635 17.8833 10.8042 11.19% 12.61% 20.1366 12.3555 13.4987 8.1688 32.96% 33.89% 3.6796 3.6054 2.02% 1.0947 1.0363 5.33%
0.025
0.05
0.1
13.5387 7.9365 1.0420 10.2673 5.9412 1.0006 24.16% 25.14% 3.97%
The variations of plastic buckling stress parameter with respect to b/h ratio are shown in Figs. 6.2 and 6.3. The thin plate buckling curves are also shown as dashed lines for comparison purposes. It can be seen that the plastic buckling stresses approach the elastic buckling stresses as the plate thickness decreases. But for larger plate thicknesses i.e. smaller b/h ratios, the plastic buckling stresses differ appreciably from the corresponding elastic buckling stresses. Generally, the IT buckling curves are higher than the corresponding DT buckling curves and the difference increases as b/h ratio decreases.
Effect of aspect ratio a/b The variations of plastic buckling stress parameter with respect to the aspect ratio are shown in Figs. 6.4 and 6.5 for, respectively, simply supported and clamped skew plates under uniaxial compressive stress. As in the rectangular plate solutions,
128
Skew Plates
there are some kinks in the buckling curves. These kinks indicate the mode switching. However, for relatively thick plates (for example = 0.05), these kinks cannot be seen clearly in the buckling curves and the buckling capacity gained due to mode switching effect is negligible. It is interesting to note that the plastic buckling stress parameter
(DT) is almost constant with respect to the aspect ratio when the thickness-to-width
ratio is large ( = 0.05 ).
Effect of skew angle The effect of skew angle on the plastic buckling stress parameter is shown in Figs. 6.6 and 6.7 for, respectively, simply supported and clamped skew plates under uniaxial compressive stress. It should be noted that the plastic buckling stress parameters for skew plates with negative skew angles are the same as those for skew plates with positive skew angles since the loading conditions are symmetric. Therefore, only the positive skew angles are considered in Figs. 6.6 and 6.7. It can be seen that the plastic buckling stress parameter increases with increasing skew angles for small thicknessto-width ratios. However, for relatively large thickness-to-width ratios (for example,
= 0.05 ), the effect of skew angle on the plastic buckling stress parameter is rather
insignificant. Generally, the difference between (IT) and (DT) increases with increasing skew angles.
129
Skew Plates
1.4
1.2
=0
30
c/0.2
0.8
0.6
=1
c
a b
Elastic IT DT
0.4
0.2
c
10 20 30 40 50 60
b/h Fig. 6.2 Normalized plastic buckling stress c/0.2 versus b/h ratio for simply supported skew plate ( = 30)
1.4
1.2
=0
30
c/0.2
0.8
=1
c
a b
0.6
0.4
c
Elastic IT DT
10 20 30 40 50 60
0.2
b/h Fig. 6.3 Normalized plastic buckling stress c/0.2 versus b/h ratio for clamped skew plate ( = 30)
130
Skew Plates
10
= xhb2/(2D)
Elastic IT DT = 0.025 = 0.05
30
2 0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 6.4 Plastic buckling stress parameter versus aspect ratio a/b for simply supported skew plate ( = 30)
30
25
Elastic IT DT = 0.025 = 0.05
30
= xhb2/(2D)
20
15
10
0 0.5 1 1.5 2 2.5 3
Aspect ratio a/b Fig. 6.5 Plastic buckling stress parameter versus aspect ratio a/b for clamped skew plate ( = 30)
131
Skew Plates
12
10
= 0.001 = 0.025 = 0.050 IT DT
a b
= xhb2/(2D)
0 0 15 30 45
Skew angle Fig. 6.6 Plastic buckling stress parameter versus skew angle for simply supported skew plate (a/b = 1)
25
20
= 0.001 = 0.025 = 0.050 IT DT
a b
= xhb2/(2D)
15
10
0 0 15 30 45
Skew angle Fig. 6.7 Plastic buckling stress parameter versus skew angle for clamped skew plate (a/b = 1)
132
Skew Plates
6.3
Shear Loading
There are two kinds of in-plane shear loadings when dealing with shear buckling problem of skew plates. The first kind of shear loading is the shear loading acting along the two horizontal edges where the traction is a pure shear stress, whereas along the two other oblique edges, the traction consists of both shear and normal stresses of such magnitude that every infinitesimal rectangular element is in a state of pure shear. The first kind of shear loading is depicted in Fig. 6.8a. The second kind is shear loading being applied uniformly along the plate edges as shown in Fig. 6.8b.
xy = R
b a
xy = R
(a)
2 xy tan
xy = S
b a
xy = S
(b) 2 xy tan
Fig. 6.8 Two kinds of shear loading conditions on skew plates: (a) R shear loading and (b) S shear loading
The first kind of shear loading is referred to as R shear loading while the second kind is referred to as S shear loading. In order to differentiate between the results of R shear and S shear loads, the subscript R will be used to indicate the former loads while
133
Skew Plates
the subscript S will denote the latter kind of load. The following symbols are used to denote the plastic buckling stress parameters
xy = R , x = y = 0 xy = s , x = 2 s tan
for R shear loading: for S shear loading.
(6.4) (6.5)
Tables 6.9 to 6.12 show the convergence studies that were carried out for the skew plates with relatively large skew angles ( = 45 and = -45) because it has been found in the previous section that more polynomial terms are required for highly skew plates. Generally, it can be seen that the 14th degree of polynomials in the displacement functions are sufficient to ensure converged solutions. Therefore, the degree of polynomial p = 14 is adopted for all subsequent computations for R shear and S shear buckling stresses. Table 6.13 shows the comparison study between the present buckling results and existing shear buckling solutions. Since no plastic shear buckling solutions are available, comparisons can only be made with elastic buckling solutions by setting
= 0.001 in the plastic buckling analysis. It can be seen that the results are close
except for some cases. For example, the present result is lower than the existing results by about 11% for simply supported skew plate of = 30 under S shear loading. It is believed that some of the previously published results may not have converged. The comparison studies verify somewhat the correctness of the present formulation and the accuracy of the present Ritz results.
134
Skew Plates
Table 6.9 Convergence study of plastic buckling parameters for simply supported skew plates under shear loading ( = 45)
= 0.001
a/b p IT DT IT
= 0.050
DT
R
8 1 10 12 14 8 2 10 12 14 9.558 9.458 9.393 9.347
S
61.600 61.482 61.464 61.451
R
9.558 9.458 9.393 9.347
S
61.600 61.482 61.464 61.451
R
2.138 2.135 2.133 2.132 2.150 2.148 2.147 2.146
S
1.840 1.840 1.840 1.840 1.766 1.766 1.766 1.766
7.698 49.983 7.619 49.979 7.566 49.979 7.528 49.978 7.861 34.201 7.775 34.148 7.736 34.132 7.709 34.129
10.153 39.517 10.083 39.392 10.039 39.361 10.007 39.353
10.153 39.517 10.083 39.392 10.039 39.361 10.007 39.353
Table 6.10 Convergence study of plastic buckling parameters for simply supported skew plates under shear loading ( = -45)
= 0.001
a/b p IT DT IT
= 0.050
DT
R
8 1 10 12 14 8 2 10 12 14 26.205 23.597 23.000 22.868 14.508 12.967 12.638 12.590
S
6.411 5.978 5.853 5.794 4.823 4.089 4.053 4.043
R
26.205 23.597 23.000 22.868 14.508 12.967 12.638 12.590
S
6.411 5.978 5.853 5.794 4.283 4.089 4.053 4.043
R
2.414 2.382 2.373 2.371 2.248 2.212 2.203 2.201
S
1.387 1.367 1.361 1.358 1.317 1.303 1.300 1.300
19.981 5.879 17.738 5.495 17.150 5.379 17.009 5.325 10.994 4.093 9.709 9.336 9.265 3.913 3.881 3.871
135
Skew Plates
Table 6.11. Convergence study of plastic buckling parameters for clamped skew plates under shear loading ( = 45)
= 0.001
a/b p IT DT IT
= 0.050
DT
R
8 1 10 12 14 8 2 10 12 14
R
2.366 2.365 2.365 2.365 2.310 2.308 2.307 2.307
S
1.897 1.897 1.897 1.897 1.836 1.836 1.836 1.836
24.108 89.471 24.051 89.241 24.042 89.170 24.039 89.161 19.555 61.626 19.198 61.394 19.170 61.380 19.168 61.376
24.108 89.470 24.051 89.241 24.042 89.170 24.039 89.161 19.555 61.626 19.198 61.394 19.170 61.380 19.168 61.376
16.459 61.769 16.422 61.769 16.418 61.769 16.417 61.769 13.823 49.709 13.655 49.709 13.638 49.685 13.637 49.685
Table 6.12. Convergence study of plastic buckling parameters for clamped skew plates under shear loading ( = -45)
= 0.001
a/b p IT DT IT
= 0.050
DT
R
8 10 1 12 13 14 8 10 2 12 13 14 36.873 32.990 31.742 31.738 22.587 20.725 19.996 19.995
S
9.856 9.376 9.375 9.285 7.628 6.892 6.872 6.739
R
36.873 32.990 31.742 31.738 22.587 20.725 19.996 19.995
S
9.856 9.376 9.375 9.285 7.628 6.892 6.872 6.739
R
2.500 2.458 2.442 2.437 2.436 2.389 2.345 2.324 2.318 2.317
S
1.463 1.441 1.433 1.432 1.431 1.445 1.404 1.386 1.385 1.381
54.049 11.940
54.049 11.940
27.904 8.968 23.528 8.258 22.176 8.080 21.636 8.067 21.588 8.047 18.887 7.768 15.836 6.490 14.600 6.109 14.181 6.081 14.148 6.032
28.826 20.205
28.826 10.205
136
Skew Plates
Table 6.13 Comparison study of elastic buckling stress parameter of skew plates under shear (a/b = 1) Skew angle -45 -30 -15 15 Researchers Present Others Present Others Present Others Present Others Present Others Present Others SSSS CCCC
R
22.868 27.7
a
S
5.794 9.348
c
R
31.738 32.56
b
S
9.285 __ 9.432 10.76d 11.0167 11.00d 22.292 22.28d 39.886 39.92d __ 89.161 __ __
16.080 17.1a 12.11 12.4a 7.539 7.61a 7.150 7.54a 7.07f 9.347 10.7a 9.04f
5.778 6.686c 6.835 7.063c 14.616 14.830c 26.873 27.674c __ 61.451 64.241c __
c
21.903 23.64b 16.963 17.24b 14.354 14.39b 16.596 16.66b __ 24.039 24.324e 24.08b
d
30
45
Durvasula (1971); b Durvasula (1970); e Wittrick (1954); fArgyis (1965)
Yoshimura (1963);
Hamada (1959);
Effect of transverse shear deformation and thickness-to-width ratio Tables 6.14 to 6.17 show the effect of transverse shear deformation on the plastic buckling stress parameter R and S for simply supported and clamped skew plates. In order to examine the effect of transverse shear deformation, the classical thin plate (Kirchhoff) solutions are presented as well. These thin plate solutions are obtained by setting the shear correction factor 2 = 1000 as discussed in Chapter 4. It can be seen that the Mindlin plate solutions are smaller than their corresponding classical thin plate solutions due to the effect of transverse shear deformation. The effect is more pronounced for the skew plates with larger thickness-to-width ratios and more
137
Skew Plates
restrained boundary conditions. Generally, the buckling stress parameters, R (Elastic and IT) and S (Elastic and IT) are more influenced by the effect of transverse shear deformation than their DT counterparts.
Table 6.14 Effect of transverse shear deformation on R for simply supported skew plates (a/b = 1) h/b Plate theory
= 30 o
IT 7.1497 7.1496 0% 6.0083 5.9823 0.43% 5.3887 5.2548 2.48% DT 7.1497 7.1496 0% 5.7139 5.6957 0.32% 2.0399 2.0296 0.50% Elastic 9.3471 9.3470 0% 9.3471 9.2590 0.94% 9.3471 9.0132 3.57%
= 45o
IT 9.3471 9.3470 0% 7.8683 7.8051 0.80% 7.7807 7.5278 3.25% DT 9.3471 9.3470 0% 6.4708 6.4505 0.31% 2.1442 2.1316 0.59%
Elastic Kirchhoff 7.1496 0.001 Mindlin 7.1496 Difference 0% Kirchhoff 7.1496 0.025 Mindlin 7.0953 Difference 0.76% Kirchhoff 7.1496 0.050 Mindlin 6.9371 Difference 2.97%
Table 6.15 Effect of transverse shear deformation on S for simply supported skew plates (a/b = 1) h/b Plate theory
= 30 o
Elastic 26.8742 26.8731 0% 26.8742 26.2551 2.30% 26.8742 24.5714 8.57% IT DT 26.8742 26.8742 26.8731 26.8731 0% 0% 22.8008 7.0310 22.3488 7.0033 1.98% 0.39% 22.7996 2.1029 21.1046 2.0761 7.43% 1.27% Elastic 61.4549 61.4509 0.01% 61.4549 59.0707 3.88% 61.4549 53.002 13.75%
= 45o
IT 61.4549 61.4509 0.01% 56.0180 54.0199 3.57% 54.2074 48.6943 10.17% DT 61.4549 61.4509 0.01% 6.4049 6.3663 0.60% 1.8792 1.8400 2.09%
138
Skew Plates
Table 6.16 Effect of transverse shear deformation on R for clamped skew plates (a/b = 1) h/b Plate theory
= 30 o
Elastic 16.5968 16.5960 0% 16.5968 16.1519 2.68% 16.5968 14.9718 9.79% IT DT 16.5968 16.5968 16.5960 16.5960 0% 0% 13.0423 7.4363 12.7553 7.3948 2.20% 0.56% 13.0382 2.3052 11.9687 2.2715 8.20% 1.46% Elastic 24.0404 24.0389 0.01% 24.0404 23.1649 3.64% 24.0404 20.9019 13.06%
= 45o
IT 24.0404 24.0389 0.01% 18.6331 18.0123 3.33% 18.6309 16.4171 11.88% DT 24.0404 24.0389 0.01% 7.9704 7.9122 0.73% 2.4160 2.3654 2.09%
Table 6.17 Effect of transverse shear deformation on S for clamped skew plates (a/b = 1) h/b Plate theory
= 30 o
Elastic 39.8888 39.8860 0.01% 39.8888 38.2375 4.14% 39.8888 34.1733 14.33% IT DT 39.8888 16.5968 39.8860 16.5960 0.01% 0% 35.4582 7.4363 34.1627 7.3948 3.65% 0.56% 35.4578 2.3052 30.8872 2.2715 12.89% 1.46% Elastic 89.1713 89.1609 0.01% 89.1713 83.3027 6.58% 89.1713 70.2337 21.24%
= 45o
IT 89.1713 89.1609 0.01% 82.8029 77.7362 6.12% 71.2645 61.8897 13.15% DT 89.1713 89.1609 0.01% 6.7018 6.6344 1.01% 1.9607 1.8966 3.27%
The variations of plastic buckling stresses, R shear and S shear, with respect to b/h ratios are shown in Figs. 6.9 and 6.10, respectively, for simply supported and clamped skew plates with = 30 o . The buckling curves based on the classical thin plate theory are presented as dashed lines for comparison purposes. Owing to the fact that the S shear loading is equivalent to the R shear loading plus a uniaxial tensile loading for
> 0 , the plastic buckling strength of plate in S shear case should be higher than that
139
Skew Plates
in R shear case. This fact is always correct for elastic and IT buckling curves. However, it can be seen from Fig. 6.10 that, S (DT) may be smaller than R (DT) for small b/h ratios. This is due to the presence of high tensile stress.
Effect of aspect ratio a/b The variations of shear plastic buckling stress parameters, R and S , with respect to aspect ratios are shown in Figs. 6.11 and 6.12. For R shear case, the kinks (that indicate buckling mode switching) along the curves are obvious for both IT and DT buckling curves. However, these kinks for DT buckling curves fade away as the thickness-to-width ratio increases. On the other hand the locations of these kinks for S shear case are not as obvious as in R shear case. It can be seen that R and S based on DT are more sensitive to the thickness-towidth ratios than their IT counterparts. Generally, as discussed before, the shear plastic buckling stress parameters based on IT are found to be higher than those based on DT. The difference is more pronounced for the skew plates with large thickness-to-width ratios and small aspect ratios (i.e. a / b < 1 )
Effect of skew angle Figures 6.13 to 6.16 show the variations of R and S with respect to the skew angle for simply supported and clamped skew plates. It can be seen that R (IT) decreases first and then increases when the skew angle increases from -45 to +45. Note that the plastic buckling results for positive shear loading of skew plates with negative skew angles will be same as those for negative shear loading of skew plates with positive skew angles. For a given aspect ratio and thickness ratio, R (IT) for a
140
Skew Plates
skew plate with negative skew angle are higher than those for a skew plate with a corresponding positive skew angle. In another word, for a skew plate with positive skew angle, R (IT) in the case of positive shear loading is lower that its negative shear loading counterpart. Therefore, the direction of the shear stresses that cause a lower plastic buckling stress parameter R (IT) is the one that tends to increase the skewness of the plate. On the other hand, the values of S (IT) increase with increasing skew angles from -45 to +45. The variation of R (DT) and S (DT) with respect to the skew angle is the same as their IT counterparts for small thickness-to-width ratios. However, for larger thickness-to-width ratios, the variations of R (DT) and S (DT) with respect to the skew angle are almost constant.
S shear
30
0.75
a b Elastic IT DT
xy/0.2
0.5
R shear
0.25
a b
30
0 20 40 60 80 100 120
b/h Fig. 6.9 Normalized buckling stress xy/0.2 versus b/h ratio for simply supported skew plate (a/b =1)
141
Skew Plates
1
30 a b
S shear
0.75
xy/0.2
0.5
Elastic IT DT
R shear
a
0.25
b 30
0 25 50 75 100 125 150
b/h Fig. 6.10 Normalized buckling stress xy/0.2 versus b/h ratio for clamped skew plate (a/b = 1)
14
40
12
= 0.015 = 0.025 = 0.050 a b
IT DT
30
30
a b
10
xy
30
20
10
2
0 0.5 1 1.5 2 2.5 3 3.5 4
0 0.5 1 1.5 2 2.5 3 3.5 4
Aspect ratio a/b
Aspect ratio a/b
Fig. 6.11. Variation of plastic buckling stress parameters with respect to aspect ratio for SSSS skew plate with = 30
142
Skew Plates
20
= 0.01 = 0.02 = 0.03

15
60
IT DT
30 a b
50
R
10
40
30
20
5
a b
0 0.5 1 1.5 2 2.5
10
30
3 3.5 4
0 0.5 1 1.5 2 2.5 3 3.5 4
Aspect ratio a/b
Aspect ratio a/b
Fig. 6.12. Variation of plastic shear buckling stress parameters with respect to aspect ratio for CCCC skew plate with = 30
25
20
= 0.001 = 0.025 = 0.050 IT DT
a b +
15
10
0 -45 -30 -15 0 15 30 45
Skew angle Fig. 6.13. Variation of plastic (R shear) buckling stress parameters with respect to skew angle of SSSS skew plate
143
Skew Plates
30
= 0.001 = 0.025 = 0.050

20
a b +
IT DT
S
10
0 -45 -30 -15 0 15 30 45
Skew angle Fig. 6.14. Variation of plastic (S shear) buckling stress parameters with respect to skew angle of SSSS skew plate
35 30
= 0.001 = 0.02 = 0.03 IT DT
a b +
25
20
15 10
0 -45 -30 -15 0 15 30 45
Skew angle Fig. 6.15. Variation of plastic (R shear) buckling stress parameters with respect to skew angle of CCCC skew plate
144
Skew Plates
30
a b
20
= 0.001 = 0.02 = 0.03 IT DT
S
10
0 -45 -30 -15 0 15 30 45
Skew angle Fig. 6.16. Variation of plastic (S shear) buckling stress parameters with respect to skew angle of CCCC skew plate
6.4
Concluding Remarks
In this chapter, plastic buckling analysis of skew plates was carried out using the Ritz method based on the skew coordinate system. In order to include the effect of transverse shear deformation, the Mindlin plate theory was adopted. The plastic buckling behaviour of skew plates under unaxial, biaxial and shear stresses are examined. Generally, the effect of transverse shear deformation lowers the buckling stress parameters and the effect is more pronounced in the case of the plastic buckling stress parameters obtained by using IT than in the case of their DT counterparts. The effects of aspect ratios, thickness to width ratios, and skew angles on the plastic buckling stress parameters were discussed.
145
CHAPTER 7
CIRCULAR AND ANNULAR PLATES
7.1
Introduction
Circular and annular plates are used in turbine disks, submarines, instrument mounting bases for space vehicles and microelectromechanical systems (MEMS) devices (Su and Spearing, 2004). Their buckling capacities need to be calculated when applied in situations where in-plane compressive of forces act on the plates. The elastic axisymmetric buckling problems (where the deflection surface is assumed to be axisymmetric) of circular and annular plates have been studied by many researchers. (Raju and Rao, 1989; Wang et al., 1993c; Liew et al., 1994a; Laura et al., 2000). The assumption of axisymmetric buckling, however, does not necessarily lead to the desired lowest critical load. It had been found that the buckling mode shape for an
146
Circular and Annular Plates
annular plate, which corresponds to the lowest critical load, may be axisymmetric or asymmetric depending on the inner radius to outer radius ratio and edge boundary conditions (Yamaki, 1958; Majumdar, 1971). Therefore, the asymmetric buckling problems of circular and annular plates have also been studied (Gerard, 1978; Thevendran and Wang, 1996; Wang, 2003). All the aforementioned studies, however, are confined to the elastic buckling problems. The elastoplastic buckling of annular plate under pure shear was studied by Ore and Durban (1989). Recently, Kosel and Bremec (2004) investigated the elastoplastic buckling of annular plates under uniform loading by using the finite different method. These studies are based on the classical thin plate (Kirchhoff) theory which neglects the effect of transverse shear deformation. When dealing with relatively thick plates, it is necessary to use thick plate theories such as the Mindlin plate theory for accurate buckling load. Otherwise the buckling load will be overestimated because the neglect of the effect of transverse shear deformation stiffens the plates. In this chapter, the plastic buckling of circular and annular Mindlin plates is studied. The Ritz method using the polar coordinate system is employed for the analysis and the program is coded in MATHEMATICA language (see Appendix A for the program called BUCKRITZPOLAR). In order to verify the numerical plastic buckling results, an analytical approach is presented for the asymmetric plastic buckling problems of Mindlin circular and annular plates There are hitherto no analytical solution is available for the asymmetric buckling problems of Mindlin circular and annular plates, although an analytical method for elastic buckling was presented by Yamaki (1958) based on the classical thin plate theory. Readers who are interested in the analytical approach for elastic buckling problems of such Mindlin plates may refer to the paper by Wang and Tun
147
Myint Aung (2005). In the next section, the analytical approach for solving plastic buckling of circular and annular plates under uniform stress is featured.
7.2
Analytical Method
Consider a circular or annular Mindlin plates subjected to a uniform radial compressive stress i.e. rr = = . The governing equations for the buckling of Mindlin plate under uniform compressive stresses shown in Eqs. (2.36a-c) can be expressed in the polar coordinate system as
1 Gh h 2 w + 2 Gh r + r + r r r
= 0
(7.1a)
Eh 3 r
1 1 r 2r Eh 3 + 12 r 2 r 2 r r r 2 12 w Gh 3 + Gh r + r 12
2
1 2 r r = 0
1 2 r
2r 1 2 1 2 + r r r 2
(7.1b)
Eh 3 1 2 12 r 2 2
Gh 3 + 12
1 r Eh 3 + 2 r + 12
1 2r 1 w 2 r Gh + r r
1 2r 1 r 2 1 + 2 =0 + 2 + r r r 2 r r r r
(7.1c)
where 2 () = 2 () / r 2 + (1 / r ) () / r + 1 / r 2 2 () / 2 . Following the work of Mindlin (1951), the rotations ( r , ) can be defined in terms of potentials ( , H ) as
r =
1 H + r r
(7.2a)
148
1 H r r
(7.2b)
In view of Eqs. (7.2a-b) and after some simplifications, Eqs. (7.1a-c) become h 2 + 1 2 w = 0 Gh
2 12 2 G 2 ( + w) 1 2G 1 2 2 E r r Eh r r
(7.3a)
1 r
G 2 12 2 G 2G 2 H H + H 2 =0 E r 2 h E E
(7.3b)
1 2 12 2 G 2G 2 ( + w) + 2 E r r Eh 2 12 2 G 1 2G 1 2 H G 2 H H 2 r r E r 2 = 0 r E h E
(7.3c)
For plates under uniform radial compressive stresses, i.e. rr = = , the value of the expression 2G / E = 0 for both IT and DT. In view of this, Eqs. (7.3b) and (7.3c) become, respectively,
2 2 12 2 G ( + w) + G 2 12 rE r Eh 2 h2
H = 0
(7.4a)
and
1 r
2 2 12 2 G ( + w) G 2 12 Eh 2 h2 E r
H = 0
(7.4b)
The potential H may be separated from and w by differentiation, addition, and subtraction of Eqs. (7.4a) and (7.4b). These equations become 149
12 2 2 2 2 h
H = 0
(7.5a)
12 2 G ( + w) = 0 2 2 2 Eh
(7.5b)
It now remains to separate and w between Eq. (7.3a) and Eq. (7.5b). It is observed that these equations are satisfied by
= ( 1) w
(7.6)
where is a constant. The constant can be obtained by substituting Eq. (7.6) into Eqs. (7.3a) and (7.5b)
1 = 0
(7.7a)
2 =
h 2 Gh
(7.7b)
Therefore, Eq. (7.5) can be rewritten as
= ( 1 1) w1 + ( 2 1) w2
In view of Eqs. (7.7a-b) and (7.8), Eq. (7.5b) can be written as
2 w1 = 0 2 Eh 3 12 h w2 = 0 h 1 2 Gh
(7.8)
(7.9a)
(7.9b)
For clarity, w1 , w2 and H will be denoted as 1 , 2 and 3 , respectively. For generality and convenience, the following nondimensional terms are introduced 150
w=
r=
w R
r R h R
(7.10a)
(7.10b)
(7.10c)
hR 2 = 2 D
(7.10d)
where w is the normalized deflection, r the normalized radial coordinate and the thickness-to-radius ratio. Therefore, in view of Eqs. (7.2a-b), the transverse deflection w and the rotations ( r , ) can be defined in terms of three potential functions Q1, Q2, Q3 as
2 1 3 1 2 E + + 1 2 2 r 12(1 ) G r r 1 2 3 1 1 2 E + 1 2 2 r 12(1 ) G r r
r =
(7.11a)
(7.11b) (7.11c)
w = 1 + 2
The governing equations, ie. Eqs. (7.5a), (7.9a-b) can be rewritten in terms of nondimendional terms 1 , 2 and 3 :
2 1 = 0
(7.12a) (7.12b)
(7.12c)
+ 22 2 = 0
+ 32 3 = 0
where
151
22 =
2 E 1 1 12(1 2 ) 2 G
(7.13a)
=
2 3
12 2
(7.13b)
It can be seen that Eqs. (7.12a-c) are Bessel equations and the general solutions of these equations are given by (Chiang, 1995) log r 1 = A1 r n cos(n ) + B1 n cos(n ) r 2 = A2 Rn ( 2 r ) cos(n ) + B2 S n ( 2 r ) cos(n ) 3 = A3 Rn ( 3 r ) sin( n ) + B3 S n ( 3 r ) sin(n )
(7.14a) (7.14b) (7.14c)
where the upper form of Eq. (7.14a) is used for n = 0 (axisymmetric buckling) and the lower form for n 0 (asymmetric buckling), n is the number of nodal diameters. The functions Rn ( j ), S n ( j ) , and the variables 1 , 2 are given by
j = j Im( j )
if j2 0 if j2 < 0
j = 2,3
(7.15)
J n ( j r ) if j2 0 Rn ( j r ) = 2 I n ( j r ) if j < 0
2 Yn ( j r ) if j 0 S n ( j r ) = 2 K n ( j r ) if j < 0
j = 2,3
(7.16)
j = 2,3
(7.17)
The parameters A1 , A2 , A3 , and B1 , B2 , B3 are unknown constants, J n () and I n () are Bessel functions of the first kind and the modified first kind of order n, respectively, and Yn () and K n () are Bessel functions of the second kind and the modified second
152
kind of order n, respectively. In the sequel, plastic buckling problems of circular and annular plates with and without internal ring supports will be solved by using the Ritz method and the aforementioned analytical method.
7.3
Circular plates
Consider a moderately thick, isotropic, circular plate of radius R, uniform thickness h, Youngs modulus E, shear modulus G, and Poissons ratio , as shown in Figure 7.1. The plate is subjected to a uniform compressive in-plane stress along its circular edge.
Fig. 7.1 Clamped circular plate under in-plane compressive stress
153
For numerical results, an aluminum alloy Al 2024 T6 is adopted with the following properties: Poissons ratio = 0.32 , E = 10.7 10 6 psi, 0 = 0.2 = 61.4 ksi, and the Ramberg-Osgood parameters k = 0.3485, c = 20. The following typical boundary conditions at the edges are considered: Simply supported edge
w = 0 ; M rr = 0 ; = 0
(7.18a-c)
Clamped edge
w = 0 ; r = 0 ; = 0
(7.19a-c)
Free edge w =0
M rr = 0 ; M r = 0 ; Qr + h
(7.20a-c)
where the bending moment M rr , twisting moment M r and shear force Qr are given by
Eh 3 12 r r + r r +
M rr =
(7.21a)
M r =
Gh 3 1 r r + r 12
(7.21b)
w Qr = 2 Gh + r r
(7.21c)
154
The Ritz method and the developed analytical method will be used to determine the plastic buckling stress parameter so as to provide an independent check on the solutions.
Ritz Method
In order to avoid singularity problems in the case of circular plate, needs to be set to a very small number (e.g. 10-7) and the inner edge is made free. The inputs needed in the program code are the material properties such as c, k, E / 0 , geometric properties such as h/R and the basic functions (boundary conditions). The basic
functions bw , br , and b , which are the product of boundary equations with
appropriate power, are given by
bw = ( 1)1 ( + 1) 0 br = ( 1) 0 ( + 1) 0 b = ( 1)1 ( + 1) 0 bw = ( 1)1 ( + 1) 0 br = ( 1)1 ( + 1) 0 b = ( 1)1 ( + 1) 0
for simply supported circular plates
(7.22)
for clamped circular plates
(7.23)
Analytical Method
For the circular plate, the arbitrary constants B1 , B2 , B3 in Eq. (7.14a-c) must vanish in order to avoid singularity for the displacement fields w , r , at the center of the plate, since log (r ) , Yn ( 2 r ) and K n ( 2 r ) are infinite as r approaches zero.
155
In view of Eqs. (7.14a-c), a set of homogeneous equations can be derived by implementing the foregoing boundary conditions and these equations can be expressed in the following matrix form
[K ]33 {}31 = {0}3 x1

where {}31 = [ A1
A2
T
(7.24)
A3 ] . It should be noted that, in the case of axisymmetric
buckling, and the derivative with respect to are zero everywhere and Eq. (7.24) becomes:
[K ]22 {}21 = {0}2 x1

where {}31 = [ A1
T
(7.25)
A2 ] . The plastic buckling stress parameter is evaluated by
setting the determinant of [K ] in Eq. (7.24) or Eq. (7.25) to be zero. The elements of the [K ] matrix are given in the Appendix B. Therefore one can easily write the algorithm in other programming languages by using these matrix elements.
Results
The convergence and comparison studies are shown in Table 7.1 for simply supported and clamped circular plates. It can be seen that the degree of polynomial p = 8 is found to furnish the converged Ritz results and these results are in total agreement with the analytical solutions.
156
Table 7.1. Convergence and comparison study of plastic buckling stress parameter for circular plates under uniform radial compression
BC
4 6 Simply 8 supported Analytical 0.4305 Method 4 6 Clamped 8 Analytical Method 1.5031 1.4876 1.4876 1.4876
= 0.001 IT DT 0.4306 0.4306 0.4305 0.4305 0.4305 0.4305

0.4305 1.5031 1.4876 1.4876 1.4876
= 0.05 IT DT 0.4292 0.4292 0.4291 0.4291 0.4291 0.4291

0.4291 1.4837 1.4715 1.4715 1.4715 0.4291 1.4837 1.4715 1.4715 1.4715
= 0 .1 IT DT 0.4240 0.4240 0.4240 0.4240 0.4240 0.4240

0.4240 0.6331 0.6321 0.6321 0.6321 0.4240 0.6130 0.6125 0.6125 0.6125
7.4
Annular plates
Consider a moderately thick, isotropic annular plate of outer radius R and inner radius
b, uniform thickness h. The plate is subjected to a uniform compressive in-plane stress
along its edges as shown in Fig. 7.2. The problem at hand is to determine the
plastic buckling stress of such plates. Owing to the practical importance thereof, an extensive study on the plastic buckling behaviour of annular Mindlin plates is conducted for all possible combinations of clamped (C), simply supported (S), free (F) boundary conditions at the outer and inner edges. For brevity, the boundary conditions of annular plates are denoted by two letter symbols, eg. C-S denotes an annular plate with clamped (C) outer edge and simply supported (S) inner edge.
157
Fig. 7.2 C-S annular plate under in-plane loading
Ritz Method
The computer program called BUCKRITZPOLAR is used to determine the plastic buckling stress parameter of the aforementioned buckling problem of annular plate. In
order to demonstrate the input expressions for the basic functions bw , br , and b , the
sample expressions of basic functions for C-S annular plates are given by
bw = ( 1)1 ( + 1)1 = ( 1)( + 1) br = ( 1)1 ( + 1) 0 = ( 1)

b = ( 1)1 ( + 1)1 = ( 1)( + 1)
(7.24a) (7.24b) (7.24c)
158
Analytical Method
In view of Eqs. (17.14), a set of homogeneous equations can be derived by implementing the boundary conditions (7.18a-c, 7.19a-c, 7.20a-c) at the both edges of annular plate and the system of equations can be expressed in the following matrix form
[K ]66 {}61 = {0}6 x1

where {}61 = [A1
A2 A3 B1 B2
T
(7.25)
B3 ] . The elements of the [K ] matrix are
given in the Appendix B. The plastic buckling stress parameter is evaluated by setting the determinant of [K ] in Eq. (7.25) to be zero.
Results and Discussions

Convergence studies of the Ritz results are shown in Table 7.2 for annular plate with
= b / R = 0.4 . It can be seen that converged results are obtained when the degree of
polynomial p = 10. Table 7.3 compares the results obtained by using the Ritz method and the analytical method with the existing elastic buckling solutions. It can be seen that the analytical results agree well with the corresponding Ritz buckling solutions. Moreover, it is found that these plastic buckling solutions agree well with the existing elastic solutions when the thickness to outer radius ratio = 0.001.
159
Table 7.2 Convergence study of plastic buckling stress parameter for annular plates ( = b / R = 0.4 )
BC
p
4 6 8 10 12 4 6 8 10 12 4 6 8 10 12 4 6 8 10 12
SS
SF
CC
FC
= 0.001 IT DT 2.9700 2.9700 2.9652 2.9652 2.9652 2.9652 2.9652 2.9652 2.9652 2.9652 0.2138 0.2138 0.2137 0.2137 0.2137 0.2137 0.2137 0.2137 0.2137 0.2137 11.8197 10.3372 10.3370 10.3356 10.3345 10.3343 10.3343 10.3343 10.3343 10.3343 0.7688 0.7688 0.6829 0.6829 0.6803 0.6803 0.6802 0.6802 0.6802 0.6802
= 0.05 IT DT 2.3549 2.3241 2.3534 2.3227 2.3534 2.3227 2.3534 2.3227 2.3534 2.3227 0.2135 0.2135 0.2133 0.2133 0.2133 0.2133 0.2133 0.2133 0.2133 0.2133 3.8111 3.1168 3.7826 3.1166 3.7821 3.1166 3.7821 3.1166 3.7821 3.1166 0.6677 0.6677 0.6632 0.6632 0.6623 0.6623 0.6622 0.6622 0.6622 0.6621
= 0 .1 IT DT 1.0255 0.7995 1.0251 0.7993 1.0251 0.7993 1.0251 0.7993 1.0251 0.7993 0.2125 0.2125 0.2124 0.2124 0.2124 0.2124 0.2124 0.2124 0.2124 0.2124 3.2344 0.9390 3.2123 0.9390 3.2119 0.9390 3.2119 0.9390 3.2119 0.9390 0.5758 0.5622 0.5742 0.5609 0.5741 0.5608 0.5741 0.5608 0.5741 0.5608
160
Table 7.3 Comparison study of plastic buckling stress parameter = hR 2 /( 2 D ) for annular plates ( = b / R = 0.4 ). Thickness-to-outer radius ratio, 0.05 0.1 IT DT IT 3.407 (3) 2.586 (3) 3.407 (3) 2.586 (3) __ __ 3.120 (3) 3.127 (3) __
BC Reference Analytical Method Ritz Method Wang et al. (1994b) Yamaki (1958) Analytical Method Ritz Method Wang et al. (1994b) Yamaki (1958) Analytical Method Ritz Method Wang et al. (1994b) Yamaki (1958) Analytical Method Ritz Method Wang et al. (1994b) Yamaki (1958) Analytical Method Ritz Method Wang et al. (1994b) Yamaki (1958)
0.001 Elastic 10.335 (3) 10.334 (3) 10.315 (3) 10.336 (3) 6.465 (0) 6.465 (0) 6.486 (0) 6.485 (0) 1.089 (0) 1.089 (0) 1.103 (0) 1.103 (0) 2.961 (0) 2.961 (0) 2.964 (0) 2.965 (0) 5.302 (0) 5.302 (0) 5.289 (0) 5.296 (0)
DT 0.708 (3) 0.708 (3) __
C-C
2.700 (1) 2.497 (1) 2.700 (1) 2.498 (1) __ __ __ __
2.386 (1) 2.387 (1) __ __ 0.648 (0) 0.648 (0) __ __ 0.948 (0) 0.949 (0) __ __ 1.427 (0) 1.428 (0) __ __
0.696 (1) 0.696 (1) __ __ 0.610 (0) 0.611 (0) __ __ 0.654 (0) 0.654 (0) __ __ 0.676 (0) 0.676 (0) __ __
C-S
1.080 (0) 1.080 (0) 1.080 (0) 1.080 (0) __ __ __ __
C-F
2.232 (0) 2.224 (0) 2.233 (0) 2.224 (0) __ __ __ __
S-S
2.432 (0) 2.399 (0) 2.432 (0) 2.399 (0) __ __ __ __
S-C
161
Table 7.3 Continued Thickness-to-outer radius ratio, 0.05 0.1 IT DT IT 0.212 (0) 0.212 (0) 0.212 (0) 0.212 (0) __ __ __ __ 0.211 (0) 0.211 (0) __ __ 0.543 (1) 0.543 (1) __ __ 0.193 (1) 0.193 (1) __ __
BC
Reference
0.001 Elastic 0.212 (0) 0.212 (0) 0.214 (0) 0.213 (0) 0.671 (2) 0.672 (2) 0.680 (2) 0.685 (2) 0.196 (1) 0.196 (1) 0.199 (1) 0.199 (1)
DT 0.211 (0) 0.211 (0) __ __ 0.542 (1) 0.542 (1) __ __ 0.193 (1) 0.193 (1) __ __
Analytical Method Ritz Method S-F Wang et al. (1994b) Yamaki (1958) Analytical Method Ritz Method F-C Wang et al. (1994b) Yamaki (1958) Analytical Method Ritz Method F-S Wang et al. (1994b) Yamaki (1958)
0.653 (2) 0.653 (2) 0.653 (2) 0.653 (2) __ __ __ __
0.195 (1) 0.195 (1) 0.195 (1) 0.195 (1) __ __ __ __
The variations of plastic buckling stress parameter with respect to aspect ratios
are shown in Figs. 7.3 to 7.10 for various boundary conditions. It can be seen that
plastic buckling stress parameters increase with increasing value of except for S-F and F-S annular plates. In the S-F annular plate, the plastic buckling stress parameters decreases monotonically while in F-S annular plates, the buckling stress increases first and later decreases as increases. It can be seen that the number of nodal diameters of the buckled surface of the plate increases in C-C annular plate as the aspect ratio increases, while in C-S, S-S, and F-S annular plates, the plates first buckle in asymmetric modes with one nodal
162
diameter (i.e. n = 1) and after a particular value of ratio, the buckle mode takes on an axisymmetric forms (i.e. n = 0). The sample critical buckling mode shapes are shown in Fig. 7.11 for C-C annular plates of various aspect ratios . Apart from F-C annular plate, the remaining types of annular plates always buckle in an axisymmetric mode. In F-C annular plate, the number of nodal diameter n changes from one to two and then changes back to one and followed by zero as the ratio increases. It is interesting to note that, as far as axisymmetric buckling (n = 0) is concerned, the buckling stress parameters are identical for S-F and F-S annular plates for a particular thickness ratio, irrespective of the hole size i.e. . The differences between the results from DT and IT plasticity theories are also examined. It has been founded that the discrepancies increase with increasing aspect ratio and thickness-to-outer radius ratio . This is because the buckling stress parameter increases with increasing aforementioned ratios and consequently the effective modulus used in IT and DT progressively differ each other. However, in S-F and F-S annular plates, the buckling stress parameters from both theories are identical for the considered thickness-to-outer radius ratios. This is because the buckling stresses for these boundary conditions are much lower than nominal yield stress for the considered thickness-to-outer radius ratios due to the less restrained boundary conditions. However, for greater thickness-to-outer radius ratios, the results obtained by IT and DT may differ each other since the plastic buckling stress increases with increasing thickness-to-outer radius ratio. It is interesting to note that the buckling stress parameters obtained by DT do not much vary with respect to ratio when the boundary conditions are either simply supported or clamped and thickness-to-outer radius ratio is large.
163
12
n = 3
10 8 6 4 2 0 0.0
n
DT & IT ( = 0.001)
n=
= hR2/2D
n=
IT ( = 0.1)
5
n=
=1
IT ( = 0.05)
n
3
5
DT ( = 0.05)
n =1
n =1
n=2
n=
=4
n=4
n=5
=
DT ( = 0.1)
n=2
n=3
n =4 n=5
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Fig. 7.3 Variation of buckling stress parameter with respect to ratio for C-C annular plate
12 10 8
IT & DT ( = 0.001)
IT ( = 0.1)
= hR2/2D
n =1
DT ( = 0.05)
n =1
1 n=
n=
n=0
2
n =1
DT ( = 0.1)
n=0
0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Fig. 7.4 Variation of buckling stress parameter with respect to ratio for C-S annular plate
n=
IT ( = 0.05)
164
DT ( = 0.05) IT & DT ( = 0.001) IT ( = 0.05) IT ( = 0.1)

n=0
n=0
= hR2/2D
DT ( = 0.1)
0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Fig. 7.5 Variation of buckling stress parameter with respect to ratio for C-F annular plate
IT & DT ( = 0.001)
IT ( = 0.05) IT ( = 0.1)
= hR2/2D
DT ( = 0.05)
2
n =1
n=0 n=0
DT ( = 0.1)
n =1
0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Fig. 7.6 Variation of buckling stress parameter with respect to ratio for S-S annular plate
165
12
10
IT & DT ( = 0.001)
IT ( = 0.05) IT ( = 0.1)
= hR2/2D
6
n=
n =1
DT ( = 0.05) DT ( = 0.1)
0.4 0.5 0.6 0.7 0.8
n=0 n=0
2
n =1
0 0.0 0.1 0.2 0.3
Fig. 7.7 Variation of buckling stress parameter with respect to ratio for S-C annular plate
0.5
IT & DT ( = 0.001, 0.05)
0.4
= hR2/2D
0.3
IT & DT ( = 0.1)
n=
0.2
0.1
0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Fig. 7.8 Variation of buckling stress parameter with respect to ratio for S-F annular plate
166
3.0
DT ( = 0.05) IT ( = 0.05) IT & DT ( = 0.001)

= 1
2.5
n=
2.0
= hR2/2D
1.0
n= 2
n= 2
n=0
0.5
n =1
n=2
n =1
0.0 0.0 0.1 0.2 0.3 0.4 0.5
0.6
=1
1.5
IT ( = 0.1)
DT ( = 0.1)
0.7 0.8
Fig. 7.9 Variation of buckling stress parameter with respect to ratio for F-C annular plate
0.25
IT & DT ( = 0.001)
n= 1
0.20
n= 0
= hR2/2D
0.15
0.10
IT & DT ( = 0.05) IT & DT ( = 0.1)
0.05
0.00 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Fig. 7.10 Variation of buckling stress parameter with respect to ratio for F-S annular plate
167
= 0.05
= 0.2
= 0.4
= 0.5
= 0.6
= 0.65
Fig. 7.11. Critical buckling mode shapes for C-C annular plates of various ratios (DT, = 0.05)
7.5
Circular and annular plates with intermediate ring supports
Consider a circular of radius R and thickness h or an annular plate of outer radius R, inner radius b and thickness h under uniform compressive stress . The plate may have simply supported or clamped edge support conditions and m concentric ring supports with radii b1, b2,, bm, as shown in Fig. 7.12. The rings are assumed to be rigid so that zero deflection is imposed along the ring supports. The problem at hand is to determine the plastic buckling stress of such ring supported circular or annular plates. The analysis will be carried out by using both Ritz method and developed analytical method.
168
h b1 R
(a)
b R bm
bm b2 b1
(b)
Fig. 7.12 Simply supported (a) annular plate (b) circular plate with multiple concentric ring supports
Ritz Method
The plastic buckling stress parameter for the circular and annular plates with internal ring supports can be easily obtained by modifying the basic function for transverse deflection in the Ritz program called BUCKRITZPOLAR. The modification for internal supports in the basic functions of the Ritz method for general shapes will be discussed in Chapter 9. The sample modified basic functions for the C-F annular plate with one internal ring support at 1 = b1 / R are given by:
bw = ( 1)1 ( + 1) 0 [ (2 1 1 ) /(1 )]
(7.26)
169
Analytical Method
The circular or annular plate with m ring support can be divided into m+1 segments. The governing equations of each segment are given by Eq. (7.1a-c) and the equations for each segments will be denoted by a superscript i, where i is the integer number which counts the number of segments starting from the outer edge. Not that for
m circular plates with internal ring supports, the arbitrary constants B1m +1 , B2 +1 , B3m +1
must vanish in order to avoid singularity for the displacement fields w m +1 , rm +1 , m +1 at the center of the m+1th circular segment. The boundary conditions for the periphery edge of the circular plates or the inner and outer edges of the annular plates are given in Eqs. (7.18 to 7.20). In order to satisfy the ring support conditions, and to ensure the continuity requirements of deflections, rotations and moments along the interface of the ith segment and the (i+1)th segment, the following essential and natural boundary conditions must be satisfied between the two segments.
wi = 0
(7.27a) (7.27b) (7.27c) (7.27d) (7.27e) (7.27f)
w i +1 = 0
i i +1 = 0 ri ri +1 = 0
i i M rr M rr+1 = 0
+ M ri M ri 1 = 0
In views of Eqs. (7.14a-c), a homogeneous system of equations can be derived by implementing the boundary conditions at the edges and the interface conditions between the segments as follow: 170
for annular plates with internal ring supports
[K ]6( m+1)6( m+1) {}6( m+1)1 = {0}6( m+1)1

for circular plates with internal ring supports
(7.28a)
[K ]6( m+1)36( m+1)3 {}6( m+1)31 = {0}6( m+1)31
(7.28b)
The plastic buckling stress parameter can be determined by setting the determinant of matrix [K ] in Eq. (7.28a) or (7.28b) to be zero.
Results and Discussion

Tables 7.4 and 7.5 show the convergence studies of plastic buckling stress parameter obtained by using the Ritz method for annular and circular plates with an internal ring support and the results are compared against the analytical solutions. Note that the bracketed values shown in the tables indicate the number of nodal diameter for the critical buckling mode shape. It can be seen that the converged Ritz results agree well with the analytical solutions. Although the formulations were presented to cater for multiple supports, only the plastic buckling behaviour of circular and annular plates with a single ring support will be studied in this section.
171
Table 7.4 Convergence study of plastic buckling stress parameter for annular plates with one concentric ring support ( = 0.2, 1 = 0.6 )
BC
p
4 6 8 10 12 Analytical Method 4 6 8 10 12 Analytical Method 4 6 8 10 12 Analytical Method 4 6 8 10 12 Analytical Method
= 0.001
IT 8.6983 6.5984 6.5603 6.5592 6.5589 6.5586 (0) 1.8805 1.8695 1.8681 1.8679 1.8678 1.8677 (0) 13.4819 12.7939 12.7793 12.7792 12.7792 12.7792 (0) 0.9893 0.9886 0.9874 0.9872 0.9871 0.9870 (0) DT 8.6983 6.5984 6.5603 6.5592 6.5589 6.5586 (0) 1.8805 1.8695 1.8681 1.8679 1.8678 1.8677 (0) 13.4819 12.7939 12.7793 12.7792 12.7792 12.7792 (0) 0.9893 0.9886 0.9874 0.9872 0.9871 0.9870 (0)
= 0.05
IT 3.2727 2.5880 2.5821 2.5818 2.5817 2.5815 (0) 1.8410 1.8312 1.8299 1.8296 1.8295 1.8293 (0) 4.4140 4.2005 4.1960 4.1960 4.1960 4.1960 (0) 0.9675 0.9666 0.9655 0.9652 0.9651 0.9647 (0) DT 2.5727 2.4789 2.4768 2.4766 2.4766 2.4765 (0) 1.8398 1.8301 1.8288 1.8285 1.8284 1.8282 (0) 2.6430 2.6326 2.6323 2.6323 2.6323 2.6323 (0) 0.9675 0.9666 0.9655 0.9652 0.9651 0.9647 (0)
= 0 .1
IT 3.0124 2.0489 2.0307 2.0292 2.0287 2.0275 (0) 0.8849 0.8823 0.8822 0.8821 0.8821 0.8821 (0) __ 3.6363 3.6361 3.6361 3.6361 3.8040 (0) 0.5693 0.5692 0.5691 0.5691 0.5691 0.5690 (0) DT 0.7067 0.6911 0.6907 0.6907 0.6907 0.6907 (0) 0.6453 0.6451 0.6451 0.6451 0.6451 0.6451 (0) 0.7182 0.7165 0.7165 0.7165 0.7165 0.7165 (1) 0.5681 0.5680 0.5679 0.5679 0.5679 0.5678 (0)
S-S
S-F
C-C
F-C
Effect of the location of ring support

The variations of plastic buckling stress parameters with respect to ring support radius are shown in Figs. 7.13 to 7.15 for circular plates and annular plates. It can be seen that the critical buckling mode shape of circular and annular plates with an internal ring support may not be the same as that of corresponding plates without
172
internal ring supports. For example, in the presence of an internal ring support of small radius, the critical buckling mode shape of a clamped circular plate change from an axisymmetric buckling mode shape (without ring support) to an asymmetric mode shape with one nodal diameter (with ring support). The hollow circles along the buckling curves represent the ring support radius where the two buckling mode shapes give the same buckling load. Therefore, these hollow circles indicate the change in the critical buckling mode shape with respect to the ring support radius. Sample 3D pictures of the critical buckling mode shapes are shown in Fig. 7.16 for C-C annular plate of = 0.2 . Note that the simply supported edge becomes clamped edge when the internal ring support is close to that edge. For example, when the internal ring support is close to the inner edge of S-S annular plate, the buckling curve approaches that of S-C annular plate.
Table 7.5 Convergence study of plastic buckling stress parameter for circular plates with one concentric ring support ( 1 = 0.6 )
BC
p
4 6 8 10 12 Analytical Method 4 6 8 10 12 Analytical Method
= 0.001
IT 2.7354 2.6664 2.6644 2.6640 2.6638 2.6635 (0) 3.1804 3.0483 3.0406 3.0390 3.0385 3.0377 (0) DT 2.7354 2.6664 2.6644 2.6640 2.6638 2.6635 (0) 3.1804 3.0483 3.0406 3.0390 3.0385 3.0377 (0)
= 0.05
IT 2.1863 2.1725 2.1721 2.1720 2.1720 2.1719 (0) 2.2572 2.2385 2.2373 2.2370 2.2369 2.2367 (0) DT 2.1801 2.1669 2.1664 2.1664 2.1663 2.1662 (0) 2.2469 2.2295 2.2283 2.2281 2.2280 2.2278 (0)
= 0 .1
IT 0.8726 0.8525 0.8521 0.8520 0.8520 0.8519 (0) 1.0358 0.9949 0.9922 0.9916 0.9914 0.9907 (0) DT 0.6488 0.6474 0.6474 0.6474 0.6474 0.6473 (0) 0.6573 0.6551 0.6550 0.6549 0.6549 0.6549 (0)
S-S
S-F
173
Optimal location of ring support

The presence of an internal ring support increases the value of plastic buckling stress parameters. For example, the buckling stress parameter for a simply supported circular plate increases from 0.4291 to 2.2192 by introducing a ring support. Of interest in the design of supported circular plate is the optimal location of the internal ring support for the maximum buckling load. The locations of optimal ring support are denoted by black dots in the Figs. 7.13 to 7.15. It can be seen that these optimal locations are influenced by the boundary conditions, , and the type of plasticity theory used. However, the nature of calculation of optimal location of a ring support is iterative and thus the calculation is computationally expensive. Chou at el. (1999) found that the optimal locations of internal ring supports for vibration of circular plates are found at the nodal circles of the corresponding internally unsupported plates. This optimality condition is examined for plastic buckling of circular and annular Mindlin plates in Table 7.6. It can be seen that the radii of nodal circles of the second buckling modes of internally unsupported circular and annular plates are not exactly the same as but are close to the corresponding iteratively computed optimal ring support radii. Therefore, it can be concluded that the nodal circles of the second buckling modes of internally unsupported circular and annular plates can be used as optimal location of internal ring support in order to achieve the maximum buckling load in design. It is interesting to note that all plates, except S-C annular plate, buckle in an axisymmetric mode shape (i.e. n = 0) when the plates are supported by a ring support at the optimal locations.
174
4.5
C-C
4
n=
n=
= hR2/2D
n=
C-S
3.5
n=
n=0
n =
n=
n=1
n=
n=
2.5
S-C
n=
0
n=1
n=0
S-S
2 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95
Ring support radius 1 = b1 / R Fig. 7.13 Variation of buckling stress parameter with respect to ring support location 1 for annular plate ( = 0.2, IT)
2.8
C-C
2.6
C-S
n=0
= hR2/2D
n 2 n=
=2
1 n=
n=1
0
n=2
n=
n=
n=0
n=
n=0
n= 1 n=1
2.4
S-C
S-S
2.2
2 0.25
0.35
0.45
0.55
0.65
0.75
0.85
0.95
Ring support locations 1 = b1 / R Fig. 7.14 Variation of buckling stress parameter with respect to ring support location 1 for annular plates ( = 0.2, DT)
175
2.6
n=0
n =
2.4
2.2
Clamped
n=
n=1
= hR2/2D
Simply supported
1.8
1.6
1.4
1.2
IT DT
1 0 0.2 0.4 0.6 0.8 1
Ring support locations 1 = b1 / R Fig. 7.15 Variation of buckling stress parameter with respect to ring support location 1 for circular plates
176
1 = b1 / R = 0.3
1 = b1 / R = 0.5
1 = b1 / R = 0.65
1 = b1 / R = 0.7
Fig.7.16 Critical Buckling mode shapes for C-C annular plates with various ring support locations (b/R = 0.2, DT, h/R = 0.05)
177
Table 7.6 Comparison study between the optimal ring support radius and radius of nodal circle of second mode shape Elastic BC Optimal IT Optimal DT Nodal circle radius 0.7064 0.6016 0.7406 0.5544 0.7841 0.7499 0.8298 0.6949 0.4699 0.2663 Optimal
Nodal circle radius 0.6547 0.6011 0.7238 0.5174 0.7625 0.7496 0.8230 0.6853 0.4618 0.2663
Nodal circle radius 0.6902 0.6015 0.7340 0.5445 0.7782 0.7499 0.8280 0.6926 0.4687 0.2663
Annular plates = 0.2 S-S C-C S-C C-S S-S C-C S-C C-S 0.6563 0.6031 0.7086 0.5180 0.7641 0.7500 0.8215 0.6855 0.4625 0.2672 0.7101 0.6022 0.7258 0.5547 0.7859 0.7504 0.8285 0.6953 0.4692 0.2688 0.6922 0.6031 0.7208 0.5504 0.7797 0.7508 0.8269 0.6934 0.4681 0.2703
Annular plates = 0.5
Simply supported circular plates Clamped circular plates
Plastic buckling analyses of Mindlin circular and annular plates with and without an internal ring support under a uniform stress were carried out by using the Ritz method. The program code for the Ritz methods using the polar coordinate system are written in MATHEMATICA and it is given in Appendix A. Moreover, the buckling problems were also solved analytically. The matrix elements for the analytical method are also provided in Appendix B for readers who want to use other programming
178
languages so that programs can be easily obtained. The analytical solutions will be useful to researchers who wish to verify their numerical methods and solutions. The Ritz results and analytical results are found to agree well. Moreover, these plastic buckling solutions are in good agreement with the existing elastic buckling solution when the plate is thin. Generally, the plastic buckling stress parameters obtained by using DT were consistently lower than their IT counterparts. It was found that the locations of the nodal circles of the second axisymmetric buckling modes of internally unsupported circular and annular plates are useful in guiding designers for maximum buckling capacity in designing such plates with an internal ring support.
179
CHAPTER 8
PLATES WITH COMPLICATING EFFECTS
8.1
Introduction
This chapter considers Mindlin plates with complicating effects such as the presence of internal line/curved supports, point supports, mixed boundary conditions, elastic foundations, internal line hinges and intermediate in-plane compressive stress. So far, the application of the Ritz method for plastic buckling analysis has been presented for Mindlin plates with various edge supports conditions. In this chapter, we show how the Ritz method may be used to accommodate the aforementioned complicating effects in plastic buckling analysis. Although the treatments are presented in the rectangular coordinate system only, one can readily recast the treatments in the polar coordinates or the skew coordinates by using suitable coordinate transformation relations. For numerical results, we adopt the aerospace material Al-7075-T6 with the properties:
180
Plates with Complicating Effects
Poissons ratio = 0.33 , E = 10.5 10 6 psi, 0.7 = 70 ksi and the Ramberg-Osgood parameters k = 3 / 7 , c = 9.2 .
8.2
Internal line/curved/loop supports
Consider a plate with internal line/curved/loop supports as shown in Fig. 8.1. These internal supports are assumed to impose a zero deflection (w = 0) along its length. This addition geometrical constraint can be taken care of in the Ritz method by the modification of the basic function for the transverse deflection as follows
bw = [i ( , )] j ( , )
w j
ne
ns
(8.1)
i =1
where the underlined term is contributed by the presence of internal supports, in which
j being the equation of the jth support and ns the number of these internal supports.
The rest of the formulations remain unchanged as presented in Chapter 4. This method can handle any number of arbitrarily oriented/shaped internal supports as long as the locations of these supports can be represented by analytical expressions except for line/curved supports whose ends do not terminate at the plate boundary. The line/curved supports whose ends do not terminate at the plate edges can be simulated by a series of point constraints along the supports as shown in Fig. 8.2. The treatment of these point supports in Ritz method will be discussed in Section 8.3.
181
Line support
Curve support
Loop support
(a)
(b)
(c)
Fig. 8.1 Plates with internal (a) line, (b) curved and (c) loop supports
Line support
Curve support
Point supports to simulate internal support
Fig. 8.2 Simulation of internal supports using point supports
For sample numerical results, consider a rectangular plate which is supported internally by a rigid internal line support as shown in Fig. 8.4. The plate is subjected to a uniaxial in-plane compressive stress x . For the considered internal line support, the underlined term in Eq. 8.1 is given by
182
ns
( , ) = ( j )
j
ns
(8.2)
where j is the coordinate of jth line support in direction as shown in Fig. 8.3.
Internal line support
ja x
a
Fig. 8.3 Rectangular plate with an internal line support
The convergence and comparison studies are shown in Table 8.1 and the some sample critical buckling mode shapes are shown in Fig. 8.4 for simply supported square plates with internal line supports. It can be seen that the satisfactory convergence is achieved when the degree of polynomial p = 12 is used. It is found that the converged results agree well with the analytical elastic buckling solutions derived by Xiang (2003) when the plate is thin ( = 0.001 ).
183
Table 8.1 Convergence and comparison studies of plastic buckling stress parameter for rectangular plate with internal line supports
p
6 8 10 12 Xiang (2003) 6 8 10 12 Xiang (2003) 6 8 10 12 Xiang (2003) 6 8 10 12 Xiang (2003)
= 0.001
IT 5.3217 5.3177 5.3174 5.3168 5.3165 6.2501 6.2499 6.2499 6.2499 6.2500 8.6996 8.6235 8.6154 8.6154 8.6119 10.7289 10.3974 10.3883 10.3878 10.386 DT 5.3217 5.3177 5.3174 5.3168 5.3165 6.2501 6.2499 6.2499 6.2499 6.2500 8.6996 8.6235 8.6154 8.6154 8.6119 10.7289 10.3974 10.3883 10.3878 10.386
= 0.025
IT 5.2703 5.2664 5.2661 5.2655 __ 6.1340 6.1338 6.1338 6.1338 __ 7.9021 7.8599 7.8555 7.8554 __ 8.7659 8.6478 8.6445 8.6443 __ DT 5.2684 5.2645 5.2642 5.2637 __ 6.1281 6.1279 6.1279 6.1279 __ 7.8658 7.8249 7.8206 7.8205 __ 8.7033 8.5901 8.5869 8.5867 __
= 0.05
IT 3.5287 3.5287 3.5287 3.5287 __ 3.5470 3.5469 3.5469 3.5469 __ 4.1644 4.1406 4.1401 4.1401 __ 4.4337 4.3686 4.3669 4.3669 __ DT 2.8770 2.8769 2.8769 2.8769 __ 2.9192 2.9192 2.9192 2.9192 __ 3.0765 3.0721 3.0718 3.0718 __ 3.1495 3.1366 3.1363 3.1363 __
-0.6
-0.7,0.7
-0.5,0.5
184
line support
line support
1 = 0.6
line supports
1 = 0
1 = 0.5, 2 = 0.5
1 = 0.7, 2 = 0.7
Fig. 8.4 Critical buckling mode shapes of square plates with internal line support
8.3
Point supports
Consider a Mindlin plate with point supports which are located at the edges or inside the plate. The transverse deflections are zero at the location of point supports, i.e.
w( i , i ) = 0
i = 1,2,....n p
(8.2)
where n p is the number of point supports and i and i are coordinates of the ith point support. These geometric constraints due to point supports can be taken into
185
account by several ways. The following are some common methods that are used to simulate the point supports in the Ritz method.
(1) Lagrangian multiplier method The point constraint may be incorporated into the total energy functional by using Lagrangian multipliers as
np
= U + V + Li w( i , i )
i =1
(8.3)
where Li is the ith Lagrangian multiplier. This yields the eigenvalue equation as
K cc H c 0 d =0 0 e 0 L
K cd K
dd
K ce K
de
K ee
(8.4)
where the elements of the H submatrix are given by

hij = iw ( j , j ) i = 1,2,......., N , j = 1,2,....., n p
(8.5)
where N is the number of polynomial terms given by Eq. (3.14).
(2) Using constraint functions This approach was first used by Liew et al. (1994) in order to solve the vibration problems of Mindlin plate on point supports. In this approach, the constraint conditions (Eq. 8.2) is used to reduce the dimensions of the eigenfunction while the matrix remains positive definite. By substituting the coordinates of internal point
186
supports in to the approximate function for transverse deflection (Eq. 3.12a), the constraint functions (Eq. 8.2) can be expressed as
[C]n N {c}n 1 = 0
p p
(8.6)
Assuming that n p is less than N, then the Eq. (8.6) can be rearranged as c1 c n p +1 c c [S] :2 = [R ] n p + 2 = Rc : c n p cN
(8.7)
where S is the positive definite matrix of n p n p , R the matrix of n p ( N n p ) , ci ( i = 1 to N) the elements of c with the order rearranged by a position matrix P of N N with elements value 0 or 1 satisfying the relation c = Pc . From Eq. (8.7), we have
S 1 R c= c I
(8.8)
where I is the identity matrix of ( N n p ) ( N n p ) . The unknown coefficient c can now be expressed in terms of modified unknown coefficient c as
S 1 R = P c = Pc c = Bc I
(8.9)
The substitution of Eq. (8.9) into the usual eigenvalue equation (Eq. 3.20) yields
187
c [K ][B ]d = 0 e
(8.10)
where
B 0 0 [B ] = 0 I 0 0 0 I
(8.11)
(3) Using elastic spring supports The rigid point supports can be simulated by elastic springs with a sufficiently large spring constant. The strain energy due to these springs is given by
np
1 2 Q = S k [w( k , k )] k =1 2
(8.12)
where S k is the spring constant of the kth point support, which needs to be sufficiently large number in order to simulate rigid point support. The augmented total energy functional becomes
= U + Q +V
(8.13)
The minimization with respect to the Ritz coefficients yields

c ([K ] + [G ])d = 0 e
(8.14)
where the matrix [G ] is contributed by the strain energy due to the springs and is given by
188
G cc [G ] = 0 0 where
np
0 0 I 0 0 I
(8.15)
G cc = S k iw ( k , k ) jw ( k , k ) i = 1,2,.., N j = 1,2,.., N
k =1
(8.16)
where I is the identity matrix of order N. Note that the first approach increases the dimensions of the eigen matrix that results in the increase of computational time whereas the second approach reduces the dimensions of the matrix but it needs more effort in forming the matrix [B ] . The last approach is the simplest and easy to implement. So far the treatment for point supports in the direction of transverse deflection is presented and the treatments for the point supports of normal and tangential rotation can also be done in the same way. It is to be remarked that the difficulty of implementing the simple support (hard type) for oblique line and curved edges in rectangular coordinates system may be overcome by using the combination of soft type simple support (S*) and point constraints on the tangential rotation t on the concerned edges as in Wang et al. (1995). Note that the mixed boundary conditions can also be simulated by using suitable point restraints such as simple point supports and clamped point supports along the boundary as in Liew and Wang (1992). For sample buckling results, the uniaxially loaded square plate supported by four symmetrically placed internal point supports as shown in Fig. 8.5 is considered. In order to compare with the elastic buckling results of Venugopal et al. (1989), Poissons ratio = 0.3 is adopted. The convergence and comparison studies are
189
shown in Table 8.2 for squares plate resting on four internal point supports. It is found that the converged results agree well with existing solutions. It can be seen that the plastic buckling stress parameters obtained by IT agree well with their DT counterparts for = 0.05 . This is due to the less restrained boundary condtion.
point supports point supports

a
Fig. 8.5 Four point supported square plate
Table 8.2 Convergence and comparison study of plastic buckling stress parameter for four points supported square plate
= 0.001
= 0.05
Uniaxial loading IT 0.9157 0.9152 0.9151 __ DT 0.9157 0.9152 0.9151 __ Biaxial loading IT 0.7351 0.7350 0.7349 __ DT 0.7351 0.7350 0.7349 __
Uniaxial loading 0.9221 0.9217 0.9217 0.9217
Biaxial loading 0.7391 0.7390 0.7390 0.7390
6 8 0.5 10 Venugopal (1989)
190
8.4 Elastically restrained edges and mixed boundary conditions
The elastically restrained edge support can be modeled by combining three different spring models whose stiffness are given by
S kw : spring stiffness of kth plate edge for the transverse elastic support, S kx : spring stiffness kth plate edge for the rotational elastic support about y-axis and S ky : spring stiffness kth plate edge for the rotational elastic support about x-axis.
Note that the arbitrary boundary conditions can be expressed by the combination of free edges and aforementioned spring supports. For example, the clamped boundary can be expressed by the combination of the free edge and the springs supports with the very large values of S kw , S kx and S ky .
The potential energy stored in the elastic supports can be derived as

1 ne 2 2 S kw w2 + Skx (x ) + Sky ( y ) dk 2 k =1 k
Q=
(8.17)
where ne is the number of edge and k refers to kth the plate boundary. The augmented total potential energy becomes
= U + Q +V
(8.18)
The minimization of the energy functional yields the following augmented eigenvalue equation
c ([K ] + [Q])d = 0 e
(8.19)
191
where
Q cc [Q] = 0 0
in which
ne
0 Q dd 0
0 0 Q ee
(8.20)
Q = S kw iw ( , ) jw ( , ) dk i = 1,2,.., N and j = 1,2,.., N

cc k =1 k
{ [
]}
(8.21a)
Q dd = S kx ix ( , ) jx ( , ) dk i = 1,2,.., N and j = 1,2,.., N

k =1 k
ne
{ [
]}
(8.21b)
Q ee = S ky iy ( , ) jy ( , ) dk i = 1,2,.., N and j = 1,2,.., N

k =1 k
ne
{ [
]}
(8.21c)
Note that the elements of the matrix [K ] are the same as those given by Eqs.(3.232a-f) in which the basic functions are chosen by assuming that all edges are free edges. The aforementioned procedure can also handle mixed boundary conditions such as those shown in Fig. 8.6 by separating the boundary integral Eq. (8.17) into several parts according to the numbers of boundary types along the edge of the plate. For example, the energy stored in the elastic support for the plate edge = 1 (denoted as
edge no. 1) of case 1 in Fig. 8.6 is given by
Q =
cc 1
{S [
1 w 11 1 x 11
w i
( , ) jw ( , )
]}
= 1
(8.22a)
dd Q1 =
{S [
1
y 12
x i
( , ) jx ( , )
]}
= 1
(8.22b)
ee Q1 =
{S [
( , ) jy ( , )
]}
=1
(8.22c)
192
Note that the first subscript of spring constant S refers to the edge number while the
y second subscript refers to the segment number along the edge, for example, S12 refers
to the rotational spring constant of second segment of the edge no. 1. It is to be remarked that, since most of the edges are simply supported for cases shown in Fig. (8.6), the mixed boundary condition can also be achieved by combining simply supported edges and rotational spring about x-axis along the portion of the edge that are clamped.
a a a
Case 1
Case 2
Case 3
Fig.8.6 Square plates with mixed boundary conditions under uniaxial compression
In order to verify the correctness of the method, comparison studies are shown in Table 8.3 for three cases shown in Fig. 8.6. Note that = 0.001 and = 0.3 are adopted in the analysis in order to compare with the existing elastic buckling results. The present results are slightly higher than those of Mizusawa and Leonard (1990) when the value of is somewhere between 0 and 2. This may be due to the fact that the boundary conditions are imposed by using the boundary integral in the present
193
treatment while in spline element method (Mizusawa and Leonard, 1990) the boundary condition are imposed only at the nodes.
Table 8.3 Comparison studies of buckling stress parameter for square plates with mixed boundary conditions Mizusawa and Leonard (1990) (Spline element method) 4.000 4.209 5.198 5.628 5.740 4.000 5.265 5.684 5.732 5.740 4.000 6.524 7.430 7.640 7.690
Case 1 0 1/3 1 4/3 2 0 1/3 1 4/3 2 0 1/3 1 4/3 2
Present (Ritz method) 4.0000 4.3573 5.4654 5.7059 5.7401 4.0000 5.5240 5.7252 5.7388 5.7401 4.0000 6.9300 7.6217 7.6837 7.6911
Case 2
Case 3
8.5 Elastic foundations
The elastic buckling of plates resting on elastic foundations has been studied by many researchers (Naidu et al., 1990; Shen, 1999; Wang, 2005). This elastic foundation effect can be taken into account in the proposed Ritz plastic buckling analysis by augmenting strain energy given in Eq. (3.9) with U F
194
UF =
1 2 S k w dA 2 A
(8.23)
where S k is the modulus of subgrade reaction for the foundation. The resulting eigenvalue equation is given by:
K cc + K cc K cd F K dd symmetric K ce c K de d = 0 K ee e
(8.24)
where the elements of K cc are given by F

cc k Fij = S k iw jw dd A
(8.25)
Note that, in above formulations, it is assumed that foundation reacts in compression as well as in tension. However, in many practical applications, the foundation cannot provide tensile reaction. For example, in the case of plates resting on soils that lacking both adhesive and cohesive properties. Other applications of the tensionless foundation are found in the delamination of laminated plates and retrofitting metal plates which are attached to concrete or steel plates in composite structures. In these cases, the foundations can be assumed as rigid foundation and the buckling problem of plates in such cases is also known as unilateral buckling. The buckling problem of these unilateral constrained plates is complicated because the location and extent of the contact region are not known at the outset. In order to demonstrate the ability of the Ritz plastic buckling analysis, unilateral buckling problems of rectangular plates are considered in this study. One of the ways of incorporating a unilateral constraint condition in the buckling analysis is to model the foundation as a bed of springs that exhibits a deformation
195
sign-dependent relationship. Such behaviour of foundation can be modeled by introducing a contact function X that is defined by
X = 0 (separation)
(8.26a) (8.26b)
X =1
(contact)
In applying contact condition, the plate area is discretized into a series of grids, and the contact condition is assessed at the grid locations. Sufficiently large spring stiffness is adopted to simulate a rigid surface. In this study, a grid spacing of 30 30
and the foundation stiffness of 104 are adopted. The matrix elements of K cc given in F Eq. (8.25) become
Np
cc Fij
= XS k iw jw
k =1
(8.27)
where N is the number of grid points. The nature of solution procedure is iterative since the final contact region between the plate and the foundation surface is unknown before the solution is sought. In order to define the contact region and sole the problem, the following iterative procedure is adopted; Step (1) Solve the eigenvalue equation by assuming that there is no contact between plate and foundation. Step (2) Check the contact status at the grid points. The elastic foundation is activated or deactivated by using the contact function X and then solve the eigenvalue equation. Step (3) Compare the eigenvalue with the previous iteration. If the convergence is satisfactory, then stop and adopt the last eigenvalue as solution. If not, go to Step 2 and repeat.
196
Table 8.4 shows the comparison study between the present results ( = 0.001 ) and the existing elastic unilateral buckling solutions. It can be seen that the results agree well. Table 8.4 Comparison study of unilateral buckling stress parameter for rectangular plates under shear Aspect ratio a/b 1 SSSS 2 3 1 CCCC 2 3 Bilateral Buckling 9.325 6.546 5.840 14.642 10.248 9.535 Unilateral Buckling Present Method 9.355 6.719 6.683 15.242 12.025 12.269 Smith et al. (1999) 9.355 6.720 6.689 15.26 12.03 12.21 Influence of unilateral buckling (%) 0.32 2.64 14.43 4.10 17.34 28.67
BC
Tables 8.5 and 8.6 show the plastic buckling stress parameters for simply supported and clamped rectangular plates under pure shear. Both bilateral and unilateral buckling stress parameters are presented and denoted as bil and uni respectively. The increase in the buckling stress parameter due to the presence of rigid restraint is also shown in the percentage of bil . It is found that the difference between bil and uni increases as the aspect ratio increases. This is because, at larger aspect ratios, the plate buckles in multiple half-wavelengths in bilateral buckling case and as a result the contact area between the plate and rigid foundation increases. Typical buckling mode shapes for bilateral and unilateral buckling cases are presented in Fig. 8.7 for the clamped rectangular plate.
197
Table 8.5 Unilateral plastic buckling stress parameters for rectangular plates under shear (IT)
BC
a/b 1
= 0.025
= 0.05
%diff 0.07 3.52 8.00 3.62 19.06 28.83
bil
6.782 5.595 5.151 11.177 8.059 7.604
uni
6.787 5.792 5.563 11.582 9.595 9.796
bil
6.139 4.594 4.275 10.554 7.618 7.151
uni
6.145 5.043 5.591 10.967 9.134 9.344
%diff 0.10 9.77 30.78 3.91 19.90 30.67
SSSS
2 3 1
CCCC
2 3
Table 8.6 Unilateral plastic buckling stress parameters for rectangular plates under shear (DT)
BC
a/b 1
= 0.025
= 0.05
%diff 0.06 2.26 8.00 0.86 5.20 8.13
bil
6.201 5.432 5.151 7.189 6.562 6.421
uni
6.205 5.555 5.563 7.251 6.903 6.943
bil
2.090 1.991 1.964 2.236 2.138 2.119
uni
2.091 2.016 2.056 2.248 2.192 2.199
%diff 0.05 1.26 4.68 0.54 2.53 3.78
SSSS
2 3 1
CCCC
2 3
198
Bilateral (a/b = 1)
Unilateral (a/b = 1)
Bilateral (a/b = 2)
Bilateral (a/b = 3)
Fig. 8.7 Typical buckling mode shapes for bilateral and unilateral buckling of clamped rectangular plates
8.6 Internal line hinges
The internal hinges may be used in the modeling of folding gates and hatches. In order to cater the slope discontinuity along the hinge line, the domain decomposition method can be employed. In the domain decomposition method, the plate is divided
199
into several subdomains along the hinges depending on the number of hinges. In order to illustrate the treatment of internal line hinge in the Ritz method, a rectangular plate with one internal hinge is considered as shown in Fig. 8.8. The adopted admissible pRitz functions for the deflection and rotations of the subdomain 1 are given by
p q
w w1 (1 ,1 ) = c 1 m (1 ,1 ) m q =0 i =0
(8.28a)
1 x x1 (1 ,1 ) = d m m (1 ,1 )
q =0 i =0
(8.28b)
y y1 (1 ,1 ) = e1 m (1 ,1 ) m q =0 i =0
(8.28c)
and the similar functions for subdomain 2 are given by

p q
2 w w2 ( 2 , 2 ) = c m m ( 2 , 2 )
q =0 i =0
(8.29a)
2 x x 2 ( 2 , 2 ) = d m m ( 2 , 2 )
q =0 i =0
(8.29b)
2 y y 2 ( 2 , 2 ) = em m ( 2 , 2 )
q =0 i =0
(8.29c)
in which 1 ,1 and 2 , 2 are the normalized coordinate systems, which have the integration limits from -1 to 1, for each subdomain. Along the internal hinge, the compatibility conditions for the transverse deflections are given by
w1 (1,1 ) = w2 (1, 2 )
(8.30)
200
y = x
internal line hinge
1 1
b
2 2
x a
(1 )a
Fig. 8.8 A rectangular plate with an internal hinge under in-plane compressive stress
In view of Eq. (8.30), the Ritz approximate function for transverse deflection for subdomain 2 can be redefined as follows (Su and Xiang, 1999):
w2 ( 2 , 2 ) = w1 (1, )
bw ( 2 , 2 ) + ( 2 1) w2 ( 2 , 2 ) bw (1, 2 )
(8.31)
The energy functional for each domain can be expressed as in Eq. (3.11) by substituting corresponding displacement fields, i.e. w1 , x1 , y1 and w2 , x 2 , y 2 . The corresponding eigenvalue equation is given by K cc K cd K dd symmetric in which
K ce c K de d = 0 K ee e
(8.32)
201
K cc K cc = 1 0 K cd K cd = 1 0
K
ce ce K 1 = 0 dd K 1 = 0
0 + K cc 2 0 2 N 2 N
[ ] [
2 N 2 N
(8.33)
0 + 0 K cd 2 0 2 N 2 N 0 + 0 K ce 2 0 2 N 2 N
2 N 2 N
(8.34)
2 N 2 N
(8.35)
dd
0 dd K 2 2 N 2 N 0 de K 2 2 N 2 N 0 K ee 2 N 2 N 2
(8.36)
K de K de = 1 0 K ee K ee = 1 0
(8.37)
(8.38)
Note that the expressions for K icc , K icd , K ice , K idd and K iee (i = 1, 2) are the same as those in the usual formulation for rectangular plates which are given by Eqs. (3.23a-f), except the fact that the corresponding Ritz functions need to be adopted accordingly, i.e. Eqs.(8.28a-c) for subdomain 1, Eqs.(8.29b-c) and Eq.(8.31) for subdomain 2. For sample results, a square plate with an internal hinge under uniaxial loading is considered. Table 8.6 shows a comparison study against existing elastic buckling results. Note that, in order to compare with existing elastic buckling results, the thickness-to-width ratio and Poissons ratio are set to be 0.001 and 0.3, respectively. It can be seen that the results agree well with the existing solutions.
202
Table 8.7 Buckling stress parameter for square plate with an internal hinge under uniaxial loading
p 4 6 SSSS 8 Xiang et al. (2001) 4 6 CCCC 8 10 Wang et al. (2001b) 0.1 1.5773 1.5764 1.5761 1.5770 8.3647 7.9073 7.8868 7.8858 7.8857 0.2 1.7481 1.7473 1.7472 1.7474 6.1853 6.0684 6.0582 6.0556 6.0545 0.3 1.8976 1.8969 1.8968 1.8969 4.6843 4.6281 4.6166 4.6125 4.6108 0.4 2.0042 2.0035 2.0035 2.0035 4.3519 4.2919 4.2790 4.2749 4.2727 0.5 2.0436 2.0429 2.0428 2.0429 4.3432 4.2486 4.2366 4.2330 4.2306
8.7 Intermediate in-plane loads
Consider a Mindlin plate which is subjected to an end compressive stress 1 and an intermediate compressive stress 2 as shown in Fig. 8.8. Note that a positive value of
implies a compressive load while negative value implies a tensile uniaxial load.
Since all the parameters , , , G depend on the stress level, the plate needs to be divided into two domains along the location of intermediate load in order to handle the different levels of stresses. The first domain is to the left of the vertical line defined by x = a (see Fig. 8.8) and the second domain is to the right of this line. The Ritz approximate functions for deflection and rotations can be adopted as usual for the whole plate. The resulting eigenvalue equation is given by
203
cc cd K1 K1 dd K1 symmetric
ce K 1 K cc K cd 2 2 dd de K1 + K2 ee K 1 symmetric
K ce c 2 de K 2 d = 0 K ee e 2
(8.39)
where subscript 1 and 2 refer to the domain 1 and 2 respectively.
1 + 2
(1 )a
Fig. 8.9 Rectangular plate under edge and intermediate compressive stress
The sample results of 1 = 1hb 2 /( 2 D) are shown in Table 8.7 for a rectangular plate of = 0.04 and for a given value of 2 = ( 1 + 2 )hb 2 /( 2 D) = 50 . The same material properties E = 10700 ksi, 0 = 61.4 ksi, k = 0.3485, c = 20 are adopted as in Wang et al. (2004) for comparison purpose. It can be seen that the present results are found to be slightly lower than existing results. This is due to the effect of transverse shear deformation which is neglected in Wang et al. (2004).
204
Table 8.8 Convergence and comparison of plastic buckling stress parameters 1 of rectangular plates under edge and intermediate loads. ( 2 = 50, = 0.04 ) BC p IT 8 10 SSSS 12 14 Wang et al. (2004) 8 10 CSCS 12 14 Wang et al. (2004) 8 10 FSFS 12 14 Wang et al. (2004) 4.4568 4.4010 4.3769 4.3769 4.470 6.3389 6.0809 5.9898 5.9458 6.214 2.3641 2.3322 2.3182 2.3092 2.451 a/b = 1 DT 3.9141 3.9021 __ __ 4.002 4.1166 4.0946 4.0847 __ 4.198 2.2225 2.1872 __ __ 2.451 IT 3.8084 3.7869 3.7809 3.7809 3.799 4.2094 4.1114 4.0871 4.0777 4.108 2.2454 2.2333 2.2194 2.2105 2.289 a/b = 2 DT 3.6628 3.6566 __ __ 3.712 3.8374 3.8150 3.8074 __ 3.878 2.1438 2.1299 __ __ 2.289 IT 3.7939 3.7528 3.7361 3.7361 3.738 4.3579 4.0434 3.9486 3.9195 3.917 2.2467 2.2237 2.2133 2.2028 2.266 a/b = 3 DT 3.6577 3.6389 __ __ 3.674 3.7954 3.7744 3.7609 __ 3.802 2.2146 2.1995 __ __ 2.266
This chapter demonstrates the treatments for the complicating effects such as the presence of internal line/point/curved/loop supports, elastically restrained edges, mixed boundary conditions, elastic foundations, internal line hinges, intermediate inplane loads when using the Ritz method for plastic buckling analysis. The various treatments are demonstrated by solving sample problems and comparing the results with existing elastic buckling solutions. It can be seen that the Ritz method is very
205
versatile and powerful to handle all the aforementioned complicating effects for plastic buckling analysis of Mindlin plates.
206
CHAPTER 9
CONCLUSIONS AND RECOMMENDATIONS
9.1 Conclusions An automated Ritz method is presented for the plastic buckling analysis of Mindlin plates of various shapes, boundary and loading conditions. The effect of transverse shear deformation was taken in to account by adopting the Mindlin plate theory. In the Ritz method, the assumed displacement functions were defined by the product of basic functions and mathematically complete two-dimensional polynomial functions whose degree may be increased until the desired accuracy is achieved. The basic functions are defined by the product of boundary equations raised to the appropriate power (either 0 or 1) that ensure automatic satisfaction of the geometric boundary conditions. As it is more convenient to handle certain plates in their natural coordinate systems, three versions of the Ritz software codes are given. The software code BUCKRITZRE 207
Conclusions and Recommendations
is based on the Cartesian coordinate system and it can be used for determining the plastic buckling stress of plates of any shape whose edges are defined by polynomial functions. Another software code, called BUCKRITZSKEW, is written based on the skew coordinate system and it is to compute the plastic buckling stress of skew plates. The last software code, called BUCKRITZPOLAR, is to solve the plastic buckling problems of circular and annular plates and it is based on the polar coordinate system. It is worth noting that the presented Ritz program codes can be used to study not only the plastic buckling of plates but also the elastic buckling of plates. One can calculate the elastic buckling stress parameter based on the classical thin plate theory by setting the thickness-to-width ratio to a small value. If one wish to calculate the elastic buckling stress parameter based on Mindlin plate theory, it can be easily done by setting the tangent modulus and secant modulus equal to Youngs modulus. The capability of the Rayleigh-Ritz method was demonstrated by solving the plastic buckling problems of rectangular, skew, triangular, elliptical, circular and annular plates. In addition to conventional loading and boundary conditions, various treatments were presented for the presence of complicating effects such as internal line/curved/point supports, internal line hinges, intermediate in-plane loads and elastic foundation. In all these situations, it has been shown that Ritz method is simple and easy to formulate since it does not require mesh discretization of the plate. Being an approximate continuum method, it eliminates the need for mesh generation and thus large numbers of degrees of freedom. Therefore it requires less memory and is easy to implement when compared with the finite element method. Since the eigen problem needs to be solved in an incremental and iterative fashion in plastic buckling problems, the computational superiority of the Ritz method is even more convincing.
208
The plastic buckling stress parameters for various shapes such as rectangular, skew, triangular, elliptical, circular and annular Mindlin plates were calculated for various aspect ratios, thickness-to-width ratios and boundary conditions. So far, most of the plastic buckling solutions available focus mainly on the rectangular plates. Therefore, the extensive plastic buckling data presented for various plate shapes will be useful as a useful reference source for designers and researchers who wish to verify their plastic buckling results. It was found that the plastic buckling stress parameters approach their elastic counterparts when the plate is thin or the plate is supported by less restrained boundary conditions. But for plates with larger thickness-to-width ratios and more restrained boundary conditions, the plastic buckling parameters differ appreciably from their elastic counterparts. Generally, the plastic buckling results obtained by IT are larger than their DT counterparts and the difference increases with increasing thickness-towidth ratios. As expected, the effect of transverse deformation lowers the plastic buckling stress parameters. The effect is more pronounced in the cases of thicker plates and plates with more constrained boundary conditions. Generally, the reductions in the plastic buckling stress parameters based on IT are larger than their DT counterparts. Owing to the lack of analytical buckling solutions for asymmetric buckling problems of circular and annular Mindlin plates, an analytical method was featured in this thesis for the first time. By following the procedure used by Mindlin (1951), the governing equations of the plates were transformed into the standard forms of partial differential equations in terms of three potential functions. These transformed governing equations were solved analytically and the exact plastic buckling solutions were obtained. The validity of the results was confirmed by comparing with existing
209
analytical elastic buckling thin plate results (by setting the plate thickness to very small value) and with plastic buckling results obtained by the Ritz method. It was found that the results agree well. However, this analytical method is not applicable for the cases of non-uniform stress distributions. If the variations of stresses are considered, the general solutions of governing equations are not possible due to the complexity arising from the variations of stresses. Nonetheless, since the buckling problems of circular and annular plate under uniform compression are fundamental in structural analysis, the developed analytical method will be useful for researchers who wish to check the validity and convergence of their numerical methods.
9.2 Recommendations There are some interesting directions for future works in the area of research presented in this thesis. One of the interesting areas is to combine the finite element analysis with the Ritz method. The major set back in the Ritz method is that it requires prebuckling stress distribution. In the case of plastic buckling analysis, it is hard to obtain the analytical expression for stress distribution. The assumption of uniform stress distribution, i.e. no stress dissipation in the x-y plane, was adopted in the present study. Although the effect of stress dissipation is very marginal in the cases of uniform loadings, the effect is more pronounced in the cases of plates under patch loading i.e. locally distributed loading and point loading (Liu and Pavlovic, 2005) and plates with holes (Uenoya and Redwood, 1978). Uenoya and Redwood (1978) combined the Ritz method and the finite element method in their elasto-plastic bifurcation analysis of perforated plates, with the Ritz method being used for the bifurcation analysis and the finite element
210
being used for determining in-plane stress. However, their study is confined to the classical thin plate theory (Kirchhoff) and the rectangular plates with simply supported or clamped boundary conditions. In future, it is desirable to extend the Ritz plastic buckling analysis of Mindlin plate of various shapes and boundary conditions. Another possible area for future work is the extension of presented Ritz method to postbuckling analysis and allowing for the effect of imperfections. This study mainly focused on the plastic bifurcation analysis i.e. small deflection theory was adopted. Therefore, the effect of initial imperfections cannot be taken into consideration in the present study. In order to include the effect of imperfections, a postbuckling analysis (with large deflection theory) needs to be carried out. The applicability of the Ritz method to the postbuckling analysis was shown in Jane et al. (2003). Therefore, the application of the Ritz method to the plastic postbuckling analysis of Mindlin plates, with the consideration of effect of imperfections, should be made in future.
211
REFERENCES
Aitchison, C.S., Miller, J.A. (1942). Tensile and Pack Compressive Tests of Some Sheets of Aluminum Alloy, 1025 Carbon Steel, and Chromium-Nickel Steel. NACA Technical Note No. 840. Anderson, R.A. (1951). Charts giving critical compressive stress of continuous flat sheet divided into parallelogram-shaped panels. NACA Technical Note No. 2392, Washington, D.C. Anderson, R.A. and Anderson, M.S. (1956). Correlation of Crippling Strength of Plate Structures with Material Properties. NACA Technical Note 3600, Washington, D.C. Argyris, J.H. (1965). Continua and dicontinua. Proceedings of the Conference on Matrix Methods in Structural Mechanics, Air Force Flight Dynamics Lab., AFFDL-TR, 66-80. Ashton, J.E. (1969). Stability of clamped skew plates under combined loads. Journal of Applied Mechanics, 36(1), 139-140. Bassily, S.F. and Dickinson, S.M. (1975). On the use of beam functions for problems of plates involving free edges. Trans. ASME, Journal of Applied Mechanics, 42, 858-864. Bassily, S.F. and Dickinson, S.M. (1978). Buckling and Vibration of in-plane loaded plates treated by a unified Ritz approach. Journal of Sound and Vibration, 59, 114. Batdorf, S.B., and Stein, M. (1947). Critical combinations of shear and direct stress for simply supported rectangular flat plates. NACA Technical Note 1223, Washington, D.C. Betten, J., Shin, C.H., (2000). Elastic-plastic buckling analysis of rectangular plates subjected to biaxial loads, Forschung Im Ingenieurwesen, 65 (9), 273-278. Bhat, R.B. (1985). Natural frequencies of rectangular plates using characteristic orthogonal polynomials in the Rayleigh-Ritz method. Journal of Sound and Vibration, 102, 493-499. Bhat, R.B. (1987). Flexural vibration of polygonal plates using characteristics orthogonal polynomials in two variables. Journal of Sound and Vibration, 114, 65-71. Bijlaard, P.P. (1947). Some contributions to the theory of elastic and plastic stability. International Association for Bridge and Structural Engineering, Zurich, 8, 17-80.
212
References
Bijlaard, P.P. (1949). Theory and tests on the plastic stability of plates and shells. Journal of Aeronautical Science, 16 (9), 529-541. Bradford, M.A. and Azhari, M. (1994). Local buckling of I-sections bent about the minor axis. Journal of Constructional Steel Reasearch, 31, 73-89. Bradford, M.A. and Azhari, M. (1995). Inelastic local buckling of plates and plate assemblies using bubble functions. Engineering Structures, 17 (2), 95-103. Bruhn, E.F. (1965). Analysis and Design of Flight Vehicle Structures, Tri-State Offset Co., Cincinnati, Ohio. Bryan, G.H. (1891). On the stability of a plane plate with thrusts in its own plane with applications to the buckling of the sides of a ship. Proceedings of London Mathematical Society, 22, 54-67. Chakrabarty, J. (2000). Applied Plasticity, Springer, New York. Chiang, C.M. (1995). Mathematical analysis in engineering: how to use basic tools, Cambridge, New York. Dickinson, S.M. and Li, E.K.H. (1982). On the use of simply supported plate functions in the Rayleigh-Ritz method applied to the flexural vibration of rectangular plates. Journal of Sound and Vibration, 80, 292-297. Dietrich, L., Kawahara, W., and Phillips, A. (1978). An experimental study of plastic buckling of a simply supported plate under edge thrusts, Acta Mechanica, 29, 257-267. Dubey, R.N. and Pindera, M.J. (1977). Effect of rotation of principal axes on effective shear modulus in elastic-plastic solids. Journal of Structural Mechanics, 5, 77-85. Durban, D., Zuckerman, Z. (1999). Elastoplastic buckling of rectangular plates in biaxial compression/tension. International Journal of Mechanical Sciences, 41, 751-765. Durvasula, S. (1970). Buckling of clamped skew plates, AIAA Journal, 8(1), 178181. Durvasula, S. (1971). Buckling of simply supported skew plates. ASCE Journal of Engineering Mechanics, 30 (3), 363-366. El-Ghazaly, H.A., Dubey, R.N. and Sherbourne, A.N. (1984). Elasto-plastic buckling of stiffner plates in beam-to-column flange connections. Computers and Structures, 18 (2), 201-213.
213
References
EL-Ghazalay, H.A., and Sherbourne, A.N. (1986). Deformation theory for elasticplastic buckling analysis of plates under non-proportional planar loading. Computers and Structures, 22 (2), 131-149. Forray, M.J. (1968). Variational Calculus in Science and Engineering, McGraw-Hill, New York. Fried, I., and Schmitt, K. (1972). Numerical results from the application of gradient iterative techniques to the finite element vibration and stability analysis of skew plates. Aeronautical Journal, 76, 166-169. Geannakakes, G.N. (1995). Natural frequencies of arbitrarily shaped plates using the Rayleigh-Ritz method together with natural co-ordinate regions and normalized characteristic orthogonal polynomials. Journal of Sound and Vibration, 182(3), 441-478. Gerard, G. (1948). Critical shear stress of plates above the proportional limit. Journal of Applied Mechanics, 15 (1), 7-12. Gerard, G., Becker, H. (1957). Handbook of structural stability: part I buckling of flat plates. NACA Technical Note No. 3781, Washington, D.C. Gerard, C.P. (1978). Asymmetric vibration and stability of circular plates. Computers & Structures, 9, 89-95. Gjelsvik, A. and Lin, G.S. (1987). Plastic Buckling of Plates with Edge frictional Shear effects. Journal of Engineering Mechanics, 113 (7), 953-964. Gupta, K.K. (1978). On a numerical solution of the plastic buckling problem of structures. International for Numerical Methods in Engineering, 12, 941-947. Han, L.S. (1960). The buckling and deflection of isosceles triangular plates. ASME, Journal of Applied Mechanics, 27, 207-208. Hamada, H. (1959). Compressive and shearing buckling load on a rhomboidal plate with all edges clamped. Transaction of Japan Society of Mechanical Engineering, 25(151), 174-181. Handelman, G.H. and Prager, W. (1948). Plastic buckling of rectangular plates under edge thrusts. NACA Technical Note No. 1530, Washington, D.C. Hencky, H.Z. (1924). Theorie Plastischer Deformationen und der hierdurch im Material hervorgerufenen Nachspannungen. Zeitschrift fr Angewandte Mathematik und Mechanik (Journal of Applied Mathematics and Mechanics), 4, 323-334. Hill, H.N. (1944). Determination of stress-strain relations from Offset yield strength values. NACA Technical Note No. 927, Washington, D.C.
214
References
Hill, R. (1950). The mathematical theory of plasticity, Oxford, London. Hou, Y., Wei, G.W., Xiang, Y. (2005). DSC Ritz method for the free vibration analysis of Mindlin plates. International Journal for Numerical Methods in Engineering, 62, 262-288. Huyton, P., York, C.B. (2001). Buckling of skew plates with continuity or rotational edge restraint. Journal of Aerospace Engineering, 14(3), 92-101. Hutchinson, J.W. (1974). Plastic buckling. Advances in Applied Mechanics, Vol. 14, 67-144. Hutchinson, J.W., and Budiansky, B. (1974). Analytical and numerical study of the effects of initial imperfections on the inelastic buckling of a cruciform column. Proceedings of the IUTAM Symposium on Buckling of Structures, 98-105. Illyushin, A.A. (1946). The elastic plastic stability of plates, NACA Technical Memorandum No. 1188, Washington, D.C. Irie, T., Yamada, G. and Takagi, K. (1982). Natural Frequencies of Thick Annular Plates. Journal of Applied Mechanics, 49, 633-638. Kirchhoff, G. (1850). Uber das gleichgwich und die bewegung einer elastischen scheibe. J. Reine Angew. Math., 40, 51-88. Kitipornchai, S., Xiang, Y., Wang, C.M., Liew, K.M. (1993). Buckling of thick skew plates. International Journal for Numerical Methods in Engineering, 36, 12991310. Kosel, F., and Bremec, B. (2004). Elastoplastic buckling of circular annular plates under uniform in-plane loading. Thin-Walled Structures, 42, 101-117. Laura, P.A.A. Gutierrez, R.H. and Bhat, R.B. (1989). Transverse vibrations of a trapezoidal cantilever plate of variable thickness. AIAA Journal, 27, 921-922. Laura, P.A.A., Gutirrez, R.H., Sanzi, H.C., Elvira, G. (2000). Buckling of circular, solid and annular plates with an intermediate circular support. Ocean Engineering, 27, 749-755. Leissa, A.W. (1973). The free vibration of rectangular plates. Journal of Sound and Vibration, 31, 257-293. Leissa, A.W. (2005). The historical bases of the Rayleigh and Ritz methods. Journal of Sound and Vibration, 287, 961-978. Liew, K.M. (1990). The Development of 2-D Orthogonal Polynomials for Vibration of Plates, Ph.D. Thesis, National University of Singapore, Singapore.
215
References
Liew, K.M. and Wang, C.M. (1992). Vibration analysis of plates by pb-2 RayleighRitz method: mixed boundary conditions, reentrant corners and curved internal supports. Mechanics of Structures and Machines, 20(3), 281-292. Liew, K.M. and Wang, C.M. (1993). pb-2 Rayleigh-Ritz method for general plate analysis. Engineering Structures, 15(1), 55-60. Liew, K.M., Xiang, Y., Kitipornchai, S., and Wang, C.M. (1994a). Buckling and vibration of annular Mindlin plates with internal concentric ring supports subject to in-plane radial pressure. Journal of Sound and Vibration, 177 (5), 689-707. Liew, K.M., Xiang, Y., Kitipornchai, S., and Lim, M.K. (1994). Vibration of Mindlin plates on point supports using constraint functions. Journal of Engineering Mechanics, 120 (3), 499-513. Liew, K.M. and Wang, C.M. (1995). Elastic buckling of regular polygonal plates. Thin-Walled Structures, 21, 163-173. Liew, K.M., Wang, C.M., Xiang, Y. and Kitipornchai, S. (1998). Vibration of Mindlin Plates: Programming the p-Version Ritz Method, Elsevier, Oxford, U.K. Liew, K.M. Chen, X.L. Reddy, J.N. (2004). Mesh-free radial basis function method for buckling analysis of non-uniformly loaded arbitrarily shaped shear deformable plates. Computer Methods in Applied Mechanics and Engineering, 193, 205-224. Liu, Y.G. and Pavlovic, M.N. (2005). Elastic Stability Analysis of Simply Supported Rectangular Plates under Arbitrary Loads. Proceedings of the Tenth International Conference on Civil, Structural and Environmental Engineering Computing, B.H.V. Topping, (Editor), Civil-Comp Press, Stirling, United Kingdom, paper 131. Majumdar, S. (1971). Buckling of a thin annular plate under uniform compression. AIAA Journal, 9 (9), 1701-1707. Meirovitch, L., and Kwak, M.K. (1990). Convergence of the classical Rayleigh-Ritz method and the finite element method. AIAA Journal, 28 (8), 1509-1516. Mendelson, A. (1968). Plasticity: Theory and Application, Macmillan Company, New York. Mindlin, R.D. (1951). Influence of rotatory inertia and shear on flexural motions of isotropic, elastic plates. Transaction of ASME, Journal of Applied Mechanics, 18, 31-38. Mizusawa, T., Kajita, T., and Naruoka, M. (1980). Analysis of skew plate problems with various constraints. Journal of Sound and Vibration, 73(4), 575-584. Mizusawa, T. (1986). Natural frequencies of rectangular plates with free edges. Journal of Sound and Vibration, 105, 451-459.
216
References
Mizusawa, T., and Leonard, J.W. (1990). Vibration and buckling of plates with mixed boundary conditions. Engineering Structures, 12, 285-290. Naidu, N.R., Raju, K.K., and Rao, G.V. (1990). Post-buckling of a square plate resting on an elastic foundation under biaxial compression. Computers and Structures, 37 (3), 343-345. Narita, Y. and Leissa, A.W. (1990). Buckling studies for simply supported symmetrically laminated rectangular plates. International Journal of Mechanical Sciences, 32, 909-924. Neale, K.W. (1975). Effect of imperfections on the plastic buckling of rectangular plates. Journal of Applied Mechanics, 42, 115-120. Onat, E.T. and Drucker, D.C. (1953). Inelastic instability and incremental theories of plasticity. Journal of the Aeronautical Science 20, 181-186. Ore, E. and Durban, D. (1989). Elastoplastic buckling of annular plates in pure shear. Journal of Applied Mechanics, Transaction of ASME, 56, 644-651. Pearson, C.E. (1950). Bifurcation criterion and plastic buckling of plates and columns. Journal of Aeronautical Sciences, 17 (7), 417-424. Peters, R.W. (1954). Buckling of long square tubes in combined compression and torsion and comparison with flat-plate buckling theories. NACA Technical Note 3184, Langley Aeronautical Laboratory. Prandtl, L. (1925). Spannungsverteilung in plastischen Koerpern. Proceedings of the 1st International Congress on Applied Mechanics, Delft, Technische Boekhandel en Druckerij, J. Waltman, Jr., 43-54. Pride, R.A. and Heimerl, G.J. (1949). Plastic buckling of simply supported compressed plates, NACA Technical Note No. 1817, Washington, D.C. Raju, K.K., Rao, G.V. (1989). Stability of moderately thick annular plates subjected to a uniform radial compressive load at the outer edge. Computers and Structures, 33 (2), 477-482. Ramberg, W. and Osgood, W. (1943). Description of stress-strain curves by three parameters. NACA Technical Note No. 902, Washington, D.C. Rayleigh, L. (1877). The Theory of Sound, Vol. 1, The Macmillan Company (reprinted 1945 by Dover Publications, New York). Rayleigh, L. (1911). On the calculation of Chladnis figures for a square plate. Philosophical Magazine Sixth Series, 22, 225-229. Reissner, E. (1945). The effect of transverse shear deformation on the bending of elastic plates. Journal of Applied Mechanics, 12, A69-A77.
217
References
Reuss, E. (1930). Beruecksichtigung der elastischen Formaenderungen in der Plastizitaetstheorie. Zeitschrift fr Angewandte Mathematik und Mechanik (Journal of Applied Mathematics and Mechanics), 10, pp. 266-274, 1930. Ritz, W. (1909). ber eine neue Methode zur Lsung gewisser Variationsprobleme der mathematischen Physik. Journal fuer die Reine und Angewandte Mathematik, 135, 1- 61. Saadatpour, M.M., Azhari, M., and Bradford, M.A. (1998). Buckling of arbitrary quadrilateral plates with internal supports using Galerkin method. Computer Methods in Applied Mechanics and Engineering, 164, 297-306. Saadatpour, M.M., Azhari, M., and Bradford, M.A. (2002). Analysis of general quadrilateral orthotropic thick plates with arbitrary boundary conditions by the Rayleigh-Ritz method. International Journal for Numerical Methods in Engineering, 54, 1087-1102. Sewell, M.J. (1964). A general theory of elastic and inelastic plate failure-II. Journal of the Mechanics and Physics of Solids, 12, 279-297. Shanley, F. R. (1947). Inelastic column theory Journal of Aeronautical Science, 14(5), 261-267. Shen, H.S. (1999). Postbuckling of Reissner-Mindlin plates under biaxial compression and lateral pressure and resting on elastic foundations, Computer Methods in Applied Mechanics Engineering, 173, 135-146. Shiaoka, Y. (1956). On the buckling of an elliptical plate with clamped edge I. Journal of Physics Society of Japan, 11(10) 1088-1091. Shrivastava, S.C. (1979). Inelastic buckling of plates including shear effects. International Journal of Solids and Structures, 15, 567-575. Shrivastava, S.C. (1995). Inelastic buckling of rectangular sandwich plates. International Journal of Solids and Structures, 32, 1099-1120. Smith, S.T., Boyle, J.M., Garbow, B.S. Ikebe, Y., Klema, V.C. and Moler, C.B. (1974). Matrix Eigensystem Routines-EISPAC Guide. Springer-Verlag, New York. Smith, S.T., Bradford, M.A. and Oehlers, D.J. (1999). Elastic buckling of unilaterally constrained rectangular plates in pure shear. Engineering Structures, 21, 443-453. Smith, S.T., Bradford, M.A. and Oehlers, D.J. (2003). Inelastic buckling of rectangular steel plates using a Rayleigh-Ritz method. International Journal of Structural Stability and Dynamics, 3(4), 503-521. Stowell, E.Z. (1948). A unified theory of plastic buckling of columns and plates, NACA Technical Note No. 1556, Washington, D.C.
218
References
Su, G.H., and Xiang, Y. (1999). Elastic buckling of rectangular plates with mixed boundary conditions. Proceedings of 16th Australasian Conference on the Mechanics of Structures and Materials, Sydney, Australia, 653-658. Su, Y.H., and Spearing, S.M. (2004). Nonlinear buckling of microfabricated thin annular plates. Thin-Walled Structures, 42, 1543-1565. Surendran. N. (2001). Buckling of Rectangular Mindlin Plates via Ritz Method. M. Eng. Thesis, National University of Singapore, Singapore. Thevendran, V. and Wang, C.M. (1996). Buckling of annular plates elastically restrained against rotation along edges. Thin-Walled Structures, 25 (3), 231-246. Timoshenko, S. P. and Gere, J. M. (1961). Theory of Elastic Stability. Mc-Graw-Hill, New York. Tugcu, P. (1991). Plastic buckling in the plastic range. International Journal of Mechanical Science, 33 (1), 1-11. Tugcu, P. (1996). Effect of axial loading on plastic buckling of long strip under pure shear. Computers and Structures, 66, 155-161. Uenoya, M. and Redwood, R.G. (1989). Elasto-plastic shear buckling of square plates with circular holes. Computers and Structures, 8, 291-300. Venugopal, N., Shastry, B.P., and Rao, G.V. (1989). Stability of square plates resting on four symmetrically placed point supports on diagonals. Computers and Structures, 31 (2), 293-295. Walker, A.C. (1984). A brief review of plate buckling research, In Behaviour of thin-walled structures, ed. by Rhodes, J. and Spence, J. Elsevier, London, 375-397. Wang, C.M., Liew, K.M., Alwis, W.A.M. (1992). Buckling of skew plates and corner condition for simply supported edges. Journal of Engineering Mechanics, 118(4), 651-662. Wang, C.M., Kitipornchai, S., Xiang, Y., Liew, K.M. (1993a). Stability of skew Mindlin plates under isotropic inplane pressure, Journal of Engineering Mechanics, ASCE, 119(2), 393-401. Wang, C.M., and Liew, K.M. (1993b). Buckling of elliptical plates under uniform pressure. Journal of Structural Engineering, 119 (11), 3418-3425. Wang, C.M., Xiang, Y., Kitipornchai, S., Liew, K.M. (1993c). Axisymmetric buckling of circular Mindlin plates with ring supports. Journal of Structural Engineering, 119(3), 782-793.
219
References
Wang, C.M., Wang, L., and Liew, K.M. (1994a). Vibration and buckling of super elliptical plates. Journal of Sound and Vibration, 171 (3), 301-314. Wang, C.M., Xiang, Y., Kitipornchai, S. and Liew, K.M. (1994b). Buckling solutions for Mindlin plates of various shapes. Engineering Structures, 16 (2), 119-127. Wang, C.M., Xiang, Y. and Kitipornchai, S. (1995). Vibration frequencies for elliptical and semi-elliptical Mindlin plates, Structural Engineering and Mechanics, 3 (1), 35-48. Wang, C.M., Xiang, Y. and Chakrabarty, J. (2001a). Elastic/plastic buckling of thick plates. International Journal of Solids and Structures, 38, 8617-8640. Wang, C.M., Xiang, Y., and Wang, C.Y. (2001b). Buckling and vibration of plates with an internal line-hinge via Ritz method. Proceedings of the First Asian-Pacific Congress on Computational Mechanics, Sydney, 1663-1672. Wang, C.M. (2004). Plastic buckling of simply supported, polygonal Mindlin plates, Journal of Engineering Mechanics, 130(1), 117-122. Wang, C.M., Chen, Y., Xiang, Y. (2004). Plastic buckling of rectangular plates subjected to intermediate and end inplane loads. International Journal of Solids and Structures, 41, 4279-4297. Wang, C.M. and Tun Myint Aung. (2005). Buckling of circular Mindlin plates with an internal ring support and elastically restrained edge. Journal of Engineering Mechanics, 131 (4), 359-366. Wang, C.Y., and Wang, C.M. (2001). Buckling of circular plates with an internal ring support and elastically restrained edges, Thin-Walled Structures, 39, 821-825. Wang, C.Y. (2003). Buckling of a circular plate with internal elastic ring support, Mechanics Based Design of Structures and Machines, 31 (1), 93-102. Wang, C.Y. (2005). On the buckling of a circular plate on an elastic foundation. Journal of Applied Mechanics, 72, 795-796. Wang, X., Tan, M., Zhou, Y. (2003). Buckling analyses of anisotropic plates and isotropic skew plates by the new version differential quadrature method. ThinWalled Structures, 41, 15-29. Wittrick, W.H. (1953). Buckling of oblique plates with clamped edges under uniform compression. Aeronautical Quarterly, 4, 151-163. Wittrick, W.H. (1954). Buckling of oblique plates with clamped edges under uniform shear. Aeronautical Quarterly, 5, 39-51. Woinovsky-Krieger, S. (1937). The stability of a clamped elliptic plate under uniform compression. Journal of Applied Mechanics, 4(4), 177-178.
220
References
Wolfram, S. (1999). Mathematica book, 4th ed. Cambridge University Press, New York. Xiang, Y., Wang, C.M., Kitipornchai, S., and Liew, K.M. (1994). Buckling of triangular Mindlin plates under isotropic inplane compression. Acta Mechanica, 102, 123-135. Xiang, Y., Wang, C.M., Kitipornchai, S. (1995). Buckling of skew Mindlin plates subjected to in-plane shear loadings, International Journal of Mechanical Science, 37(10), 1089-1101. Xiang, Y., Wang, C.M. and Wang, C.Y. (2001). Buckling analysis of rectangular plates with internal hinge. International Journal of Structural Stability and Dynamics, 1 (2), 169-179. Xiang, Y. (2003). Exact solutions for buckling of multispan rectangular plates. Journal of Engineering Mechanics, 129 (2), 181-187. Yamaki, N. (1958). Buckling of a thin annular plate under uniform compression. Journal of Applied Mechanics, ASME, 25, 267-273. York, C.B., Williams, F.W. (1995). Buckling analysis of skew plate assemblies: classical plate theory results incorporating lagrangian multipliers. Computers and Structures, 56(4), 625-635. Yoshimura, Y., and Iwata, K. (1963). Buckling of simply supported oblique plates. ASME Journal of Applied Mechanics, 30 (3), 363-366.
221
Appendix A: Ritz Program Codes

This Appendix is prepared in MATHEMATICA since it is more convenient to print the program code directly from MATHEMATICA than copying to word file. In MATHEMATICA, comments on the code can be written in (*.....*) or seperate cells. In oder to explain the program code properly, comments are written in seperate cells as a text.
A.1 Ritz Program Code with Catesian Coordinate System
BUCKRITZRE
This program is based on the Catesian coordinate formulations of Ritz method and is to solve the plastic buckling problems of Mindlin plates under any combinations of in-plane and shear stresses with various boundary conditions. Firstly basic functions need to be defined and the sample basic functions for a SSSS rectangular plate are given by fw1 = Hx + 1L Hh + 1L Hx - 1L Hh - 1L; fx1 = Hx + 1L0 Hh + 1L Hx - 1L0 Hh - 1L; fy1 = Hx + 1L Hh + 1L0 Hx - 1L Hh - 1L0 ;
222
Appendix A
The mathematically complete polynomial functions are formed as a list. Hence the displacement functions w, fx, fy are obtained by multiplying the corresponding basic function with polynomial terms. p = 6; H p is the degree of polynomial equation L H# + 1L H# + 2L Nmax = &@pD; 2 polyfunction = Flatten@Reverse @ Table@i qi , 8q, 0, p<, 8i, 0, q<DD; w = polyfunction@@#DD w1 &; x = polyfunction@@#DD x1 &; y = polyfunction@@#DD y1 &; i 1 h j j j j x xh j j j The integration of the matrix j j j j j Aj j j N-1 N-1 x h kx hN-1 y z z z N-1 z z xh z z A are z z z z z z z N-1 hN-1 z x {
carried out over the plate domain for one time only and store it in a variable
named inte as an array. Then a user defined integration function is written in such a way that the exponents of x and h in all the integrands are tested
and use the coorsponding integrated values in inte. It is more computationally efficient since the integration of polynomial are repeatedly involved in the calculation.
223
Appendix A
Table@i q , 8i, 0, 2 p + 5<, 8q, 0, 2 p + 5<D; co = Coefficient@Coefficient@#, , Exponent@#, DD, , Exponent@#, DD &; Hco is the function to get coefficient of the polynomial functionL preintefunct = If@TrueQ@# 0D, 0, co@#D Part@inte, Exponent@#, D + 1, Exponent@#, D + 1DD &; intefunction = If@TrueQ@Head@#D PlusD, Map@preintefunct, #D, preintefunct@#DD &; The matrix elements are formed according to the Eqs. (3.23a-f) Table@ w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kcc1 = Map@intefunction, %, 82<D; Table@ w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kcc2 = Map@intefunction, %, 82<D; Table@ w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kcc3 = Map@intefunction, %, 82<D; Table@ w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kcc4 = Map@intefunction, %, 82<D; Table@ w@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kcd1 = Map@intefunction, %, 82<D; inte = # & @ %;
1 1 1 1
Table@ w@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kce1 = Map@intefunction, %, 82<D; Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd1 = Map@intefunction, %, 82<D;
224
Appendix A
Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd2 = Map@intefunction, %, 82<D; Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd4 = Map@intefunction, %, 82<D; Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd5 = Map@intefunction, %, 82<D;
Table@x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd3 = Map@intefunction, %, 82<D;
Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kde1 = Map@intefunction, %, 82<D; Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kde2 = Map@intefunction, %, 82<D;
Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kde3 = Map@intefunction, %, 82<D; Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kde4 = Map@intefunction, %, 82<D;
Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee1 = Map@intefunction, %, 82<D; Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee2 = Map@intefunction, %, 82<D;
225
Appendix A
Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee3 = Map@intefunction, %, 82<D; Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee4 = Map@intefunction, %, 82<D; Table@y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee5 = Map@intefunction, %, 82<D;
The stiffness matrix can be obtained as a function of lx, ly and lxy. The algorithm shown here is for the deformation theory of plasticity. << LinearAlgebra`MatrixManipulation` stiffness@x_, y_, xy_D := ModuleA 8t, s, c11, c12, c13, c22, c23, c33, c44<, "################################ ################### # # e = x2 x y + y2 + 3 xy2 ; ly = 1; = lx ly; = h ly; c1 1 i y y i j He e 2 Lz z; t = 1 j1 + c k j z z j j z { { k 12 H1 2 L k c1 2 i y y j z z; j j1 + k i e e z z j z s=1 j j j z 2L z 12 H1 { z k k { xy2 y i y2 z; z j c11 = 1 3 H1 t sL j + j z 2 2 e { k 4 e 1 c12 = 2 2 i j1 H1 2 L t 3 H1 t sL i x y + xy j j j j j 2 e2 k k 2 e 3 i 2 x y y xy z c13 = H1 t sL j ; j z 2 e k { e xy2 y i x2 z; j z j + c22 = 1 3 H1 t sL j z 2 2 4 e e { k 3 i 2 y x y xy z c23 = H1 t sL j ; j z 2 e k { e
yy zz; zz zz {{
226
Appendix A
E;
H3 t s H1 2 L tL; = c22 c33 c232 ; = c13 c23 c12 c33; = c12 c23 c13 c22; = c11 c33 c132 ; = c12 c13 c11 c23; = c11 c22 c122 ; = 1 t Det@88c11, c12, c13<, 8c12, c22, c23<, 8c13, c23, c33<<D; t i1 y j kcc1 + kcc2z; kc = j z c44 h2 k { 1 i 1 x j gcc = kcc1 + j 12 H1 2 L ly2 k y xy y z kcc2 + Hkcc3 + kcc4Lz; { kcc = kc; ly t kcd = kcd1; 2 c44 h2 ly t kce = kce1; 2 c44 h2 1 kdd = kdd1 + kdd2 + 12 12 t kdd3 + kdd4 + kdd5; 4 c44 2 12 12 kde = kde1 + kde2 + 12 12 1 kde3 + kde4; 12 12 1 kee = kee1 + kee2 + 12 12 t kee5; kee3 + kee4 + 12 12 4 c44 2 rowc = AppendRows@kcc, kcd, kceD; rowd = AppendRows@Transpose@kcdD, kdd, kdeD; rowe = AppendRows@ Transpose@kceD, Transpose@kdeD, keeD; mat = Map@Chop, AppendColumns@ rowc, rowd, roweD, 82<D; zeromat = ZeroMatrix@NmaxD; matkm = BlockMatrix@88gcc, zeromat, zeromat<, 8zeromat, zeromat, zeromat<, 8zeromat, zeromat, zeromat<<D; Return@mat matkmD; c44 = 1
c33 = 3 t s H1 2 L t + 9 H1 t sL
xy2 e2
227
Appendix A
The input data for material properties, geometrical properties and loading conditions need to be defined here (sample input data are for AL 7075 T6 aluminum alloy). Then an angorithm (the detailed procedure was discussed in Chapter 3) is written in oder to determine the plastic buckling stress parameter l. Since, the stiffness matrix was formed without having input data, the analysis can be perform for another material properties or aspect ratio without having to recalculate stiffness matrix. x := ; y := 0; xy := 0;
= 5 6; = 0.33; e = 150; He means E0L k = 3 7; c = 9.2; h = 0.05; lx = 1; BucklingLoadFinder@0_D := Module@8n, num, W, W1, Wmid, d<, = 0; d = 5; Do@ Do@ n = Eigenvalues@ Chop@stiffness@x, y, xyD NDD; num = Length@Select@n, Sign@#D 1 &DD; If@num > 0, Print@num, " buckling load is less than ", ND; Break@D, = + d D, 8i, 50< D; If@num > 1, = d; d = d 2; = + d; Print@D, Break@D D, 8a, 10< D; If@num < 1, Print@"Initial Value of is too small"D; Abort@D
228
Appendix A
D;
W2 = Det@stiffness@x, y, xyDD; = d; W1 = Det@stiffness@x, y, xyDD;
If@Sign@W1D Sign@W2D, Print@"Limit is not OK"D; Abort@D, Print@"Limit is OK"D D; Do@d = d 2; mid = + d; If@d < 0.00005, Break@D D; = mid; Wmid = Det@stiffness@x, y, xyDD; Print@"=", " ", SetPrecision@mid, 8D, "W1=", " ", W1, "Wmid=", " ", WmidD; If@Sign@W1D Sign@WmidD, = mid dD, 8i, 40< D; Print@"h=", h, " ", "c=", c, " ", "=", " ", mid Pi2 D D; H0 is the initial value of L HThe following command will load the algorithm and determine L BucklingLoadFinder@1.0D
229
Appendix A
The buckling mode shape can be ploted as follow:
Clear@D; = mid; eigen = Eigenvalues@ 8Chop@mat ND, Chop@matkm ND<D; eigenew = If@# 0, , #D & @ eigen; po = Ordering@eigenew, 1D; Print@"=", eigenew@@po@@1DDDDD; v = Eigenvectors@ 8Chop@mat ND, Chop@matkm ND<D; Co = v@@po@@1DDDD; defl = HTable@Co@@iDD polyfunction@@iDD w1, 8i, Nmax<DL . List Plus; Plot3D@defl, 8, 1, 1<, 8, 1, 1<, Axes False, BoxRatios > 8lx, 1, 0.3<, PlotRange AllD Clear@"Global`"D
A.2 Ritz Program Code with Skew Coordinate System

BUCKRITZSKEW
This program is based on the skew coordinate formulations of Ritz method and is to solve the plastic buckling problems of skew Mindlin plates under any combinations of in-plane and shear stresses with various boundary conditions. The choice of basic functions and the use of integration function are similar to those in BUCKRITZRE. Firstly basic functions need to be defined and the sample basic functions for a SSSS skew plate are given by
230
Appendix A
The mathematically complete polynomial functions are formed as a list. Hence the displacement functions w, fx, fy are obtained by multiplying the corresponding basic function with polynomial terms. p = 6; H p is the degree of polynomial equation L H# + 1L H# + 2L Nmax = &@pD; 2 polyfunction = Flatten@Reverse @ Table@i qi , 8q, 0, p<, 8i, 0, q<DD; w = polyfunction@@#DD w1 &; x = polyfunction@@#DD x1 &; y = polyfunction@@#DD y1 &; i 1 h j j j j x xh j j j The integration of the matrix j j j j j Aj j N-1 N-1 j x h kx hN-1 y z z N-1 z z z xh z z A are z z z z z z z N-1 hN-1 z x {
fw1 = Hx + 1L Hh + 1L Hx - 1L Hh - 1L; fx1 = Hx + 1L0 Hh + 1L Hx - 1L0 Hh - 1L; fy1 = Hx + 1L Hh + 1L0 Hx - 1L Hh - 1L0 ;
carried out over the plate domain for one time only. Then an integration function is written in such a way that the exponents of x and h in all the
integrands are tested and use the coorsponding integrated values. It is more computationally efficient since the integration of polynomial are repeatedly involved in the calculation.
231
Appendix A
Table@i q , 8i, 0, 2 p + 5<, 8q, 0, 2 p + 5<D; inte = # & @ %;

1 1 1 1
The matrix elements are formed according to the Eqs.(3.34a-f) HElements for kccL
co = Coefficient@Coefficient@#, , Exponent@#, DD, , Exponent@#, DD &; Hco is the function to get coefficient of the polynomial functionL preintefunct = If@TrueQ@# 0D, 0, co@#D Part@inte, Exponent@#, D + 1, Exponent@#, D + 1DD &; intefunction = If@TrueQ@Head@#D PlusD, Map@preintefunct, #D, preintefunct@#DD &; Hintefunction is the function to integrate the polynomial by using known value inteL
Table@ w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k1 = Map@intefunction, %, 82<D; Table@ w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k3 = Map@intefunction, %, 82<D; Table@ w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k4 = Map@intefunction, %, 82<D; Table@ w@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k5 = Map@intefunction, %, 82<D; HElements for kcdL Table@ w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k2 = Map@intefunction, %, 82<D;
Table@ w@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k6 = Map@intefunction, %, 82<D;
232
Appendix A
Table@ w@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k7 = Map@intefunction, %, 82<D; Table@ w@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k8 = Map@intefunction, %, 82<D; Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k9 = Map@intefunction, %, 82<D; HElements for kddL
HElements for kceL
Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k10 = Map@intefunction, %, 82<D; Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k11 = Map@intefunction, %, 82<D; Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k12 = Map@intefunction, %, 82<D; HElements for KdeL Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k13 = Map@intefunction, %, 82<D;
Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k14 = Map@intefunction, %, 82<D; Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k15 = Map@intefunction, %, 82<D;
233
Appendix A
Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k17 = Map@intefunction, %, 82<D; HElements for keeL
Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k16 = Map@intefunction, %, 82<D;
Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k19 = Map@intefunction, %, 82<D;
Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k18 = Map@intefunction, %, 82<D;
Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k20 = Map@intefunction, %, 82<D; Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k21 = Map@intefunction, %, 82<D; Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k22 = Map@intefunction, %, 82<D;
The stiffness matrix can be obtained as a function of lx,ly and lxy.The algorithm shown here is for the deformation theory of plasticity. << LinearAlgebra`MatrixManipulation` stiffness@x_, y_, xy_D :=
Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; k23 = Map@intefunction, %, 82<D;
234
Appendix A
ModuleA8t, s, c11, c12, c13, c22, c23, c33, , , , , , <, "################################ ################### # # e = x2 x y + y2 + 3 xy2 ; ly = 1; = ly lx; = h ly; c1 2 i y y j j z z; j1 + c k i e e j z z t= 1 j j z z z j 12 H1 2 L { { k k c1 2 i y y j j z z; j1 + k i e e j z z z s=1 j j j z 2L z 12 H1 { { k k xy2 y i y2 z; j z + c11 = 1 3 H1 t sL j j z 2 2 e { k 4 e 1 c12 = 2 2 i j1 H1 2 L t 3 H1 t sL i x y + xy j j j j j 2 e2 k k 2 e 3 i 2 x y y xy z c13 = H1 t sL j ; j z 2 e k { e xy2 y i x2 j z; j z c22 = 1 3 H1 t sL j + z 2 2 4 e e { k 3 i 2 y x y xy z c23 = H1 t sL j ; j z 2 e k { e xy2 c33 = 3 t s H1 2 L t + 9 H1 t sL ; e2 1 c44 = H3 t s H1 2 L tL; = c22 c33 c232 ; = c13 c23 c12 c33; = c12 c23 c13 c22; = c11 c33 c132 ; = c12 c13 c11 c23; = c11 c22 c122 ; = 1 t Det@88c11, c12, c13<, 8c12, c22, c23<, 8c13, c23, c33<<D; kcc1 = Sec@D c44 h2 1 i y j k1 + k2 Sin@D k3 Sin@D k4z; j z k { 1 gcc = 12 H1 2 L ly2 x y i 1 i j j k1 + Sec@D j Cos@D j k2 k k t
yy zz; zz zz {{
y Sin@D k3 Sin@D k4 + Sin@D2 k1z z {
235
Appendix A
y Hk3 + k4 2 Sin@D k1Lz z; { kcc = kcc1; t ly i 1 j k5 Sin@D k6 y z kcd = j z; 2 c44 h2 k { xy kce = t ly 12 2 c44 h2 kdd1 = Cos@D3 k9; i 1 y j k7 Sin@D k8z j z; k {
i Tan@D Sin@D j jSin@D2 k9 + 12 k 1 y k10 Sin@D k11 Sin@D k12z z; { kdd3 = Sin@D Cos@D 12 H2 Sin@D k9 k11 k12L; i kdd4 = Cos@D j4 Sin@D2 k9 j 12 k 1y z 2 Sin@D k11 2 Sin@D k12 + k10 z; { kdd5 = Cos@D2 Hk11 + k12 4 Sin@D k9L; 12 i 2 j kdd6 = Sin@D j k10 + 12 k y z 3 Sin@D Hk11 + k12L 4 Sin@D2 k9z; { t 1 kdd7 = Cos@D k13; 4 c44 2 kdd = kdd1 + kdd2 + kdd3 + kdd4 + kdd5 + kdd6 + kdd7; kdd2 = i Tan@D j Sin@D2 k14 + j k 1y Sin@D k15 + Sin@D k16 k17 z z; { kde2 = Cos@D H Sin@D k14 + k15L; 12 kde3 = Cos@D H 2 Sin@D k14 + k16L; 12 kde4 = Cos@D2 k14; 12 i j Sin@D2 k14 + kde5 = j3 12 k 1 y k17 2 Sin@D Hk15 + k16Lz z; { t 1 kde6 = Sin@D Cos@D k18; 4 c44 2 kde1 = 12
236
Appendix A
kde = kde1 + kde2 + kde3 + kde4 + kde5 + kde6; 1 i Sec@D j + jk19 12 k y Sin@D2 k20 Sin@D k21 Sin@D k22z; z { kee2 = Cos@D k20 ; 12 kee3 = H2 Sin@D k20 + k21 + k22L; 12 t 1 kee4 = Cos@D k23; 4 c44 2 kee = kee1 + kee2 + kee3 + kee4; kee1 = rowc = AppendRows@kcc, kcd, kceD; rowd = AppendRows@Transpose@kcdD, kdd, kdeD; rowe = AppendRows@ Transpose@kceD, Transpose@kdeD, keeD; mat = Map@Chop, AppendColumns@ rowc, rowd, roweD, 82<D; zeromat = ZeroMatrix@NmaxD; matkm = BlockMatrix@88gcc, zeromat, zeromat<, 8zeromat, zeromat, zeromat<, 8zeromat, zeromat, zeromat<<D; Return@mat matkmD;
The input data or material properties and geometrical properties need to be defined here (sample input data are for AL 7075 T6 aluminum alloy and S shear loading).The angorithm for the calculation of buckling stress parameter is the same as in BUCKRITZRE. = 30 Pi 180; Hskew angle in radianL x = 2 Tan@D; y = 0; xy = ; = 33 100; = 5 6; e = 150; He means E0L k = 3 7; BucklingLoadFinder@0_D := Module@8n, num, W, W1, Wmid, d<, = 0; d = 5;
E;
237
Appendix A
Do@ Do@ n = Eigenvalues@ Chop@stiffness@x, y, xyD NDD; num = Length@Select@n, Sign@#D 1 &DD; If@num > 0, Print@num, " buckling load is less than ", ND; Break@D, = + d D, 8i, 50< D; If@num > 1, = d; d = d 2; = + d; Print@D, Break@D D, 8a, 10< D; If@num < 1, Print@ "Initial Value of is too small"D; Abort@D D; W2 = Det@stiffness@x, y, xyDD; = d; W1 = Det@stiffness@x, y, xyDD; If@Sign@W1D Sign@W2D, Print@"Limit is not OK"D; Abort@D, Print@"Limit is OK"D D; Do@d = d 2; mid = + d; If@d < 0.00005, Break@D D; = mid; Wmid = Det@stiffness@x, y, xyDD; Print@"=", " ", SetPrecision@mid, 8D, ", WmidD; "W1=", " ", W1, "Wmid=", " If@Sign@W1D Sign@WmidD, = mid dD, 8i, 40< D; Print@"h=", h, " ", "c=", c, " ", "=", " ", mid Pi2 D D; H0 is the initial value of L HThe following command will load the algorithm and determine L BucklingLoadFinder@1.0D
238
Appendix A
Clear@D; = mid; stiffness@x, y, xyD; eigen = Eigenvalues@ 8Chop@mat ND, Chop@matkm ND<D; eigenew = If@# 0, , #D & @ eigen; po = Ordering@eigenew, 1D; Print@"=", eigenew@@po@@1DDDDD; v = Eigenvectors@ 8Chop@mat ND, Chop@matkm ND<D; Co = v@@po@@1DDDD; defl = HTable@Co@@iDD polyfunction@@iDD w1, 8i, Nmax<DL . List Plus; Needs@"DrawGraphics`DrawingMaster`"D IteratorSubstitution@8, , defl<, 8, 1 + Tan@D, 1 + Tan@D<, xD Simplify; function = %@@1DD; ParametricPlot3D@function, 8x, 0, 1<, 8, Cos@D, Cos@D<, PlotPoints 825, 25<, BoxRatios 81.5773, 1, 0.3<, ViewPoint 80, 3, 2<, PlotRange All, Boxed False, Axes FalseD Clear@"Global`"D
A.3 Ritz Program Code with Polar Coordinate System
BUCKRITZPOLAR
This program is based on the polar coordinate formulations of Ritz method and is to solve the plastic buckling problems of circular and annular Mindlin plates under in-plane radial stress with various boundary conditions.
239
Appendix A
Firstly basic functions need to be defined and the sample basic functions for a simply supported annular plate are given by w1 = H + 1L H 1L ; HSSSSL x1 = H + 1L0 H 1L0 ; y1 = H + 1L H 1L;
The mathematically complete one dimensional polynomial functions are formed as a list. Hence the displacement functions w, fx, fy are obtained by multiplying the corresponding basic function with polynomial terms. p = 10; H p is the degree of polynomial equation L = 0.4; H=bRL = 2 Pi; polyfunction = Table@i1 , 8i, p<D; Nmax = p; w = polyfunction@@#DD w1 &; x = polyfunction@@#DD x1 &; y = polyfunction@@#DD y1 &; i 1 y i 1 y z z j j j j z z j x z j x z z z j j z z j j H1+zL+H1-zL j z A, z z z j The integration of the matrix j x j j j z 2 z z j j z A, z z j j z z j j j A j 2 p+5 z A z z j j 2 p+5 z { { kx kx i 1 y j z j j x z z j z j z 2 z x j j z H1+zL+H1-zL j z A are carried out over the plate domain for one j z j z j z j A j 2 p+5 z z kx { time only. Then an integration function is written in such a way that the
exponents of x and h in all the integrands are tested and use the coorsponding integrated values. It is more computationally efficient since the integration of polynomial are repeatedly involved in the calculations.
240
Appendix A
function = Table@i , 8i, 1, 2 p + 5<D; co = Coefficient@#, , Exponent@#, DD &; Hco is the function to get coefficient of the polynomial functionL preintefunct = If@TrueQ@# 0D, 0, co@#D Part@inte, Exponent@#, D + 1DD &; inte = # & @ function;
1 1
intefunction = If@TrueQ@Head@#D PlusD, Map@preintefunct, #D, preintefunct@#DD &;
function2 = TableA
inte1 = # & @ function1;

1 1
function1 = H1 + L + H1 L i TableA , 8i, 0, 2 p + 5<E; 2 i , 8i, 0, 2 p + 5<E;
inte2 = # & @ function2;

1 1
H1 + L + H1 L 2
preintefunct1 = If@TrueQ@# 0D, 0, co@#D Part@inte1, Exponent@#, D + 1DD &; intefunction1 = If@TrueQ@Head@#D PlusD, Map@preintefunct1, #D, preintefunct1@#DD &;
preintefunct2 = If@TrueQ@# 0D, 0, co@#D Part@inte2, Exponent@#, D + 1DD &; intefunction2 = If@TrueQ@Head@#D PlusD, Map@preintefunct2, #D, preintefunct2@#DD &;
241
Appendix A
The matrix elements are formed according to the Eqs.(3.47a-f) Table@ w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kcc1 = Map@intefunction1, %, 82<D; Table@w@iD w@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kcc2 = Map@intefunction2, %, 82<D; Table@ w@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kcd1 = Map@intefunction1, %, 82<D; Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd1 = Map@intefunction1, %, 82<D; Table@w@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kce1 = Map@intefunction, %, 82<D;
Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd2 = Map@intefunction2, %, 82<D; Table@x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd3 = Map@intefunction1, %, 82<D; Table@x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd4 = Map@intefunction2, %, 82<D;
Table@x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd6 = Map@intefunction, %, 82<D;
Table@ x@iD x@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kdd5 = Map@intefunction, %, 82<D; Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kde1 = Map@intefunction, %, 82<D;
242
Appendix A
Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kde2 = Map@intefunction, %, 82<D; Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kde3 = Map@intefunction2, %, 82<D; Table@ x@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kde4 = Map@intefunction2, %, 82<D;
Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee2 = Map@intefunction1, %, 82<D;
Table@y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee1 = Map@intefunction2, %, 82<D;
Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee3 = Map@intefunction2, %, 82<D;
Table@ y@iD y@jD , 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee5 = Map@intefunction, %, 82<D; Table@y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee6 = Map@intefunction1, %, 82<D;
Table@ y@iD y@jD, 8i, 1, Nmax<, 8j, 1, Nmax<D; Expand @ %; kee4 = Map@intefunction, %, 82<D;
243
Appendix A
The stiffness matrix can be obtained as a function of lx,ly and lxy.The algorithm shown here is for the deformation theory of plasticity. dropZeros[m_List?MatrixQ] is to drop the matrix comlumns and rows whose elements are zeros. This can be happen when n = 0 because all the elements of K ce , K de , K ee are zeros for n = 0. << LinearAlgebra`MatrixManipulation` S1 := 2 S1 := 1 S2 := 0 S2 := 1 ; n 0 ; n > 0 ; n 0 ; n > 0
notZero@ro_ListD := Or @@ H# 0 &L @ ro dropZeroes@m_List ? MatrixQD := Transpose@Select@ Transpose@Select@m, notZeroDD, notZeroDD stiffness@x_, y_, xy_D := ModuleA 8t, s, c11, c12, c13, c22, c23, c33, c44<, "################################ ################### # # e = x2 x y + y2 + 3 xy2 ; ly = 1; = h ly; c1 1 y i j1 + c k i 2 y j z z; z j t= 1j He e Lz z j k 12 H1 2 L { { k c1 2 i y y j j z z; j1 + k i e e j z z s=1 j j z z j z 12 H1 2 L { { k k xy2 y i y2 j z; j z c11 = 1 3 H1 t sL j + z 2 2 4 e e { k 1 c12 = 2 xy2 i i x y j j j1 H1 2 L t 3 H1 t sL j + j j 2 e2 k k 2 e 3 i 2 x y y xy z c13 = H1 t sL j ; j z 2 e k { e xy2 y i x2 z; j z c22 = 1 3 H1 t sL j + j z 2 2 4 e e { k 3 i 2 y x y xy z c23 = H1 t sL j ; j z 2 e k { e
yy zz; zz zz {{
244
Appendix A
H3 t s H1 2 L tL; = c22 c33 c232 ; = c13 c23 c12 c33; = c12 c23 c13 c22; = c11 c33 c132 ; = c12 c13 c11 c23; = c11 c22 c122 ; = 1 t Det@88c11, c12, c13<, 8c12, c22, c23<, 8c13, c23, c33<<D; HThis is for Deformation theory of plasticityL t i j kc = S1 kcc1 + j c44 2 k H1 L H1 L 2 y n S2 kcc2z z; 4 { 1 x i j gcc = S1 kcc1 + j 12 H1 2 L k H1 L y 2 H1 L y n S2 kcc2z z; 4 { kcc = kc; t kcd = S1 kcd1; c44 2 2 H1 L t kce = n S2 kce1; c44 2 4 kdd = H1 L 2 S1 kdd1 + n S2 kdd2 + 12 H1 L 12 4 H1 L H1 L t S1 kdd3 + 4 12 4 c44 2 S1 kdd4 + S1 kdd5 + S1 kdd6; 24 24 kde = n S1 kde1 n S2 kde2 + 2 12 2 12 12 H1 L H1 L n S1 kde3 + n S2 kde4; 4 12 4 H1 L 2 kee = n S1 kee1 + 12 4 H1 L S2 kee2 + S2 kee3 12 H1 L 12 4 S2 kee4 S2 kee5 + 12 2 12 2 t H1 L S2 kee6; c44 2 4 c44 = 1 rowc = AppendRows@kcc, kcd, kceD;
c33 = 3 t s H1 2 L t + 9 H1 t sL
xy2 e2
245
Appendix A
The input data or material properties and geometrical properties need to be defined here (sample input data are for AL 2024 T6 aluminum alloy).The angorithm for the calculation of buckling stress parameter is the same as in BUCKRITZRE. Note that asymetric buckling mode shape (n 0) is possible and the numbers of nodal diameter n is unknown. So, the lowest buckling stress parameter and corresponding nodal number n are determined by an iteration fashion. Clear@n, D; x := ; y := ; xy = 0; 5 = ; 6 = 32 100; e = 10700 61.4; He means E0L k = 0.3485; c = 20; h = 0.05; BucklingLoadFinder@0_D := Module@8n, num, W, W1, Wmid, d<, = 0; d = 5; Do@ Do@ n = Eigenvalues@ Chop@stiffness@x, y, xyD NDD; num = Length@Select@n, Sign@#D 1 &DD; If@num > 0, Print@num, " buckling load is less than ", ND; Break@D, = + d D, 8i, 50<
246
E;
rowd = AppendRows@Transpose@kcdD, kdd, kdeD; rowe = AppendRows@ Transpose@kceD, Transpose@kdeD, keeD; mat = AppendColumns@rowc, rowd, roweD; matkm = PadRight@dropZeroes@gccD, 8Length@matD, Length@matD<D; Return@mat matkmD;
Appendix A
D; If@num > 1, = d; d = d 2; = + d; Print@D, Break@D D, 8a, 10< D; If@num < 1, Print@ "Initial Value of is too small"D; Abort@D D; W2 = Det@stiffness@x, y, xyDD; = d; W1 = Det@stiffness@x, y, xyDD; If@Sign@W1D Sign@W2D, Print@"Limit is not OK"D; Abort@D, Print@"Limit is OK"D D; Do@d = d 2; mid = + d; If@d < 0.00005, Break@D D; = mid; Wmid = Det@stiffness@x, y, xyDD; Print@"=", " ", SetPrecision@mid, 8D, "W1=", " ", W1, "Wmid=", " ", WmidD; If@Sign@W1D Sign@WmidD, = mid dD, 8i, 40< D; Print@"h=", h, " ", "c=", c, " ", "=", 2 " ", mid Pi ND; Return@midD; D;
n = 0; t = BucklingLoadFinder@1D; n = 1; n = BucklingLoadFinder@1D; While@t n, t = n; n = n + 1; n = BucklingLoadFinder@1D; D; Print@"n = ", n 1, " ", " = ", SetPrecision@t Pi2 , 10DD;
247
Appendix A
Clear@"Global`"D
stiffness@t, t, 0D; Eigenvalues@8mat, matkm<D If@# 0, , #D & @ % po = Ordering@%, 1D v = Eigenvectors@8mat, matkm<D; HEigenvectors@%D;L Co = %@@po@@1DDDD Table@Co@@iDD polyfunction@@iDD w1, 8i, Nmax<D; defl = % . List Plus; pic = ParametricPlot3D@ 8H1 2 HH1 + L + H1 L LL Cos@1 2 D, H1 2 HH1 + L + H1 L LL Sin@1 2 D, defl Cos@n 1 2 D<, 8, 1, 1<, 8, 1, 1<, BoxRatios > 81, 1, 0.25<, Axes > False, Boxed > False, PlotPoints 820, 60<D
248
Appendix B: Elements of Matrices
B.1: Plastic buckling of circular plate under uniform in-plane compression The non-zero elements of the matrix [K ]33 in Eq. (7.22) are given by (i) for simply supported edge
k11 = 1
(B.1) (B.2) (B.3)
k12 = J n ( 2 )
k 21 = (1 n)n( )
k 22 =
2 E 1 2 2 2 12(1 2 ) 2 G 1 2 J n 2 ( 2 ) + 2 2 J n 1 ( 2 ) 2 2 + 2n J n ( 2 ) 4 + 2 J n+1 ( 2 )}+ 22 J n+2 ( 2 ) (B.4)
{(
k 23 = n( ){(n 1)I n ( 3 ) + 3 I n +1 ( 3 )} k 31 = n
2 E k 32 = 1 12(1 2 ) 2 G nJ n ( 2 )
for n = 0 0 k 33 = nI n ( 3 ) 3 I n +1 ( 3 ) for n 0
(B.5) (B.6)
(B.7)
(B.8)
(ii) for clamped edge The first row and third row of matrix elements of [K ]33 in Eq. (7.22) are same as the elements in corresponding rows for simply supported edge. The non-zero elements in second row are for n = 0 0 k 21 = n for n 0 (B.9)
249
Appendix B
2 E k 22 = 12(1 2 ) 2 G 1{nJ n ( 2 ) 2 J n + ( 2 )}
(B.10)
k 23 = nI n ( 3 )
(B.11)
B.2: Plastic buckling of annular plate under uniform in-plane compression
In this appendix, the non-zero elements of the matrix [K ]66 in Eq. (7.25) are presented. The elements in the first three rows depend on the boundary condition of the outer edge while the elements in the next three rows depend on the inner edge condition.
If the outer edge is (i). clamped edge

k11 = 1
(B.12) (B.13) (B.14) (B.15) (B.16)
k12 = J n ( 2 ) 0 k14 = 1 k15 = Yn ( 2 ) for n = 0 0 k 21 = n for n 0 for n = 0 for n 0
k 22 = 2 1{nJ n ( 2 ) 2 J n + ( 2 )} G k 23 = nI n ( 3 ) 1 k 24 = n for n = 0 for n 0
(B.17) (B.18) (B.19)
250
Appendix B
k 25 = 2 1{nYn ( 2 ) 2Yn + ( 2 )} G k 26 = nK n ( 3 ) k 31 = k 34 = n
(B.20) (B.21) (B.22)

(B.23)
k 32 = 1 2 nJ n ( 2 ) G
for n = 0 0 k 33 = nI n ( 3 ) 3 I n +1 ( 3 ) for n 0
(B.24)
k 35 = 1 2 nYn ( 2 ) G
for n = 0 0 k 36 = nK n ( 3 ) + 3 K n +1 ( 3 ) for n 0
(B.25)
(B.26)
(ii). simply supported edge The first row and third row of matrix elements are same as the elements in corresponding rows for clamped edge (i.e. Eqs. B.12 to B.26). The non-zero elements in second row are
k 21 = (1 n)n( )
(B.27)
k 22 =
1 2 2 2 2 1 2 J n 2 ( 2 ) + 2 2 J n 1 ( 2 ) 2 2 + 2n J n ( 2 ) 4 G
{(
+ 2 J n +1 ( 2 )} + 22 J n + 2 ( 2 )
(B.28) (B.29) (B.30)
k 23 = n( ){(n 1)I n ( 3 ) + 3 I n +1 ( 3 )} k 24 = n(1 + n)( ) k 25 = for n = 0 for n 0
1 2 2 2 2 1 2 Yn 2 ( 2 ) + 2 2Yn 1 ( 2 ) 2 2 + 2n Yn ( 2 ) 4 G
{(
2Yn+1 ( 2 )} + 22Yn+ 2 ( 2 )]
(B.31)
251
Appendix B
k 26 = n( ){(n 1)K n ( 3 ) 3 K n +1 ( 3 )}
(B.32)
iii. free edge The elements in the second row of the matrix are same as simply supported case (Eqs. B.27-B.32). The remaining elements in the first and thirds rows are given by
k11 = n 2G
(B.33) (B.34)
for n = 0 for n 0
k13 = nI n ( 3 )
2 k14 = G n 2 G
(B.35)
k16 = nK n ( 3 ) k 31 = 2n(n 1) k 32 = 2n 2 1{(1 n) J n ( 2 ) + 2 J n +1 ( 2 )} G 0 = 2n ( 2 + 2n 2 2)I ( ) (2n(n 1) + 2 )I ( ) n +1 3 3 n+2 3 3 3
(B.36) (B.37) (B.38) for n = 0 for n 0 (B.39)
k 33
k 34 = 2n(n + 1) k 35 = 2n 2 1{(1 n)Yn ( 2 ) + 2Yn +1 ( 2 )} G
(B.40) (B.41)
for n = 0 0 k 36 = 2n ( 2 + 2n 2 2)K ( ) (2n(n 1) + 2 )K ( ) for n 0 n +1 3 3 n+2 3 3 3
(B.42)
252
Appendix B
If the inner edge is i. clamped edge

k 41 = b n k 42 = J n ( 2 ) log(b) k 44 = n b k 45 = Yn ( 2 ) k 51 = nb n 1 k 52 =
(B.43) (B.44) for n = 0 for n 0 (B.45) (B.46) (B.47) (B.48)
2 2 1{J n 1 (b 2 ) J n +1 (b 2 )} 2 G
n I n (b 3 ) b
for n = 0 for n 0
k 53 =
(B.49)
k 54
b 1 = ( n +1) nb
(B.50)
k 55 =
2 2 1{Yn 1 (b 2 ) Yn +1 (b 2 )} 2 G
n K n (b 3 ) b
(B.51)
k 56 =
(B.52) (B.53) (B.54)
k 61 = nb n 1
n k 62 = 1 2 J n (b 2 ) Gb
for n = 0 0 k 63 = 3 { } 2 I n +1 (b 3 ) + I n 1 (b 3 ) for n 0
(B.55)
k 64 = nb ( n +1)
(B.56) (B.57)
n k 65 = 1 2 Yn (b 2 ) Gb
253
Appendix B
for n = 0 0 k 66 = 3 2 {K n +1 (b 3 ) + K n 1 (b 3 )} for n 0
(B.58)
ii. simply supported edge The fourth row and sixth row of matrix elements are same as the elements in corresponding rows for clamped edge (i.e. Eqs. B.43-B.58). The non-zero elements in fifth row are
k 51 = b n 2 (1 n)n( ) k 52 =
(B.59)
1 1 b 2 22 J n 2 (b 2 ) + 2b 2 J n 1 (b 2 ) 2 b 2 22 + 2n 2 J n (b 2 ) 2 2 4b G
{(
+ 2 b 2 J n +1 (b 2 )} + b 2 22 J n + 2 (b 2 ) k 53 =
(B.60) (B.61)
n ( ){(n 1)I n (b 3 ) + b 3 I n +1 (b 3 )} b2
k 54
b 2 ( ) = ( n + 2) n(1 + n)( b
for n = 0 for n 0
(B.62)
k 55 =
1 4b 2
2 2 2 2 2 2 1 b 2 Yn 2 (b 2 ) + 2b 2Yn 1 (b 2 ) 2 b 2 + 2n Yn (b 2 ) G
{(
+ b 2Yn +1 (b 2 )} + b 2 22Yn+ 2 (b 2 ) k 56 =
(B.63) (B.64)
n ( ){(n 1)K n (b 3 ) b 3 K n +1 (b 3 )} b2
iii. free edge The elements in the fifth row of the matrix are same as simply supported case (Eqs. (B.59-B.64). The remaining elements in the fourth and sixth rows are given by
254
Appendix B
k 41 = k 43 =
b n 1 n 2G
(B.65) (B.66)
n I n (b 3 ) b
for n = 0 for n 0
k 44
b 2 G = n n +1 2 b G
n K n (b 3 ) b
(B.67)
k 46 =
(B.68) (B.69) (B.70)
k 61 = 2b n 2 (n 1)n k 62 =
2n 1{(1 n) J n (b 2 ) + b 2 J n +1 (b 2 )} 2 2 b G
for n = 0 0 k 63 = 2n b 2 2 + 2n 2 2 I (b ) 2n (n 1) + 2 I (b ) for n 0 (B.71) 2 3 n +1 3 3 n+2 3 b 3 3 b
k 64 = 2nb ( n + 2) (n + 1) k 65 =
(B.72) (B.73)
2n 1{(1 n)Yn (b 2 ) + b 2Yn +1 (b 2 )} 2 2 b G
k 66
for n = 0 0 2n 2n = b 2 32 + 2n 2 2 K n +1 (b 3 ) 2 (n 1) + 32 K n + 2 (b 3 ) for n 0 (B.74) b 3 3 b
255
List of Authors Publications
Journal Papers
1. Wang, C.M. and Tun Myint Aung, Plastic buckling analysis of thick plates using p-Ritz method. International Journal of Solids and Structures, submitted for possible publication. 2. Tun Myint Aung, Wang, C. M. and Chakrabarty, J., (2005). Plastic buckling of moderately thick annular plates. International Journal of Structural Stability and Dynamics, 5 (3), 337-357. 3. Tun Myint Aung and Wang, C.M, (2005). Buckling of circular plates under intermediate and edge radial loads. Thin-Walled Structures, Vol. 43, 19261933. 4. Wang, C.M. and Tun Myint Aung, (2004). Buckling of circular Mindlin plates with an internal ring support and elastically restrained edge. Journal of Engineering Mechanics, ASCE, Vol. 131 (4), 359-366. 5. Wang, C.Y., Wang, C.M. and Tun Myint Aung, (2004). Buckling of a weakened column. Journal of Engineering Mechanics, ASCE, Vol. 130 (11), 1373-1376.
Conferences Papers 1. Tun Myint Aung and Wang, C. M. (2005). Plastic buckling skew plates via Ritz method. Proceedings of the Eighteenth KKCNN Symposium on Civil Engineering, Kaohsiung, Taiwan, 471-476.
256
List of Authors Publications
2. Wang, C. M. and Tun Myint Aung, (2005). Ritz method for plastic buckling analysis of thick plates. Third International Conference of Structural Stability and Dynamics, Kissimmee, Florida, USA. 3. Wang, C. M., Tun Myint Aung and Vo, K.K. (2005). Plastic buckling analysis of Mindlin plates using the Ritz method. Proceedings of the Tenth International Conference on Civil, Structural and Environmental Engineering Computing, Rome, Italy. 4. Tun Myint Aung and Wang, C. M. (2004). Plastic buckling of annular plates. Proceedings of the 17th KKCNN Symposium on Civil Engineering, Ayutthaya, Thailand, 367-372. 5. Wang, C.M. and Tun Myint Aung, (2003). Buckling Of Circular Plates With An Internal Ring Support: Asymmetric Mode And Optimal Ring Support Radius. Proceedings of the 16th KKCNN Symposium on Civil Engineering, Kyungju, Korea.
257

Plastic Buckling of Mindlin Plates

Încărcat de

Informații document

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Plastic Buckling of Mindlin Plates

Încărcat de

Drepturi de autor:

Formate disponibile

PLASTIC BUCKLING OF MINDLIN PLATES

TUN MYINT AUNG

NATIONAL UNIVERSITY OF SINGAPORE 2006

PLASTIC BUCKLING OF MINDLIN PLATES

TUN MYINT AUNG B. Eng, (Yangon Institute of Technology)

ACKNOWLEDGEMENT ................................................................................... TABLE OF CONTENTS SUMMARY .. NOMENCLATURE . LIST OF FIGURES .

Derivation of Stress-Strain Relations Based on Henckys Deformation Theory ... 23 25 29 33 34 35 37 43 49 59 60 61 61 63 88 97

Material Modeling ......

Derivation of Energy Functional and Governing Equations ..

Concluding Remarks 103

Introduction ..... 146 Analytical Method .. 148

Concluding Remarks ... 178 180

strain energy potential energy contact function

S 0 0.2 0.7 , 0.85 c m x , y , z

Fig. 4.25 Fig. 4.26

Normalized plastic buckling stress xy / 0.2 versus b/h ratio for

Fig. 6.1 Fig. 6.2

130 131 131 132 132 133

Fig. 7.12 Fig. 7.13

Fig. 7.16 Fig. 8.1 Fig. 8.2 Fig. 8.3

177 182 182 183

185 190 193 199 201 204

Table. 4.1 Table. 4.2 Table. 4.3

Table. 4.4 Table. 4.5

Table. 4.6 Table. 4.7 Table. 4.8 Table. 4.9

108 110 110 114

Table. 5.3 Table. 5.4 Table. 5.5 Table. 5.6

114 115 115

Table. 5.7 Table. 5.8 Table. 6.1

Table. 6.5 Table. 6.6

Table. 6.7 Table. 6.8 Table. 6.9

127 128 135 135 136 136 137

178 184 190 194 197 198 198 203

Plastic Buckling Theories of Plates

Inclusions of Effect of Transverse Shear Deformation

Objectives and Scope of Study

GOVERNING EQUATIONS FOR PLASTIC BUCKLING OF PLATES

Governing Equations for Plastic Buckling Analysis of Plates

Governing Equations for Plastic Buckling Analysis of Plates

in which u, v are in-plane displacements in the x and y directions, respectively, w

Governing Equations for Plastic Buckling Analysis of Plates

Fig. 2.1 Deformation of cross-section of Mindlin plates

Governing Equations for Plastic Buckling Analysis of Plates

In view of Eqs. (2.1a-c), the strain-displacement relations can be expressed as

Stress-Strain Relations in Plastic Range

Governing Equations for Plastic Buckling Analysis of Plates

Derivation of Stress-Strain Relations Based on Prandtl-Reuss Equation

Governing Equations for Plastic Buckling Analysis of Plates

The corresponding equivalent or effective plastic strain increment d

Governing Equations for Plastic Buckling Analysis of Plates

A straightforward differentiation of Eq. (2.14) gives

(2 p q)d x + (2q p)d y + 6sd xy 2 2

Governing Equations for Plastic Buckling Analysis of Plates

c11 c12 d x c 22 d y Et d xy = d xz d yz sym where

Governing Equations for Plastic Buckling Analysis of Plates

& x &y & xy = E & xz & yz sym

& x & y & xy 0 & xz 2 G & yz E 0 0 0

(c13 c 23 c12 c33 )

(c12 c23 c13 c 22 ) (c12 c13 c11c23 )

Governing Equations for Plastic Buckling Analysis of Plates

Derivation of Stress-Strain Relations Based on Henckys Deformation Theory

where p / can be written as

Governing Equations for Plastic Buckling Analysis of Plates

Governing Equations for Plastic Buckling Analysis of Plates