Ch17 Arches

CHAPTER 17
ARCHES
17.1 INTRODUCTION
An arch is a structure curved in elevation, loaded in its plane, with spreading of the
supports prevented, and with its rib primarily in compression. The stability can be
characterized by buckling in or out of the plane of the arch. In-plane buckling occurs
when the arch is substantially braced against out-of-plane deformations, while
out-of-plane buckling occurs for arches with significant free-standing portions.
In-plane buckling is associated with combined compression and bending while
out-of-plane buckling is associated with compression, biaxial bending, and torsion.
Arch-type structures are most efficient if they carry their load in such a way
that the funicular curve coincides with the centroidal axis of the rib, which results
in axial compression and no bending of the arch rib. Examples of arches under
pure axial compression include circular arches subjected to uniform normal pres-
sure, commonly called hydrostatic loading, parabolic arches subjected to uniform
load on a horizontal projection, and catenary arches with load uniformly dis-
tributed along the arch axis. Arches can further be classified by their articulation: a
three-hinged arch is statically determinate; two hinged and fixed arches are hyper-
static. In bridges, a distinction can be made between a through arch with the bridge
deck between the supports and suspended from the arch rib and a deck arch with
the bridge deck above the arch rib and supported by spandrel columns. The hor-
izontal reaction forces (or thrusts) can be resisted in two ways: In the abutment
arch, also called a true arch, the horizontal reactions are resisted by the abutments
or supports, while in the tied arch, the reactions are carried by a tie between the
two supports. The deck arch is an example of the former and the through arch
the latter.
As shown in Fig. 17.1, arches are subject to two fundamentally different
types of loading, including gravity-type loading and tilting loads, also referred
762 Guide to Stability Design Criteria for Metal Structures, Sixth Edition Edited by Ronald D. Ziemian
Copyright © 2010 John Wiley & Sons, Inc.
INTRODUCTION 763
Bridge deck
Buckled
shape
Hor. component
Buckled Hor. comp.
shape
Buckled
shape
Bridge deck
In-plane: Tilting loads Out-of-plane: Tilting loads Gravity load
FIGURE 17.1 Gravity and tilting loads in arches.
to as conservative and nonconservative loading, respectively. In a through-arch

bridge that is restrained to buckle in a sway mode, the hangers tilt and develop a
horizontal-load component that supports the arch rib against further buckling both
in and out of plane. Generally, the buckling loads associated with tilting loads are
considerably higher than those of gravity loading cases. In a deck arch bridge,
the in-plane horizontal-load component of the tilting spandrel columns tends to
push the arch rib away from its funicular shape. The out-of-plane horizontal-load
component drives the arch out of its plane and causes bending and torsion. Hence,
the buckling loads for deck arch bridges are usually smaller than those of through
arch bridges when considering gravity load conditions.
This chapter treats in-plane stability (Section 17.2) and out-of-plane stability
(Section 17.3) for various shapes of arches and types of loading. In both cases,
specific deign criteria are presented. Furthermore, bracing recommendations for
arches are presented in Section 17.4 and stability design criteria for arch bridges
are given in Section 17.5.
Because it is not feasible to give a complete overview of all research on the
stability of arches within a single chapter, the more important and recent findings
will be presented. There are, however, several excellent resources on arches. An
extensive bibliography on the stability of arches prior to 1970 is given by DaDeppo
and Schmidt (1970). The Handbook of Structural Stability (Hayashi, 1971) gives an
overview of results of stability research of arches in which either the equations or
graphs of the quoted literature are reproduced. Petersen (1982) dedicates a chapter
to the in-plane and out-of-plane stability of arches and arch bridges. In Stability of
Metal Structures, a World View (Beedle 1991), international building standards with
provisions for in-plane and out-of-plane stability are compared with each other.
An extensive state-of-the-art report on elastic and inelastic stability of arches is
given in Fukumoto (1996). Singer et al. (1998) provides a chapter on experimental
research that has been conducted on arches. King and Brown (2001) present a
comprehensive study for the practical design of steel curved beams and arches,
which includes several worked examples.
764 ARCHES
17.2 IN-PLANE STABILITY OF ARCHES
The analysis and design of arches may be broadly classified according to their
response to load and their in-plane mode of failure. Based on these characteristics,
the types of arches that will be treated in this chapter include the following:
• Slender arches, generally of solid-web rolled or built-up sections, subject pri-

marily to axial force in the arch rib (analogous to axially loaded columns)
• Slender arches subject to significant bending and deformation, due primarily
to asymmetric loading (analogous to beam-columns)
• Stocky arches, frequently of truss form, used because of high bending loads,
in which strength limit states will be due primarily to chord or flange failure
arising from axial load and excessive bending (similar to any truss subject to
axial load and moment)
• Arches composed of arch ribs and deck-stiffening girders
When the loads acting on an arch are increased proportionally, it loses its stabil-
ity as a certain critical value of the load is attained. In the case of elastic structures
under conservative loads the critical load always corresponds to either a birfurca-
tion point (see Section 2.2) or a stability limit point (Section 2.3). Various possible
load–deformation plots for arches, called equilibrium paths, are shown in Fig. 17.2.
Each point on a path represents an equilibrium configuration of the structure.
Figure 17.2a shows an equilibrium path for an unsymmetric problem in which
the arch boundary conditions and/or the manner of loading may be unsymmetrical.
The point on the equilibrium path at which the load is a relative maximum is called
a limit point.
Figure 17.2b illustrates possible load–deformation relationships for a symmet-
rical and symmetrically loaded arch. Here the primary or fundamental equilibrium
path is intersected by a secondary path. The primary branch represents a symmetric
deflected mode and the secondary branch an antisymmetric mode, with bifurcation
points defined where equilibrium paths intersect. If an antisymmetrical mode does
not become dominant, the arch eventually will become unstable in a symmetrical
mode with the load–deflection curve gradually reaching a limit point. On the other
hand, the limit load may be significantly reduced if an antisymmetrical buckling
mode dominates, as shown in Fig. 17.2b. Although the load capacity usually drops
continuously beyond the bifurcation point, as is the behavior when the arch buckles
in sidesway, it is possible for the load to increase slightly after bifurcation. In this
case and as shown in Fig. 17.2c, the maximum load is attained at a limit point that
corresponds to large in-plane displacements.
In the case of an unstable, rapidly descending postbuckling path, the arch is
sensitive to geometric imperfections or load eccentricity. The associated drop in
critical load is shown in Fig. 17.2b (dashed curve), and the buckling load at the
bifurcation point is replaced by the snap load at the limit point on the dashed curve.
The degeneration of a bifurcation point into a limit point due to the presence of
imperfections suggests that bifurcation is an exception rather than the rule.
IN-PLANE STABILITY OF ARCHES 765
FIGURE 17.2 Some possible load–deformation paths for arches. (a) Nonlinear stability:
unsymmetrical or unsymmetrically loaded arch. (b) Nonlinear stability: symmetrical and
symmetrically loaded arch. (c) Nonlinear stability: symmetrical and symmetrically loaded
arch. (d ) Linear stability: negligible prebuckling deformation.
Arch ribs under pure compression experience only very small displacements
before buckling. In the ideal situation such structures undergo birfurcation buckling.
These bifurcation problems are far more amenable to analysis because they permit a
linearization of the equilibrium equations for the prebuckling, or fundamental, state.
As in small-deflection column theory, this linearization is based on the assumption
of very small displacements and a horizontal equilibrium path (Fig. 17.2d ).
For practical design loadings, however, the funicular curves normally do not
coincide with the centroidal axis and in such cases the arch experiences substantial
bending moments and displacements before buckling. Studies that take into account
these prebuckling deformations are nonlinear stability problems, as illustrated by
Figs. 17.2a –c.
17.2.1 Linear Stability

Early papers on arch stability were devoted to linear stability problems. This work
by Gaber, Stüssi, Kollbrunner, Hilman, Dischinger, and Dinnik is summarized by
Austin (1971) and Timoshenko and Gere (1961).
The buckling of arches of constant cross section (herein termed uniform arches)
in which the arch is the funicular curve for the loading are first considered.
766 ARCHES
TABLE 17.1 Critical-Load Parameter and Critical Horizontal Reaction Parameter

for Uniform Elastic Arches in Pure Compressiona
h Three-Hinged Arch Two-Hinged Arch Fixed Arch

L qL3 /EI HL2 /EI qL3 /EI HL2 /EI qL3 /EI HL2 /EI
Parabolic arches subjected to vertical load uniformly distributed on a horizontal projection

0.10 22.5 28.1 29.1 36.3 60.9 76.2
0.15 39.5 32.9 85.1 70.9
0.20 39.6 24.8 46.1 28.8 103.1 64.5
0.25 49.2 24.6 114.6 57.3
0.30 49.5 20.6 49.5 20.6 120.1 50.0
0.35 47.8 17.1 120.6 43.1
0.40 45.0 14.1 45.0 14.1 117.5 36.7
0.50 38.2 9.6 38.2 9.6 105.3 26.3
Catenary arches subjected to vertical load uniformly distributed along the arch axis
0.10 28.7 36.3 60.1 76.2
0.15 38.3 32.8 82.7 70.9
0.20 43.5 28.5 98.0 64.3
0.25 44.8 24.1 105.9 56.9
0.30 43.2 19.8 107.4 49.4
0.35 39.7 16.1 104.0 42.1
0.40 35.3 12.9 97.2 35.4
0.50 26.5 8.2 79.3 24.5
Circular arches subjected to normal load uniformly distributed along the arch axis
0.10 22.2 26.7 28.4 34.1 58.9 70.7
0.20 33.5 17.6 39.3 20.6 90.4 47.5
0.30 34.9 9.3 40.9 10.9 93.4 24.9
0.40 30.2 3.4 32.8 3.7 80.7 9.1
0.50 24.0 0 24.0 0 64.0 0
a h,
rise; L, span; q, critical intensity of distributed load; H , critical horizontal reaction at supports; E ,
Young’s modulus of elasticity; I , moment of inertia of the cross section.
The critical values of distributed load and horizontal reaction of fixed two-hinged
and three-hinged symmetric arches are summarized in Table 17.1 for the follow-
ing cases: (1) parabolic arches subjected to vertical load uniformly distributed on
a horizontal projection, (2) catenary arches under uniform vertical load along the
arch axis, and (3) circular arches subjected to uniform normal pressure, commonly
called hydrostatic loading. In each of these cases, the arch assumed is geometrically
perfect and the loading results in pure axial compression (no bending) at every cross
section of the arch. The arches are free to buckle in their plane without restraint.
Axial compressive strain has been neglected in the analyses reported because it has
been shown to have a very small effect on the behavior slender arches. In all cases,
buckling is defined by bifurcation from the undeflected position, as repsented in
Fig. 17.2d . The critical values in Table 17.1 are given for a range of rise-to-span
ratios from 0.10 to 0.50.
In Table 17.1, the critical values for two-hinged and fixed parabolic and catenary
arches are taken from Austin and Ross (1976). The remaining values are from
Timoshenko and Gere (1961). Kollbrunner (1936, 1942) showed that the elastic
critical values obtained experimentally are in close agreement with the theoretical
values reported in Table 17.1.
The fundamental modes that correspond to the critical values given in Table 17.1
are shown in Fig. 17.3. The fixed and two-hinged arches always buckle into an
antisymmetrical mode in which the arch sways laterally, with the crown moving
horizontally and becoming a point of contraflexure (Figs. 17.3a and b). In the
case of a three-hinged arch, buckling is symmetric for low rise-to-span ratios, as
shown in Fig. 17.3c, with the crown moving downward. For high rise-to-span
ratios, the three-hinged arch buckles in an antisymmetrical mode similar to that of
the two-hinged arch and at the same critical load.
The sidesway buckling behavior of uniform arches of low and moderate span-to-
depth ratios under pure axial compression is similar to the buckling behavior of
straight columns. For example, a fixed arch that buckles into two waves with a
FIGURE 17.3 Fundamental buckling mode shapes.

768 ARCHES
point of contraflexure at the crown (Fig. 17.3a) has a mode shape from support
to crown similar to that of a fixed-hinged column, and the critical axial thrust at
the quarter-point of the arch agrees fairly well with the critical compressive force
in the end-loaded, fixed-hinged column whose length is equal to the arc length of
the arch from support to crown. This interpretation is developed fairly extensively
in the third edition of this guide. It is not repeated here because it is not useful,
except as a very broad concept, for studying the elastic buckling of slender arches
under more general loadings that do not produce pure axial compression. It is used,
however, in ultimate-strength studies and as a design tool, as will be described.
The elastic buckling of nonuniform parabolic arches subjected to vertical loading
distributed uniformly on a horizontal projection has been studied for moment-of-
inertia variations I = Ic sec φ and I = Ic sec3 φ, in which Ic is the moment of inertia
at the crown and φ the angle between the tangent to the arch axis and the horizontal.
These studies are reported in the third edition of the guide. A simple approximate
procedure for applying uniform arch-buckling theory to cases of nonuniform arches
with antisymmetrical buckling is to assume the critical load on a nonuniform arch
equals that of a similar uniform arch with an “equivalent” moment of inertia. First,
one-half of the nonuniform arch is straightened out to form a simply supported
beam that is then loaded with a concentrated force at midspan. From the deflection
at midspan, an equivalent uniform moment of inertia can be computed and used to
represent a nonuniform arch. The result of such a computation is discussed in detail
in the third edition of the guide. Similar procedures developed by Aas-Jakobsen
are described by Forrester (1972) in a discussion of Austin’s (1971) paper. More
recent results on the stability of tapered arches are reported by Wolde-Tinsaie and
Foadian (1989).
Stiffened arches are deck structures that consist of an arch connected to an
overhead horizontal girder by closely spaced columns. As discussed in the third
edition of this guide, the critical load and critical horizontal reaction for two- and
three-hinged, low-rise parabolic arches subjected to uniform load on a horizontal
projection can be closely approximated by use of the values in Table 17.1 with the
moment of inertia I equal to the sum of the arch and girder moments of inertia.
This simple concept recognizes that the arch and girder buckle together.
17.2.2 Nonlinear Elastic Stability

Section 17.2.1 addressed arches subjected to special loadings that produced only
axial compression. Because the arches were considered to be inextensible, there
were no prebuckling deflections. Actual design loadings on arches, however, usually
produce both an axial compression and bending moment on a general cross section
of the arch rib. These internal forces cause a change in shape of the arch before
buckling occurs and results in the problem being nonlinear even though the material
is deforming elastically.
Symmetrical Loadings on Symmetrical Arches The general behavior of

symmetrical arches that are symmetrically loaded is shown in Fig. 17.2b and has
TABLE 17.2 Elastic Buckling Coefficients for Uniform Arches with Vertical Load
Uniformly Distributed Along Arch Axisa
Two-Hinged Arch Fixed Arch
θ (deg) h i /L qL3 /EI HL2 /EI h c /h i qL3 /EI HL2 /EI h c /h t
Parabolic arches—antisymmetrical modes

0.10 28.6 36.3 1.002 60.4 76.1 1.003
0.15 38.2 32.9 1.004 83.4 70.8 1.006
0.20 43.4 28.8 1.006 99.3 64.3 1.010
0.25 44.9 24.5 1.008 107.7 57.0 1.013
0.30 43.5 20.5 1.009 109.6 49.5 1.015
0.35 40.4 16.96 1.009 106.4 42.5 1.017
0.40 36.5 13.98 1.009 100.0 36.2 1.017
0.50 28.5 9.51 1.008 83.0 25.9 1.017
Circular arches—antisymmetrical modes

50 0.1109 31.2 35.6 0.994 64.8 75.2 0.992
70 0.1577 39.5 32.0 0.989 83.4 70.0 0.983
90 0.2071 44.0 27.4 0.981 95.5 63.3 0.969
106.26 0.2500 44.5 23.2 0.974 99.9 57.0 0.966
120 0.2887 42.8 19.34 0.968 99.8 51.2 0.938
140 0.3501 37.1 13.75 0.959 93.8 42.1 0.910
160 0.4196 28.9 8.72 0.953 81.9 32.6 0.880
180 0.5000 20.0 4.78 0.950 66.0 23.4 0.854
Circular arches—symmetrical modes

50 0.1109 63.0 74.6 0.88 90.9 110.6 0.94
70 0.1577 77.6 66.8 0.81 110.0 98.5 0.91
90 0.2071 85.6 57.9 0.75 119.4 85.7 0.88
106.26 0.2500 87.7 50.5 0.69 120.3 75.2 0.85
120 0.2887 86.7 44.3 0.64 117.2 66.7 0.82
140 0.3501 81.7 35.1 0.57 107.5 54.2 0.78
160 0.4196 73.8 27.1 0.45 93.5 43.1 0.73
180 0.5000 64.4 19.07 0.34 77.2 32.5 0.67
ah
i , initial rise of arch; h c , height of arch at crown at instant of buckling; θ, angle of opening of the
circular arch.
been described previously. Buckling data are given in Table 17.2 for two-hinged and
fixed parabolic and circular arches subjected to vertical load uniformly distributed
along the arch axis. These data are from Austin and Ross (1976). Similar data
are also available in this paper for a single concentrated load at midspan. In these
studies axial strains have been neglected because their influence is very small for
slender arches.
The antisymmetrical buckling loads for circular arches are shown to be less than
the symmetrical buckling loads in Table 17.2. The same is true for parabolic arches,
770 ARCHES
although the symmetrical buckling critical loads are not available. In general, anti-
symmetrical buckling governs, although there are few cases in which symmetrical
buckling controls, such as fixed circular arches with a single concentrated load at
the crown and fixed parabolic arches with a rise-to-span ratio less than about 0.40
that are also subjected to a concentrated load at the crown. It can be observed from
Table 17.2 that the arches tend to buckle in a sidesway mode when the ratio hc /hi
is close to unity, which indicates that for these cases the profile at the instant of
buckling is nearly the same as the unloaded profile.
In the fifth edition of the guide, it was shown that the critical conditions for
antisymmetrical buckling can be expressed in terms of horizontal reactions. It was
further shown that the critical horizontal reactions are insensitive to the arch shape
and to a lesser extent to the loading; they vary primarily with the rise-to-span ratio.
With this in mind, close estimates of the critical horizontal reaction are provided
as a function of rise-to-span ratio.
Unsymmetrical Loading The general response behavior of unsymmetrically

loaded arches is shown in Fig. 17.2a and has been described previously. Several
studies have been made for the most important practical loading of parabolic arches
which comprise a uniformly distributed dead load q (on a horizontal projection)
combined with a uniformly distributed live load p extending a variable distance
from one abutment, as shown in Fig. 17.4.
The first studies by Deutsch (1940), Harries (1970), and Kuranishi and Lu (1972)
used half-span live load. Kuranishi and Lu noted that for elastic buckling the total
dead plus live load intensity, w = p + q, at buckling seemed roughly equal to the
buckling value for uniform load over the entire span. Studies have been made by
Chang (1973) and Harrison (1982) in which the length of the live load s was
varied. The results by Chang are given in Tables 17.3 and 17.4. Parabolic hinged
and fixed arches are considered with span-to-depth ratios of 0.15 and 0.25 and live
load–dead load ratios p/q of ∞, 0.50, 0.15, and 0, which should cover the range
of practical values. Because a live load–dead load ratio p/q = 0 corresponds to
uniform symmetrical load, the arch buckling is by bifurcation, and the values are
the same as those reported in Table 17.1.
FIGURE 17.4 Unsymmetrical loading.

TABLE 17.3 Minimum Elastic Buckling Coefficients for Uniform Hinged- and
Fixed-Ended Parabolic Arches Under Distributed Dead and Live Loads
Hinged Ends Fixed Ends
h/L p/q s/L wL3 /EI HL2 /EI s/L wL3 /EI HL2 /EI
0.15 ∞ 0.72 34.0 26.6 0.72 71.7 58.1

0.50 0.67 35.5 29.4 0.66 75.9 62.2
0.15 0.63 36.8 30.8 0.63 78.4 66.1
0 — 39.5 32.9 — 85.1 70.9
0.25 ∞ 0.77 44.7 22.5 0.73 99.3 49.5
0.50 0.68 46.0 23.3 0.66 104.0 51.9
0.15 0.64 47.0 23.9 0.64 107.5 54.4
0 — 49.2 24.6 — 114.6 57.3
TABLE 17.4 Elastic Buckling Coefficients for Uniform Hinged- and Fixed-Ended
Parabolic Arches Under Distributed Dead and Half-Span Live Load (s = 0.50L)
Hinged Ends Fixed Ends
h/L p/q wL3 /EI HL2 /EI wL3 /EI HL2 /EI
0.15 ∞ 42.4 22.4 96.2 46.7

0.50 37.1 28.4 80.2 60.4
0.15 37.3 30.6 80.0 65.2
0 39.5 32.9 85.1 70.9
0.25 ∞ 56.4 19.3 131.4 41.0
0.50 48.3 22.8 109.8 50.4
0.15 47.7 23.8 109.1 53.7
0 49.2 24.6 114.6 57.3
In Table 17.3, the minimum total load intensity w is provided along with the
corresponding horizontal reaction H and length of live load s. These results were
obtained with a numerical solution using 24 equal divisions in the arch rib, with the
load length being incremented one panel at a time. Although the values of w and H
are accurate, the loaded length s is only approximate. Table 17.4 gives the values of
w and H corresponding to the specific case of half-span live loading (s = 0.50L).
Several conclusions can be drawn for practical loading ranges. The minimum
total buckling load is obtained when the live load acts over from 63 to 68% of the
span. The buckling value of the total load s is minimized for high live load–dead
load ratios but does not drop below about 89% of the full-span buckling value
for the practical range, which is remarkably constant. Half-span loading is a good
approximation for a partial loading case, as the corresponding elastic buckling load
is not greater than about 6% more than the absolute minimum.
It should be noted that high live loads acting over partial spans produce
very large moments and deformations. As a result, the elastic limit load is
reached only after the arch is significantly distorted (Harrison, 1982; Yabuki and
772 ARCHES
Vinnakota, 1984). When inelastic behavior is considered, as will be discussed in

the next section, the limit loads for these cases are much lower than the elastic
buckling values reported above. One approach to greatly reducing the bending
moments in the arch rib is to include hangers that are inclined in two directions in
its plane; such a system is often referred to as a network arch (Tveit, 1966, 1999).
17.2.3 In-Plane Ultimate Load

Limit Analysis of Stocky Arches The limit analysis of arches was presented
by Onat and Prager (1953) and Galli and Franciosi (1955). Following their
investigations several researchers reported on the collapse load of arches (Stevens,
1957; Coronforth and Childs, 1967). In these early studies the collapse loads were
obtained by the upper and lower bound methods of plasticity, which assumes
rigid–plastic behavior. Based on a localized plastic hinge concept these methods
cannot consider the impact of longitudinal spreading of yield zones or the
second-order effects of deflections on moments. Hence, the studies are valid only
for structures with very stocky cross sections subjected to predominant flexural
moments. Stevens (1957) experimentally checked the validity of this plastic hinge
approach on models of circular, elliptic, and parabolic arches with two-hinged
and fixed support conditions. He found that a reasonably accurate estimate of the
collapse load can be obtained by limit analysis for arches under a concentrated
load. On the other hand, the experimental collapse load for an arch under half-span
loading is considerably less than the calculated value. More recent experimental
studies and comparisons with finite element analyses of the plastic collapse load
of stocky arches are reported by Jukes et al. (1983), Mitri and Hassani (1990),
and Khan et al. (1996).
Slender Arches in Pure Compression As indicated earlier, slender two-

hinged and fixed arches with loadings that produce pure axial compression buckle
in a sidesway mode with a point of contraflexure at the crown (Figs. 17.3a and b).
Because the behavior is very much like that of a column, it is common to express
the buckling strength of such arches in terms of the axial thrust at the quarter point
of the arch P. The elastic critical value can be expressed as
π 2 EI
Pc = (17.1)
(KS)2
where S is the length of the curved centroidal axis of the arch rib from support to
crown, E the Young’s modulus, and I the moment of inertia of the arch rib. The
effective-length factor K depends on the support fixity condition, the rise-to-span
ratio, and the shape of the arch profile. Values of K for arches are tabulated in the
third edition of this guide and have values close to the well-known corresponding
values for a column with a hinge at one end (crown) and a fixity condition at the
other end that corresponds to the arch support condition.
Komatsu and Shinke (1977) have made inelastic ultimate-strength studies for
two-hinged parabolic arches subjected to uniform load. Residual stress and initial
crookedness were considered, and rise-to-span ratios of 0.10, 0.15, and 0.20 were
used along with a wide range of other parameters. It was found that the ultimate
value of the thrust at the quarter point of the arch was accurately predicted with
the usual column curves adjusting for yield points, effective lengths, and so on.
Thus, the standard relationships employed for columns relating the inelastic ultimate
strength to the elastic critical load values can be used for arches.
Slender Arches under Symmetric and Asymmetric Load In this case, the
loading produces combined axial compressive force and bending moment on a gen-
eral cross section, which results in the arch behaving similar to a beam-column. It
has been shown by Shinke et al. (1975) that the most demanding practical loading
for low bridge arches of uniform cross section is the unsymmetrical loading dis-
cussed in the preceding section and shown in Fig. 17.4 with s = 0.50. The response
behavior is similar to that shown in Fig. 17.2a, in which the buckling failure is a
limit-point phenomenon. Arches of practical proportions develop extensive inelas-
tic action before failure. To accurately predict the load-carrying capacity of steel
arch structures, it is necessary to consider the effects of the spread of yielding zones
in the cross section and along the longitudinal axial direction of the rib, the effects
of initial residual stresses, and, of course, the amplification of the moments caused
by the displacements. The following studies considered uniform arches subjected
to uniform dead load over the entire span and half-span live load (as in Fig. 17.4,
except as noted).
Harries (1970) reported analytical results on the ultimate strength of two-hinged
parabolic steel arches in which the effects of prebuckling deformations and the
spread of yielding zones were considered. Elastic–plastic deformations were cal-
culated. Harries considered rectangular sections and circular tubes.
Kuranishi and Lu (1972) reported extensive parametric studies on the load-
carrying capacity of two-hinged parabolic steel arches having either rectangular
or idealized sandwich cross sections (approximating a wide-flange or box cross
section). Residual stresses and strain hardening were taken into account. They
found, as did Harries, that when the effects of partial yielding are considered,
the strength of an arch can be drastically reduced under unsymmetrical loads; the
buckling load is less when the live load acts over only one-half the span instead of
over the entire span, and the larger the live load–dead load ratio, the smaller the
buckling load becomes.
In a study that reevaluated the solutions of the previous paper, Kuranishi
(1973) examined the ratio of the ultimate load to the elastic limit load found by a
second-order analysis that uses a sandwich cross section to approximate I- and box-
shaped cross sections. For a practical range of live load–dead load ratios, he
showed that the ultimate load lies between that which produces initial yielding
and about 93% of this value. He concluded that one could design efficiently
by using second-order elastic analyses and keeping the maximum stress below
90% of the yield stress. The same could not be concluded for a solid rectangular
section because the ultimate load always exceeded the load at first yielding.
774 ARCHES
Shinke et al. (1975) studied analytically the behavior of two-hinged and fixed
parabolic arches of solid rectangular, pipe, and box cross sections. In this ini-
tial study residual stresses and strain hardening were not considered. Comparisons
were made of the ultimate strength of several cases of uniform live load symmet-
rically distributed about the centerline and with arches having an unsymmetrically
distributed load in the pattern of Fig. 17.4. For the same dead- and live-load intensi-
ties, it was found that unsymmetrically distributed load always governs. They also
investigated the effect of length of the live load s and found that the buckling load is
least when the length of live load s is roughly equal to one-half the span. With this
loading configuration, a parametric study was also made of the ultimate strength of
a box section. In a follow-up study, Shinke et al. (1977) considered the effects of
residual stresses and initial crookedness of the arch rib. Residual stress effects were
shown to be significant, while initial crookedness did not impact the results for the
controlling unsymmetrical loading cases. An extensive series of experiments on
two-hinged and fixed parabolic arches of solid rectangular cross section were also
reported in this study, with experimental and analytical results agreeing well.
Kuranishi and Yabuki (1979) and Yabuki (1981) have published numerical
studies on the load-carrying capacity of two-hinged parabolic steel arches with
thin-walled box cross sections. These works have been summarized by Yabuki and
Vinnakota (1984).
The strength of arches with stiffening girders has been studied by Shinke et al.
(1980), Kuranishi et al. (1980), and Yabuki (1981). Within these papers formulas
are proposed for estimating the ultimate loads.
Studies of the effects of the unsymmetrical distributed load of Fig. 17.4 supple-
mented by a single concentrated load at the quarter point of the span (a requirement
of Japanese highway specifications) have been made by Yabuki and are summarized
by Yabuki and Vinnakota (1984).
As indicated in the literature, the limit load depends on many factors, including
rise-to-span ratio, live load–dead load ratio, slenderness ratio, yield stress, type of
cross section (box section, wide-flange, rectangle, etc.), and residual-stress magni-
tudes and patterns. Only a brief review of the major points presented by Yabuki
and Vinnakota (1984) will be given below.
The effects of several of these parameters on the ultimate load-carrying capac-
ity of two-hinged, parabolic, uniform arches subject to uniform dead load plus
half-space live load are shown in Figs. 17.5 through 17.7. In these figures, the
ultimate or maximum load is expressed in terms of the ordinate wmax /wy , in which
wy is the magnitude of a uniformly distributed load (p = 0) that would cause the
arch to yield at the springings (supports) under the direct axial thrust produced.
Based on this definition, wy can be calculated from
2Aσy
wy = (17.2)
1
L 16 (L/h)2 + 1
FIGURE 17.5 Variation of load-carrying capacity as a function of slenderness ratio.
where σy is the yield stress of the material. In these diagrams, c = q/w (see
Fig. 17.4) and the slenderness ratio λ = Ls /r, in which Ls is the curved length of
the arch axis.
Figure 17.5 shows the variation of maximum strength wmax /wy of steel arch ribs
as a function of the slenderness ratio λ for different load ratios c. The load-carrying
capacity decreases dramatically with the slenderness ratio. The capacity also
decreases markedly as c decreases (i.e., as the live-load component increases)
because the live load causes large bending moments and large displacements.
The influence of rise-to-span ratio on the load-carrying capacity of arch ribs
with a slenderness ratio of 200 is shown in Fig. 17.6. It can be observed that the
ultimate load, expressed in the nondimensional form with wy given by Eq. 17.2,
is not significantly affected by the rise-to-span ratio of the arch rib, especially for
high values of c.
The reduction in the ultimate strength of arch ribs due to the effect of residual
stresses can be as much as 20%. The reduction becomes especially noticeable
when the arch is subjected to uniformly distributed load covering the entire span.
The maximum variation of the ultimate strength using various distribution patterns
for the residual stresses is within 10%. The reduction in ultimate strength becomes
more significant as the level of compressive residual stress increases, and especially
so when the compressive residual stress is greater than 0.4σy .
776 ARCHES
FIGURE 17.6 Variation of load-carrying capacity as a function of rise-to-span ratio.
The initial deformations of arch ribs, such as out-of-straightness, can result in a

reduction of their ultimate strength. This reduction is important when the arch rib
is subjected to uniformly distributed load covering the entire span. In the case of
unsymmetrical loading, which is the critical loading condition for arch rib struc-
tures, the deformations due to the loading become large so that the effects of the
initial deformations are not significant.
The nondimensional ultimate-load intensity wmax /wy decreases in proportion to
the square root of the yield-stress level of the material, as shown in Fig. 17.7.
The ultimate strength of stiffened two-hinged arch structures, with the arch and
stiffening girder behaving as an entirely integrated structure, is analogous to that
of two-hinged arches. The slenderness ratio of the stiffened arch structure can be
defined by
Ls
λT = (17.3)
(I + Ig )/A
where Ig is the moment of inertia of the stiffening girder. The ultimate strength of
a stiffened arch structure with a slenderness ratio λT is always somewhat greater
than that of a two-hinged arch using a slenderness ratio λ = λT . Judging from
the analytical results, however, the two may be considered equal for all practical
purposes. It is generally not required that attention be given to the local failure
of arch rib members (buckling between the columns connecting the arch and the
stiffening girder) for the unsymmetrical loading case, if a check is made for their
FIGURE 17.7 Influence of yield-stress level of material on the load-carrying capacity of

arches.
strength for constant, uniformly distributed load. The local failure strength of arch
rib members can be determined by the basic column strength curve when they have
straight members between columns. For curved members, however, it is advisable
to reduce this strength by 15%.
17.2.4 Design Criteria for In-Plane Stability

Arches of considerable span are normally not of uniform cross section but are com-
posed of segments with different cross-sectional properties. Each of the segments
must be designed for the applied loads that produce the maximum stresses in that
segment. Many arch structures are complex structural systems such as deck bridge
arches, in which the arch rib is connected to the roadway girder by rigidly attached
columns. Studies of the behavior of uniform free-standing arches may be of lim-
ited usefulness in these cases. A general design approach is to factor the loads,
use a second-order elastic analysis for the entire system, and keep the maximum
combined stresses for each segment below some reference stress. Kuranishi (1973)
recommends this procedure with the maximum stress less than 90% of the yield
stress, as stated previously.
Design procedures for arches based on the ultimate inelastic strength studies
reviewed in the previous section have been proposed by Kuranishi (1973), Komatsu
and Shinke (1977), and Kuranishi and Yabuki (1984). Kuranishi proposed an
interaction-type formula similar to beam-column formulas for two-hinged parabolic
arches under unsymmetrical loading. Komatsu and Shinke presented practical for-
mulas for the planar ultimate-load intensity of two-hinged and fixed parabolic steel
arches that are a function of the normal thrust calculated at a quarter point of the
arch rib by first-order elastic analysis. They also recommended that the ultimate-
load capacity of arch ribs with varying and/or hybrid cross section can be evalu-
ated accurately by using mean values of the cross-sectional area and/or yield-stress
level of the material, which would be calculated by averaging along the entire
axial length of the rib. Kuranishi and Yabuki also presented accurate practical
778 ARCHES
formulas for the in-plane ultimate strength of parabolic two-hinged steel arch
ribs and steel arch bridge structures with a stiffening girder. These formulas are
expressed in terms of bending moment and axial thrust (or stresses provided by
these cross-sectional forces) at a quarter point of the rib, with all calculations based
on a first-order elastic analysis. More recently, Sakimoto and Watanabe (1995) pro-
posed a design procedure based on nonlinear analysis that proportions each member
automatically so as to meet the ultimate-strength requirements of the complete
structural system. The procedure is based on the tangent-modulus method and
eigenvalue analysis.
Circular Arches Loaded in Bending and Compression Pi and Trahair

(1999) and Pi and Bradford (2004) studied the in-plane inelastic stability of hinged
and fixed circular arches with I-shape cross sections. These studies included initial
geometric imperfections and residual stresses. The load cases investigated included
uniform compression, uniformly distributed load over half the span or over the
full span, a concentrated load at the crown or at quarter span, and uniform bending
(Table 17.5). The subtended angles studied ranged from 10◦ for shallow to 180◦ for
deep arches. In Table 17.5, the slenderness of the arch is defined as λs = S 2 /(4rx R),
where S and R are the developed length and radius of the arch and rx the radius of
gyration of the cross section (Fig. 17.8). To check in-plane stability, the following
interaction equation was proposed for a lower bound prediction (the notations are
slightly changed from the cited papers to combine results for fixed and hinged
arches):
N∗ M∗
+ ≤φ (17.4)
αanx · Nacx αamx · Mp
where N* is the maximum axial compression based on first-order elastic analysis,

M* the maximum moment, αanx and αamx the axial compression and moment
modification factors (Table 17.5), φ a capacity reduction factor with a value of 0.9
suggested for design, Nacx the axial compression capacity for uniform compression,
and Mp the plastic section capacity about the major axis.
The axial compression capacity for uniform compression is defined as Nacx =
αacx Ny for hinged and Nacx = φαacx Ny ≤ Ny for fixed arches, with Ny equaling the
yield (squash) load of the cross section and αacx the arch slenderness reduction fac-
tor (see Eq. 17.5a). The maximum moment is defined as M ∗ = δb M , where M is the
maximum moment based on a first-order elastic analysis and δb = 1/(1 − N ∗ /NA )
the moment amplification factor.
The in-plane buckling load of an arch in uniform compression, NA, is defined
as NA = kax NEx , in which kax is a factor for in-plane buckling and NEx is
the elastic in-plane buckling load for uniform compression. For hinged arches
1
kax = 0.95 for 50 ≤ λs ; kax = 0.7 + 0.1(λs − 9.9) /4 for 10 < λs < 50; kax =
0.15 + 0.006λs for λs ≤ 10, and for fixed arches kax = 0.36 + 0.0011λ2s for
2
9.87 ≤ λs ≤ 18.6; kax = 0.6 + 0.4 1 − 3.109π 2 /λ2s for 18.6 < λs < 50; kax =
1 + 0.049(θ/π )2 for λs > 50. The elastic in-plane buckling load for uniform
TABLE 17.5 Factors for the In-Plane Design of Circular Arches
q L/2
Q L/4 Q
q q M
M
q q q q q q
α amx 1.0 1.0 1.0 1.1 1.1 —
λs < 18.6 — 2.0 + 1.2 + 1.5 + 1.8 + —

Fixed α amx 0.1(18.6 – λs )θ /π 0.5(18.6 – λs )θ /π 0.05(18.6 – λs )θ /π 0.05(18.6 – λs )θ /π
λs > 18.6 — 2.0 1.2 1.5 1.8 —
λs < 10 — 1.13 – 0.137λs 1 + 0.02λs 1.35 + 0.065λs 1.1 + 0.05λs 1.0

Hinged αamx
λs > 10 — 2.5 1.2 2.0 1.6 1.0
779
780 ARCHES
s
z
x R
h
y y
L
q
Hinged Fixed
FIGURE 17.8 Geometry of arches studied by Pi and Trahair (1999).
compression is defined as NEx = bπ 2 EIx /S 2 with b = 4 for hinged and b ≈ 8.16

for fixed arches. The remaining variables used are
⎡ ⎤
2
90
αacx = ξax ⎣1 − 1 − ⎦ (17.5a)
ξax · λagx
2
λagx /90 + 1 + ηax
ξax = 2 (17.5b)
2 λagx /90

c·S 1 σy
λagx = (17.5c)
rx kax 250

In Eq. 17.5b, ηax = 0.00326 λagx − 13.5 ≥ 0 is an imperfection parameter and
in Eq. 17.5c, c equals 0.5 for hinged and 0.35 for fixed arches, and σy is the yield
strength. Pi and Trahair (1999) and Pi and Bradford (2004) recommend that M* and
N* should be based on a second-order elastic analysis if the moment amplification
factor δb exceeds 1.4. In Fig. 17.9, the results from extensive finite-element analyses
are compared to the interaction equation 17.4 for φ = 1.0 (continuous line) and
φ = 0.9 (dashed line). The majority of the analysis results lie above the prediction
of the interaction equation, especially when φ = 0.9 is used.
Design Standards A comparison of design provisions for in-plane stability

between international standards can be found in Beedle (1991). Some of the more
recent provisions are listed below.
Eurocode 3, Part 2 (CEN, 2006a) provides charts with effective length factors
for the elastic in-plane buckling of circular, parabolic, and caternary arches with
unmovable supports and several articulations. For tied arches with vertical hangers,
effective lengths are also given, as is a criterion which indicates if the arch is prone
to snap-through buckling. Furthermore, the standard gives imperfections to be used
for the in-plane analyses.
AASHTO (2004) provides effective-length factors for fixed, two-hinged and
three-hinged arches with rise-to-span ratios of 0.1 to 0.4.
FIGURE 17.9 Interaction equation for in-plane stability compared to finite element anal-
ysis results (Pi and Trahair, 1999; Pi and Bradford, 2004).
782 ARCHES
FIGURE 17.9 (Continued )
17.3 OUT-OF-PLANE STABILITY OF ARCHES
When applied forces acting in the plane of a curved member such as an arch reach
a certain critical level, a combination of twisting and lateral bending will cause
the member to deform out of its original plane. The critical load is influenced
by the nature and distribution of the loads, the shape of the axis of the arch,
the variation of the flexural and torsional stiffness of the cross sections along
OUT-OF-PLANE STABILITY OF ARCHES 783
the length axis, and restraint available at the supports and other brace points.
For a steel arch, residual stresses either resulting from nonuniform cooling of the
material during fabrication or caused by welding or cold working (roller bending)
the material into an arch are also important factors. The multiplicity of impacting
parameters prevents the development of simple and widely applicable rules for the
determination of buckling loads.
Subjects of research have progressed with time from elastic linear buckling
of a single arch to ultimate strength of total systems—that is, from simple and
idealized theoretical models to more practical models to be encountered in actual
structures. Some of the available solutions that may be useful as a guide to the
design of practical structures are presented in formulas and tables. Results from
elastic linear buckling theory are shown in the next sections while solutions for
inelastic stability under general loading are presented in Section 17.3.3. Results
from ultimate-strength analyses are presented in Section 17.5.
Figure 17.8 shows the notation and coordinate system that will be used in the
following sections. Unless stated otherwise, the results presented apply to sym-
metric arches of constant cross section. The effect of warping torsional restraint is
included in all results presented. Note that warping torsional resistance is negligi-
ble for closed-profile cross sections such as box sections, but it is of significant
importance for open-profile cross sections such as I-shapes. In general, results
obtained by ignoring warping torsional restraint are conservative for open-profile
sections.
Isolated arches supported by ends that are free to rotate out of plane (y axis)
have very little lateral stability and should be avoided. The practical significance
of this should be emphasized, because in actual arches, completely rigid supports
are difficult, if not impossible, to construct.
17.3.1 Elastic Buckling of Circular Arches

Uniform Compression An arch is loaded in uniform compression when uni-
formly distributed radial forces p are directed inward along the arch’s centroidal
axis and the ends are prevented from translation. Although this load case rarely
exists in practice, it is useful as reference for design procedures.
For arches with rectangular cross sections and in- and out-of-plane hinged sup-
ports that prevent a twisting rotation about the length, or z , axis, Timoshenko
and Gere (1961) obtained a closed-form solution for elastic out-of-plane buckling.
Vlasov (1963) extended the solution to include warping. More refined solutions
were then obtained by Papangelis and Trahair (1987b), Rajasekaran and Padman-
abhan (1989), and Yang and Kuo (1986, 1987, 1991). The most recent solutions
for the elastic flexural–torsional buckling load of an arch in uniform compression
have been obtained by Pi and Trahair (1998) for hinged arches and Pi and Bradford
(2005) for fixed arches and are given by
Nays = ka NEy (17.6a)

784 ARCHES
where, for hinged arches,

1 a 2 NT 2
ka = − 1+ 2 − 1 − a2
2 b NEy
⎤
2
a 2 2 N2 a N
+ 1+ 2 + 1 − a2 ⎦
2 T 4
T
2
+2 − 1 1 − a2
b NEy b2 NEy
(17.6b)
and, for fixed arches,

1 a 2 NT 2
ka = 1+ 2 + 1 − a2
2 b NEy
⎤
2 2 2
a 2 N a 2 NT
− 1+ 2 + 1 − a2 ⎦
4
T
2
+ − 1 1 − a2
b NEy b2 NEy
(17.6c)
with NEy equaling the minor axis buckling load of a column of length S , NEy =
π 2 EIy /(S /c)2 where
c = 1 for hinged and c = 2 for fixed arches, a = (S /c)/(π R),
and b = π Myz / NEy (S /c) . The torsional buckling load NT of a column and the
lateral–torsional buckling moment Myz for a beam in uniform bending are given by

A π 2 EIw
NT = GJ + (17.7)
Ix + Iy (S /c)2

π 2 EIw
Myz = NEy GJ + (17.8)
(S /c)2
Uniform Moment An arch is loaded in uniform bending when equal couples are
applied about the x axis and the ends are fixed against translation in the x direction
but are free to rotate about the x and y axes. Although this load case rarely exists
in practice, it is useful as a reference load. Uniform bending can cause buckling
out of the y–z plane if the flexural stiffness about the x axis is large in comparison
to both the flexural stiffness about the y axis B(= EIy ) and the St.-Venant torsional
rigidity of the cross section about the z axis C (= GJ). Solutions for the critical
moment were obtained by Timoshenko and Gere (1961) and were extended by
Vlasov (1963) and Kollár and Iványi (1966) to include the warping effects. More
refined solutions were obtained by Papangelis and Trahair (1987b), Rajasekaran
and Padmanabhan (1989), and Yang and Kuo (1986, 1987, 1991). The most recent
solutions are given by
Mays = kas Myz (17.9a)

where, for hinged arches (Pi and Trahair 1998),

2
a a 3b a a 3b 2
kas = −ab − + + −ab − + + 1 − a2 (17.9b)
2b 2 2b 2
and, for fixed arches (Pi and Bradford 2005),

ab a ab a 2
kas = − − ± + + 1 − a2 (17.9c)
2 2b 2 2b
The parameters a and b are the same as those defined for Eq. 17.6.
End Forces Directed along the Chord Two collinear forces applied to the
ends of circular arches can cause lateral buckling (Klöppel and Protte, 1961; Ojalvo
et al., 1969). Forces directed away from each other (pull loads) will generally
cause antisymmetric buckling about the center of the arch, whereas forces directed
toward each other (push loads) cause a symmetric buckling. The buckling load for
ends hinged about the x axis may be expressed in the form (Klöppel and Protte,
1961)
B
Fc = m (17.10)
(Rα/2)2
When the central angle α is less than π and the warping torsional stiffness
is negligible, the value m (negative for pull loads) may be approximated by the
following equation: For pull loads
a1 a2
m = a0 + + 2 (17.11a)
α α
and for thrust loads
m = 2.47 + a3 α + a4 α 2 (17.11b)
Coefficients a0 through a4 are given in Table 17.6 for two C /B ratios. The
effect of the warping torsional stiffness was also studied by Klöppel and Protte
(1961).
TABLE 17.6 Coefficients for Use in Eq. 17.11

C/B a0 a1 a2 a3 a4
0.01 0.052 −0.217 −0.36 −0.24 0.09

0.5 0.815 −6.313 −7.411 −0.034 0.069
786 ARCHES
Uniformly Distributed Vertical Forces When the arch is fixed against rotation
about the y axis at its ends, the uniform load wc at buckling may be expressed in
the form (Demuts, 1969)
B
wc = m (17.12)
L3
with several values of m provided in Table 17.7 for circular arches.
The usual closed-profile cross section used in arches has a C/B ratio in the
range 0.5 to 1.5 and for the open-profile cross section C/B is in the range 0.01
to 0.001. If the arch cross section has a thin-walled open profile, the warping
torsional rigidity, C1 , also has a primary effect on the buckling loads, as shown in
Fig. 17.10 (Fukasawa, 1963; Namita, 1968). In this figure the buckling coefficient
m = wc α 2 R 3 /B for uniform vertical load w is shown by solid lines and the buckling
coefficient mr = pc α2 R 3 /B for uniform radial load p is shown by dashed lines.
When the height-to-span ratio h/L is not large, h/L ≤ 0.2, the values m and mr
are approximately related by mr = 2m sin (α/2)/α.
TABLE 17.7 Approximate Values of m for Use in Eq. 17.12

Circular Arch Parabolic Arch
Hanger Loads Column Loads
Vertical Loads Vertical Loads (Symmetric) (Antisymmetric)
C/B 0.01 0.5 0.01 0.5 0.01 0.5 0.01 0.5
H/L
0.1 18 28 16 28 39 70 12 18
0.2 17 39 15 39 35 110 12 29
0.3 13 38 13 37 28 116 10 32
0.4 9 30 11 30 24 104 8 31
0.5 5 20 9 24 20 87 6 28
FIGURE 17.10 Effects of the warping rigidity C 1 .

17.3.2 Buckling of Parabolic Arches

Uniformly Distributed Vertical Loads The uniformly distributed buckling
load wc for parabolic arches may also be expressed in the form of Eq. 17.12.
Representative values of m are given in Table 17.7 (Tokarz and Sandhu, 1972).
The ratio C /B = 0.01 is typical of thin-walled open-profile cross sections for which
warping torsional restraint is significant. The effect of the warping torsional rigidity
for these sections is similar to what is observed for circular arches (Fig. 17.10).
Additional information on this case is presented by Stüssi (1943), Tokarz (1971),
Kee (1961), and Papangelis and Trahair (1987a,b).
Parabolic Arches with Tilting Loads It has been recognized by Stüssi (1943)
and others that the buckling load of an arch is increased if the arch is loaded by
a system of vertical hangers connected to a laterally stiff girder at the elevation of
the chord, as in the case of the through-type arch bridge shown in Fig. 17.1. If the
loads are applied by means of columns connected to a laterally stiff girder above
the arch and the arch is connected directly to the girder at the crown or by very
short columns, the lateral deformations of the buckled arch will be antisymmetric
about the crown. Both hanger- and column-loaded parabolic arches were studied
theoretically by Östlund (1954), Godden (1954), and Shukla and Ojalvo (1971).
The buckling load wc may be expressed by Eq. 17.12 using appropriate values
of m from Table 17.7. Again, it is assumed that the ends of the arch are fixed
against rotation about the x axis. Östlund (1954) and Almeida (1970) reported that
certain braced arches subjected to hanger loads are also controlled by antisymmetric
buckling.
17.3.3 Design Criteria for Out-of-Plane Stability

Gravity Loading of Circular Arches Pi and Trahair (1998, 2000) and Pi and
Bradford (2005) studied the out-of-plane stability of freestanding circular arches of
I-sections including imperfections, residual stresses, and inelasticity. The support
conditions studied included in- and out-of-plane hinged arches, which are prevented
to twist about the longitudinal axis at the supports, and in- and out-of-plane fixed
arches. The load cases investigated were the same as those for the in-plane stability,
which is discussed in Section 17.2.4, and are included in Table 17.8. To check the
out-of-plane stability, the following interaction equation, similar to the one for
in-plane stability (Eq. 17.4), was proposed (to combine both fixed and hinged
arches, notations are slightly changed from the original papers):
N∗ M∗
+ ≤φ (17.13)
αany · Nacys αamy · Mamys
where N* and M* are the maximum axial compression moments based on a

first-order elastic analysis, αany and αamy axial compression and moment modi-
fication factors (Table 17.8), φ the capacity reduction factor with a value of 0.9
788 ARCHES
TABLE 17.8 Factors for the Out-of-Plane Design of Circular Arches
q L/2
Q L/4 Q
q q M
M
q q q q q q
α any — 1.2 1.6 1.4 2.2 —

Fixed
α amy — 2.2 2.2 1.5 1.8 —
α any 1.0 1.1 1.5 1.4 2.7 —

Hinged
α amy — 1.1 1.1 1.2 1.2 1.0
suggested for design, and Nacys and Mamys the out-of-plane strengths for uniform
compression and uniform bending.
The out-of-plane strength for uniform compression is defined as Nacys = αacys Ny
for hinged arches and Nacys = φαacys Ny ≤ Ny for fixed arches, where Ny is the axial
yield (squash) load of the cross section and αacys the arch slenderness reduction
factor for uniform compression (Eq. 17.14). Here, M* is the nominal maximum
moment, where for hinged arches M ∗ = M and for fixed arches M ∗ = δby M , in
which M is the maximum moment based on a first-order
elastic analysis and the
moment amplification is δby = 1/ 1 − N ∗ /Nays with Nays given by Eq. 17.6a.
Remaining variables used are
⎡ ⎤
2
90
αacys = ξays ⎣1 − 1− ⎦ (17.14a)
ξays · λagys
2
λagys /90 + 1 + ηays
ξays = 2 (17.14b)
2 λagys /90

c·S 1 σy
λagys = (17.14c)
ry kacys 250

In Eq. 17.14b, ηays = 0.00326 λagys − 13.5 ≥ 0 is an imperfection parameter,
and in Eq. 17.14c, c equals 1.0 for hinged arches and 0.5 for fixed arches, σy is
the yield strength, and ry is the radius of gyration about the minor axis.
The out-of-plane strength of an arch under uniform bending is defined as
Mamys = αsays Mpx for hinged arches and Mamys = φαsays Mpx ≤ Mpx for fixed
arches, with Mpx the full major axis plastic moment capacity and αsays = 0.6

λ4amys + 3 − λ2amys the arch slenderness reduction factor for uniform bending.

The modified slenderness for uniform bending is defined as λamys = Mpx /Mays
in which Mays is the elastic flexural–torsional buckling moment for an arch in
uniform bending (Eq. 17.9a). Pi and Trahair (1998, 2000) and Pi and Bradford
(2005) both recommend that M* and N* should be based on a second-order elastic
analysis when δby exceeds 1.4.
Figure 17.11 shows comparison of the interaction equation 17.13 and the
finite-element analysis results from Pi and Trahair (2000) and Pi and Bradford
(2005). In general, the interaction equation is accurate, noting that it seems to
be conservative for fixed arches. Finally, it should be noted that the in- and
out-of-plane load-bearing capacity for hinged arches vanishes when the subtended
angle approaches 180◦ because the y axes of the hinges line up. Naturally, this is
not the case for the fixed arches.
Circular Arches with Tilting Loads La Poutré (2005) studied both experi-
mentally and numerically the inelastic out-of-plane stability of deep circular steel
arches, with subtended angles ranging from 90◦ to 180◦ loaded by a single tilting
load at the crown. The arches were in-plane hinged and out-of-plane fixed. The
cross sections used were standard European wide-flange beams HEA 100 and HEB
600, with the latter tested to scale. The arches were produced by roller bending,
which cold worked the material and changed the stress–strain characteristic of the
steel and the residual-stress pattern.
To design for out-of-plane stability, the general method of Eurocode 3 (CEN,
2006b) was used. The plastic strength of the arch was determined by a material
nonlinear analysis (MNA) and the elastic critical load by an eigenvalue analysis,
also termed a linear buckling analysis (LBA). With these two parameters, the
nondimensional slenderness of the arch was determined by

FMNA
λGM = (17.15)
FLBA
The actual buckling load can then be determined by either an experimental

FEXP or a finite element analysis FGMNIA that include large displacements, material
nonlinearity, imperfections, and residual stresses. The nondimensional reduction
factor is determined by dividing the actual buckling load by the plastic strength
FEXP
χEXP = (17.16a)
FMNA
FGMNIA
χGM = (17.16b)
FMNA
Figures 17.12a and 17.12b show that the experimental reduction factors almost
all lie above the most favorable buckling curve “a” of Eurocode 3 and all of the
numerical results are above the same curve. Differences can be attributed to
the actual imperfections in the tests being larger than the imperfections given by
Eurocode 3, Part 2 (CEN, 2006a) for use with the numerical simulations. For
design purposes, however, buckling curve “b” seems to be more appropriate.
790 ARCHES
FIGURE 17.11 Interaction equation for out-of-plane stability compared to finite-element

analysis results (Pi and Trahair, 2000; Pi and Bradford, 2005).
FIGURE 17.11 (Continued )

792 ARCHES
+ xexp HEA 100

xGM Sim. of exp.
1 + 1 +
+ ++ + x Additional
xexp ++
+ × xexp HEB 600*
xGM
++++ Δ Δ Δ GM
sim.
×× ++ + Δ Δ
[.] [.]
× curve “a” Δ
curve “d” curve “a”

curve “d”
0 0
0 0.4 0.8 1.2 0 0.5 1 1.5
lGM [.] lGM [.]
(a) Experimental reduction factors (b) Numerical reduction factors
FIGURE 17.12 Design for out-of-plane stability by the general method.
The experiments and numerical simulations by la Poutré (2005) have shown that
the general method of Eurocode 3 is an accurate method for assessing the inelastic
stability of arches under general loading, a condition for which no closed-form
buckling equations exist.
Design Standards For the elastic out-of-plane buckling of free-standing

arches of circular, parabolic, or caternary shaped arches with unmovable supports
and different articulation, Eurocode 3, Part 2 (CEN, 2006a) provides charts with
effective-length factors. Furthermore, the standard defines imperfections that
should be used for out-of-plane buckling analyses. For through-arch bridges with
wind bracing, only the stability of the end portal needs to be checked. Beedle
(1991) provides a comparison of design methods for out-of-plane stability found
in several international standards.
17.4 BRACED ARCHES AND REQUIREMENTS

FOR BRACING SYSTEMS
17.4.1 Bracing by Transverse Bars

Most arches used in practice are braced against lateral movement either contin-
uously or at regularly spaced intervals. An arch rib that is always connected to
a curved roof is an example of continuous bracing. Unless such arch ribs are
unusually deep, lateral buckling will not be a problem.
Twin-arch ribs with intermittent lateral bracing often appears in arch bridges.
The elastic lateral buckling of twin-arch ribs braced with transverse bars nor-
mal to the plane of the ribs has been studied by several investigators, including
Östlund (1954), Kuranishi (1961a), Almeida (1970), Tokarz (1971), and Sakimoto
and Namita (1971). The following properties of the bracing system are important
in preventing lateral buckling from occurring: (1) the location and spacing of the
transverse bars; (2) the distance between the arch ribs b; (3) the flexural stiffness
BRACED ARCHES AND REQUIREMENTS FOR BRACING SYSTEMS 793
of the bars about the x axis, Dx ; and (4) the flexural stiffness of the bars about the
z axis, Dz .
It has been found that increasing the flexural stiffness Dx is a more effective
way of suppressing lateral buckling than increasing the flexural stiffness Dz . The
flexural stiffness Dz is less important except when the arch ribs have an open-profile
cross section and the stiffness Dx cannot be provided. When either of the flexural
stiffnesses Dx or Dz is increased independently, the buckling load of twin arches
increases and tends to attain an asymptotic value. While the magnitude of the
asymptotic value depends on the properties of the bracing system listed above,
the asymptotic value of the buckling load for a closely braced arch system can
be obtained when Dx ≥ 10b(B/R) or Dz ≥ b(B/R), respectively (Östlund, 1954;
Sakimoto and Namita, 1971).
When pairs of arch ribs are braced closely with such stiff transverse bars, the
elastic critical load for one arch rib of a set of braced arches can be increased up
to 250% or more of that for the identical isolated single arch. For actual bridge
arches, however, this increase in buckling strength will be limited by yielding of
the material (as demonstrated later in Fig. 17.16).
Wästlund (1960) and others have suggested a simple approximate method for
the determination of the lateral buckling load of braced arches by utilizing a planar
system that is obtained by straightening out the arch in a horizontal plane. The
compressive force required to buckle the longitudinal ribs of the assumed planar
system would be computed as for a column with batten lacing, and these compres-
sive forces would be the approximations to the rib forces required to buckle the
actual arch structure. The approximate method can be accurate for the arches of a
torsionally stiff (closed) cross section, but it is unconservative for the arches of an
open-profile cross section (Östlund, 1954).
The strength of braced arches in the plane of bracing is similar to the buckling
strength of columns with lacing bars and battens, which is given by Timoshenko
and Gere (1961) as
1 1 1
= + (17.17)
Pcr PE S
where Pcr is the critical load of the column, PE the critical load of the column
with stiff lacing, and S the shear stiffness provided by the bracing bars or battens.
While increasing the shear stiffness of the arch in the lateral plane increases the
critical load, the increase is limited by the flexural buckling strength in the verti-
cal plane. Therefore, there is a critical value for the shear stiffness above which
the buckling capacity is determined by the flexural strength. Based on extensive
parametric studies using finite-element analysis, Kuranishi (1993) presented the
following relationship for the critical value of the shear stiffness:
π2 1
μS ,cr = (17.18)
48 (Ki ρi /Ko ρo )2 − 1
794 ARCHES
where μS ,cr is the critical value for the ratio of the shear rigidity of the rib and
bracing system in the plane containing the two ribs to the flexural rigidity of the
two arch ribs, ρo the out-of-plane slenderness ratio of the rib, Ko the out-of-plane
effective length factor, ρi the in-plane slenderness ratio of an arch rib, and Ki the
in-plane effective length factor.
Although twin arches braced with transverse bars might be a favorable structure
from an aesthetic point of view, a more effective bracing system for suppressing lat-
eral buckling is to use diagonal members (Sakimoto, 1979; Sakimoto and Komatsu,
1982). If the lateral bracing consists of diagonals, diagonals in combination with
transverse bars, or a K system, out-of-plane buckling of the arch tends not to be
a problem except for through-type arch bridges, which cannot have lateral brac-
ing over the entire arch length. Through-type steel arches with a double-diagonal
bracing system will be discussed in the next section.
The stiffness of diagonal bracing members required to provide lateral stability
of an arch system was studied by Sakimoto and Komatsu (1977a,b) and Kuranishi
and Yabuki (1981), both focusing on the ultimate strength of braced steel arches
subjected to the combination of vertical and horizontal uniform loads. More recent
results on the effects of discrete elastic restraints on the out-of-plane buckling
loads of circular arches loaded in uniform compression or uniform bending are
given by Bradford and Pi (2002). Threshold values are given for the stiffness of
the bracing at which the out-of-plane buckling mode changes from symmetrical to
asymmetrical.
17.4.2 Ultimate Strength of Steel Arches Braced by Double-Diagonal

Bracing System
Since the arch is basically a compressive member, initial out-of-plane deflections
and residual stresses resulting from welding can have a significant influence on the
lateral stability of steel arches. Numerical results obtained for typical theoretical
models of square-box cross sections show that (1) the residual stress may cause
at most 20% reduction in strength of mild-steel arches and 10% for high-strength
steel arches and (2) the initial out-of-plane deflections may reduce the strength
of a perfectly plane arch by a maximum of 15% (Sakimoto and Komatsu, 1977b;
Komatsu and Sakimoto, 1977). These results are illustrated in Figs. 17.13 and 17.14,
with the vertical axis representing the ultimate unit strength σu normalized by the
yield stress σy and the horizontal axis defined by the slenderness ratio λ = ls /rx ,
in which ls is the total curved length of the arch rib and rx is the radius of
gyration of the rib section about the x axis. The term σrc denotes the maximum
compressive residual stress assumed in a trapezoidal distribution pattern and υ 0
denotes the amplitude of initial out-of-plane deflection that is assumed in the shape
of a half-sine wave.
As shown in Fig. 17.15 and described in the introduction to this chapter
(Fig. 17.1), the effect of load direction is also significant for steel arch bridges.
In these figures, elastic buckling curves of the identical arches are also shown
(Shukla and Ojalvo, 1971; Tokarz and Sandhu, 1972).
h
= 0.1
su/sy L
h ELASTIC
1.0 = 0.2
L
h src
1.9 = 0.1, = 0.4
L sy
h src
0.8 = 0.1, =0
L sy
h src
0.7 = 0.2, = 0.4
L sy
0.6 h s rc
= 0.2, =0
L sy
0.5
0.4 PARABOLIC ARCH

VERTICAL LOAD
0.3 v0 1
= ,
L 1000
0.2
E l 1 sy (l/2)
= 875, l = ,l =
sy rx p√ E rx
0.1
l
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
l
50 100 150 200 250 300 350
FIGURE 17.13 Effects of residual stresses.
Extensive parametric studies (Sakimoto, 1979; Sakimoto and Komatsu, 1982)

have resulted in simple approximate methods for determining the strength of braced
or unbraced steel arches that fail by lateral instability. For example, Sakimoto and
Komatsu (1983a) use an analogy between an arch and a column to obtain the
following equivalent slenderness parameter λa , which can be used to determine of
the ultimate strength of through-type steel arches of box cross sections:

1 σy l
λa = Ke Kl Kβ (17.19)
π E rx
where σy is the lowest yield stress of the different steel grades used in the arch
rib and Ke , Kl , and Kβ are effective-length factors. The coefficient Ke relates to
the rotational fixity of an arch rib at its ends with respect to the centroidal x axis.
For the clamped condition Ke is 0.5 and for the hinged condition it is 1.0. The
coefficient Kl corresponds to the direction of the loads and is taken as 0.65 for
the tilting hanger case and 1.0 for nontilting hangers (i.e., for vertical loads). The
coefficient Kβ is related to the lateral restraint supplied by the bracing system and
given by Kβ = 1 − β + (2rx β/Ke b), where the term β is the ratio of the length of
796 ARCHES
su/sy
ELASTIC
1.0 v0 1
=
L 5000
0.9
v0 1
0.8 =
L 1000
0.7 v0 1
=
L 500
0.6
0.5
0.4 PARABOLIC ARCH

VERTICAL LOAD
0.3 h src
= 0.1, s = 0.4,
L y
0.2
E
= 875, l =
l 1 sy (l/2)
sy ,l = p
rx √ E rx
0.1
l
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
l
50 100 150 200 250 300 350
FIGURE 17.14 Effects of initial lateral deflections.
the braced portion to the total length of the arch rib. Since β is equal to zero for
arches without bracing, Kβ equals 1.0 for an isolated single arch.
The ultimate unit strength σu is computed for the column with the equivalent
slenderness, and this ultimate unit strength would be an approximation of the ulti-
mate stress required to buckle the actual arch structure. The ultimate unit strength
σu for arches is defined as the tangential thrust at the support Nu divided by the
cross-section area A; that is, Nu = A · σu . The thrust Nu is determined from a lin-
ear analysis of the loaded arch. For a parabolic arch, for example, the uniformly
distributed load per unit length of the arch span at the ultimate state wu can be
computed from Eq. 17.2, where wy = wu and A is the cross-sectional area of an
individual arch rib. When the rib cross section varies, A and rx are weighted average
values for the entire curved length l .
Although the Japanese column curve was used in this investigation (Sakimoto
and Komatsu, 1983a), SSRC curve 2 (Eq. 3.8), curve C of the ECCS multiple col-
umn curves, or similar column curves are suggested as counterparts. Applicability
of this approximate method has been examined by use of computer simulations for
parabolic and circular steel arches composed of box section arch members with
h/L ratios of 0.1 and 0.2. The results of one of these comparisons are shown in
Fig. 17.16, noting that SSRC curve 2 was used in the approximate method. Prac-
tical applications for actual steel bridges are presented in Sakimoto and Komatsu
su/sy
1.0
ELASTIC
0.9
0.8
0.7
VERTICAL LOAD
0.6
HANGER LOAD
0.5
0.4
PARABOLIC ARCH
0.3
h src v 1
= 0.1, s = 0.4, 0 = ,
0.2 L y L 1000
E
= 875, l =
I 1 sy (l/2)
sy ,l =p
0.1 rx √ E rx
0 l
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
l
50 100 150 200 250 300 350
FIGURE 17.15 Effects of load directions.
su/sy
1.0
BRACED ARCH (b = 0.59, L/b = 10)
l
bl
h
b
L = 100 m
0.5 SINGLE ARCH
HANGER LOADED ARCH (h/L = 0.1)

PROPOSED FORMULA
ANALYSIS l/rx
0 100 200 300 400 500 600 700
FIGURE 17.16 Accuracy of the approximation method.
(1983b), Sakimoto and Yamao (1983), Sakimoto et al. (1989), and Sakimoto and
Sakata (1990).
When overall lateral buckling is suppressed, only buckling of rib segments
between bracing points need be considered. The slenderness limit for which local
instability of the arch segment may occur prior to overall lateral instability can
798 ARCHES
be determined approximately from Eq. 17.19 as follows (Sakimoto and Komatsu,

1983a):
√ √
σy lp σy l
≥ Ke Kl Kβ (17.20)
E rx max E rx
where (lp /rx )max denotes the slenderness ratio for the most slender arch segment
having an unbraced length lp and the term σy in the left side is the yield stress of
the arch segment material where (lp /rx )max occurs.
Nazmy (1997) studied long-span half-through and deck arch bridges with
K-bracing in vertical and inclined arches and found that there is a threshold for
the stiffness of the lateral bracing system at which the critical buckling load is
governed by the stiffness of the arch rib itself and not by the stiffness of the
bracing elements. Nazmy also showed that the critical buckling load of through-
arch bridges can be increased by changing the connections between the arch rib
and the deck from pinned to fully restrained (rigid).
17.5 ULTIMATE STRENGTH OF STEEL ARCH BRIDGES
17.5.1 Bridges Subjected to Uniform Vertical Loads

In the previous section,
Eq. 17.19
defines slenderness as a function of effective
length Le , with Le = Ke Kl Kβ l . The following procedure, suggested by Sakimoto
et al. (1992), is based on using spatial eigenvalue (three-dimensional buckling)
analyses, available in many finite element software packages, to determine effective
lengths and in turn slenderness. Assuming linearity of the prebuckling state (i.e.,
the relative internal force distribution remains constant when scaling the applied
load), the critical axial force of a member Ncr can be expressed as
π 2 EI
Ncr = αNo = (17.21)
L2e
where α is the smallest eigenvalue (buckling load factor), No the axial force of
the member corresponding to the reference applied vertical load Po , and Le the
effective length of the member. From Eq. 17.19, the slenderness can be defined as

1 σy Le Aσy L2e Aσy
λa = = 2
= (17.22)
π E rx π EI αNo
If the reference load is taken as the axial yield (squash) load of the arch rib,
No = Aσy , where A is the cross-sectional area, the slenderness becomes a function
of the smallest eigenvalue for the system:

1
λa = (17.23)
α
ULTIMATE STRENGTH OF STEEL ARCH BRIDGES 799
By further substituting this slenderness into a column strength formula, as dis-

cussed in Section 17.4.2, an estimate of the ultimate strength of the arch bridge
can be obtained.
To verify the validity of this procedure, the ultimate strengths of the arch bridge
models shown in Fig. 17.17, with properties described in Table 17.9, were computed
by an elastoplastic finite displacement analysis. The ultimate normal stress σNu can
be given as σNu = Nu /A, where Nu is the axial force at the springing (support)
calculated by linear theory for the ultimate load Pu .
h
A L h
A L
b (c) Deck type

(a) Through type b
arch bridge
arch bridge
h src = 0.4 sy
srt = 0.9 sy
A. L
Arch rib Floor girder
b (b) Half-through type (d ) Cross-sections and assumed

arch bridge residual stresses
FIGURE 17.17 Numerical models of arch bridges.
TABLE 17.9 Structural Properties of the Numerical Models

Through-Type Half-Through-Type Deck-Type
Arch Bridge Arch Bridge Arch Bridge
Span length, L (m) 150 150 100, 200, 300

Rise, h (m) 22.5 22.5 20, 40, 60
Rise-to-span ratio, h/l 0.15 0.25 0.20
Arch rib distance, a (m) 20, 10, 5 20, 10, 5 5.2
Bending stiffness ratio, I 0.1−3.0 0.1−3.0 0.1−1.0
Bracing ratio, β 0.48−0.86 0.64−0.88 1.0
Cross-sectional area, A (cm2 ) 448 448 298.8−903.1
800 ARCHES
sNu /sy ~ sNu /sy ~

I 0.1 1.0 3.0 I 0.1 1.0 3.0
b b
1.0 0.364 1.0 0.875
0.733 0.756
0.480 0.542
0.5 0.5
Column curve Column curve
la la
0 1.0 2.0 0 1.0 2.0
(a) Through type arch bridge (b) Half-through type arch bridge
~
I 0.1 0.3 0.5 1.0

sNu /sy 100
200
1.0 300
0.5
~
Column curve I = the ratio of the lateral flexural
stiffness of the floor system
la to that of the twin arch ribs
0 1.0 2.0 as a whole
(c) Deck type arch bridge
FIGURE 17.18 Validity of the slenderness procedure using the eigenvalue.
In Fig. 17.18, the σNu values obtained are plotted as a function of the slenderness
λa ,which is given by Eq. 17.22 for the computational models. The close agreement
between the calculated values and the column strength curves demonstrates the
validity of this procedure.
17.5.2 Bridges Subjected to Vertical and Lateral Uniform Loads

The spatial elastic–plastic behavior and the ultimate-load capacity of the
through-type braced arch bridges of box cross sections was studied for a
combination of the vertical and horizontal uniform loads by Sakimoto (Sakimoto
and Komatsu, 1977a, 1979; Sakimoto et al., 1979). Based on computer analyses
for various theoretical models, a simple approximate method for determining the
ULTIMATE STRENGTH OF STEEL ARCH BRIDGES 801
ultimate lateral strength of bridge arches braced partly over the central portion
was developed. By utilizing an analogy between a laterally loaded arch and a
beam-column, the following interaction formula is proposed
N M
+ ≤ 1.0 (17.24)
Nu My [1 − (N /Ne )]
Although Eq. 17.23 has a form similar to an interaction equation for a

beam-column, attention should be placed on the precise definition of each
of the terms, which include N being the tangential end thrust computed by
linear theory for the uniformly distributed design loads; Nu the tangential end
thrust corresponding to the uniformly distributed vertical loading that produces
inelastic lateral buckling of the arch (which can be determined from Nu = Aσu as
demonstrated in the previous section); M the lateral end moment of the individual
arch rib computed by linear theory for the uniformly distributed horizontal
loads [this can be approximated by the value computed for the planar system
which is obtained by straightening out the arch in a horizontal plane; a simple
approximation formula for the determination of this moment is also given by
Sakimoto and Komatsu (1979)]; My the yield moment of the arch rib at the ends
with respect to the x axis; and Ne the Euler buckling load computed for a hinged
column of which the length and cross section are identical to those of the arch rib
for the unbraced segment.
The validity of this interaction equation has been verified by extensive computer
simulations for various braced bridge arches of practical proportions. If the twin
ribs are braced closely over the entire length of the arch, the moment ratio term
of Eq. 17.23 becomes negligible, which indicates that typical lateral loads are not
significant for such braced arches.
Deflections and stresses due to forces normal to the plane of the arch are
increased by the second-order effects associated with the combination of large axial
loads and lateral deflections. Stüssi (1943), Östlund (1954), and Wästlund (1960)
suggest that such increases may be estimated by multiplying the first-order deflec-
tions and stresses due to transverse loads by the amplification factor 1/(1 − w /wc ),
in which w is the intensity of the uniform load in the plane of the arch and wc is the
intensity of the in-plane load that would cause lateral buckling. While this ampli-
fication factor is at best only a rough approximation of the second-order effects
(Layrangues, 1959; Donald and Godden, 1961a,b), it is convenient for use in a
preliminary design.
The effect of lateral horizontal force on the in-plane strength of arch bridges
was studied by Kuranishi (1961b), Yabuki and Kuranishi (1973), and Kuranishi
and Yabuki (1977). They determined that the in-plane strength of arch bridges with
a close brace over the entire length of the arch is not significantly affected by
lateral loads typically encountered in actual structures. Hence, it was concluded
that through-type steel arch bridges can be designed with a lateral bracing system
(between arch ribs) of sufficient out-of-plane stiffness so that the arch rib design will
be determined primarily by the in-plane loads. With this in mind, the effect of lateral
802 ARCHES
loads can be taken into account for practical purposes as a set of additional in-plane
vertical loads. Thus, the arch rib design can be made on the basis of a quasi-planar
model subjected to the principal vertical loads slightly modified to include the
effects of wind loads. Kuranishi and Yabuki (1981, 1984) and Yabuki et al. (1983)
present the lateral bracing stiffness required to ensure that two-hinged arch bridges
behave basically as in-plane structures. Ju (2003) studied parabolic steel arch deck
and through bridges with vertical and inclined ribs braced by transverse bars. Using
a statistical analysis of a large set of computational results, he derived effective
lengths for these bridges and found that the effective lengths used by AASHTO
(2004) are conservative.
REFERENCES
AASHTO (2004). LRFD Bridge Design Specifications, American Association of State High-
way and Transportation Officials, Washington, DC.
Almeida, P. N. (1970), “Lateral Buckling of Twin Arch Ribs with Transverse Bars,” disser-
tation, Ohio State University, Columbus, OH.
Austin, W. J. (1971), “In-Plane Bending and Buckling of Arches,” ASCE J. Struct. Div.,
Vol. 97, No. ST5, pp. 1575–1592.
Austin, W. J., and Ross, T. J. (1976), “Elastic Buckling of Arches Under Symmetrical
Loading,” ASCE J. Struct. Div., Vol. 102, No. ST5, pp. 1085–1095.
Beedle, L. S. (Ed.) (1991), Stability of Metal Structures, a World View , 2nd ed., Structural
Stability Research Council, Bethlehem, PA.
Bradford, M. A., and Pi, Y.-L. (2002), ”Elastic Flexural-Torsional Buckling of Discretely
Restrained Arches,” ASCE J. Struct. Eng., Vol. 128, No. 6, pp. 719–727.
CEN (2006a) Eurocode 3: Design of Steel Structures, Part 2: Steel Bridges, EN 1993-2,
Comité Européen de Normalisation (CEN), European Committee for Standardization
Brussels, Belgium.
CEN (2006b), Eurocode 3: Design of Steel Structures, Part 1-1: General Rules and Rules
for Buildings, EN 1993-1-1, Comité Européen de Normalisation (CEN), European Com-
mittee for Standardization Brussels, Belgium.
Chang, C. K. (1973), “Effect of Loaded Length on the Buckling Strength of Slender Arches,”
thesis, Rice University, Houston, TX.
Coronforth, R. C., and Childs, S. B. (1967), “Computer Analysis of Two Hinged Circular
Arches,” ASCE J. Struct. Div., Vol. 93, No. ST2, pp. 319–338.
DaDeppo, D. A., and Schmidt, R. (1970). ”Stability of an Arch Under Combined Loads;
Bibliography on Stability of Arches,” Ind. Math., Vol. 20, Part 2, pp. 71–89.
Demuts, E. (1969), “Lateral Buckling of Circular Arches Subjected to Uniform Gravity Type
Loading,” dissertation, Ohio State University, Columbus, OH.
Deutsch, E. (1940), “Das Knicken von Bogenträgern bei unsymmetrischer Belastung,”
Bauingenieur, Dec., pp. 353–360.
Donald, P. T. A., and Godden, W. G. (1961a), “A Numerical Solution to the Curved Beam
Problem,” Struct. Eng., Vol. 41, pp. 179–186.
Donald, P. T. A., and Godden, W. G. (1961b), “The Transverse Behavior of Laterally
Unsupported Parabolic Arches,” Struct. Eng., Vol. 41, pp. 187–191.
REFERENCES 803
Forrester, K. (1972), Discussion of “In-Plane Bending and Buckling of Arches” by W. J.

Austin, ASCE J. Struct. Div ., Vol. 98, No. ST1, pp. 378–379.
Fukasawa, Y. (1963), “Buckling of Circular Arches by Lateral Flexure and Torsion Under
Axial Thrust,” Trans. Jpn. Soc. Civ. Eng., No. 96, pp. 29–47 (in Japanese).
Fukumoto, Y. (1996), Structural Stability Design, Steel and Composite Structures, Pergamon,
Amsterdam.
Galli, A., and Franciosi, G. (1955), “Limit Analysis of Thin Arch Bridges with Stiffening
Girders,” G. Genio Civ., Vol. 93, No. 11.
Godden, W. G. (1954), “The Lateral Buckling of Tied Arches,” Proc. Inst. Civ. Eng., Vol.
3, Part III, pp. 496–514.
Harries, H. (1970), “Traglast stahlerner Zweigelenkbogen mit ausgebreiteten Fliesszonen,”
Stahlbau, Vol. 6, pp. 170–177; Vol. 8, pp. 248–257.
Harrison, H. B. (1982), “In-Plane Stability of Parabolic Arches,” ASCE J. Struct. Div., Vol.
108, No. ST1, pp. 195–205.
Hayashi, T. (Ed.) (1971), Handbook of Structural Stability, Corona Publishing, Tokyo.
Ju, S.-H. (2003), ”Statistical Analyses of Effective Lengths in Steel Arch Bridges,” Comput.
Struct., Vol. 81, pp. 1487–1497.
Jukes, S. G., Hassani, F. P., and Whittaker, B. N. (1983), ”Characteristics of Steel Arch
Support Systems for the Mine Roadway. Part I, Modelling Theory, Instrumentation and
Preliminary Results,” Mining Sci. Technol., Vol. 1, pp. 43–58.
Kee, C. F. (1961), “Lateral Inelastic Buckling of Tied Arches,” ASCE J. Struct. Div., Vol.
87, No. ST1, pp. 23–39.
Khan, U. H., Mitri, H. S., and Jones, D. (1996), ”Full Scale Testing of Steel Arch Tunnel
Supports,” Int. J. Rock Mech. Min., Vol. 33, No. 3, pp. 219–232.
King, C., and Brown, D. (2001), Design of Curved Steel , Steel Construction Institute, Ascot
Berks UK.
Klöppel, K., and Protte, W. (1961), “A Contribution to the Buckling Problem of Circular
Curved Bars,” Stahlbau, No. 30, pp. 1–15 (in German).
Kollár, L. U., and Iványi, G. (1966), “Buckling of Shell Arches by the Energy Method,”
Bautech. Arch., pp. 1–23 (in German).
Kollbrunner, C. F. (1936), “Versuche über die Knicksicherheit und die Grundschwingungzahl
vollwandiger Bogen,” Bautechnik , Mar., pp. 186–188.
Kollbrunner, C. F. (1942), “Versuche über die Knicksicherheit und die Grundschwingungzahl
vollwandiger Dreigelenkbogen, Schweiz. Bauzg., Vol. 120, No. 10, pp. 113–115.
Komatsu, S., and Sakimoto, T. (1977), “Ultimate Load Carrying Capacity of Steel Arches,”
ASCE J. Struct. Div., Vol. 103, No. ST12, pp. 2323–2336.
Komatsu, S., and Shinke, T. (1977), “Practical Formulas for In-Plane Load Carrying Capacity
of Arches,” Proc. Jpn. Soc. Civ. Eng., No. 267, pp. 39–51 (in Japanese).
Kuranishi, S. (1961a), “The Torsional Buckling Strength of Solid Rib Arch Bridge,” Trans.
Jpn. Soc. Civ. Eng., No. 75, pp. 59–67 (in Japanese).
Kuranishi, S. (1961b), “Analysis of Arch Bridge Under Certain Lateral Forces,” Trans. Jpn.
Soc. Civ. Eng., No. 73, pp. 1–6 (in Japanese).
Kuranishi, S. (1973), “Allowable Stress for Two-Hinged Steel Arch,” Proc. Jpn. Soc. Civ.
Eng., No. 213, pp. 71–75.
804 ARCHES
Kuranishi, S. (1993), “Stiffening Effect of Lateral Bracing of Steel Arch Bridges on Their
In-Plane Strength,” Proc. SSRC Ann. Techn. Session, Milwaukee, WI.
Kuranishi, S., and Lu, L. W. (1972), “Load Carrying Capacity of Two-Hinged Steel Arches,”
Proc. Jpn. Soc. Civ. Eng., No. 204, pp. 129–140 (in English).
Kuranishi, S., Sato, T., and Otsuki, M. (1980), “Load Carrying Capacity of Two-Hinged Steel
Arch Bridges with Stiffening Deck,” Proc. Jpn. Soc. Civ. Eng., No. 300, pp. 121–130.
Kuranishi, S., and Yabuki, T. (1977), “In-Plane Strength of Arch Bridges Subjected to
Vertical and Lateral Loads,” paper presented at the 2nd Int. Colloq. Stabl. Steel Struct.,
Prelim. Rep., Liege, Belgium, Apr., pp. 551–556.
Kuranishi, S., and Yabuki, T. (1979), “Some Numerical Estimations of Ultimate In-Plane
Strength of Two-Hinged Steel Arches,” Proc. Jpn. Soc. Civ. Eng., No. 287, pp. 155–158.
Kuranishi, S., and Yabuki, T. (1981), “Required Out-of-Plane Rigidities of Steel Arch
Bridges with Two Main Ribs Subjected to Vertical and Lateral Loads,” Techn. Rep.
Tohoku Univ., Sendai, Japan, Vol. 46, No. 1.
Kuranishi, S., and Yabuki, T. (1984), “Lateral Load Effect on Arch Bridge Design,” ASCE
J. Struct. Eng., Vol. 110, No. 9, pp. 2263–2274.
la Poutré, D. B. (2005), ”Inelastic Spatial Stability of Circular Steel Wide Flange Arches,”
Ph.D. thesis, Technische Universiteit Eindhoven, Eindhoven, The Netherlands.
Layrangues, P. (1959), “Elastic Deformations and Forces in a Fixed End Symmetrical Cir-
cular Arch Subjected to a Constant Normal Force and a Uniformly Distributed Lateral
Force,” Ann. Ponts Chaussees, Vol. 129, pp. 323–343 (in French).
Mitri, H. S., and Hassani, F. P. (1990), ”Structural Characteristics of Coal Mine Steel Arch
Supports,” Int. J. Rock Mech. Min., Vol. 27, No. 2, pp. 121–127.
Namita, Y. (1968), “Second Order Theory of Curved Bars and Its Use in the Buckling
Problem of Arches,” Trans. Jpn. Soc. Civ. Eng., No. 155, pp. 32–41 (in German).
Nazmy, A. S. (1997), ”Stability and Load-Carrying Capacity of Three-Dimensional
Long-Span Steel Arch Bridges,” Comput. Struct., Vol. 65, No. 6, pp. 857–868.
Ojalvo, M., Demuts, E., and Tokarz, F. J. (1969), “Out-of-Plane Buckling of Curved Mem-
bers,” ASCE J. Struct. Div., Vol. 95, No. ST10, pp. 2305–2316.
Onat, E. T., and Prager, W. (1953), “Limit Analysis of Arches,” J. Mech. Phys. Solids, Vol.
1, pp. 71–89.
Östlund, L. (1954), “Lateral Stability of Bridge Arches Braced with Transverse Bars,” Trans.
R. Inst. Technol., Stockholm, No. 84.
Papangelis, J. P., and Trahair, N. S. (1987a), “Flexural–Torsional Buckling Tests on Arches,”
ASCE J. Struct. Eng., Vol. 113, No. ST7, pp. 1433–1443.
Papangelis, J. P., and Trahair, N. S. (1987b), “Flexural–Torsional Buckling of Arches,”
ASCE J. Struct. Eng., Vol. 113, No. ST4, pp. 889–906.
Petersen, C. (1982), Statik und Stabilität der Baukonstruktionen, 2nd ed. Vieweg & Sohn,
Braunschweig (in German).
Pi, Y.-L., and Bradford, M. A. (2004), ”In-Plane Strength and Design of Fixed Steel I-Section
Arches,” Eng. Struct., Vol. 26 . No. 3, pp. 291–301.
Pi, Y.-L., and Bradford, M. A. (2005), ”Out-of-Plane Strength Design of Fixed Steel
I-Section Arches,” ASCE J. Struct. Eng., Vol. 131, No. 4, pp. 560–568.
Pi, Y.-L., and Trahair, N. S. (1998), ”Out-of-Plane Inelastic Buckling and Strength of Steel
Arches,” ASCE J. Struct. Eng., Vol. 124, No. 2, pp. 174–183.
REFERENCES 805
Pi, Y.-L., and Trahair, N. S. (1999), ”In-Plane Buckling and Design of Steel Arches,” ASCE
J. Struct. Eng., Vol. 125, No. 11, pp. 1291–1298.
Pi, Y.-L., and Trahair, N. S. (2000), ”Inelastic Lateral Buckling Strength and Design of Steel
Arches,” Eng. Struct., Vol. 22, pp. 993–1005.
Rajasekaran, S., and Padmanabhan, S. (1989), ”Equations of Curved Beams,” ASCE J. Eng.
Mech., Vol. 115, No. 5, pp. 1094–1110.
Sakimoto, T. (1979), “Elasto-Plastic Finite Displacement Analysis of Three Dimensional
Structures and Its Application to Design of Steel Arch Bridges,” dissertation, Osaka
University, Osaka, Japan.
Sakimoto, T., and Komatsu, S. (1977a), “A Possibility of Total Breakdown of Bridge Arches
Due to Buckling of Lateral Bracings,” paper presented at the 2nd Int. Colloq. Stab. Steel
Struct., Final Rep., Liege, Belgium, Apr., pp. 299–301.
Sakimoto, T., and Komatsu, S. (1977b), “Ultimate Load Carrying Capacity of Steel Arches
with Initial Imperfections,” paper presented at the 2nd Int. Colloq. Stabil. Steel Struct.,
Prelim. Rep., Liege, Belgium, Apr., pp. 545–550.
Sakimoto, T., and Komatsu, S. (1979), “Ultimate Strength of Steel Arches Under Lateral
Loads,” Proc. Jpn. Soc. Civ. Eng., No. 292, pp. 83–94.
Sakimoto, T., and Komatsu, S. (1982), “Ultimate Strength of Arches with Bracing Systems,”
Sakimoto, T., and Komatsu, S. (1983a), “Ultimate Strength Formula for Steel Arches,” ASCE
J. Struct. Div., Vol. 109, No. 3, pp. 613–627.
Sakimoto, T., and Komatsu, S. (1983b), “Ultimate Strength Formula for Central-Arch-Girder
Bridges,” Proc. Jpn. Soc. Civ. Eng., No. 33, pp. 183–186.
Sakimoto, T., and Namita, Y. (1971), “Out-of-Plane Buckling of Solid Rib Arches Braced
with Transverse Bars,” Proc. Jpn. Soc. Civ. Eng., No. 191, pp. 109–116.
Sakimoto, T., and Sakata, T. (1990), “The Out-of-Plane Buckling Strength of Through-Type
Arch Bridges,” J. Constr. Steel Res., Vol. 16, pp. 307–318.
Sakimoto, T., Sakata, T., and Tsuruta, E. (1989), “Elasto-Plastic Out-of-Plane Buckling
Strength of Through Type and Half-Through Type Arch Bridges,” Struct. Eng., Earth-
quake Eng., Vol. 6, No. 2, pp. 307s–318s (Proc. Jpn. Soc. Civ. Eng. No. 410/I-12 ).
Sakimoto, T., Sakata, T., and Kobori, T. (1992), “An Effective Length Procedure for
Out-of-Plane Buckling Strength Estimation of Steel Arch Bridges,” Proc. 3rd Pac. Struct.
Steel Conf., Tokyo, pp. 251–258.
Sakimoto, T., and Watanabe, H. (1995), “A New Procedure for Frame Design,” Proc. SSRC
Annual Tech. Session, Pittsburgh, PA, pp. 23–31.
Sakimoto, T., and Yamao, T. (1983), “Ultimate Strength of Deck-Type Steel Arch Bridges,”
paper presented at the 3rd Int. Colloq. Stab. Met. Struct., Prelim. Rep., Paris, Nov.
Sakimoto, T., Yamao, T., and Komatsu, S. (1979), “Experimental Study on the Ultimate
Strength of Steel Arches,” Proc. Jpn. Soc. Civ. Eng., No. 286, pp. 139–149.
Shinke, T., Zui, H., and Nakagawa, T. (1980), “In-Plane Load Carrying Capacity of
Two-Hinged Arches with Stiffening Girder,” Trans. Jpn. Soc. Civ. Eng., No. 301, pp.
47–59.
Shinke, T., Zui, H., and Namita, Y. (1975), “Analysis of In-Plane Elasto-Plastic Buckling
and Load Carrying Capacity of Arches,” Proc. Jpn. Soc. Civ. Eng., No. 244, pp. 57–69
(in Japanese).
806 ARCHES
Shinke, T., Zui, H., and Namita, Y. (1977), “Analysis of Experiment on In-Plane Load Car-
rying Capacity of Arches,” Proc. Jpn. Soc. Civ. Eng., No. 244, pp. 11–23 (in Japanese).
Shukla, S. N., and Ojalvo, M. (1971), “Lateral Buckling of Parabolic Arches with Tilting
Loads,” ASCE J. Struct. Div., Vol. 97, No. ST6, pp. 1763–1733.
Singer, J., Arbocz, J., and Weller, T. (1998), Buckling Experiments: Experimental Methods
in Buckling of Thin-Walled Structures, J Wiley, New York.
Stevens, L. K. (1957), “Carrying Capacity of Mild Steel Arches,” Proc. Inst. Civ. Eng., Vol.
6, pp. 493–514.
Stüssi, F. (1943), “Lateral-Buckling and Vibration of Arches,” Proc. Int. Assoc. Bridge Struct.
Eng. Publ., Vol. 7, pp. 327–343 (in German).
Timoshenko, S. P., and Gere, J. M. (1961), Theory of Elastic Stability, 2nd ed., McGraw-Hill,
New York.
Tokarz, F. J. (1971), “Experimental Study of Lateral Buckling of Arches,” ASCE J. Struct.
Div., Vol. 97, No. ST2, pp. 545–559.
Tokarz, F. J., and Sandhu, R. S. (1972), “Lateral–Tosional Buckling of Parabolic Arches,”
Tveit, P. (1966), ”Desing of Network Arches,” Struct. Eng., Vol. 44, No. 7, pp. 247–259.
Tveit, P. (1999), ”Comparison of Steel Weights in Narrow Arch Bridges with Medium
Spans,” Stahlbau, Vol. 68, No. 9, pp. 753–757.
Vlasov, V. Z. (1963), Thin Walled Elastic Beams, 2nd ed., Israel Program for Scientific
Translations, Jerusalem.
Wästlund, G. (1960), “Stability Problems of Compressed Steel Members and Arch Bridges,”
Wolde-Tinsaie, A. M., and Foadian, H. (1989), ”Asymmetrical Buckling of Prestressed
Tapered Arches,” ASCE J. Eng. Mech., Vol. 115, No. 9, pp 2020–2034.
Yabuki, T. (1981), “Study on Ultimate Strength Design of Steel Arch Bridge Structures,”
dissertation, Tokoku University, Sendai, Japan (in Japanese).
Yabuki, T., and Kuranishi, S. (1973), “Out-of-Plane Behavior of Circular Arches Under Side
Loadings,” Proc. Jpn. Soc. Civ. Eng., No. 214, pp. 71–82.
Yabuki, T., and Vinnakota, S. (1984), “Stability of Steel Arch-Bridges, A State-of-the-Art
Report,” Solid Mech. Arch., Vol. 9, No. 2, Martinus Nijhoff Publishers, The Hague, The
Netherlands.
Yabuki, T., Vinnakota, S., and Kuranishi, S. (1983), “Lateral Load Effect on Load Carrying
Capacity of Steel Arch Bridge Structures,” ASCE J. Struct. Div., Vol. 109, No. 10, pp.
2434–2449.
Yang, Y. B., and Kuo, S. R. (1986), ”Static Stability of Curved Thin-Walled Beams,” ASCE
J. Eng. Mech, Vol. 112, No. 8, pp. 821–841.
Yang, Y. B., and Kuo, S. R. (1987), ”Effect of Curvature on Stability of Curved Beams,”
ASCE J. Struct. Eng., Vol. 113. No. 6, pp. 1185–1202.
Yang, Y. B., and Kuo, S. R. (1991), “Use of Straight-Beam Approach to Study Buckling of
Curved Beams,” ASCE J. Struct. Eng., Vol. 117, No. 7, pp. 1963–1978.

Ch17 Arches

Uploaded by

Copyright:

Available Formats

Ch17 Arches

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ch17 Arches

Uploaded by

Copyright:

Available Formats

CHAPTER 17

FIGURE 17.1 Gravity and tilting loads in arches.

to as conservative and nonconservative loading, respectively. In a through-arch

17.2 IN-PLANE STABILITY OF ARCHES

• Slender arches, generally of solid-web rolled or built-up sections, subject pri-

17.2.1 Linear Stability

TABLE 17.1 Critical-Load Parameter and Critical Horizontal Reaction Parameter

h Three-Hinged Arch Two-Hinged Arch Fixed Arch

Parabolic arches subjected to vertical load uniformly distributed on a horizontal projection

FIGURE 17.3 Fundamental buckling mode shapes.

17.2.2 Nonlinear Elastic Stability

Symmetrical Loadings on Symmetrical Arches The general behavior of

Parabolic arches—antisymmetrical modes

Circular arches—antisymmetrical modes

Circular arches—symmetrical modes

Unsymmetrical Loading The general response behavior of unsymmetrically

FIGURE 17.4 Unsymmetrical loading.

0.15 ∞ 0.72 34.0 26.6 0.72 71.7 58.1

0.15 ∞ 42.4 22.4 96.2 46.7

Vinnakota, 1984). When inelastic behavior is considered, as will be discussed in

17.2.3 In-Plane Ultimate Load

Slender Arches in Pure Compression As indicated earlier, slender two-

FIGURE 17.5 Variation of load-carrying capacity as a function of slenderness ratio.

FIGURE 17.6 Variation of load-carrying capacity as a function of rise-to-span ratio.

The initial deformations of arch ribs, such as out-of-straightness, can result in a

FIGURE 17.7 Influence of yield-stress level of material on the load-carrying capacity of

17.2.4 Design Criteria for In-Plane Stability

Circular Arches Loaded in Bending and Compression Pi and Trahair

where N* is the maximum axial compression based on first-order elastic analysis,

α amx 1.0 1.0 1.0 1.1 1.1 —

λs < 18.6 — 2.0 + 1.2 + 1.5 + 1.8 + —

λs < 10 — 1.13 – 0.137λs 1 + 0.02λs 1.35 + 0.065λs 1.1 + 0.05λs 1.0

FIGURE 17.8 Geometry of arches studied by Pi and Trahair (1999).

compression is defined as NEx = bπ 2 EIx /S 2 with b = 4 for hinged and b ≈ 8.16

Design Standards A comparison of design provisions for in-plane stability

FIGURE 17.9 (Continued )

17.3 OUT-OF-PLANE STABILITY OF ARCHES

17.3.1 Elastic Buckling of Circular Arches

Nays = ka NEy (17.6a)

where, for hinged arches,

Mays = kas Myz (17.9a)

where, for hinged arches (Pi and Trahair 1998),

TABLE 17.6 Coefficients for Use in Eq. 17.11

0.01 0.052 −0.217 −0.36 −0.24 0.09

TABLE 17.7 Approximate Values of m for Use in Eq. 17.12

FIGURE 17.10 Effects of the warping rigidity C 1 .

17.3.2 Buckling of Parabolic Arches

17.3.3 Design Criteria for Out-of-Plane Stability

where N* and M* are the maximum axial compression moments based on a

TABLE 17.8 Factors for the Out-of-Plane Design of Circular Arches

α any — 1.2 1.6 1.4 2.2 —

α any 1.0 1.1 1.5 1.4 2.7 —

The actual buckling load can then be determined by either an experimental

FIGURE 17.11 Interaction equation for out-of-plane stability compared to finite-element

FIGURE 17.11 (Continued )

+ xexp HEA 100

curve “d” curve “a”

FIGURE 17.12 Design for out-of-plane stability by the general method.

Design Standards For the elastic out-of-plane buckling of free-standing

17.4 BRACED ARCHES AND REQUIREMENTS