Föppl–von Kármán equations

The Föppl–von Kármán equations, named after August Föppl and Theodore von Kármán, are a set of nonlinear partial differential equations describing the large deflections of thin flat plates. With applications ranging from the design of submarine hulls to the mechanical properties of cell wall, the equations are notoriously difficult to solve, and take the following form:
"Theory of Elasticity". L. D. Landau, E. M. Lifshitz, (3rd ed. )

:$\begin (1) \qquad & \frac\nabla^4 w-h\frac\left(\sigma_\frac\right)=P \\ (2) \qquad & \frac=0 \end$

where is the Young's modulus of the plate material (assumed homogeneous and isotropic), is the Poisson's ratio, is the thickness of the plate, is the out–of–plane deflection of the plate, is the external normal force per unit area of the plate, is the Cauchy stress tensor, and are indices that take values of 1 and 2 (the two orthogonal in-plane directions). The 2-dimensional biharmonic operator is defined as
:
\nabla^4 w := \frac\left frac\right = \frac + \frac + 2\frac \,.

Equation (1) above can be derived from kinematic assumptions and the constitutive relations for the plate. Equations (2) are the two equations for the conservation of linear momentum in two dimensions where it is assumed that the out–of–plane stresses () are zero.

Validity of the Föppl–von Kármán equations

While the Föppl–von Kármán equations are of interest from a purely mathematical point of view, the physical validity of these equations is questionable. Ciarlet states: ''The two-dimensional von Karman equations for plates, originally proposed by von Karman 910 play a mythical role in applied mathematics. While they have been abundantly, and satisfactorily, studied from the mathematical standpoint, as regards notably various questions of existence, regularity, and bifurcation, of their solutions, their physical soundness has been often seriously questioned.'' Reasons include the facts that
# the theory depends on an approximate geometry which is not clearly defined
# a given variation of stress over a cross-section is assumed arbitrarily
# a linear constitutive relation is used that does not correspond to a known relation between well defined measures of stress and strain
# some components of strain are arbitrarily ignored
# there is a confusion between reference and deformed configurations which makes the theory inapplicable to the large deformations for which it was apparently devised.
Conditions under which these equations are actually applicable and will give reasonable results when solved are discussed in Ciarlet.

Equations in terms of Airy stress function

The three Föppl–von Kármán equations can be reduced to two by introducing the Airy stress function $\varphi$ where
:$\sigma_ = \frac ~,~~ \sigma_ = \frac ~,~~ \sigma_ = - \frac \,.$
Equation (1) becomes
:$\frac\Delta^2 w-h\left(\frac\frac+\frac\frac-2\frac\frac\right)=P$
while the Airy function satisfies, by construction the force balance equation (2). An equation for $\varphi$ is obtained
enforcing the representation of the strain as a function of the stress. One gets "Theory of Elasticity". L. D. Landau, E. M. Lifshitz, (3rd ed. )
:$\Delta^2\varphi+E\left\=0 \,.$

Pure bending

For the pure bending of thin plates the equation of equilibrium is $D\Delta^2\ w=P$, where

:$D :=\frac$
is called flexural or ''cylindrical rigidity'' of the plate.

Kinematic assumptions (Kirchhoff hypothesis)

In the derivation of the Föppl–von Kármán equations the main kinematic assumption (also known as the Kirchhoff hypothesis) is that surface normals to the plane of the plate remain perpendicular to the plate after deformation. It is also assumed that the in-plane (membrane) displacements are small and the change in thickness of the plate is negligible. These assumptions imply that the displacement field in the plate can be expressed as

:$u_1(x_1,x_2,x_3) = v_1(x_1,x_2)-x_3\,\frac ~,~~ u_2(x_1,x_2,x_3) = v_2(x_1,x_2)-x_3\,\frac ~,~~ u_3(x_1, x_2, x_3) = w(x_1,x_2)$
in which is the in-plane (membrane) displacement. This form of the displacement field implicitly assumes that the amount of rotation of the plate is small.

Strain-displacement relations (von Kármán strains)

The components of the three-dimensional Lagrangian Green strain tensor are defined as
:
E_ := \frac\left frac + \frac
+ \frac\,\frac\right\,.

Substitution of the expressions for the displacement field into the above gives
:
\begin
E_ & = \frac
+ \frac\left left(\frac\right)^2
+ \left(\frac\right)^2
+ \left(\frac\right)^2\right\
&= \frac - x_3\,\frac
+ \frac\left _3^2\left(\frac\right)^2
+ x_3^2\left(\frac\right)^2
+ \left(\frac\right)^2\right\
E_ & = \frac
+ \frac\left left(\frac\right)^2
+ \left(\frac\right)^2
+ \left(\frac\right)^2\right\
&= \frac-x_3\,\frac
+ \frac\left _3^2\left(\frac\right)^2
+ x_3^2\left(\frac\right)^2
+ \left(\frac\right)^2\right\
E_ & = \frac
+ \frac\left left(\frac\right)^2
+ \left(\frac\right)^2
+ \left(\frac\right)^2\right\
&= \frac\left left(\frac\right)^2
+ \left(\frac\right)^2
\right\
E_ & = \frac\left frac + \frac
+ \frac\,\frac
+ \frac\,\frac
+ \frac\,\frac\right\
& = \frac\frac + \frac\frac -x_3\frac
+ \frac\left _3^2\left(\frac\right)\left(\frac\right)
+ x_3^2\left(\frac\right)\left(\frac\right)
+ \frac\,\frac\right\
E_ & = \frac\left frac + \frac
+ \frac\,\frac
+ \frac\,\frac
+ \frac\,\frac\right\
& = \frac\left _3\left(\frac\right)\left(\frac\right)
+ x_3\left(\frac\right)\left(\frac\right)
\right\
E_ & = \frac\left frac + \frac
+ \frac\,\frac
+ \frac\,\frac
+ \frac\,\frac\right\\
& = \frac\left _3\left(\frac\right)\left(\frac\right)
+ x_3\left(\frac\right)\left(\frac\right)
\right \end

For small strains but moderate rotations, the higher order terms that cannot be neglected are
:$\left(\frac\right)^2 ~,~~ \left(\frac\right)^2 ~,~~ \frac\,\frac \,.$
Neglecting all other higher order terms, and enforcing the requirement that the plate does not change its thickness, the strain tensor components reduce to the von Kármán strains
:$\begin E_ & = \frac + \frac\left(\frac\right)^2 -x_3\,\frac \\ E_ & = \frac + \frac\left(\frac\right)^2 -x_3\,\frac \\ E_ & = \frac\left(\frac+\frac\right) + \frac\,\frac\,\frac -x_3\frac \\ E_ & = 0 ~,~~ E_ = 0 ~,~~ E_ = 0 \,. \end$

The first terms are the usual small-strains, for the mid-surface. The second terms, involving squares of displacement gradients, are non-linear, and need to be considered when the plate bending is fairly large (when the rotations are about 10 – 15 degrees). These first two terms together are called the membrane strains. The last terms, involving second derivatives, are the flexural (bending) strains. They involve the curvatures. These zero terms are due to the assumptions of the classical plate theory, which assume elements normal to the mid-plane remain inextensible and line elements perpendicular to the mid-plane remain normal to the mid-plane after deformation.

Stress–strain relations

If we assume that the Cauchy stress tensor components are linearly related to the von Kármán strains by Hooke's law, the plate is isotropic and homogeneous, and that the plate is under a plane stress condition,Typically, an assumption of zero out-of-plane stress is made at this point. we have = = = 0 and
:$\begin\sigma_ \\ \sigma_ \\ \sigma_ \end = \cfrac \begin 1 & \nu & 0 \\ \nu & 1 & 0 \\ 0 & 0 & 1-\nu \end \begin E_ \\ E_ \\ E_ \end$
Expanding the terms, the three non-zero stresses are
:
\begin
\sigma_ &= \cfrac\left[\left(\frac
+ \frac\left(\frac\right)^2
-x_3\,\frac \right) +
\nu\left(\frac
+ \frac\left(\frac\right)^2
-x_3\,\frac \right) \right] \\
\sigma_ &= \cfrac\left[\nu\left(\frac
+ \frac\left(\frac\right)^2
-x_3\,\frac \right) +
\left(\frac
+ \frac\left(\frac\right)^2
-x_3\,\frac \right) \right] \\
\sigma_ &= \cfrac\left frac\left(\frac+\frac\right)
+ \frac\,\frac\,\frac
-x_3\frac \right\,.
\end

Stress resultants

The stress resultants in the plate are defined as

:$N_ := \int_^ \sigma_\, d x_3 ~,~~ M_ := \int_^ x_3\,\sigma_\, d x_3 \,.$

Therefore,

:
\begin
N_ &= \cfrac\left \frac + \left(\frac\right)^2
+ 2\nu\frac + \nu\left(\frac\right)^2 \right\\
N_ &= \cfrac\left \nu\frac + \nu\left(\frac\right)^2
+ 2\frac + \left(\frac\right)^2 \right\\
N_ &= \cfrac\left frac + \frac + \frac\,\frac \right \end

the elimination of the in-plane displacements leads to

\begin
\frac\left (1 + \nu)\frac - \frac + \nu\frac
- \frac + \nu\frac\right
= \left frac\frac - \left(\frac\right)^2\right\end

and

:
\begin
M_ &= -\cfrac\left frac +\nu \,\frac \right\\
M_ &= -\cfrac\left nu \,\frac +\frac \right\\
M_ &= -\cfrac\,\frac \,.
\end

Solutions are easier to find when the governing equations are expressed in terms of stress resultants rather than the in-plane stresses.

Equations of Equilibrium

The weak form of the Kirchhoff plate is

:$\int_\int_^\rho \ddot_i\delta u_i \,d\Omega dx_3+ \int_\int_^ \sigma_\delta E_\,d\Omega dx_3 + \int_\int_^ p_i \delta u_i \,d\Omega dx_3 =0$

here Ω denotes the mid-plane. The weak form leads to

:$\begin \int_\rho h \ddot_1 \delta v_1 \,d\Omega &+ \int_ N_\frac + N_\frac\,d\Omega = -\int_ p_1 \delta v_1 \,d\Omega \\ \int_\rho h \ddot_2 \delta v_2 \,d\Omega &+ \int_ N_\frac + N_\frac\,d\Omega = -\int_ p_2 \delta v_2 \,d\Omega \\ \int_\rho h \ddot \delta w \,d\Omega &+ \int_ N_\frac\frac - M_\frac \,d\Omega\\ &+ \int_ N_\frac\frac - M_\frac \,d\Omega\\ &+ \int_ N_\left(\frac\frac + \frac\frac\right) - 2M_\frac \,d\Omega = -\int_ p_3 \delta w \,d\Omega\\ \end$

The resulting governing equations are

$\begin &\rho h \ddot - \frac - \frac - 2\frac - \frac\left(N_\,\frac + N_\,\frac\right) - \frac\left(N_\,\frac + N_\,\frac\right) = -p_3 \\ & \rho h \ddot_1 - \frac- \frac = -p_1\\ & \rho h \ddot_2 - \frac- \frac = -p_2 \,. \end$

Föppl–von Kármán equations in terms of stress resultants

The Föppl–von Kármán equations are typically derived with an energy approach by considering variations of internal energy and the virtual work done by external forces. The resulting static governing equations (Equations of Equilibrium) are

:$\begin &\frac + \frac + 2\frac + \frac\left(N_\,\frac + N_\,\frac\right) + \frac\left(N_\,\frac + N_\,\frac\right) = P \\ & \frac = 0 \,. \end$

When the deflections are small compared to the overall dimensions of the plate, and the mid-surface strains are neglected,

$\begin \frac \approx 0 ,\frac \approx 0, v_1 \approx 0, v_2\approx 0 \end$.

The equations of equilibrium are reduced ( pure bending of thin plates) to

:$\frac + \frac + 2\frac = P$.

References

See also

* Plate theory

{{DEFAULTSORT:Foppl-Von Karman Equations
Partial differential equations
Continuum mechanics