Journal of Constructional Steel Research 140 (2018) 191–207

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

Review

Behaviour of thin-walled curved steel plates under generalised in-plane

stresses: A review
J.P. Martins ⁎, F. Ljubinkovic, L. Simões da Silva, H. Gervásio
ISISE, Department of Civil Engineering, University of Coimbra, Portugal

a r t i c l e i n f o a b s t r a c t

Article history: Cylindrically curved plates are increasingly used in steel construction. In particular, there is a clear trend for their
Received 2 June 2017 use in box-girder steel bridges with curved bottom flanges. However, there is a gap in standards dealing accurate-
Received in revised form 5 September 2017 ly with these types of structural elements under several arrangements of loadings and boundary conditions. This
Accepted 15 October 2017
paper provides a state-of-the-art on the stability behaviour and design of cylindrically curved panels under gen-
Available online 5 November 2017
eralised in-plane loading. A detailed review of the behaviour of curved panels subject to uniaxial compressive
Keywords:
stresses, circumferential stresses, shear stresses and combined in-plane compressive stresses is presented,
Curved steel plates followed by a comparison of the design provisions of DNV and DNVGL standards with FEM numerical results
Stability obtained by the authors.
Local buckling © 2017 Elsevier Ltd. All rights reserved.
Global buckling

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
2. Geometry of the curved panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
3. Uniaxial compressive stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
3.1. Linear elastic buckling behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
3.1.1. Unstiffened panels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
3.1.2. Stiffened panels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
3.2. Postbuckling behaviour and ultimate strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
3.2.1. Unstiffened panels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
3.2.2. Stiffened panels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
4. Circumferential stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
4.1. Linear elastic buckling behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
4.2. Postbuckling behaviour and ultimate strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5. Shear stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5.1. Linear elastic buckling behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5.2. Postbuckling behaviour and ultimate strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6. Combined in-plane compression and shear stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6.1. Linear elastic buckling behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6.1.1. Biaxial loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6.1.2. Axial compression and shear loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6.2. Postbuckling behaviour and ultimate strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
6.2.1. Biaxial loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
6.2.2. Axial compression and shear loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
7. Comparison of results from selected standards and fem calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
7.1. Preliminary remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
7.2. Axial compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
7.3. Shear stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
7.4. Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202

⁎ Corresponding author at: Civil Engineering Department, University of Coimbra, R. Luís Reis Santos - Pólo II, Pinhal de Marrocos, 3030-290 Coimbra, Portugal.
E-mail address: jpmartins@dec.uc.pt (J.P. Martins).

https://doi.org/10.1016/j.jcsr.2017.10.018
0143-974X/© 2017 Elsevier Ltd. All rights reserved.
192 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207

8. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
Appendix A. Annex A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

1. Introduction cylindrically curved panels (subject to uniaxial compressive stresses,

circumferential compressive stresses, shear stresses and a combination
Despite the limited guidance on the structural behaviour of curved of stresses, respectively). Subsequently, a review of the postbuckling
steel plates under generalised in-plane stresses and the lack of practi- and ultimate behaviour for cylindrically curved panels is provided,
cal design rules from codes of practice, structural engineers are including the consideration of boundary conditions other than simply
increasingly using this type of elements. Recently, Reis et al. [1] supported and loading conditions other than uniform compressive
provided an exhaustive survey of steel bridges using curved bottom stresses. Finally, current design methodologies for the assessment of
flanges. In their report, a total of 18 bridges with curved cross- the resistance of structures incorporating curved panels are identified
sectional parts are described. In these box-girder bridges, the global in Section 7. In particular, the design methods found in DNVGL [6]
curvature parameter (see Section 2) of the bottom flange varies and DNV [7] are compared with numerical results obtained by the
from 1 up to 600. As an example, in Fig. 1, the cable-stayed Ebro authors.
River Pedestrian Bridge is illustrated during its erection process,
showing a highly curved unstiffened bottom flange (global curvature 2. Geometry of the curved panel
parameter varying between 300 and 600).
The use of curved shapes in cross-sections (which are predomi- Fig. 2 defines the relevant geometrical variables for the characterisa-
nantly under compressive stresses but may also be subject to shear tion of a cylindrically curved panel is given as follows
stresses) raises a problem: how can designers cope with mandatory/ where
standardised safety levels when current design codes do not provide
• a is the length of the panel;
guidance on the design of curved plates? The only possible solution
• aloc is the length of the subpanel;
is the use of advanced Finite Element Analysis. However, its use
• b is the width of the panel;
requires experienced designers and a high level of understanding of
• bloc is the width of the subpanel;
computational tools to tackle with confidence complex stability driven
• t is the thickness of the panel;
problems.
• R is the radius of curvature of the panel;
Other authors have presented literature reviews of shell structures.
• φ is the sectorial angle of the panel;
For example, Schmidt [3] reviewed the main features of the stability
• bs is the width between longitudinal stiffeners;
design of shell structures. However, this state-of-the-art report is limited
• hs is the height of the stiffener;
to shells of revolution. For the design of shells of revolution the reader is
• ts is thickness of the stiffener.
also referred to the ECCS publication no. 125 [4] and for a comparison of
standards to [5]. The aspect ratio is defined by the ratio between the panel's length by
Therefore, this paper presents a review of the stability behaviour its width as shown in Eq. (1).
of simply supported cylindrically curved panels under generalised
in-plane stresses. After an introductory description of the geometry of
cylindrically curved panel in Section 2, Sections 3 to 6 characterise a
a¼ ð1Þ
the elastic critical and postbuckling behaviour of simply supported b

Fig. 1. Ebro River pedestrian bridge (cable-stayed bridge), Zaragoza, Spain [2]
 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207 193

a) b)

Fig. 2. Geometry of a cylindrically curved panel: a) unstiffened b) stiffened.

Additionally, using these parameters, it is common to describe the Similarly, it is possible to define a curvature parameter for the longi-
curvature of the panel as shown in Eq. (2). This curvature parameter tudinal direction, Eq. (7).
was first proposed by Batdorf [8,9,10].

a2
b pffiffiffiffiffiffiffiffiffiffiffiffiffi2
2 Za ¼ ð7Þ
Zb ¼ 1−ν ð2Þ R:t
R:t

3. Uniaxial compressive stresses

Recently, some authors adopted a similar parameter, removing
the Poisson's coefficient (Eq. (3)). In particular, Martins et al. [11, 3.1. Linear elastic buckling behaviour
12] have shown that the parameter given by Eq. (3) is suitable to
define the mechanical behaviour of cylindrically curved panels 3.1.1. Unstiffened panels
under axial compression and preliminary results from Ljubinkovic Fig. 3 shows an unstiffened curved panel subject to longitudinal
[13] showed that the same parameter is suitable to describe compressive stresses.
the behaviour of shallow cylindrically curved panels under shear One possible definition for the elastic critical behaviour of cylindri-
stresses. cally curved panels is the one given by Gerard & Becker in 1957:
“A curved plate loaded in axial compression buckles in the same manner
2 as cylinder when the plate curvature is large, and when the plate curvature
b φ2 :R
Z¼ ¼ ð3Þ is small it buckles essentially as flat plate. Between these two limits, there is
R:t t a transition from one type of behaviour to the other” (p. 54, Gerard &
Becker [15]). This statement was believed to be correct for a long period
Although not as common as the previous ones, the following of time and its main idea was reflected by the expressions proposed at
mathematical definition for the curvature parameter, Eq. (4), is also that time (Redshaw [16] and Marguerre [17] – used by other authors
found in the literature [14]:

  
4 3 1−ν2 b 4
θ ¼ ð4Þ
4R2 t 2 π

The three curvature parameters are related by:

 
3 1−ν2 2 3
θ4 ¼ Z ¼ 4 Z 2b ð5Þ
4π 4π

Generally, it may be useful to distinguish the global curvature

parameter for a stiffened panel from the local curvature parameter for
unstiffened sub-panel. The local curvature parameter is obtained by
replacing b with bloc in Eq. (3):

2
bloc Fig. 3. Unstiffened cylindrically curved panel subject to uniaxial compressive and in-plane
Z loc ¼ ð6Þ
R:t bending stresses.
194 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207
[18]). Recently, with the availability of numerical tools, it was
re-shown [19] (similarly to what Stowell's formulae [20] predicted
16 years before the work of Gerard & Becker) that the behaviour of
cylindrically curved panels differs considerably from the behaviour
of full revolution cylindrical shells. In fact, there is no development
of tensile circumferential membrane stresses along the longitudinal
edges, or at least not in the same way, which leads to different buck-
ling modes from those observed in full revolution cylindrical shells.
The main reason for this distinction is the fact that the longitudinal
edges of the curved panel are independent (i.e. they are constrained
by independent boundary conditions) while for a full revolution
shell there are no longitudinal edges. This affects naturally the magni-
tude of the elastic critical stresses, which are reduced to almost 50% of
the elastic critical stress of the full revolutions cylindrical shells [21],
and also the buckling shape. On the other hand, the higher the radius
of panel, the more the behaviour of a panel becomes similar to the behav-
iour of flat plates.
The authors that have proposed expressions or developed a method-
ology to compute/characterise the elastic critical stress of simply sup-
ported unstiffened cylindrically curved panels are, in chronological
order: Redshaw [16], in 1934, Marguerre [17], in 1939, (also given in
[23]), Stowell [20], in 1943, Volmir [24], in 1963, Domb & Leigh [25],
in 2001, Martins et al. [11,26], in 2013 and 2016 respectively. It should
be noticed that different authors have considered slightly different
cases of simply supported curved panels. For example, while Stowell's
expression is valid for simply supported square curved panels with all
edges constrained to remain straight, the expression by Martins et al.
[11] is valid for simply supported short curved panels, with only the
curved edges constrained to remain straight, and the expression by
Martins et al. [26] is valid for simply supported short curved panels
with the curved edges constrained and unconstrained to remain
straight. In addition, the work of Martins et al. [11] enables the calcula-
tion of the elastic critical stress from pure compression to pure in-plane
bending.
In 1947, Batdorf [8,9,10] proposed design curves fitting theoret-
ical curves to experimental data using an equivalent Donnell's
equation and proposed a solution given by Fourier series that was
subsequently calibrated with test data. These design curves,
which were based on the erroneous assumption that the elastic
critical stress was equivalent to the ultimate stress, still could not
fully explain the discrepancies between theoretical and experimen-
tal results. These inconsistencies would be solved with the intro-
duction of the concept of imperfection sensitivity and with the
awareness that the elastic critical stress was in fact different from
the ultimate stress. In 1971, Snyder [27] developed a computer
program to solve several stability problems, among which axially
loaded curved panels with and without cutouts. One year later,
Durban & Singer [28], in 1972, based on the Donnell's equations
and using the Galerkin method, co-authored a report on the buck-
ling of curved panels under various axial compression conditions.
These included in-plane bending and a concentrated force applied
at the curved edges. In 1976, Sobel et al. [21] presented a numerical
study on cylindrically curved panels where the effects of boundary
conditions on the elastic critical stress were studied. The authors
considered two sets of boundary conditions for the straight edges
(four sub-sets of simply supported and four sub-sets of clamped
boundary conditions). In 1994, Bismarck-Nasr [22], using a two-
field variables variational principle (transverse displacement and
Airy stress function as field variables), derived the elastic critical
stress of curved plates with different aspect ratios under several
boundary conditions (clamped edges, simply supported edges and
freely supported edges). Magnucki & Mackiewicz [29] in 2006 and
Wilde et al. [30] in 2007 studied the problem of an axially compressed
cylindrical panel with three edges simply supported and one edge
free. The range of validity of these studies is limited to values between
π/2 and π rad of the sectorial angle. For smaller values of the sectorial
angle, Chu & Krishnamoorthy [31] had previously performed a similar
study in 1967. Magnucki & Mackiewicz [29] solved the Donnell's
equations for linear buckling of shells that were reduced to a general-
ised eigenvalue problem with the use of the Galerkin method.
Wilde et al. [30] developed an analytical model to predict the elastic
buckling stress and performed a numerical study using the finite
element method and the results were compared to those obtained
by Magnucki & Mackiewicz one year earlier. The comparison of the
solutions obtained with the analytical and numerical procedures
showed that the differences between these results were small.
Eipakchi & Shariati [32], in 2011, presented a study concerning the
buckling stresses of cylindrically curved panels subjected to axial
stress with the two opposite sides simply supported and the two
other edges free or simply supported. The authors proposed an ana-
lytical method based on perturbation methods to find the buckling
axial stress of a cylindrically curved panel.
Recently, in 2015, Afkhami et al. [33], using both numerical tools and
Donnell's equations together with the Galerkin method, studied the
effects of variations of geometry on the elastic stability of axially
compressed cylindrically curved panels where the curved edges are
simply supported and the longitudinal edges are free.
Table A.1 summarises several available formulae and/or methodolo-
gies to compute the elastic critical stress of unstiffened curved panels
under compressive stresses.
3.1.2. Stiffened panels
Similarly to longitudinally stiffened flat plates, stiffened cylindri-
cally curved panels may present four different buckling modes: global
buckling of the cylindrically curved panel, beam-column-type buck-
ling of the panel-stiffener system, local buckling of the unstiffened
panel between longitudinal stiffeners and local buckling of the
stiffeners.
In 1948, Batdorf & Schildcrout [34] derived the theoretical critical
stress of a simply supported curved rectangular panel with one middle
transverse stiffener under uniform axial compression. Later, in 1949,
Schildcrout & Stein [35] carried out a similar study for a curved panel
with one longitudinal stiffener. In both cases, it was concluded that
the addition of one stiffener was responsible for a considerable increase
in the elastic critical stress of the curved panel. Then, in 1961, Peterson &
Whitley [36] prepared a report on the local buckling behaviour of
Z-stiffened curved plates. In this report, the authors tried to account
for the effect of curvature by applying the empirical relationship first de-
rived by Stowell [20] for unstiffened panels. In 1976, Sobel et al. [37]
presented a numerical study on stringer-stiffened curved panels,
where the effects of boundary conditions and panel width were
explored. The main conclusions were that using outside stiffeners led
to an increase of 40–50% in the buckling load. In 2010, Khedmati &
Edalat [38] performed a numerical parametric study on curved stiffened
plates with parabolic curvature in which the elastic critical stress was
obtained from a linear buckling analysis. Generally, the authors
concluded that the parabolic curvature had a strong strengthening
effect.
3.2. Postbuckling behaviour and ultimate strength
3.2.1. Unstiffened panels
The initial postbuckling behaviour of cylindrically curved panels
was studied and discussed by Marguerre [17], Leggett [39], Cox &
Pribram [40], Koiter [41], Gerard & Becker [15], Volmir [24], Pope
[42], Tamate & Sekine [43], Knightly & Sather [44] and Zhang
& Mathews [45]. Koiter [41] considered the radial displacements
supressed along the longitudinal edges, i.e. simply supported
unloaded edges. He demonstrated that as a consequence of increasing
curvature the elastic buckling stress would have also to increase. On
the other hand, Koiter also showed that increasing curvature led to
highly unstable responses. For large postbuckling deformations, Koiter
 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207
stated: “it would appear to be not too bold a conjecture that the be-
haviour of a narrow curved panel in the advanced post-buckling
stage approaches the behaviour of a flat panel of equal width”
(p. 76, [41]). A similar description of the postbuckling behaviour of
curved panels was given by Tvergaard [46] in 1977. The postbuckling
equilibrium path for a simply supported square panel without imper-
fections with all edges constrained (to remain straight after buckling)
can be derived using the Galerkin method, as in Volmir's work [24]
[47].
In 2000, Tvergaard & Needleman [48] co-authored a study where
the effects of curvature and material nonlinearity on the buckling
localization (on the elastic and plastic ranges) of a cylindrical panel are
addressed. One year later, Hilburger et al. [49] reviewed the studies on
the effect of a cutout on the pre- and postbuckling behaviour of axially
compressed curved panels. Additionally, the authors concluded that the
edge restraints may be responsible for introducing a highly nonlinear
pre-buckling behaviour in shallow curved panels with a cutout and, con-
sequently, increase the buckling load.
As for plates, the ultimate strength of unstiffened cylindrically
curved panels is governed by several parameters such as the
post-buckling behaviour, boundary conditions, constitutive law of
the material, geometric imperfections and material imperfections.
Additionally, it depends on the curvature, expressed by the curva-
ture parameter Z.
Several authors, seeking to establish the buckling load, have
performed experimental tests on the assumption that the ultimate
strength was the same as the critical stress. However, as proven
by von Karman et al. [60] in 1932 and later by Koiter [61] and
Budiansky & Hutchinson [62] in 1945 and 1964, respectively, the
unavoidable imperfections influence greatly the postbuckling and
ultimate load of any structural system. Therefore, these experimen-
tal tests could not confirm the elastic buckling load given by earlier
theoretical studies. This is well patent in [63], where Sechler attribut-
ed the scatter of experimental results only to statistical dispersion
of material data and to the test layout. Among the researchers
performing early experimental tests on unstiffened curved panels
with several types of boundary conditions are Sechler [63], Newell
& Gale [64], Wenzek [65], Cox & Clenshaw [66], Crate & Levin [67],
Jackson & Hall [68], Welter [69,70,71], Schuette [72] and Tenerelli
& Holmes [73]. The work of Thomas [74] presents a complete descrip-
tion of the majority of the cited experimental studies.
Several authors have studied specifically the ultimate strength
of unstiffened cylindrically curved panels under pure compression:
Gerard & Becker [15], Gerard [75], Parks et al. [52], Verolme
[54], Featherston & Ruiz [76], Yumura et al. [77], Park et al.
[78,80], Featherston [80], Tran [81], Tran et al. [19] and Martins
et al. [12,82].
In 2005, Yumura et al. [77] investigated the buckling/plastic
collapse behaviour of compressed cylindrically curved plates. The
authors performed an elastic eigenvalue analysis, an elastic large
displacement analysis to investigate the characteristics of the
postbuckling behaviour and a series of elastic plastic large deflection
analysis to clarify the buckling/plastic collapse behaviour. Park et al.
[78,79], in 2009, performed some studies to predict the buckling
strength and ultimate strength of a simply supported cylindrically
curved plate. The buckling strength and ultimate strength formulae
are empirically derived based on finite element analyses. In 2012,
Tran [81] and Tran et al. [19] developed two alternative methods to
compute the ultimate strength of simply supported cylindrically
curved panels subject to uniform axial compression. These approaches
provided a reduction factor that is applied to the plastic resistance
of the panel giving its ultimate strength. The first one is based on a
modification of the Ayrton-Perry approach and the second one is a mod-
ification of the methodology prescribed in EN 1993–1-6:2007 [83].
Also in 2012, Featherstone [80] presented a study on the imperfection
sensitivity of curved panels under combined non-uniform compression
195
and bending. Additionally, tests were conducted and an automated
technique was proposed, based on topographic data obtained
from optical readings to generate finite element meshes ready to be
implemented.
Recently, in 2014, Martins et al. [12] proposed an approach
based on the effective width concept and on a modification of the
reduction curves given in EN 1993–1-5:2006 [84]. The main
outcome is the possibility to compute the ultimate strength of
curved plated elements under a combination of pure compression
and in-plane bending. In 2015, Martins et al. [82] studied the effect
of geometric imperfection on the ultimate strength of curved
panels. By means of a large parametric study, where the initial
imperfection covered the ten first buckling modes with several
amplitudes, the authors concluded that using the first buckling
mode as initial imperfection may be an unsafe choice and also
that panels with larger curvature are more sensitive to initial
geometric imperfections.
Table A.2 summarises several available formulae and/or methodolo-
gies to compute the ultimate strength of unstiffened curved panels
under compressive stresses.
3.2.2. Stiffened panels
The amount of published works dealing with the study of the
postbuckling behaviour of stiffened curved panels is very low. In
1982, Agarwal [50] developed an experimental method to study
the postbuckling behaviour of curved panels stiffened with closed
trapezoidal stiffeners. In 1987, Hui & Chen [51] applied Koiter's
theory to study the influence of geometric imperfections on
the postbuckling behaviour of integrally stiffened curved plates
under compression. In 1990, Sheinman & Frostig [53] studied
the postbuckling behaviour of stiffened curved panels with arbi-
trary boundary conditions. Their main conclusion was that the
buckling behaviour of panels where the stiffener is placed on the
convex side presented a bifurcation point only for low values of
curvature; the buckling behaviour of panels with higher curvature
or where the stiffener is placed on the concave side presented a
limit point.
To the extent of the authors' knowledge, the earliest experi-
mental report on stiffened curved panels is due to Gall [85], in
1930, who performed tests on aluminum curved panels. Subse-
quently, several researchers performed experimental testing on
stiffened curved panels [64,86,87,88] [89,90,91,92,93]. The works
of Becker [59], in 1958, and of Thomas [71], in 1972, present a
complete description of the cited works. In 1995, Verolme [54]
Fig. 4. Unstiffened cylindrically curved panel subject to circumferential compressive
stresses.
196 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207
developed a design tool to assess the ultimate load of unstiffened
and stiffened curved panels under compression. The developed
methodology was calibrated with own experimental data on
fiber metal laminates and aluminum panels. In 2007, Cho et al.
[94] performed also an experimental study on six stiffened curved
panels. The results were the basis to calibrate a nonlinear FEM
model.
More recently, Park et al. [55] studied numerically the ultimate
strength and the progressive collapse of unstiffened and stiffened
curved panels. More specifically, besides the curvature, the authors
studied the influence of several parameters like the aspect ratio,
slenderness, and amplitude of initial imperfections. In 2014, Tran
et al. [56] [57] presented two alternative methods to compute the
ultimate strength of stiffened curved panels under pure compression.
The first method [56] is developed using a “design of experiment
method”, which was statistically calibrated. The second one is an ex-
tension of the design methodology presented in EN 1993-1-5:2006
[84]. In 2016, Seo et al. [58] assessed numerically the influence of
the geometry of the panel and stiffeners on the buckling behaviour.
It was concluded that the geometry of the stiffeners significantly influ-
ences the collapse modes, although they did not affect considerably
the ultimate strength.
The most recent works dealing with the stability and strength of
stiffened curved panels are almost exclusively devoted to composite
materials. Exceptions are made to the already cited works of Park
et al. [55], Tran et al. [56] [57] and Seo et al. [57].
4. Circumferential stresses
4.1. Linear elastic buckling behaviour
Fig. 4 shows an unstiffened curved panel subject to circumferential
compressive stresses.
In steel construction, it is common to find unstiffened curved
panels under compressive circumferential stresses. Due to the curva-
ture, the linear behaviour of the curved panel is immediately changed.
This means that even for stockier panels (low values of slenderness),
the (effective) width is reduced according to a reduction factor which
has an upper limit given by an expression proposed by Jetteur &
Maquoi [95]. The same authors, relying on the curvature parameter
Za and on the proposed equation for the reduction factor, derived an
expression to obtain the elastic critical stress of cylindrically curved
panels under circumferential compression. According to the authors, the
expression is only valid for values of Za lower than 3.5 (very low values
of curvature).
Table A.3 summarises the available formulae and/or methodologies
to compute the elastic critical strength of unstiffened curved panels
under circumferential compressive stresses.
4.2. Postbuckling behaviour and ultimate strength
To the extent of the authors' knowledge, the literature on the
postbuckling behaviour of thin cylindrically curved panels under
circumferential compressive stresses is very scarce. The only relevant
document is from Jetteur & Maquoi [95]. In this work, the post-
buckling behaviour of a cylindrically curved panel under circumferen-
tial compressive stresses is completely defined by a set of equations
that result from a variational analysis.
For I-girders curved in elevation, studies of several authors
[96,97,98] show that, for stocky flanges, the ultimate strength is lit-
tle affected by the presence of curvature. However, in box-girders
curved in elevation, the compressed flange usually presents a sig-
nificantly higher slenderness. To control the potential instability
phenomena that might arise, several design standards and recom-
mendations, based on experimental and computational results,
indicate a limit value for the curvature [99]. Nevertheless, in
Fig. 5. Unstiffened cylindrically curved panel subject to shear stresses.
several bridge designs (e.g. bow string arch bridges), this limitation
is impossible to cope with. In fact, in order to verify the values im-
posed by EN 1993-1-5:2006 [84] for the curvature parameter, the
radii of curvature must be kept in a small range of values, resulting
in unfeasible and uneconomical thick plates. To overcome this gap
in the standards, Jacques et al. [100] examined the problem of the
buckling of unstiffened curved plates using the finite element
method; subsequently, in 1984, Jetteur & Maquoi [95] developed
an approach based on a two-field variational principle and obtain-
ed a simplified formulation that models the physical response of
cylindrically curved panels under compressive circumferential
stresses.
Another design situation where it is possible to find curved
plates under circumferential compressive stresses is in bridges
curved in-plan. In this case, the curved plate is subject to a non-
uniform stress field due to the presence of a bending moment.
Some of the most recent documents providing a review of the
state-of-the-art in global behaviour of horizontally curved I-girder
bridges are given by Nakai & Yoo [101], Zureick & Naqib [102] in
1999. Linzell et al. [104], in 2004, described in detail the panorama
and historical evolution of curved bridges in the USA, mainly
from a standards' point of view. Previously, in 1985, Schuenzel
et al. [103] performed an experimental study on a plate-girder
curved steel bridge. Later, in 1999 and 2000, Linzell [105] and
Zureick et al. [106] respectively, also performed and presented
experimental results on curved I-girder bridges. Concerning
the local buckling of curved webs, the works of Yabuki et al.
[107,108] and Bogaert [109] are highlighted. Recently, Tran [81],
in 2012, proposed a methodology to compute the ultimate
resistance of unstiffened curved panels under non-uniform circum-
ferential stresses.
Table A.4 summarises the available formulae and/or methodologies
to compute the ultimate strength of unstiffened curved panels under
circumferential compressive stresses.
5. Shear stresses
5.1. Linear elastic buckling behaviour
Fig. 5 shows an unstiffened curved panel subject to shear
stresses.
The first attempt to obtain the critical shear stress of curved
panels date back to 1937, where Leggett [110] provided a solution
for shallow panels with small curvature, based on the Dean's differ-
ential equation for curved panels [111], assuming that the displace-
ments in both axial and circumferential directions were prevented
 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207
along the edges. In this study, only long panels were considered
with either simply supported or clamped longitudinal edges. Two
years later, in 1939, Kromm [121] provided a solution for a wider
range of curvatures, but for simply supported edges only. In addi-
tion, the author examined the buckling load in case the displace-
ments in the circumferential direction are not prevented and, as
the outcome, he obtained lower critical stresses.
Batdorf et al. [112] and Schildcrout et al. [128] studied the critical
shear stress of a perfect transversally curved isotropic panel, with
both simply supported and clamped edges for a wide curvature
range. Specifically, Batdorf et al. and Schildcrout et al. analysed rectan-
gular curved panels under pure shear loads and under the combina-
tion of in-plane normal and shear loads. In both cases, the authors
used Donnell's equations to find a solution and proposed an expres-
sion for the critical shear load similar to the one for the flat plates
under pure shear loads. Later on, several authors carried out literature
reviews and provided a summary of all the works regarding isotropic
and orthotropic flat plates, as well as isotropic curved panels [15,113,114].
In 1994, Bismarck-Nasr [22] also derived the elastic critical stress for
curved plates under shear.
The most recent works on curved plater under shear belong to
Featherstone & Ruiz [76], Featherstone [123], Featherstone [124],
Domb & Leigh [125] and Amani et al. [126]. In particular, Amani
et al. [126] performed a numerical study in which he examined buckling
and postbuckling behaviour of unstiffened cylindrically curved panels
subject only to pure shear and extended the range of parameters
(curvature and aspect ratio) in comparison to previous works [123,124,
125]. Table A.5 summarises some available formulae and/or methodolo-
gies to compute the elastic critical stress of unstiffened curved panels
under shear stresses.
5.2. Postbuckling behaviour and ultimate strength
In 2011, Amani et al. [126] reviewed previous works regarding
the postbuckling behaviour of cylindrically curved panels subjected
to pure shear. Since those works were restricted to a limited range
for the curvature parameter, aspect ratio, and also imperfection ampli-
tudes and shapes, Amani et al. attempted to overcome these limits
and to obtain more reliable conclusions regarding the postcritical
behaviour. Therefore, Amani et al. performed a series of geometrical
and material nonlinear analysis both on perfect and imperfect curved
panels subjected to shear stresses. The authors analysed the effect of
curvature and aspect ratio on the postbuckling behaviour of perfect
curved panels using an arc-length method to obtain convergence.
Additionally, a parametric study was performed in which the influ-
ence of the shape and amplitude of the imperfections on the
Fig. 6. Unstiffened cylindrically curved panel subject to biaxial loading.
197
postbuckling behaviour and collapse load was investigated. This para-
metric study allowed the authors to conclude that the higher the cur-
vature the more unstable was the postbuckling behaviour of perfect
curved panels. Another important conclusion was that initial geomet-
ric imperfections (shape and amplitude) had more influence on
the postbuckling behaviour than on the prebuckling behaviour. The
effect of amplitude is also relevant for the overall behaviour. In
fact, the amplitude of the imperfections affects the buckling load
and initial postbuckling response, whereas the ultimate load remains
practically unchanged, which is in agreement with the results provided
by [124].
Table A.6 summarises some available formulae and/or methodologies
to compute the ultimate strength of unstiffened curved panels under
shear stresses.
6. Combined in-plane compression and shear stresses
6.1. Linear elastic buckling behaviour
6.1.1. Biaxial loading
Fig. 6 shows an unstiffened curved panel subject to biaxial stresses.
Among the few authors [115,116] that have studied the linear
elastic buckling behaviour of biaxially loaded curved panels, the
work of Hilburger et al. [117] is highlighted. The authors have
reached several important conclusions in what concerns the linear
elastic buckling behaviour of curved panels under biaxial loading,
highlighting that only panels under compression along the curved
edges (Ny/Nx = 0) exhibit a bifurcation point; for the remaining
cases, curved panels exhibit a limit point. This is due to the
fact that the loading in the transverse direction is responsible for
pre-buckling out of plane deformations that act as geometric
imperfections, forcing the structural system to adopt equilibrium
configurations different from the initial one for load values lower
than the bifurcation point for a panel under uniaxial compression. The
same conclusion was later reached by Girish & Ramachandra [118] in
2008.
Recently, in 2012, Kasaeian et al. [119] published a paper where
the elastic critical behaviour of unstiffened aluminum curved plates
with and without thickness-tapered sections under biaxial compres-
sion is studied using a virtual work formulation and a finite strip
model.
6.1.2. Axial compression and shear loading
Among the authors that have studied the buckling behaviour of
curved panels under shear and compression are Kromm [121], Leggett
[122], Schildcrout & Stein [128], Featherston & Ruiz [76], Featherston
[123], Featherston [124] and Domb [127].
In 1939, Kromm [121] derived analytically the elastic critical
load for simply supported long cylindrically curved panel under
the combination of compression and shear. In 1943, Leggett proposed
an interaction equation which was later modified by Batdorf et al.
[112]. To the extent of the authors' knowledge, the first experimental
tests performed in curved panels under the combination of compres-
sion and shear are due to Schildcrout & Stein [128]. These tests were
performed for several ratios of compression/shear and for several
values of aspect ratio and curvature parameter. Based on these test re-
sults and on theoretical results, the authors proposed a semi-empirical
method to be used in design. Later, Featherston & Ruiz [76], in 1998,
carried out a series of tests to determine the accuracy of the theoret-
ical elastic buckling load. The authors concluded that the existing
analytical formulations may be used in practice provided that
large safety factors are incorporated. In 2002, Domb [127] imple-
mented a nonlinear buckling analysis technique to predict the
initial buckling in simply supported curved panels under compres-
sion and shear.
198 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207

6.2. Postbuckling behaviour and ultimate strength authors concluded that curved panels under biaxial loading present
a better postbuckling response compared to panels under longitudi-
6.2.1. Biaxial loading nally compressive stresses with the same geometry. More specifically,
The postbuckling behaviour of unstiffened curved panels was stud- for a certain value of curvature, whenever the biaxial loading includes
ied by Chang & Librescu [120]. The authors used a higher-order theory tension, the postbuckling behaviour tends to be stable while for those
to analyse the postbuckling behaviour of several arrangements of cases where both loads are applied in compression the postbuckling
curved panels with combined loading. Among other outcomes, the path tends to be unstable. Similar conclusions were reached by

80.0 80.0
DNV-RP-C202 (α=1.0; b/t=50) DNV-RP-C202 (α=1.0; b/t=100)
70.0 70.0
DNVGL-CG-0128 (α=1.0; b/t=50) DNVGL-CG-0128 (α=1.0; b/t=100)

60.0 FEM (α=1.0; b/t=50) 60.0 FEM (α=1.0; b/t=100)

Buckling coefficient, k σ

Buckling coefficient, k σ
50.0 50.0

40.0 40.0

30.0 30.0

20.0 20.0

10.0 10.0

0.0 0.0
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

180.0 180.0
DNV-RP-C202 (α=1.5; b/t=50) DNV-RP-C202 (α=1.5; b/t=100)
160.0 160.0
DNVGL-CG-0128 (α=1.5; b/t=50) DNVGL-CG-0128 (α=1.5; b/t=100)

140.0 FEM (α=1.5; b/t=50) 140.0 FEM (α=1.5; b/t=100)

120.0 120.0
Buckling coefficient, k σ

Buckling coefficient, k σ

100.0 100.0

80.0 80.0

60.0 60.0

40.0 40.0

20.0 20.0

0.0 0.0
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

300.0 300.0
DNV-RP-C202 (α=3.0; b/t=50) Series1
DNVGL-CG-0128 (α=3.0; b/t=50) DNVGL-CG-0128 (α=3.0; b/t=100)
250.0 250.0
FEM (α=3.0; b/t=50) FEM (α=3.0; b/t=100)

200.0 200.0
Buckling coefficient, k σ

Buckling coefficient, k σ

150.0 150.0

100.0 100.0

50.0 50.0

0.0 0.0
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

Fig. 7. Comparison of buckling coefficients for α ≥ 1 and b/t = 50 and 100 [11,12]
 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207 199

Hilburger et al. [117] and Girish & Ramachandra [118], who concluded
that: i. the higher the value Nx/Ny, the lower the limit point; and ii.
there is a specific value of Nx/Ny, that depends on the geometry of
the curved panel that defines the transition from an unstable
postbuckling path to a stable postbuckling path.
6.2.2. Axial compression and shear loading
As already stated, apart from Schildcrout & Stein [120], Featherston
& Ruiz [76] also performed experimental tests on panels under
compression and shear. In addition, the authors performed numerical
nonlinear analyses to obtain the ultimate strength.
7. Comparison of results from selected standards and fem calculations
7.1. Preliminary remarks
Several available standards provide the design rules for cylindrically
curved panels subjected to in-plane loads (i.e. normal and shear stresses).

1.0 1.0
DNV-RP-C202 (α=1.0; b/t=100)
DNV-RP-C201 (α=1.0; b/t=100)
DNVGL-CG-0128 (α=1.0; b/t=100)
0.8 0.8
FEM (α=1.0; b/t=100)
Reduction factor, χ

Reduction factor, χ
0.6 0.6

0.4 0.4

0.2 DNV-RP-C202 (α=1.0; b/t=50) 0.2

DNV-RP-C201 (α=1.0; b/t=50)
DNVGL-CG-0128 (α=1.0; b/t=50)
FEM (α=1.0; b/t=50)
0.0 0.0
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

1.0 1.0
DNV-RP-C202 (α=1.5; b/t=100)

DNV-RP-C201 (α=1.5; b/t=100)

0.8 0.8
DNVGL-CG-0128 (α=1.5; b/t=100)

FEM (α=1.5; b/t=100)

Reduction factor, χ

Reduction factor, χ

0.6 0.6

0.4 0.4

DNV-RP-C202 (α=1.5; b/t=50)

0.2 0.2
DNV-RP-C201 (α=1.5; b/t=50)
DNVGL-CG-0128 (α=1.5; b/t=50)
FEM (α=1.5; b/t=50)
0.0 0.0
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

1.0 1.0
DNV-RP-C202 (α=3.0; b/t=100)

DNV-RP-C201 (α=3.0; b/t=100)

0.8 0.8
DNVGL-CG-0128 (α=3.0; b/t=100)

FEM (α=3.0; b/t=100)

Reduction factor, χ

Reduction factor, χ

0.6 0.6

0.4 0.4

DNV-RP-C202 (α=3.0; b/t=50)

0.2 DNV-RP-C201 (α=3.0; b/t=50) 0.2
DNVGL-CG-0128 (α=3.0; b/t=50)
FEM (α=3.0; b/t=50)
0.0 0.0
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

Fig. 8. Comparison of the ultimate strength in compression for α ≥ 1 and b/t = 50 and 100 [11,12]
200 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207

Within the scope of this paper, the standards DNVGL-CG-0128 [6]
and DNV-RP-C202 [7] are addressed and compared with numerical
results (all needed description of the numerical models is given in
[11,12,82]).
For the analysis of the mentioned standards, it may be observed
that each one provides a different design methodology to verify
the safety of cylindrically curved panels. Therefore, it is difficult to
compare the standards and to recognize their weaknesses, since
the steps needed to calculate the overall resistance are fundamentally
different. In DNV-RP-C202 [7] there is no interaction formula to
account for the simultaneous action of normal and shear stresses,
but these effects are accounted for when a reduced slenderness
parameter is calculated, which accounts for the critical stresses
for each action. At the end, it is checked that the total applied
stress does not exceed the overall resistance. On the other hand,
the safety verification of curved panel in DNVGL-CG-0128 [6] is
based on an interaction formula where the buckling coefficients
and buckling reduction factors are calculated for four different
cases (i. simply supported panels under longitudinal stresses –
shorter edge loaded; ii. transversally loaded simply supported panels;
iii. Transversally loaded panels where the loaded edges are simply
supported and the unloaded edges one is clamped and the other is
free-to-wave; and iv. simply supported curved panel subject to shear
stresses). For each case, two different expressions are provided for
the buckling coefficient depending on the panel's aspect ratio, and
furthermore, buckling reduction factors are calculated depending on
the slenderness parameter of the panel. Finally, it should be noticed
that the results from the selected standards are not reduced by safety
coefficients.
7.2. Axial compression
This section compares two standards (DNV-RP-202 and DNVGL-CG-
0128) that deal with the axial resistance of cylindrically curved panels
with FEM numerical results. Firstly, an example that compares the elas-
tic critical buckling factor was performed, in which several parameters
are kept constant (b = 1000 mm, material properties of S355 steel
and two cases of simply supported boundary conditions – all edges
are constrained to remain straight and only the curved edges are
constrained to remain straight), whereas the thickness (t = 10 mm
and t = 20 mm), aspect ratio (α = 1.0, α = 1.5, α = 3.0) and curvature
(Z = 1, Z = 3, Z = 10, Z = 20, Z = 40, Z = 60, Z = 100) were varied. The
verification of the numerical models was performed in previous studies
by the authors [11,12]. The results are given in Fig. 7.
From the analysis of these graphs, the immediate conclusion is
that DNV-RP-C202 is considerably more accurate than DNVGL-CG-
0128 in what concerns the calculation of the elastic critical stress.
The main reasons for this are: i) this standard is based on an entirely
different concept where the cylindrically curved panels are seen as
classical shell elements, and ii) it is applicable only for R/t ≤ 2500
(for higher values, the design rules for flat plates should be used
instead).

1000 1000
DNV-RP-C202 (α=1.0) DNV-RP-C202 (α=1.5)
DNVGL-RP-0128 (α=1.0) DNVGL-RP-0128 (α=1.5)
Shear buckling coeffic ie nt, kτ

Shear buckling coeffic ie nt, kτ

FEM (α=1.0) FEM (α=1.5)

100 100

10 10

1 1
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

1000 1000
DNV-RP-C202 (α=2.0) DNV-RP-C202 (α=2.5)
DNVGL-RP-0128 (α=2.0) DNVGL-RP-0128 (α=2.5)
Shear buckling coeffic ie nt, kτ

Shear buckling coeffic ie nt, kτ

FEM (α=2.0) FEM (α=2.5)

100 100

10 10

1 1
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

Fig. 9. Comparison of shear buckling coefficients [11,12]

 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207 201

Concerning the ultimate compressive strength, the results are given
in Fig. 8. Generally, both standards provide safe results. However, for
larger values of the curvature parameter, the results become unsafe.
This may be explained by the fact that the applied amplitude of the
geometric imperfections was b/200, which may be a too conservative
assumption within the scope of these standards. Additionally, it is
observed that stockier panels (b/t = 50) show better agreement than
slender panels (b/t = 100).
Finally, the most surprising observation results from the distinction
between flat panels and curved panels made by both standards. In
what concerns DNV-RP-C202, although it does not state a limit for the
curvature parameter, it is still possible to compare its results with
results from DNV-RPC201 [129], which is devoted to flat panels.
Hence, in contrast to what would be logically expected, whenever the
curvature parameter tends to 0, the reduction factor does not tend to
the reduction factor for flat panels. The same is true for DNVGL-CG-
0128 where for panels with R/t N 2500, the rules for flat panels are
applied.
7.3. Shear stresses
This section compares to compare two design standards (DNV-
RP-202 and DNVGL-CG-0128that deal with the shear resistance of
cylindrically curved panels. An example was used, in which several
parameters are kept constant (t = 3 mm- following previous study
by Amani et. al [126], b = 1000 mm, material properties of S355
steel and simply supported boundary conditions), whereas the as-
pect ratio was varied from 0.5 to 2.5 and the curvature parameter
from 0 to 500. To verify the results, numerical analyses were addi-
tionally performed, using the numerical models calibrated in [13]
with the results in [126].
Firstly, the critical shear load is compared for various aspect ra-
tios higher than 1 (α = 1.0; 1.5; 2.0; 2.5) and the results are given
in Fig. 9, where the variation of the shear buckling coefficient with
aspect ratio and curvature parameter is observed. It may be noticed
that the numerical results are much closer to the results obtained
by DNV-RP-C202, while there is a bigger difference in comparison
to DNVGL-CG-0128. In particular, in relation to DNV-RP-C202, a
very good agreement with the numerical results is observed for
lower values of the curvature parameter, regardless of the aspect
ratio, whereas a larger deviation is noticed with the increase of
curvature.
A possible reason for this discrepancy results from the use of
semi-empirical formulae for the calculation of the critical load by
this standard. Nevertheless, the results obtained with this standard
are on the safe side. However, further investigation should be per-
formed in order to improve these semi-empirical expressions. On
the other hand, a significant disagreement is obtained between the

1 1
FEM (+) (α=1.0) FEM (+) (α=1.5)
FEM (-) (α=1.0) FEM (-) (α=1.5)
0.8 DNV-RP-C202 (α=1.0) 0.8 DNV-RP-C202 (α=1.5)
DNV-RP-C201 (α=1.0) DNV-RP-C201 (α=1.5)
Reduction factor, χw
Reduction factor, χw

DNVGL-CG-0128 (α=1.0) DNVGL-CG-0128 (α=1.5)

0.6 0.6

0.4 0.4

0.2 0.2

0 0
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

1 1
FEM (+) (α=2.0) FEM (+) (α=2.5)
FEM (-) (α=2.0) FEM (-) (α=2.5)
0.8 DNV-RP-C202 (α=2.0) 0.8 DNV-RP-C202 (α=2.5)
DNV-RP-C201 (α=2.0) DNV-RP-C201 (α=2.5)
Reduction factor, χw
Reduction factor, χw

DNVGL-CG-0128 (α=2.0) DNVGL-CG-0128 (α=2.5)

0.6 0.6

0.4 0.4

0.2 0.2

0 0
1 10 100 1 10 100
Curvature parameter, Z Curvature parameter, Z

Fig. 10. Comparison of the ultimate shear capacity [11,12]

202 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207
numerical results and DNVGL-CG-0128. The main reasons for this are:
i. this standard is based on an entirely different concept where the cy-
lindrically curved panels are seen as classical shell elements, and ii. it
is applicable only for R/t ≤ 2500 (for higher values of R/t, i.e. for low
values of curvature, the design rules for flat plates should be used
instead).
Secondly, the ultimate shear resistance is compared between
the two standards and the numerical results (Fig. 10). The numer-
ical results are divided into two groups: i) with an imperfection
amplitude of 0.1 t applied in the same direction of the 1st buck-
ling mode (denoted by FEM +) and ii) with an imperfection am-
plitude of 0.1 t applied in the opposite direction of the 1st
buckling mode (denoted by FEM-). The study confirms that the
sign of the amplitude of the geometric imperfections has a non-
negligible influence on the ultimate resistance. In general terms,
the negative sign results in lower bearing capacity for all consid-
ered cases.
Regarding the results obtained using the expressions from the
standards, very good agreement is noticed for higher curvature, for
all analysed aspect ratios, whereas for the lower and intermediate
values of curvatures, the deviation in results with respect to the
numerical results is larger. The reason for the good correlation of
the results, in case of highly curved panels, is that the behaviour of
these panels asymptotically approaches the behaviour of cylindrical
shells of revolution, and the design procedure for this type of struc-
tures is well established. On the other hand, the poor correlation in
case of low and intermediate curvatures may be due to the following
reasons.
Firstly, the analysed cases fall outside the scope of DNVGL-CG-
0128, since, for this geometry, the limit of R/t ≤ 2500 is exceeded
for curvatures Z b 50. In those cases, the standards suggest the
use of the design methodology for flat plates. Albeit the results
are close to the numerical ones, the behaviour of the shallow
panels is still slightly different from the flat panels. Additionally,
the same jump in the design resistant value is seen for panels
with values of R/t around 2500.
The second possible reason for the deviation in results is the
lack of knowledge regarding the nature of the geometric imperfec-
tions. In fact, the semi-empirical expressions provided in the stan-
dards are based on the assumed level of the imperfections taken
from the tolerance requirements and not on values measured in
experimental tests. As it is stated in DNV-RP-C202, the reason for
this was that in many cases a disagreement between experimental
and analytical results was noticed. The idea was naturally to pro-
vide one general formula, which may be applied to all possible
cases, which unfortunately leads to many uncertainties, such as
shape, amplitude and sign of imperfections. The shape of the
imperfections may in reality differ significantly from the shape
affine to the 1st buckling mode, and it was also proven that the
upper eigenmode could be the governing one [124,126]. Probably,
the most correct solution would be to introduce explicitly the
shape of the imperfections in the numerical model with the ampli-
tude and sign measured on the fabricated specimen and then
to derive the analytical formulae based on the numerical results
calibrated in such a way.
Furthermore, it is also necessary to address the drop in the bearing
capacity for intermediate curvature parameters, obtained in the
numerical study. The reason for this reduction is directly related to
a change of the shape of the buckling mode from symmetric to
asymmetric.
7.4. Remarks
It is important to observe once again that both standards con-
sider the buckling of the cylindrically curved shell only as one of
the several possible failure modes of a cylinder (DNV-RP-C202)
and ship structure (DNVGL-CG-0128). It is also assumed that the
boundary conditions of the panel (longitudinal and transverse
stiffeners) remain straight, which is an acceptable assumption for
these types of structures, but cannot be generally accepted for
all cylindrically curved panels, since in reality they may have
various boundary conditions (e.g. bottom flange of a box-girder
bridge).
8. Concluding remarks
This review addressed a specific topic within the field of plate
and shell structures. In particular, a review was made of the stabil-
ity of cylindrically curved panels subject to in-plane loading
(uniaxial compression, circumferential compression, biaxial com-
pression and shear loading, isolated or in interaction).
Before the publication of the two standards, DNV-RP-C202 [7]
(2013) and DNVGL-CG-0128 [6] (2015), there were generally two
approaches to predict the buckling and post-bucking behaviour
of cylindrically curved panels under compression and shear
load, as it is stated in Featherston [124]: i. a simplified approach,
in which the buckling load is obtained from a linear buckling
analysis of a perfect structure and then the imperfections and
plasticity are accounted for by means of reduction factors; and
ii. a fully non-linear analysis with geometric imperfections and
plasticity explicitly incorporated in the numerical model, where
the shape of the imperfections is assumed to be the affine to
the first buckling mode with a predefined amplitude. Technically,
these two approaches may still be used. However, these two re-
cent standards have brought simplified design formulae, which
should spare designers from time-consuming numerical
simulations.
The validity of the standards DNV-RP-C202 [7] and DNVGL-CG-0128
[6], as well as their comparison, was examined within this work. It
was concluded that: i. for curved panels under compression, both
standards fail to accurately predict the elastic critical stress and
ultimate strength of curved panels; ii. for curved panels under shear,
both standards may be effectively used to predict the ultimate resis-
tance of the panels with higher values of curvature parameters,
whereas for the lower and intermediate values of curvatures, the
standards give conservative results. Therefore, for compression and
for curved panels under shear with low to intermediate curvatures
(Z b ~40), numerical simulations should be performed, where partic-
ular attention should be given to the size, shape and sign of the
imperfections.
Furthermore, these standards are directly applied to offshore
and/or shipbuilding industry. For other engineering applications
(like buildings and bridges) there is still a void in what concerns
standardised rules. For this reason, the authors are actively
involved in two research projects – UltimatePanel (ref.: PTDC/
ECM-EST/1494/2014) and OUTBURST (ref.: RFCS-2015-709,782) –
aiming for the development of design rules for curved panels
subject to in-plane loading.
Acknowledgements
Financial support from the Portuguese Ministry of Science,
Technology and Higher Education (Ministério da Ciência,
Tecnologia e Ensino Superior) under the project contract Grant
PTDC/ECM-EST/1494/2014 is gratefully acknowledged. This work
was also financed by FEDER funds through the Competitivity
Factors Operational Programme - COMPETE and by national funds
through FCT – Foundation for Science and Technology within the
scope of the project POCI-01-0145-FEDER-007633. Funding from
the European Commission's Research Fund for Coal and Steel
through the research project RFCS-2015-709782 is also
acknowledged.
 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207 203

Appendix A. Annex A

A.1. Elastic critical stress of cylindrically curved panels under compressive stresses

Table A.1
Formulae for obtaining the elastic critical stress of cylindrically curved panels subject to compressive stresses.

Author Year Expression Range of validity

 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Redshaw [16] 1934 t 2 2Þ 2 Uniform compression of simply supported panels
σ cr π2 E
¼ 6ð1−ν 2 Þ ð bÞ 1 þ 1 þ 12ð1−ν π4
Z
8  
Marguerre 1939 > π2 E t 2 3ð1−ν 2 Þ 2
if Z ≤ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4π 2 ffi
< 12ð1−ν2 Þ ðbÞ 4 þ π4 Z 12ð1−ν2 Þ
[17] σ cr ¼  pffiffiffiffiffiffiffiffiffiffiffiffiffiffi 
>
: π E
2
ð t
Þ
2 4 3ð1−ν 2Þ
Z if ZN p ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4π 2 ffi
12ð1−ν2 Þ b π2 12ð1−ν2 Þ
Stowell [20] 1943 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!
π2 E t 2 48ð1−ν 2 Þ 2
σ cr ¼ kσ;plate 24ð1−ν2 Þ ðbÞ 1 þ 1 þ π4 ðk Z
Þ
σ ;plate

Domb & Leigh 2001 σ cr ¼ kσ σ E ¼ kσ π2 E 2 ð t Þ2 Uniform compression of simply supported panels – Zb ≤ 200
12ð1−ν Þ b
[25] 8
< ∑c logZ
i
> 3
i b
kσ ¼ 10 i¼0 if 1≤Z b ≤ 23:15
>
:
cZ b d if 23:15 ≤Z b ≤200
where c0 ¼ 0:6021; c1 ¼ 0:005377; c2 ¼ 0:192495; c3 ¼ −0:002670;
c ¼ 0:4323; d ¼ 0:9748
Martins et al. 2013 σ cr ¼ kσ σ E ¼ kσ π2 E 2 ð t Þ2 From uniform compression to pure in-plane bending of simply
12ð1−ν Þ b
[11] ( supported panels – Z ≤ 100
A
Bþψ if 0bψ≤ 1
where kσ ¼ A
B þ Cψ þ Dψ2 if −1 ≤ψ≤ 0
A ¼ a1 þ a2 Z þ a3 Z 2 C ¼ d1 þ d2 Z þ d3 Z 2
B ¼ b1 þ b2 Z þ b3 Z 2 D ¼ d1 þ d2 Z þ d3 Z 2

DNV [7] 2013 σ cr ¼ kσ σ E ¼ kσ π2 E 2 ð t Þ2 Uniform compression of simply supported panels

12ð1−ν Þ b
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 2ffi
−0:5
0:702Z b 0:5ð1þRtÞ
where kσ ¼ 4 1 þ 4

DNVGL [6] 2015 σ cr ¼ kσ σ E ¼ kσ π2 E 2 ð t Þ2 Uniform compression of simply supported panels – R/t ≤ 2500
12ð1−ν Þ a
8 qffiffi
>
<
2
1 þ 23 aRt if Ra ≤0:5 Rt
where kσ ¼  q ffiffi qffiffi
> 2 2
: 0:267 aRt 3− Ra Rt ≥0:4 aRt if Ra N0:5 Rt

Martins et al. 2016 σ cr ¼ kσ σ E ¼ kσ π2 E 2 ð t Þ2 Uniform compression of simply supported panels – Z ≤ 100

12ð1−ν Þ b
[26]  pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2 
π 6 þϕA ðπ 2 −8ÞZ
where kσ ¼ 2 1 þ π3

12 if unloaded edges are constrained to remain straigth
ϕA ¼
9:6 if unloaded edges are free−to−wave

A.2. Ultimate strength of cylindrically curved panels under compressive stresses

Table A.2
Formulae for obtaining the ultimate strength of cylindrically curved panels subject to compressive stresses.

Author Year Expression Range of validity

Tran et al. 2012 χ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

2β Square, simply supported panels under uniform compression – Z ≤ 100
βþλþ ðβþλÞ2 −4βðλ−α z ðλ−λz;0 ÞÞ
[19]
Z
where : β ¼ 1þ0:97
2 λz;0 ¼ 0:2 þ 0:473ð0:95Z Þ

8
Tran [82] 2012 >
> 1 if λ≤λz;0
< λ−λ 0

χ¼ 1−β ð 0 z;0 Þ if λz;0 ≤ λ≤λ

λ −λz;0
>
>
: αZ
if
0
λ ≤λ
λ2
0 0
−20
; λ ¼ 0:673 þ 0:54e−Z ; λ ¼ 0:673 þ 0:54e−Z
Z 4 4
where : λz;0 ¼ 0:2 þ 0:473e
0
1þe−Z=33 2
α Z ¼ AZ λ þ BZ ; AZ ¼ ; BZ ¼ ð1−βÞ λ −AZ λ2
DNV [7] 2013 χ ¼ qffiffiffiffiffiffiffiffiffi
1 ffi Uniform compression of simply supported panels
4
1þλ

Where the slenderness is calculated with the elastic critical stress obtained
from Table A.1.
Tran et al. [57] 2014 χ ¼ ð−0:09 þ 0:326α−0:148α 2 Þ þ ð40:6 þ 0:314ZÞð t Þ−ð444 þ 8:40ZÞð t Þ2 Simply supported panels under uniform compression – Z ≤ 100
b b

(continued on next page)

204 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207

Table A.2 (continued)

Author Year Expression Range of validity

8
Martins et al. 2014 > ρ ¼ 1 if λ≤λ0;Z Square, simply supported panels under uniform compression and pure
>
< λ −λþρ0;Z ðλ−λ0;Z Þ
[12] ρ ¼ 0;p λ0;p −λ if λ0;Z bλbλ0;p in-plane bending – Z ≤ 100
0;Z
>
>
: ρ ¼ λ−0:055aZ ð3 þ ψÞ þ S if λ≥λ0;p
cZ λ2 Z
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
where : λ0;p ¼ 0:5 þ 0:085−0:055 ψ; λ0;Z ¼ ϕ0 þ ϕ1 ϕ2 Z

8
DNVGL [6] 2015 > χ ¼1 if λ≤0:25 Uniform compression of simply supported panels – R/t ≤ 2500
>
>
>
<χ ¼ 11:233−0:933λ if 0:25bλb1
χ ¼ 3
0:3
if 1bλb1
>
> λ
>
>
:χ ¼ 0:22 if λ≥1:5
λ
Where the slenderness is calculated with the elastic critical stress obtained
from Table A.1.

A.3. Elastic critical stress of cylindrically curved panels under circumferential stresses

Table A.3
Formulae for obtaining the elastic critical stress of cylindrically curved panels subject to circumferential stresses.

Author Year Expression Range of validity

Jetteur & Maquoi [95] 1984 σ cr ¼ π2 E

kσ 12ð1−ν t 2
2 Þ ðbÞ
Simply supported curved panels with Za b 3.5
Where:
kσ ¼ 4ρ3ρ−1
t
ðm α 2
α þ mÞ
t

ρt ¼ 1− 8
π6
π 2 ð1þ Þ
96ð1−ν 2 ÞZ 2
a

DNV [7] 2013 σ cr ¼ kσ σ E ¼ kσ π2 E

ðt Þ
2 Simply supported curved panels
12ð1−ν2 Þ b
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffiffi i2
2 2
h
0:61:04α1 Z b
where kσ ¼ ½1 þ ðα1 Þ  1 þ 2 2
½1þðα1 Þ 

DNVGL [6] 2015 σ cr ¼ kσ σ E ¼ π2 E

kσ 12ð1−ν t 2
2 Þ ðaÞ
Simply supported curved panels – R/t ≤ 2500
8 0:175
qffiffi
< paffiffiffiffi þ 3 ðRtÞ0:35 if a
≤1:63 R
R
where kσ ¼ Rt a
qt ffiffi
: 0:3 a þ 2:25ðR2 Þ2 if
2
R N1:63
a R
R2 at t
π2 E
σ cr ¼ kσ σ E ¼ kσ 12ð1−ν t 2
2 Þ ðaÞ
Loaded edges are simply supported, one unloaded
8 pffiffiffiffi qffiffi edge is fixed and the other is free – R/t ≤ 2500
< 0:6a
pffiffiffiffi þ Rt −0:3 Rt2 if a
≤ R
a R
where kσ ¼ Rt a
qtffiffi
: 0:3 a2 þ 0:291ðR2 Þ2
RN
a R
R2 at if t

A.4. Ultimate strength of cylindrically curved panels under circumferential stresses

Table A.4
Formulae for obtaining the ultimate strength of cylindrically curved panels subject to circumferential stresses.

Author Year Expression Range of validity

!
Jetteur & Maquoi [95] 1984 1þ 8
2 Simply supported curved panels with Za b 3.5
ρ ¼ min λ
3 ½1 þ
2 − 32ρ −5 ð1−ρt Þ; ρt
1 2
λ t

Alternative simplified approach:

 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207 205

Table A.4 (continued)

Author Year Expression Range of validity

8
< ρ ¼ ρt if λ b 0:576
ρt
:ρ ¼ ð1− 0:26Þ if λ ≥ 0:576
1:05
ρt
λ ρt λ

DNV [7] 2013 χ¼ q ffiffiffiffiffiffiffiffiffi

1 ffi Simply supported curved panels
4
1þλ

Where the slenderness is calculated with the elastic critical stress obtained from Table A.3
8
DNVGL [6] 2015 <χ ¼ 1
> if λ ≤ 0:4 Simply supported curved panels – R/t ≤ 2500
χ ¼ 1:274−0:686λ if 0:4 b λ≤ 1:2 Loaded edges are simply supported, one unloaded
>
: χ ¼ 0:652 if 1:2 N λ edge is fixed and the other is free – R/t ≤ 2500
λ
Where the slenderness is calculated with the elastic critical stress obtained from Table A.3

A.5. Elastic critical stress of cylindrically curved panels under shear stresses

Table A.5
Formulae for obtaining the elastic critical stress of cylindrically curved panels subject to shear stresses.

Author Year Expression Range of validity

Domb 2001 σ cr ¼ kσ σ E ¼ π2 E
kσ 12ð1−ν t 2
2 Þ ðbÞ
Uniform compression of simply supported panels – Zb ≤ 200
( 3
∑i¼0 ci ½ logðZ b Þi if 1 ≤ Z b ≤ 15
kσ ¼ 10
cðZ b Þd if 15 ≤ Z b ≤ 200
where c0 ¼ 0:7782; c1 ¼ −0:1088; c2 ¼ 0:2025; c3 ¼ −0:0037;
c ¼ 1:4805; d ¼ 0:6401
DNV [7] 2013 σ cr ¼ kσ σ E ¼ kσ π2 E
ðt Þ
2 Simply supported curved panels
12ð1−ν2 Þ b
h 2
irffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
h 0:600:856pffibZ 3=4 iffi2
where kσ ¼ 5:34 þ 4ðbtÞ 1þ a b
s 2
5:34þ4ðt Þ

DNVGL [6] 2015 π E

σ cr ¼ kσ σ E ¼ kσ 12ð1−ν
2
2 Þ ðaÞ
t 2 Simply supported curved panels – R/t ≤ 2500
8 pffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffi
< 3 28:3 þ 0:67a 3
if a
≤ 8:7 Rt
R1:5 t 1:5 R
qffiffi
where kσ ¼ pffiffiffi
: pffiffiffiffi if
2
3 0:28a R N 8:7
a R
R Rt t

A.6. Ultimate strength of cylindrically curved panels under shear stresses

Table A.6
Formulae for obtaining the ultimate strength of cylindrically curved panels subject to shear stresses.

Author Year Expression Range of validity

DNV [7] 2013 χ ¼ qffiffiffiffiffiffiffiffiffi

1 ffi Simply supported curved panels
4
1þλ

Where the slenderness is calculated with the elastic critical stress obtained from Table A.5
8
DNVGL [6] 2015 <χ ¼ 1
> if λ ≤ 0:4 Simply supported curved panels – R/t ≤ 2500
χ ¼ 1:274−0:686λ if 0:4 b λ≤ 1:2
>
:χ ¼ 20:65
if 1:2 N λ
λ
Where the slenderness is calculated with the elastic critical stress obtained from Table A.5

References
[1] A. Reis, J.O. Pedro, A.B. Graça, C. Hendy, P. Romoli, L. Simões da Silva, J.P. Martins,
Report on the characterization of relevant parameters of curved plated bridge
structures and identification of bridge cases where they can be found, RFCS
Research Project OUBURST (RFCS-2015-709782): Deliverable 2.1, 2017.
[2] J. Manterola, Pasarela de Peatones - Zaragoza EXPO 2008, 2008 PowerPoint
presentation.
[3] H. Schmidt, Stability of steel shell structures, J. Constr. Steel Res. 55 (2000)
159–181.
[4] Buckling of shells, European Design recommendations, 5th edition, in: J.M. Rotter,
H. Schmidt (Eds.),Brussels, European Convention for Constructional Steelwork
(ECCS), Publication Number: 125, 2008.
[5] T. Manco, C. Rigueiro, J.P. Martins, L. Simões da Silva, Comparative assessment of
the design of tubular elements according to offshore design standards and
Eurocode 3, Steel Constr. Des. Res. 9 (4) (2016) (266-178).
[6] DNVGL, DNVGL-CG-0128, Buckling Class Guideline, DNVGL, Norway, 2015.
[7] DNV, DNV-RP-C202, Buckling Strength of Shells, DNV, Norway, 2010.
[8] S.B. Batdorf, A simplified method of elastic stability analysis for thin cylindrical shells,
National Aeronautics and Space Administration. Technical Report number: 847, 1947.
[9] S.B. Batdorf, A simplified method of elastic-stability analysis for thin cylindrical
shells I – Donnell's Equation, National Aeronautics and Space Administration.
Technical Report number: 1341, 1947.
[10] S.B. Batdorf, A simplified method of elastic-stability analysis for thin cylindrical
shells II – modified equilibrium equation, National Aeronautics and Space Admin-
istration. Technical Report number: 1342, 1947.
[11] J.P. Martins, L. Simões da Silva, A. Reis, Eigenvalue analysis of cylindrically curved
under compressive stresses – extension of rules from EN1993-1-5, Thin-Walled
Struct. 68 (2013) 183–194.
[12] J.P. Martins, L. Simões da Silva, A. Reis, Ultimate load of cylindrically curved panels
under in-plane compression and bending – extension of rules from EN1993-1-5,
Thin-Walled Struct. 77 (2014) (36-4).
[13] F. Ljubinkovic, Cylindrically Curved Steel Panels in Bridge DesignProject of thesis
University of Coimbra, 2016.
[14] A. Reis, D. Camotim, Estabilidade e Dimensionamento de Estruturas, Edições Orion,
Amadora, 2012.
206 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207
[15] G. Gerard, H. Becker, Handbook of structural stability: part III – buckling of Curved
Plates and shells, National Advisory Committee for Aeronautics. Technical Report
number: 3783, 1957.
[16] S.C. Redshaw, The elastic instability of a thin curved panel subjected to an axial
thrust, its axial and circumferential edges being simply supported, British
Aeronautical Research Council. Reports and Memorandum Number: 1565, 1934.
[17] K. Marguerre, On the theory of the curved plate with large displacements, Proc.
Fifth International Congress for Applied Mechanics, Vol. 5, Massachusetts,
Cambridge 1939, p. 93.
[18] D.M.A. Leggett, The buckling behaviour of a long curved panel under axial com-
pression, Aeronautical Research Council. Report and Memoranda No. 1899, 1942.
[19] K. Tran, L. Davaine, C. Douthe, K. Sab, Stability of curved panels under uniform axial
compression, J. Constr. Steel Res. 69 (2012) 30–38.
[20] E.Z. Stowell, Critical compressive stress for a curved sheet supported along all
edges and elastically restrained against rotation along the unloaded edges, National
Advisory Committee for Aeronautics. War Report number: L-691, 1943.
[21] L.H. Sobel, T. Weller, B.L. Argawal, Buckling of cylindrical panels under axial
compression, Comput. Struct. 6 (1976) 29–35.
[22] M.N. Bismarck-Nasr, Buckling analysis of cylindrically curved panels based on a
two-field variables variational principle, Comput. Struct. 51 (4) (1995) 453–457.
[23] S.P. Timoshenko, J.M. Gere, Theory of Elastic Stability, 2nd Ed Dover Publications,
New York, 1961.
[24] A.S. Volmir, Stability of elastic systems. Fizmatig, Moscow, Trans. National Aeronau-
tics and Space Administration. Technical Memorandum number: AD628508, 1963.
[25] M.M. Domb, B.R. Leigh, Refined design curves for compressive buckling of curved
panels using nonlinear finite element analysis, 42nd AIAA/ASME/AHS/ASC
Structures, Structural Dynamics and Materials Conference, Seattle, U.S.A. Paper
1348, 2001.
[26] J.P. Martins, L. Simões da Silva, N. Silvestre, Energy-based analytical model to pre-
dict the elastic critical behaviour of curved panels, J. Constr. Steel Res. 127 (2016)
165–175.
[27] R.E. Snyder, A Stability Analysis of Cylindrical Panels Using a Finite Element
FormulationPhD Thesis Virginia Polytechnic Institute and State University, USA,
1971.
[28] D. Durban, J. Singer, Buckling of Cylindrical Panels Under Non-uniform Axial
Compression. TAE Report No. 136, 1972.
[29] K. Magnucki, M. Mackiewicz, Elastic buckling of an axially compressed cylindrical
panel with edges simply supported and one edge free, Thin-Walled Struct. 44
(4) (2006) 387–392.
[30] R. Wilde, K. Zawodny, K. Magnucki, Critical state of an axially compressed cylindri-
cal panel with three edges simply supported and one edge free, Thin-Walled
Struct. 45 (2007) 955–959.
[31] K.H. Chu, G. Krishnamoorthy, Buckling of open cylindrical shells, J. Eng. Mech. 2
(1967) 177–205.
[32] H.R. Eipakchi, M. Shariati, Buckling analysis of a cylindrical panel under axial stress
using perturbation technique, J. Appl. Math. Mech. 91 (2) (2011) 138–145.
[33] S.E. Afkhami, Maleki Vahid, M. Niazi, M. Abbasi, Analytical and numerical investiga-
tions on buckling of an axially compressed cylindrical panel with specific boundary
condition, International Review of Mechanical Engineering 9 (2) (2015).
[34] S. Batdorf, M. Schildcrout, Critical axial compressive stress of a curved rectangular
panel with a central chordwise stiffener, NACA. Technical Note No.: 1661, 1948.
[35] M. Schildcrout, M. Stein, Critical axial compressive stress of a curved rectangular
panel with a central longitudinal stiffener, NACA. Technical Note No.: 1879, 1949.
[36] J.P. Peterson, R.O. Whitley, Local buckling of longitudinally stiffened curved plates,
NASA. Technical Note: D-750, 1961.
[37] L.H. Sobel, B.L. Argawal, Buckling of eccentrically stringer-stiffened cylindrical
panels under axial compression, Comput. Struct. 6 (1976) 193–198.
[38] M. Khedmati, P. Edalat, A numerical investigation into the effects of parabolic
curvature on the buckling strength and behaviour of stiffened plates under in-
plane compression, Latin American Journal of Solids and Structures 7 (3) (2010)
249–264.
[39] D.M.A. Leggett, The Behavior of a Cylindrical Shell Under Axial Compression When
the Buckling Load Has Been Exceeded, HM Stationery Office, 1942.
[40] H.L. Cox, E. Pribram, The elements of the buckling of curved plates, Journal of the
Royal Aeronautical Society 52 (453) (1948) 551–565.
[41] W.T. Koiter, Buckling and post-buckling behaviour of a cylindrical panel under
axial compression, Reports and Transactions National Aeronautical Research
Institute, 20, 1956, pp. 71–84.
[42] G.G. Pope, On the axial compression of long, slightly curved panel, British
Aeronautical Research Council. Report and Memoranda Number: 3392, 1965.
[43] O. Tamate, H. Sekine, Postbuckling behavior of thin curved panels under axial com-
pression, Bulletin of Japan Society of Mechanical Engineering 12 (51) (1969)
415–420.
[44] G.H. Knightly, D. Sather, Nonlinear buckling and stability of cylindrical panels,
SIAM J. Math. Anal. 10 (2) (1979) 389–403.
[45] Y. Zhang, F.L. Mathews, Large deflection analysis behaviour of simply supported
laminated panels under in-plane loading, ASME Journal of Applied Mechanics 52
(1985) 553–558.
[46] V. Tvergaard, Buckling of elastic-plastic cylindrical panel under axial compression,
Int. J. Solids Struct. 13 (1977) 957–970.
[47] A. Chajes, Principles of Structural Stability Theory, Prentice-Hall, New Jersey, 1974.
[48] V. Tvergaard, A. Needleman, Buckling localization in a cylindrical panel under axial
compression, Int. J. Solids Struct. 37 (46–47) (2000) 6825–6842.
[49] M.W. Hilburger, V.O. Britt, M.P. Nemeth, Buckling behavior of compression-loaded
quasi-isotropic curved panels with a circular cutout, Int. J. Solids Struct. 38 (2001)
1495–1522.
[50] B.L. Agarwal, Post-buckling behaviour of composite stiffened curved panels loaded
in compression, Exp. Mech. 22 (6) (1982) 231–236.
[51] D. Hui, Y.H. Chen, Imperfection-sensitivity of cylindrical panels under compression
using Koiter's improved postbuckling theory, Int. J. Solids Struct. 23 (7) (1987)
969–982.
[52] M.B. Parks, C. Santaputra, W. Yu, Local buckling of curved elements, 8th Interna-
tional Specialty Conference on Cold Formed Steel Structures, Missouri, USA 1986,
pp. 277–294.
[53] I. Sheinman, Y. Frostig, Postbuckling analysis of stiffened laminated curved panels,
ASCE Journal of Engineering Mechanics 116 (1990) 2223–2237.
[54] J.L. Verolme, The Development of a Design Tool for Fiber Metal Laminate Compres-
sion PanelsPhD Thesis Technical University of Delft, Netherlands, 1995.
[55] J.S. Park, K. Iijima, T. Yao, Buckling/ultimate strength and progressive collapse
behaviour comparison of unstiffened and stiffened curved plates subjected to
axial compression, Int. J. Comput. Appl. Technol. 41 (1–2) (2011) 60–72.
[56] L.K. Tran, C. Douthe, K. Sab, J. Dallot, L. Davaine, A preliminary design formula
for the strength of stiffened curved panels by design of experiment method,
Thin-Walled Struct. 79 (2014) 129–137.
[57] L.K. Tran, C. Douthe, K. Sab, J. Dallot, L. Davaine, Buckling of stiffened curved panels
under uniform axial compression, J. Constr. Steel Res. 103 (2014) 140–147.
[58] J.K. Seo, C.H. Song, J.S. Park, Nonlinear structural behaviour and design formulae for
calculating the ultimate strength of stiffened curved plates under axial compression,
Thin-Walled Struct. 107 (2016) 1–17.
[59] Herbert Becker, Handbook of structural stability: part VI— strength of stiffened
curved plates and shells, NACA Technical Note: 3786, 1958.
[60] T. von Kármán, E.E. Sechler, L.H. Donnell, Strength of thin plates in compression,
Trans. Am. Soc. Mech. Eng. 54 (1932) 53–57.
[61] W.T. Koiter, Over de Stabiliteit van het Elastische EvenwichtPh.D. Thesis NASA.
Technical. Translation Number: TT F 10833, Delft University, 1945.
[62] B. Budiansky, J.W. Hutchinson, Dynamic buckling of imperfection sensitive
STRUCTURES, Proceedings of the 11th IUTAM Congress, Munich, Germany 1964,
pp. 636–651.
[63] E.E. Sechler, The Ultimate Compressive Strength of Thin Sheet Metal Panels,
Ph.D. thesis California Institute of Technology, 1934.
[64] J.S. Newell, W.H. Gale, A Report on Aircraft Materials, Massachusetts Institute of
Technology, USA, 1932.
[65] W.A. Wenzek, The effective width of curved sheet after buckling, NACA. Technical
Memoranda No.: 880, 1938.
[66] H.L. Cox, W.J. Clenshaw, Compression tests on curved plates of thin sheet duralumin,
Aeronautical Research Council. Reports and memoranda No.: 1849, 1941.
[67] H. Crate, L.R. Levin, Data on buckling strength of curved sheet in compression,
National Advisory Committee for Aeronautics. Technical Report No.: L-557,
1943.
[68] K.V. Jackson, A.H. Hall, Curved plates in compression, National Research Council,
AR-1, 1947.
[69] G. Welter, Curved aluminum-alloy sheets in compression for monocoque construc-
tions, J. Aerosol Sci. 12 (3) (1945) 357–369.
[70] G. Welter, Influence of different factors on buckling loads of curved thin
aluminum-alloy sheets for monocoque constructions, J. Aerosol Sci. 13 (4)
(1946) 204–208.
[71] G. Welter, The effect of radius of curvature and preliminary artificial eccentricities
on buckling loads of curved thin aluminum-alloy sheets for monocoque construc-
tions, J. Aerosol Sci. 13 (11) (1946) 593–596.
[72] E.H. Schuette, Buckling of curved sheet in compression and its relation to the
secant modulus, Journal of the Aeronautic Sciences 15 (1) (1948) 18–22.
[73] D.J. Tenerelli, A.M.C. Holmes, An Experimental Buckling Study of Skin-corrugated
Ring-stiffened Curved Panels, S.E.S.A. Spring Meeting, Cleveland, Paper No.
1993A, 1972.
[74] F.C. Thomas, A Study of Rectangular Plates Subjected to Non-uniform Axial
CompressionPhD Thesis Georgia Institute of Technology, USA, 1974.
[75] G. Gerard, Handbook of structural stability: supplement to part III – buckling of
curved plates and shells, NASA. Technical Report No.: D-163, 1959.
[76] C.A. Featherston, C. Ruiz, Buckling of curved panels under combined shear and
compression, Proc. Inst. Mech. Eng. C J. Mech. Eng. Sci. 212 (3) (1998) 183–196.
[77] K. Yumura, S. Katsura, K. Iijima, T. Yao, Simulation of buckling collapse behaviour of
cylindrically curved plates under axial compression, The Nineteenth Asian Technical
Exchange and Advisory Meeting on Marine Structures, Singapore, 2005.
[78] J.S. Park, K. Iijima, T. Yao, Estimation of buckling and collapse behaviours of stiff-
ened curved plates under compressive load, Proceedings of the 18th International
Offshore and Polar Engineering Conference, Vancouver, Canada, 2008.
[79] J.S. Park, K. Iijima, T. Yao, Characteristics of buckling and ultimate strength and
collapse behaviour of cylindrically curved plates subjected to axial compression,
Adv. Mater. Res. 33-37 (2008) 1195–1200.
[80] C.A. Featherston, Geometric imperfection sensitivity of curved panels under
combined compression and in-plane bending – a study using adaptive meshing
and DIC, Strain 48 (2012) 286–295.
[81] K. Tran, Étude de la résistance et de la stabilité des tôles courbes en acier,
Application aux ouvrages d'art, Université Paris-Est, 2012 PhD thesis.
[82] J.P. Martins, D. Beg, F. Sinur, L. Simões da Silva, Imperfection sensitivity of cylindri-
cally curved steel panels, Thin-Walled Struct. 89 (2015) 101–115.
[83] CEN, Design of steel structures: strength and stability shells structures, Brussels,
European Committee for Standardisation, 2007 EN 1993-1-6.
[84] CEN, Design of steel structures: plated structural elements, Brussels, European
Committee for Standardisation, 2006 EN1993-1-5.
[85] H. Gall, Compressive Strength of Stiffened Sheets of Aluminum AlloyPh.D. thesis
Massachusetts Institute of Technology, 1930.
 J.P. Martins et al. / Journal of Constructional Steel Research 140 (2018) 191–207
[86] E. Lundquist, Comparison of three methods for calculating the compressive
strength of flat and slightly curved sheet and stiffener combinations, NACA.
Technical Report No.: 455, 1933.
[87] I.J. Gerard, B.G. Dickens, Stressed-skin structures. Compression tests of panels with
tubular stiffeners, Aeronautical Research Council. Reports and Memoranda No.:
1830, 1936.
[88] W. Ramberg, A.E. McPherson, S. Levy, Compressive Tests of a Monocoque Box,
NACA. Technical Note No.: 721, 1939.
[89] E.E. Lundquist, Preliminary data on buckling strength of curved sheet panels in
compression, NACA. War time Report No.: L-690, 1941.
[90] M. Holt, Tests of aluminum-alloy stiffened sheet specimens cut from an airplane
wing, NACA. Technical Note No.: 883, 1943.
[91] W. Ramberg, S. Levy, K. Fienup, Effect of curvature on strength of axially loaded
sheet-stringer panels, NACA. Technical Note No.: 944, 1944.
[92] A.B.T. Soderquist, Experimental Investigation of Stability and Post Buckling Behav-
iour of Stiffened Curved Plates, Institute of Aerophysics, University of Toronto,
1960.
[93] R.C. Tennyson, An experimental investigation of the behavior of stiffened plates in
axial compression, UTIA. Technical Note No.: 57, 1961.
[94] S.R. Cho, H.Z. Park, H.S. Kim, J.S. Seo, Experimental and numerical investiga-
tions on the ultimate strength of curved stiffened plates, Proceeding 10th In-
ternational Symposium on Practical Design of Ships and Other Floating
Structures, 2007.
[95] P. Jetteur, R. Maquoi, Larguer effective d'une tôle courbe comprimée, Construction
Métallique 2 (1984) 51–57.
[96] C. Massonet, M. Save, Résistance limite d'une poutre courbe à parois minces
soumise à flexion, Mémoires AIPC 23 (1963) 245–261.
[97] D. Vandepitte, Ultimate strength of curved flanges of I-beams, J. Constr. Steel Res. 3
(1982) 22–28.
[98] D. Vandepitte, B. Verhegge, Ultimate Strength of curved flanges in box-girders,
Instability and Collapse of Steel Structures, Granada Pub., London 1983,
pp. 446–454.
[99] P. Dubas, E. Gehri, Structural stability, behaviour and design of steel plated struc-
tures, ECCS. Publication No.: 44, 1986.
[100] T.H. Jacques, R. Maquoi, G. Fonder, Buckling of unstiffened compression curved
plates, J. Constr. Steel Res. 3 (1) (1983) 28–34.
[101] H. Nakai, C.H. Yoo, Analysis and Design of Curved Steel Bridges, McGraw-Hill, New
York, 1988.
[102] A. Zureick, R. Naqib, Horizontally curved steel I-girders state-of-the-art analysis
methods, J. Bridg. Eng. 4 (1) (1999) (38-37).
[103] P.F. Schuenzel, B.T. Yen, J.H. Daniels, Strength of Horizontally Curved Steel Plate
Girders, Fritz Engineering Laboratory, 1985 Report No.: 454.1.
[104] D. Linzell, D. Hall, D. White, Historical perspective on horizontally curved I girder
bridge design in the United States, J. Bridg. Eng. 9 (3) (2004) 208–229.
[105] D.G. Linzell, Studies of a Full-scale Horizontally Curved Steel I Girder Bridge System
Under Self-weightPhD Thesis School of Civil and Environmental Engineering,
Georgia Institute of Technology, 1999.
[106] A. Zureick, D. Linzell, R.T. Leon, J. Burrell, Curved steel I girder bridges: experimen-
tal and analytical studies, Eng. Struct. 22 (2) (2000) 180–190.
[107] T. Yabuki, Y. Arizumi, T. Shimozato, Y. Nagamine, Buckling modes of plate- girders
curved in plan, Structural and Earthquake Engineering 12 (2) (1995) 67–72.
[108] T. Yabuki, Y. Arizumi, S. Yashiro, Ultimate strength and its practical evaluation of
cylindrical steel shell panels under various compression, Structural Engineering
and Earthquake Engineering 11 (1) (1994) 11–21.
207
[109] P. Van Bogaert, Local buckling of curved steel girder webs. 6th European confer-
ence on steel and composite structures, Proceedings of 6th European Conference
on Steel and Composite Structures, Brussel 2011, pp. 795–800.
[110] D.M.A. Leggett, The elastic stability of a long and slightly bent rectangular plate
under uniform shear, Proceedings of the Royal Society A (Mathematical, Physical
and Engineering Sciences) 162 (908) (1937) 62–83.
[111] W.R. Dean, Theory of elastic stability, Proceedings of the Royal Society 107 (744)
(1925) 734–760.
[112] S.B. Batdorf, M. Stein, M. Schildcrout, Critical shear stress of curved rectangular
panels, NACA. Technical note No.: 1348, 1947.
[113] D.J. Johns, Shear Buckling of isotropic and orthotropic plates: a review, Aeronauti-
cal Research Council. Reports and Memoranda No.: 3677, 1970.
[114] E.H. Baker, L. Kovalevsky, F.L. Rish, Structural Analysis of Shells, Robert E. Krieger
Pub. Company, Malabar, Florida, 1972.
[115] K. Sayers, The initial buckling of slightly curved panels under combined longitudi-
nal and circumferential direct stresses, The Journal of the Royal Aeronautical Soci-
ety 68 (640) (1964) 271–274.
[116] D.J. Dawe, Finite strip buckling analysis of curved plate assemblies under biaxial
loading, Int. J. Solids Struct. 13 (1977) 1141–1155.
[117] M.W. Hilburger, M.P. Nemeth, J.H. Starnes, Nonlinear and buckling behavior of
curved panels subjected to combined loads, 42nd AIAA/ASME/AHS/ASC Struc-
tures, Structural Dynamics and Materials Conference, Seattle, U.S.A. Paper 1398,
2001.
[118] J. Girish, L.S. Ramachandra, Stability and vibration behavior of composite cylindri-
cal shell panels under axial compression and secondary loads, Journal of Applied
Mechanics (ASME) 75 (2008) 1–11.
[119] Sh. Kasaeian, M. Azhari, A. Heidarpour, A. Hajiannia, Inelastic local buckling of
curved plates with or without thickness-tapered sections using finite strip method,
International Journal of Steel Structures 12 (3) (2012) 427–442.
[120] M.Y. Chang, L. Librescu, Postbuckling of shear deformable flat and curved panels
under combined loading conditions, Int. J. Mech. Sci. 37 (1995) 121–143.
[121] A. Kromm, The limit of stability of curved plate strip under shear and axial stresses,
NACA. Technical Note No.: 898, 1939.
[122] D.M.A. Leggett, The initial buckling of slightly curved panels under combined shear
and compression, Aeronautical Research Council. Reports and memoranda No.:
1972, 1943.
[123] C.A. Featherston, The use of finite element analysis in the examination of instability
in flat plates and curved panels under compression and shear, Int. J. Non Linear
Mech. 35 (2000) 515–529.
[124] C.A. Featherston, Imperfection sensitivity of curved panels under combined
compression and shear, Int. J. Non Linear Mech. 38 (2003) 225–238.
[125] M.M. Domb, R.L. Leigh, Refined design curves for shear buckling of curved panels
using nonlinear finite element analysis, 43rd AIAA/ASME/AHS/ASC Structures,
Structural Dynamics and Materials Conference, Denver, U.S.A, Paper #2002-1257,
2002.
[126] M. Amani, B.L.O. Edlund, M.M. Alinia, Buckling and postbuckling behaviour
of unstiffened curved plates under uniform shear, Thin-Walled Struct. 49 (8)
(2011) 1017–1031.
[127] M.M. Domb, Nonlinear buckling predictions of curved panels under combined
compression and shear loading, 23rd Congress of International Council of the
Aeronautical Sciences, ICAS2002, Toronto, 322, 2002, pp. 1–322.8.
[128] M. Schildcrout, M. Stein, Critical combinations of shear and direct axial stress for
curved rectangular panels, NACA. Technical Note No.: 1928, 1949.
[129] DNV, DNV-RP-C201. Buckling Strength of Plated Structures, Norway, DNV, 2010.