This article was downloaded by: [Memorial University of Newfoundland]

On: 01 August 2014, At: 09:38

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Mathematical and Computer Modelling

of Dynamical Systems: Methods, Tools
and Applications in Engineering and
Related Sciences
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/nmcm20

A continuum-mechanics interpretation
of Reissner's non-linear shear-
deformable beam theory
a b
Hans Irschik & Johannes Gerstmayr
a
Institute for Technical Mechanics, Johannes Kepler University of
Linz , Altenbergerstrasse 69, 4040, Linz, Austria
b
Linz Center of Mechatronics GmbH , Altenbergerstrasse 69,
4040, Linz, Austria
Published online: 25 Jan 2011.

To cite this article: Hans Irschik & Johannes Gerstmayr (2011) A continuum-mechanics
interpretation of Reissner's non-linear shear-deformable beam theory, Mathematical and Computer
Modelling of Dynamical Systems: Methods, Tools and Applications in Engineering and Related
Sciences, 17:1, 19-29, DOI: 10.1080/13873954.2010.537512

To link to this article: http://dx.doi.org/10.1080/13873954.2010.537512

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
 systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014
 Mathematical and Computer Modelling of Dynamical Systems
Vol. 17, No. 1, February 2011, 19–29

A continuum-mechanics interpretation of Reissner’s non-linear

shear-deformable beam theory
Hans Irschika* and Johannes Gerstmayrb
a
Institute for Technical Mechanics, Johannes Kepler University of Linz, Altenbergerstrasse 69,
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

4040 Linz, Austria; bLinz Center of Mechatronics GmbH, Altenbergerstrasse 69, 4040 Linz, Austria
(Received 30 November 2009; final version received 23 May 2010)

This article deals with the non-linear modelling of beams that are bent, sheared and
stretched by external forces and moments. In the following, we restrict to plane-
deformations and static conditions. Our task is to present a continuum mechanics-
based interpretation of the celebrated large displacement finite deformation structural
mechanics theory, which was presented by Eric Reissner [On one-dimensional finite-
strain beam theory: the plane problem, J. Appl. Math. Phys. 23 (1972), pp. 795–804].
The latter formulation was restricted to the notions of structural mechanics and thus
did not use the notions of stress and strain, which are fundamental for continuum
mechanics. Thus, the common continuum mechanics-based constitutive modelling at
the stress–strain level cannot be utilized in connection with Reissner’s original theory.
Instead, Reissner suggested that constitutive relations between certain generalized strains
(bending, shear and axial force strains) and generalized static entities (bending moments,
shear and normal forces) should be evaluated from physical experiments. This means that
the beam to be studied must be first built up, and the experiments must be performed for
the real beam as a whole. Although such physical experiments are indeed to be performed
in practice for safety reasons in sensible cases, for example, bridge decks or aircraft
wings, it is nevertheless felt to be a drawback that the results of simple standardized
stress–strain experiments concerning the constitutive behaviour of the materials, from
which the beam is built up, cannot be used. Moreover, relying only on physical
experiments on the whole beam means that computations (virtual experiments) can be
made only after the beam has been built up. To overcome this problem, we subsequently
present a continuum mechanics-based interpretation of Reissner’s structural mechanics
modelling, by attaching a proper continuum mechanics-based meaning to both the
generalized static entities and the generalized strains in Reissner’s theory [E. Reissner,
On one-dimensional finite-strain beam theory: the plane problem, J. Appl. Math. Phys.
23 (1972), pp. 795–804]. Consequently, these generalized static entities can be related to
the generalized strains on the basis of a constitutive modelling on the stress–strain level.
We show this in some detail in this contribution for a hyperelastic material proposed by
Simo and Hughes [Computational Inelasticity, Springer, New York, 1998]. An
illustrative numerical example is given which shows the results of large bending and
axial deformation behaviour for different constitutive relations. This article represents an
extended version of a preliminary work published in [H. Irschik and J. Gerstmayr, A
hyperelastic Reissner-type model for non-linear shear deformable beams, Proceedings of
the Mathmod 09 Vienna, I. Troch and F. Breitenecker, eds., 2009, pp. 1–7].
Keywords: Reissner’s shear-deformable beam; continuum mechanics; non-linear mate-
rial model

*Corresponding author. Email: hans.irschik@jku.at

ISSN 1387-3954 print/ISSN 1744-5051 online

© 2011 Taylor & Francis
DOI: 10.1080/13873954.2010.537512
http://www.informaworld.com
 20 H. Irschik and J. Gerstmayr

1. Introduction
This article is concerned with the non-linear modelling of plane deformations of beams that
are bent, sheared and stretched by external forces and moments. Our goal is to present a
continuum mechanics-based interpretation of the celebrated large displacement finite defor-
mation structural mechanics theory for plane shear-deformable beams, which was laid down
by Eric Reissner [1].
The principal result derived by Reissner [1] was a system of non-linear strain–displace-
ment relations, which are consistent with the exact one-dimensional local equilibrium
equations for forces and moments through an appropriate structural mechanics version of
the principle of virtual work. Reissner first derived the local structural mechanics relations of
beam equilibrium by studying the equilibrium of a deformed beam element of differential
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

length. In the plane case, this involves normal force, shear force and bending moment as
generalized static entities. Reissner then assumed constitutive relations at the structural
mechanics level to exist in the form of functional dependencies between these static entities
and a set of generalized strains, namely, a bending strain, an axial force strain and a shear
force strain. To find out the correct kinematic meaning of the latter generalized strains,
Reissner in an a priori manner postulated a virtual work expression that connects the
generalized static entities and the generalized strains. He then required that this virtual
work relation must lead to the local structural mechanics relations for beam equilibrium
that he had derived before, from which he in turn found the required kinematical meaning of
the generalized strains.
A main advantage of Reissner’s structural theory [1] is that it is variationally consistent,
due to his special construction of the generalized strains based on the principle of virtual
work. Reissner himself, however, indicated a conceptual problem that is associated with this
formulation, namely, that the constitutive relations between the generalized static entities
and the generalized strains must be evaluated from appropriate physical experiments to be
performed for the beam under consideration as a whole, that is, at the real beam. This
drawback can be attributed to the fact that the notions of stress and strain, which are basic for
continuum mechanics, were not used by Reissner in the above-sketched structural
mechanics derivations, such that the results of standardized stress–strain experiments con-
cerning the constitutive behaviour of the materials, from which the beam is built up, cannot
be used. Although physical experiments on real beams indeed are to be performed in practice
for safety reasons in sensible cases, for example, for bridge decks or aircraft wings, the
restriction to such physical experiments means that computations (virtual experiments) can
be made only after the beam has been built up.
To overcome this problem, we present a continuum mechanics-based interpretation of
Reissner’s structural mechanics modelling. We show that substituting the kinematical
assumptions of Reissner’s shear-deformable beam theory directly into the continuum
mechanics version of the principle of virtual work is equivalent to the virtual work relation
postulated by Reissner [1] between the static structural entities and the generalized strains.
This derivation then automatically attaches a continuum mechanics meaning to the general-
ized static entities and the generalized strains introduced [1] by relating them, for example, to
the work-conjugate notions of second Piola–Kirchhoff stress and Green strain. This general
formulation can be used to explain problems with the correctness and accuracy of certain
formulations in the literature, which have been encountered and resolved by Gerstmayr and
co-authors in [2] and [3]. In the light of this solution we may say that the drawback of
needing physical experiments for the real beam has led authors in the past to propose such
constitutive relations by analogy to related, often linear, theories. If, however, the exact
 Mathematical and Computer Modelling of Dynamical Systems 21

kinematical meaning of the generalized strains is not captured in a correct manner by such a
proposed constitutive relation, the latter is not variationally admissible, as it will not lead to
the correct local structural mechanics relations for beam equilibrium stated by Reissner [1].
This problem cannot arise when starting from a constitutive modelling at the stress–strain
level of continuum mechanics first, and afterwards using the below-stated relations for
computing the generalized static entities and the generalized strain from continuum
mechanics notions such as the first and the second Piola–Kirchhoff stress and Green strain.
In this contribution, we particularly derive the relations between generalized static entities
and generalized strains that are valid for a non-linear hyperelastic material law proposed by
Simo and Hughes [4]. This work represents an extension of a previous paper by the present
authors devoted to the case of beams rigid in shear, the so-called Bernoulli–Euler beams [5].
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

The relations derived in the latter paper turn out to be compatible with these results, when the
shear deformation is neglected. A preliminary version of this article was presented in [6].

2. Kinematics of shear-deformable beams

In this Section, we present a kinematical model for a beam. A beam is a solid body occupying
a certain region that extends in three-dimensional space, where one extension of this region,
the axial length of the beam, is much larger than the other two extensions, which are called
cross-sectional extensions. We make use of the material or Lagrangian description of
continuum mechanics, in which a certain reference configuration is selected to describe
the deformed configuration of the beam in space. For the sake of brevity, in the following we
adopt the kinematical situation studied in [1], namely, that the deformation is plane, that is,
the beam axes in the actual deformed and in the reference configurations are situated in a
common plane, and that no twist of the cross sections about the beam axes takes place. We
furthermore restrict this to the practical important case of beams having a straight axis in the
reference configuration.
In terms of notions that are fundamental for the kinematics of the material description of
continuum mechanics, namely, position vector r, deformation gradient F, Jacobian of the
deformation gradient J and Green strain G, see, for example, Simo and Hughes [4] and
Ziegler [7], we subsequently present the consequences of a frequent kinematical approx-
imation for the plane beam deformation, which is usually named after Stephen Timoshenko,
and which was addressed in the structural mechanics formulations by Reissner in [1] also,
see, for example, Ziegler [7] for the linear theory. This kinematical assumption approximates
the beam deformation by assuming that a cross section of the beam, which was plane and
normal to the beam axis in the reference configuration, remains plane and undistorted in the
actual configuration, and that its unit outer normal vector in the deformed configuration
encloses a certain (generally non-vanishing) shear angle with the tangent to the deformed
axis. Compared to the reference configuration, the axis in the deformed configuration in
general, however, will be stretched and curved in comparison to the reference configuration.
To allow direct transition to the notions of structural mechanics, we try to remain as
elementary as possible in our following derivations, without violating the fundamental spirit
of the material description of continuum mechanics. We therefore refer to a common
Cartesian x–y–z coordinate system to describe the plane beam deformation in connection
with the Timoshenko assumption, see Figure 1. This common Cartesian frame is fixed in the
reference configuration, and it is oriented such that x represents the coordinate of the axis in
the reference configuration. The y axis is selected such that the deformed axis remains
situated in the x–z plane. Plane cross sections of the beam in the reference configuration
therefore are spanned by the unit vectors ey and ez. The position vector of a particle, which
 22 H. Irschik and J. Gerstmayr

ξ
ϕ–χ
z
χ

ϕ
ζ
rmed
Defo s
axi
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

Figure 1. Coordinate systems and notation.

was located at the place xex þ yey þ zez in the reference configuration, having the transverse
distance |z| and out-of-plane distance |y| away from the straight reference axis, in the
deformed configuration can be written in the following form:
rðx; y; zÞ ¼ r0 ð xÞ þ y ey þ zeζ ð xÞ (1)
The position vector of an axis point in the deformed configuration is denoted by r0, see
Figure 1. Equation (1) reveals the Timoshenko assumption, namely, that the deformed cross
section is assumed to remain plane and undistorted in comparison to the reference config-
uration. The cross section in the deformed configuration is spanned by the unit vectors ey and
eζ, the latter vector being situated in the x–z plane, see also Figure 2.
Hence, the deformation gradient tensor can be obtained by partial spatial differentiation
with respect to the reference coordinates, noting that the common Cartesian frame is fixed in
the reference configuration:
F ¼ Gradr
¼ r 0 # ex þ ey # ey þ eζ # ez (2)


¼ r00 þ ze0ζ # ex þ ey # ey þ eζ # ez

χ
x
x

χ ξ
ϕ ξ
Cross section in
ζ deformed state
ζ
ϕ
s
axi

ϕ–χ
ed
m
for
De

Figure 2. Deformation gradient vectors.

 Mathematical and Computer Modelling of Dynamical Systems 23

where the prime denotes the derivative with respect to the axial coordinate x, and the symbol
# stands for the dyadic vector product.
To obtain a suitable kinematical description for the entity e0ζ , we denote the angle, by
which a reference cross section with axial coordinate x is rotated about the y axis into the
deformed configuration, as φ (x). Hence, we may write
eζ ¼ sin ’ ex þ cos ’ ez (3)
from which it follows that
e0ζ ¼ ’0 ðcos ’ ex  sin ’ ez Þ (4)
For further reference, we note that the unit vector ex is rotated by f into the unit vector e,
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

see Figures 1 and 2. The deformation gradient in Equation (2) now becomes
 
F ¼ r00  ex þ z ’0 cos ’ ex # ex þ sin ’ ex # ez
þ ey # ey (5)
 
þ r00  ez  z ’0 sin ’ ez # ex þ cos ’ ez # ez
where we have projected r00 , which is tangent to the deformed axis, onto the fixed x and z
directions to obtain a formulation which can be directly assigned with a matrix in the
common frame.
A suitable kinematical description can be obtained for the latter projections by noting
that r00 in general will not be perpendicular to the deformed cross section, and hence will not
be perpendicular to the vector eζ, because the tangent to the deformed axis generally encloses
the shear angle χ with the normal to the cross section in the deformed configuration, see
Figures 1 and 2. Would the unit vector eζ be perpendicular to the deformed axis, such that no
shear deformation is present, one has χ ¼ 0, which is commonly denoted as Bernoulli–Euler
assumption. In the following, however, we study the Timoshenko case of generally non-
vanishing shear angles χ. The tangent angle to the deformed axis, and hence to r00 , measured
from the x direction about the y axis, is given by (φ – χ). We denote the unit vector in the
direction of r00 by e, see Figure 2, such that we can write
 
r00 ¼ x0 e; x0 ¼ r00  (6)
The stretch of an element of the beam axis, that is, the ratio of the length of a differential
axis element in the deformed and in the undeformed configuration, is denoted by x0. The
projections of r00 in Equation (5) now become (see Figure 2):
r00  ex ¼ x0 cosð’  χ Þ (7)

r00  ez ¼ x0 sinð’  χ Þ (8)

From this, we obtain the following matrix representation for F in the common frame
2 3
x0 cosð’  χ Þ þ z’0 cos ’ 0 sin ’
½F  ¼ 4 0 1 0 5 (9)
ðx0 sinð’  χ Þ þ z ’0 sin ’Þ 0 cos ’
The Jacobian determinant J of F in the framework of the Timoshenko assumption
follows to
J ¼ x0 cos χ þ z’0 (10)
 24 H. Irschik and J. Gerstmayr

The symmetric right Cauchy–Green tensor, generally defined as

C ¼ FT F (11)
within the Timoshenko assumption, has the matrix representation
2
 3
2x0 þ 2z’0 x0 cos χ þ ðz ’0 Þ2 0 ðx0 sin χ Þ
6 7
½C  ¼ 4 0 1 0 5 (12)
ðx0 sin χ Þ 0 1

The Green strain tensor, which has the general definition,

Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

1
G ¼ ðC  I Þ (13)
2
is thus found to have a matrix representation that reads
2
 3
0 0 2

1 6 x0
2
þ 2z’  cos χ þ ð z ’ Þ 1 0 ðx0 sin χ Þ
x0
7
½G  ¼ 4 0 0 0 5 (14)
2
ðx0 sin χ Þ 0 0
Note that G in Equation (14) indeed reflects the Timoshenko kinematical assumption,
in the framework of which only axial strains Gxx and shear strains Gzx ¼ Gxz should be
present.
We are now in the position to relate the above continuum mechanics-based results to the
generalized strains that are fundamental in Reissner’s structural mechanics formulation [1].
Reissner introduced a bending strain as
κ ¼ ’0 (15)
and generalized normal and shear strains in the form
" ¼ x0 cos χ  1 (16)

γ ¼ x0 sin χ (17)

It is to be emphasized that the expressions for J, C and G given in Equations (10),

(12) and (14) can be completely expressed by the generalized strains presented in Equations
(15)–(17) and the transverse coordinate z. This proves that Reissner’s generalized strains are
proper strain measures also in the sense of continuum mechanics, because J, C and G are
known to be material frame indifferent deformation measures, for which the components of
proper matrix representations must not change when the reference or the deformed config-
urations are subjected to rigid body rotations. In this case, this becomes evident when
kinematically interpreting the term x0 cos χ as projection of the axial stretch to the direction
of e normal to the deformed cross section, and the term x0 sin χ as projection onto the
deformed cross section, in the direction of eζ. Moreover, the expression f0 represents a
proper curvature measure, known from the geometric analysis of curves, see, for example,
Meyberg and Vachenauer [8]. The kinematical entities in Equations (15)–(17), which
describe the deformation of the actual with respect to the reference configuration only, of
course, do not change when the rigid body rotations of the respective configurations are
superimposed. Note also that the cross-sectional coordinate z is not stretched in the frame-
work of the Timoshenko assumption.
 Mathematical and Computer Modelling of Dynamical Systems 25

3. Virtual work considerations

The application of the principle of virtual work is basic for Reissner’s structural mechanics
theory presented in [1], in which he postulated a particular simple expression for the virtual
work of the internal forces per unit length of the undeformed axis as
δWint ¼ N δ" þ Q δγ þ Mδκ (18)
where generalized static entities were introduced as normal force N, shear force Q and
bending moment M. Reissner [1] denoted these generalized static entities as stress resultants;
the exact relations to stress definitions from continuum mechanics, however, were not
addressed in [1]. In Equation (18) and in the following equations, δ denotes a variation,
that is, a virtual change in deformation.
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

In this section, we show that when using the above Timoshenko-type expressions for the
Green strain, Equation (14), and defining the generalized static entities properly, the virtual
work expression, which is known to be generally valid for the material description of
continuum mechanics, does exactly lead to Reissner’s structural mechanics postulate for
the virtual work (Equation (18)).
We introduce the symmetric second Piola–Kirchhoff stress S by writing its matrix
representation in the common frame as follows:
2 3
Sxx Syx Szx
½S  ¼ 4 Syx Syy Szy 5 (19)
Szx Szy Szz
The virtual work of S done upon a virtual change of the Green strain G is known to define
the virtual work of the internal forces in the framework of the material description of
continuum mechanics in the form:
Z
δWint ¼ S  δG dA (20)
A

where the dot indicates the scalar or double-contracted tensor product, and the integration is
to be performed over the cross-sectional area A in the reference configuration. Using the
Timoshenko-type matrix representation for F in Equation (14), and noting that S in Equation
(19) is symmetric, Sxz ¼ Szx, we obtain
Z  
1

Sxx δ 2x0 þ 2z ’0 x0 cos χ þ ðz’0 Þ 1 þ Szx δðx0 sin χ Þ dA
2
δWint ¼ (21)
A 2
Motivated by the fact that Equations (15)–(17) yield the variations,
δκ ¼ δð’0 Þ (22)

δ" ¼ δx0 cos χ  δχ x0 sin χ (23)

δγ ¼ δx0 sin χ þ δχ x0 cos χ (24)

the virtual work expression in Equation (21) is identically expanded to the form
Z
δWint ¼ ðSxx ðx0 cos χ þ z’0 Þ ðδx0 cos χ  δχ x0 sin χ Þ
A
(25)
þ ðSxx x0 sin χ þ Sxz Þðδx0 sin χ þ δχ x0 cos χ Þ
þ Sxx z ðx0 cos χ þ z’0 Þδ’0 Þ dA
 26 H. Irschik and J. Gerstmayr

Hence, defining the generalized static entities as

Z Z
N¼ Sxx ðx0 cos χ þ z’0 Þ dA ¼ Sxx J dA (26)
A A
Z Z
Q¼ ðSxx x0 sin χ þ Sxz Þ dA ¼ ðSxx γ þ Szx ÞdA (27)
A A
Z Z
M¼ Sxx ðx0 cos χ þ z ’0 ÞzdA ¼ Sxx J z dA (28)
A A

it follows that the continuum mechanics expression for the virtual work, Equation (20), in the
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

framework of the Timoshenko assumption indeed yields Reissner’s relation, Equation (18).
As is seen from Equations (26)–(28), the static entities N, Q and M represent generalized
stress resultants. The second Piola–Kirchhoff stress tensor is generally considered as a
proper stress measure to be related to the Green strain tensor in the form of constitutive
stress–strain relations, see below. A further continuum mechanics-based justification of
Equations (26)–(28) nevertheless seems to be necessary. The following arguments rest
upon the sketch given in Figure 3. As is well known, the Lagrangian stress vector, which
is taken per unit area in the undeformed state, is to be decomposed with respect to the
generally non-orthogonal and non-unit deformation gradient vectors, see, for example,
Ziegler [7] and Washizu
 [9]. For the Timoshenko case, these deformation gradient vectors
are represented by r00 þ z’0 e and eζ. The corresponding skew stress components are
formed by the second Piola–Kirchhoff stresses Sxx and Szx, respectively, see Figure 3.
Projecting onto the directions of e and eζ, and forming the resulting force components by
integration over the undeformed cross-sectional area, yields Equations (26) and (27).
Moreover, forming the resulting moment about the axis point gives Equation (28); see
Figure 3 for the corresponding application of the lever rule. These simple continuum
mechanics interpretations are felt to be results in their own right.
That the continuum mechanics expression for the virtual work, Equation (20), in the
framework of the Timoshenko assumption, yields Reissner’s structural mechanics relation,
Equation (18), proves that the definitions for the generalized static entities given in
Equations (26)–(28) are consistent with the local structural mechanics relations of beam
equilibrium that were stated by Reissner [1]. Moreover, we now have at our disposal

χ
Sxx x

χ ξ

Sxx ϕ ξ

ζ
Sxx x

Sxz

Figure 3. Graphical interpretation of stress components.

 Mathematical and Computer Modelling of Dynamical Systems 27

relations between the static entities N, Q and M and components of the second Piola–
Kirchhoff stress, which can be used to consistently introduce constitutive models at the
stress–strain level of continuum mechanics into the boundary value problems for the
generalized static entities and the generalized strains that were stated by Reissner [1]. For
the sake of brevity, we do not repeat these boundary-value problems, which would involve
several additional displacement relations. Rather, in the following we show an example for
producing proper relations for N, Q and M.

4. A special hyperelastic material law

The following constitutive stress–strain relation was suggested in the book by Simo and
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

Hughes [4] for hyperelastic isotropic materials:

λ 2   
S¼ J  1 C 1 þ μ I  C 1 (29)
2
where λ and μ denote the Lamé parameters. Using the expressions for J, C and γ stated in
Equations (10), (12) and (17) in the framework of the Timoshenko assumption, the matrix
representation of the inverse of C in the common frame can be written as
2 3

1
J2 0  Jγ2
C 1 ¼ 4 0 1 0 25 (30)
 Jγ2 0 1 þ Jγ 2
Substituting this into Equation (29) and using the resulting respective components of S,
the expressions given in Equations (26)–(28) can be evaluated in the form
Z   
λ 1
N¼ μþ J dA (31)
A 2 J
Z
Q ¼ γ μ dA (32)
A
Z   
λ 1
M¼ μþ J zdA (33)
A 2 J
Note that the shear force depends on the shear strain only linearly in this case. The
integrals in Equations (31)–(33) can be easily evaluated for a given distribution of elastic
parameters and a given form of the cross section. To avoid logarithmic expressions for the
normal force and the bending moment, numerical integration is utilized to perform
the integration over height of the beam. Note that a linearization of Equations (31)–(33)
for the case of small deformations leads to the well-known expressions of the normal force,
the shear force and the bending moment. However, it needs to be taken into account that the
deformation gradient J needs to consider the influence of the Poisson ratio. Thus, the
geometrically simplest case arises if the Poisson ratio is 0. To improve the accuracy of
the above-mentioned models, a shear correction factor needs to be added in Equation (32) to
compensate the effect of the constant approximation of shear stress along the cross section.
In the following, a large deformation cantilever beam is studied as a numerical example,
to show the difference of hyperelastic material models as compared to linear elastic ones.
The dimensions of the cantilever beam with rectangular cross section, shown in Figure 4, are
length L¼2 m, height h¼0.5 m, width w¼0.1 m, Young’s modulus E ¼ 2:07  1011 N=m2 ,
Poisson ratio v¼0 and a shear correction factor ks ¼ 10ð1þνÞ
12þ11ν .
 28 H. Irschik and J. Gerstmayr

F0 M0

Figure 4. Example of a large deformation cantilever beam with combined tip force and tip moment.
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

A combined tip force and tip moment is applied to the beam, which leads to large
deformations within the beam. The vertical tip load is F0 ¼ 1:25  108 N and the tip moment
is M0 ¼ 1:6  108 Nm. Table 1 shows a comparison of the classical St. Venant–Kirchhoff
material law,
S ¼ 2μG þ λI trðGÞ (34)
and the hyperelastic material law as formulated in Equation (29) for the case of a shear-
deformable beam (absolute nodal coordinate formulation (ANCF) shear beam) and a fully
3D computation with linear hexahedral elements. Both formulations have been implemented
into the multibody and finite-element research code HOTINT [10]. For details on the
numerical model of the shear-deformable beam element, which is based on the absolute
nodal coordinate formulation (ANCF), see Gerstmayr et al. [3]. It turns out that especially in
the case of larger strains, the classical constitutive Equation (34), which is linear in the Green
strain and in the second Piola–Kirchhoff stress, leads to considerably different results as
compared to a hyperelastic material given in Equation (29). The comparison to the 3D
hexahedral finite-element formulation shows that shear-deformable beams have the ability
to give accurate results for elastic and even for hyperelastic constitutive relations in the case
of large deformations. It shall be noted that the solution with linear standard hexahedral
elements gives slightly smaller vertical displacements as compared to the beam formulation.
This is due to the geometrically non-linear effects and combined tip force and tip moment.

5. Conclusion
In this article, we have shown how a continuum mechanics-based meaning can be attached in
a rational manner to the generalized static entities and the generalized strains in Reissner’s

Table 1. Comparison of numerical results for the shear deformable ANCF beam and 3D hexahedral
elements using a linear material and a non-linear constitutive model.
Material model Element type Number of elements ux uz
St. Venant–Kirchhoff ANCF shear beam 16 0.02932 0.09136
St. Venant–Kirchhoff ANCF shear beam 32 0.02938 0.09256
St. Venant–Kirchhoff 3D hexahedral FE 256 0.02948 0.09030
St. Venant–Kirchhoff 3D hexahedral FE 4096 0.02958 0.09043
Hyperelastic ANCF shear beam 16 0.01537 0.10500
Hyperelastic ANCF shear beam 32 0.01541 0.10610
Hyperelastic 3D hexahedral FE 256 0.01559 0.10461
Hyperelastic 3D hexahedral FE 4096 0.01565 0.10487
 Mathematical and Computer Modelling of Dynamical Systems 29

structural mechanics theory for large displacements and finite strains of beams [1]. This
enables us to bring into play a constitutive modelling at the stress–strain level of continuum
mechanics, which we have demonstrated for a special isotropic hyperelastic model proposed
by Simo and Hughes in [4]. The numerical example shows that the hyperelastic material
model, which has been applied to the shear-deformable beams in a straightforward manner,
leads to almost exactly the same results for shear beams and computationally expensive 3D
finite elements. In principle, the presented methodology should enable to introduce any
suitable continuum mechanics-based constitutive modelling at the stress–strain level. This
work considers shear deformation in the framework of the Timoshenko assumption. It
represents an extension of our previous paper [5] for beams rigid in shear in the framework
of the so-called Bernoulli–Euler assumption. The proposed formulation can be directly used
Downloaded by [Memorial University of Newfoundland] at 09:38 01 August 2014

in a dynamical model for studying, for example, the control of beam vibrations.

Acknowledgement
Support of the present work in the framework of the Comet-K2 Austrian Center of Competence in
Mechatronics (ACCM) is gratefully acknowledged.

References
[1] E. Reissner, On one-dimensional finite-strain beam theory: the plane problem, J. Appl. Math.
Phys. 23 (1972), pp. 795–804.
[2] J. Gerstmayr and H. Irschik, On the correct representation of bending and axial deformation in
the absolute nodal coordinate formulat ion with an elastic line approach, J. Snd. Vib. 318 (2008),
pp. 461–487.
[3] J. Gerstmayr, M.K. Matikainen, and A.M. Mikkola, A geometrically exact beam element based
on the absolute nodal coordinate formulation, J. Mult. Sys. Dyn. 20 (2008), pp. 359–384.
[4] J.C. Simo and T.J.R. Hughes, Computational Inelasticity, Springer, New York, 1998.
[5] H. Irschik and J. Gerstmayr, A continuum mechanics based derivation of Reissner’s large-
displacement finite-strain beam theory: the case of plane deformations of originally straight
Bernoulli–Euler beams, Acta Mechanica 206 (1–2) (2008), pp. 1–21.
[6] H. Irschik and J. Gerstmayr, A hyperelastic Reissner-type model for non-linear shear deform-
able beams, in Proceedings of the Mathmod 09 Vienna, I. Troch and F. Breitenecker, eds.,
Springer-Verlag, Wien, 2009, pp. 1–7.
[7] F. Ziegler, Mechanics of Solids and Fluids, 2nd ed., 2nd corr. print, Springer, New York, 1998.
[8] K. Meyberg and P. Vachenauer, Höhere Mathematik 1, 6th ed., 1st corr. print, Springer, Berlin,
2003.
[9] K. Washizu, Variational Methods in Elasticity and Plasticity, 2nd ed., Pergamon Press, Oxford,
1974.
[10] J. Gerstmayr, HOTINT – A C++ Environment for the simulation of multibody dynamics systems
and finite elements, Proceedings of the Multibody Dynamics 2009 Eccomas Thematic
Conference, K. Arczewski, J. Fraczek, M. Wojtyr, eds., Warsaw, Poland (CD-ROM), 2009, pp.
1–20.