Section 23.1 of Chapter 23: 3D beam elements. Linear case

3D beam element. Euler–Bernoulli model

Hypotheses:

StrainSmall
DisplacementSmall
MaterialLinear

Euler–Bernoulli beam model.

Beam element: straight, prismatic (constant cross-section), two nodes.

The local reference frame is composed of the beam centroidal axis, the local xˉ\bar x-axis, and the two principal axes of the constant cross-section, the local yˉ\bar y- and zˉ\bar z-axes. Each node has six DOFs: three displacements (u,v,w)(u,v,w) and three rotations (φx,φy,φz)(\varphi_x,\varphi_y,\varphi_z).

Figure 1
Figure 1.

In the local reference frame:

uel={uˉ1vˉ1wˉ1φxˉ1φyˉ1φzˉ1uˉ2vˉ2wˉ2φxˉ2φyˉ2φzˉ2}T. \overline{\boldsymbol{u}}_{el}= \left\{ \begin{array}{cccccccccccc} \bar u_1& \bar v_1& \bar w_1& \varphi_{\bar x1}& \varphi_{\bar y1}& \varphi_{\bar z1}& \bar u_2& \bar v_2& \bar w_2& \varphi_{\bar x2}& \varphi_{\bar y2}& \varphi_{\bar z2} \end{array} \right\}^{T}.

In the global reference frame:

uel={u1v1w1φx1φy1φz1u2v2w2φx2φy2φz2}T. \boldsymbol{u}_{el}= \left\{ \begin{array}{cccccccccccc} u_1& v_1& w_1& \varphi_{x1}& \varphi_{y1}& \varphi_{z1}& u_2& v_2& w_2& \varphi_{x2}& \varphi_{y2}& \varphi_{z2} \end{array} \right\}^{T}.

01 Notations used for the beam element

Hooke’s law:

σ=Dε, \boldsymbol{\sigma}=\boldsymbol{D}\boldsymbol{\varepsilon},

or

{NMxˉMyˉMzˉ}=[EA0000GIt0000EIy0000EIz]{ε0κxˉκyˉκzˉ}. \left\{ \begin{array}{c} N\\ M_{\bar x}\\ M_{\bar y}\\ M_{\bar z} \end{array} \right\} = \begin{bmatrix} EA&0&0&0\\ 0&GI_t&0&0\\ 0&0&EI_y&0\\ 0&0&0&EI_z \end{bmatrix} \left\{ \begin{array}{c} \varepsilon_0\\ \kappa_{\bar x}\\ \kappa_{\bar y}\\ \kappa_{\bar z} \end{array} \right\}.

The generalized strain and stress vectors are

ε={ε0κxˉκyˉκzˉ},σ={NMxˉMyˉMzˉ}. \boldsymbol{\varepsilon}= \left\{ \begin{array}{c} \varepsilon_0\\ \kappa_{\bar x}\\ \kappa_{\bar y}\\ \kappa_{\bar z} \end{array} \right\}, \qquad \boldsymbol{\sigma}= \left\{ \begin{array}{c} N\\ M_{\bar x}\\ M_{\bar y}\\ M_{\bar z} \end{array} \right\}.

The constitutive matrix is

D=[EA0000GIt0000EIy0000EIz]. \boldsymbol{D}= \begin{bmatrix} EA&0&0&0\\ 0&GI_t&0&0\\ 0&0&EI_y&0\\ 0&0&0&EI_z \end{bmatrix}.

The axial strain of the centroidal axis is

ε0=duˉdxˉ. \varepsilon_0=\frac{d\bar u}{d\bar x}.

In the Euler–Bernoulli beam model, the slopes in the two principal planes are

φyˉ=dwˉdxˉ,φzˉ=dvˉdxˉ. \varphi_{\bar y}=-\frac{d\bar w}{d\bar x}, \qquad \varphi_{\bar z}=\frac{d\bar v}{d\bar x}.

Consequently,

κxˉ=dφxˉdxˉ, \kappa_{\bar x}= \frac{d\varphi_{\bar x}}{d\bar x},
κyˉ=dφyˉdxˉ=d2wˉdxˉ2, \kappa_{\bar y}= \frac{d\varphi_{\bar y}}{d\bar x} = -\frac{d^2\bar w}{d\bar x^2},
κzˉ=dφzˉdxˉ=d2vˉdxˉ2. \kappa_{\bar z}= \frac{d\varphi_{\bar z}}{d\bar x} = \frac{d^2\bar v}{d\bar x^2}.

Thus the sign associated with bending about the local yˉ\bar y-axis is included in the geometrical definition of κyˉ\kappa_{\bar y}. The moment–curvature relations remain

Mxˉ=GItκxˉ,Myˉ=EIyκyˉ,Mzˉ=EIzκzˉ. M_{\bar x}=GI_t\kappa_{\bar x}, \qquad M_{\bar y}=EI_y\kappa_{\bar y}, \qquad M_{\bar z}=EI_z\kappa_{\bar z}.

NN is the axial effort, MxˉM_{\bar x} is the torque and Myˉ,MzˉM_{\bar y},M_{\bar z} are the bending moments.

EE is Young’s modulus and GG is the shear modulus.

A,It,Iy,IzA,I_t,I_y,I_z are geometrical properties of the cross-section: the cross-section area, the torsional geometric property and the two geometrical moments of inertia.

02 Interpolation functions in the local reference frame

The deflected centroidal axis of the beam is described by a third-degree polynomial for the two transverse displacements vˉ\bar v and wˉ\bar w, and by a linear polynomial for the axial displacement uˉ\bar u and for the twist angle φxˉ\varphi_{\bar x}.

In the Euler–Bernoulli beam model, in the local reference frame, the slopes are [1]:

φyˉ=dwˉdxˉ,φzˉ=dvˉdxˉ. \varphi_{\bar y}=-\frac{d\bar w}{d\bar x}, \qquad \varphi_{\bar z}=\frac{d\bar v}{d\bar x}.

Consequently, the six displacement components have the following variation along the beam element.

Let

ξ=xˉL, \xi=\frac{\bar x}{L},

where LL is the length of the beam element.

Axial displacement — linear polynomial

uˉ(ξ)=(1ξ)uˉ1+ξuˉ2. \bar u(\xi) = (1-\xi)\bar u_1+\xi\bar u_2.

Twist angle — linear polynomial

φxˉ(ξ)=(1ξ)φxˉ1+ξφxˉ2. \varphi_{\bar x}(\xi) = (1-\xi)\varphi_{\bar x1} + \xi\varphi_{\bar x2}.

Transverse displacement and slope in the xˉyˉ\bar x\bar y principal plane

Define:

H1=2ξ33ξ2+1, H_1=2\xi^3-3\xi^2+1,
H2=L(ξ32ξ2+ξ), H_2=L(\xi^3-2\xi^2+\xi),
H3=2ξ3+3ξ2, H_3=-2\xi^3+3\xi^2,
H4=L(ξ3ξ2). H_4=L(\xi^3-\xi^2).

Then

vˉ=H1vˉ1+H2φzˉ1+H3vˉ2+H4φzˉ2, \bar v = H_1\bar v_1 + H_2\varphi_{\bar z1} + H_3\bar v_2 + H_4\varphi_{\bar z2},

and

φzˉ=dvˉdxˉ. \varphi_{\bar z}=\frac{d\bar v}{d\bar x}.

Transverse displacement and slope in the xˉzˉ\bar x\bar z principal plane

wˉ=H1wˉ1H2φyˉ1+H3wˉ2H4φyˉ2, \bar w = H_1\bar w_1 - H_2\varphi_{\bar y1} + H_3\bar w_2 - H_4\varphi_{\bar y2},

and

φyˉ=dwˉdxˉ. \varphi_{\bar y}=-\frac{d\bar w}{d\bar x}.

Hence,

u(xˉ)=N(xˉ)uel, \overline{\boldsymbol{u}}(\bar x) = \boldsymbol{N}(\bar x)\, \overline{\boldsymbol{u}}_{el},

where

u(xˉ)={uˉvˉwˉφxˉφyˉφzˉ}. \overline{\boldsymbol{u}}(\bar x) = \left\{ \begin{array}{c} \bar u\\ \bar v\\ \bar w\\ \varphi_{\bar x}\\ \varphi_{\bar y}\\ \varphi_{\bar z} \end{array} \right\}.

The interpolation functions satisfy exactly the Euler–Bernoulli kinematic assumptions.

It results the matrix B\boldsymbol{B} [1]:

ε={ε0κxˉκyˉκzˉ}={duˉdxˉdφxˉdxˉd2wˉdxˉ2d2vˉdxˉ2}=Buel, \boldsymbol{\varepsilon} = \left\{ \begin{array}{c} \varepsilon_0\\ \kappa_{\bar x}\\ \kappa_{\bar y}\\ \kappa_{\bar z} \end{array} \right\} = \left\{ \begin{array}{c} \dfrac{d\bar u}{d\bar x}\\[2mm] \dfrac{d\varphi_{\bar x}}{d\bar x}\\[2mm] -\dfrac{d^2\bar w}{d\bar x^2}\\[2mm] \dfrac{d^2\bar v}{d\bar x^2} \end{array} \right\} = \boldsymbol{B}\, \overline{\boldsymbol{u}}_{el},

that is,

B=[1L000001L000000001L000001L0000612ξL204+6ξL0006+12ξL202+6ξL006+12ξL20004+6ξL0612ξL20002+6ξL]. \boldsymbol{B}= \begin{bmatrix} -\dfrac1L&0&0&0&0&0& \dfrac1L&0&0&0&0&0 \\[3mm] 0&0&0&-\dfrac1L&0&0& 0&0&0&\dfrac1L&0&0 \\[3mm] 0&0&\dfrac{6-12\xi}{L^2}&0& \dfrac{-4+6\xi}{L}&0& 0&0&\dfrac{-6+12\xi}{L^2}&0& \dfrac{-2+6\xi}{L}&0 \\[3mm] 0&\dfrac{-6+12\xi}{L^2}&0&0&0& \dfrac{-4+6\xi}{L}& 0&\dfrac{6-12\xi}{L^2}&0&0&0& \dfrac{-2+6\xi}{L} \end{bmatrix}.

03 Element stiffness matrix

The deformation energy for a single finite element is

Uel=12LεTσdxˉ, U_{el}= \frac12 \int_L \boldsymbol{\varepsilon}^{T} \boldsymbol{\sigma}\, d\bar x,

or

Uel=12LεTDεdxˉ. U_{el}= \frac12 \int_L \boldsymbol{\varepsilon}^{T} \boldsymbol{D} \boldsymbol{\varepsilon}\, d\bar x.

Using

ε=Buel, \boldsymbol{\varepsilon} = \boldsymbol{B}\overline{\boldsymbol{u}}_{el},

we obtain

Uel=12uelT[LBTDBdxˉ]uel=12uelTkeluel, U_{el} = \frac12 \overline{\boldsymbol{u}}_{el}^{\,T} \left[ \int_L \boldsymbol{B}^T\boldsymbol{D}\boldsymbol{B}\,d\bar x \right] \overline{\boldsymbol{u}}_{el} = \frac12 \overline{\boldsymbol{u}}_{el}^{\,T} \overline{\boldsymbol{k}}_{el} \overline{\boldsymbol{u}}_{el},

where

kel=LBTDBdxˉ. \boxed{ \overline{\boldsymbol{k}}_{el} = \int_L \boldsymbol{B}^T\boldsymbol{D}\boldsymbol{B}\,d\bar x }.

The integral can be calculated analytically. It results:

kel=[EAL00000EAL00000012EIzL30006EIzL2012EIzL30006EIzL20012EIyL306EIyL200012EIyL306EIyL20000GItL00000GItL00006EIyL204EIyL0006EIyL202EIyL006EIzL20004EIzL06EIzL20002EIzLEAL00000EAL00000012EIzL30006EIzL2012EIzL30006EIzL20012EIyL306EIyL200012EIyL306EIyL20000GItL00000GItL00006EIyL202EIyL0006EIyL204EIyL006EIzL20002EIzL06EIzL20004EIzL]. \overline{\boldsymbol{k}}_{el}= \begin{bmatrix} \frac{EA}{L}&0&0&0&0&0&-\frac{EA}{L}&0&0&0&0&0\\ 0&\frac{12EI_z}{L^3}&0&0&0&\frac{6EI_z}{L^2}&0&-\frac{12EI_z}{L^3}&0&0&0&\frac{6EI_z}{L^2}\\ 0&0&\frac{12EI_y}{L^3}&0&-\frac{6EI_y}{L^2}&0&0&0&-\frac{12EI_y}{L^3}&0&-\frac{6EI_y}{L^2}&0\\ 0&0&0&\frac{GI_t}{L}&0&0&0&0&0&-\frac{GI_t}{L}&0&0\\ 0&0&-\frac{6EI_y}{L^2}&0&\frac{4EI_y}{L}&0&0&0&\frac{6EI_y}{L^2}&0&\frac{2EI_y}{L}&0\\ 0&\frac{6EI_z}{L^2}&0&0&0&\frac{4EI_z}{L}&0&-\frac{6EI_z}{L^2}&0&0&0&\frac{2EI_z}{L}\\ -\frac{EA}{L}&0&0&0&0&0&\frac{EA}{L}&0&0&0&0&0\\ 0&-\frac{12EI_z}{L^3}&0&0&0&-\frac{6EI_z}{L^2}&0&\frac{12EI_z}{L^3}&0&0&0&-\frac{6EI_z}{L^2}\\ 0&0&-\frac{12EI_y}{L^3}&0&\frac{6EI_y}{L^2}&0&0&0&\frac{12EI_y}{L^3}&0&\frac{6EI_y}{L^2}&0\\ 0&0&0&-\frac{GI_t}{L}&0&0&0&0&0&\frac{GI_t}{L}&0&0\\ 0&0&-\frac{6EI_y}{L^2}&0&\frac{2EI_y}{L}&0&0&0&\frac{6EI_y}{L^2}&0&\frac{4EI_y}{L}&0\\ 0&\frac{6EI_z}{L^2}&0&0&0&\frac{2EI_z}{L}&0&-\frac{6EI_z}{L^2}&0&0&0&\frac{4EI_z}{L} \end{bmatrix}.

04 Transformation from the global to the local reference frame

To transform the displacement components from the global reference frame to the local reference frame, the passive rotation matrix R\boldsymbol{R} is used:

uel=Ruel. \boxed{ \overline{\boldsymbol{u}}_{el} = \boldsymbol{R}\,\boldsymbol{u}_{el} }.

The element rotation matrix is

R=[R00000R00000R00000R0], \boldsymbol{R}= \begin{bmatrix} \boldsymbol{R}_0&\boldsymbol{0}&\boldsymbol{0}&\boldsymbol{0}\\ \boldsymbol{0}&\boldsymbol{R}_0&\boldsymbol{0}&\boldsymbol{0}\\ \boldsymbol{0}&\boldsymbol{0}&\boldsymbol{R}_0&\boldsymbol{0}\\ \boldsymbol{0}&\boldsymbol{0}&\boldsymbol{0}&\boldsymbol{R}_0 \end{bmatrix},

where

R0=[l1m1n1l2m2n2l3m3n3]. \boldsymbol{R}_0= \begin{bmatrix} l_1&m_1&n_1\\ l_2&m_2&n_2\\ l_3&m_3&n_3 \end{bmatrix}.

Matrix R0\boldsymbol{R}_0 contains the direction cosines of the local reference-frame axes expressed in the global reference frame. Thus, R0\boldsymbol{R}_0 transforms vector components from the global reference frame to the local one.

Therefore,

Uel=12uelTRTkelRuel=12uelTkeluel, U_{el} = \frac12 \boldsymbol{u}_{el}^T \boldsymbol{R}^T \overline{\boldsymbol{k}}_{el} \boldsymbol{R} \boldsymbol{u}_{el} = \frac12 \boldsymbol{u}_{el}^T \boldsymbol{k}_{el} \boldsymbol{u}_{el},

and hence

kel=RTkelR. \boxed{ \boldsymbol{k}_{el} = \boldsymbol{R}^T \overline{\boldsymbol{k}}_{el} \boldsymbol{R} }.

This is the same passive transformation convention used for finite-element assembly throughout the 3D developments.

05 Definition of the local reference frame

The simplest way to define the direction of the local reference frame for each beam element is to attach a third node n3n_3 to the two beam nodes n1n_1 and n2n_2.

The plane defined by the three nodes n1,n2,n3n_1,n_2,n_3 is the principal xˉyˉ\bar x\bar y-plane of the beam element.

The third node n3n_3 must not be collinear with nodes n1n_1 and n2n_2.

A MATLAB computer code is available for download.

MATLAB3D beam element — Euler–Bernoulli modelDownload ZIP

Subprogram princ, called by gen, computes the 3×33\times3 passive rotation matrices R0\boldsymbol{R}_0 for all beam elements.

06 Numerical integration

Although the integral

LBTDBdxˉ \int_L \boldsymbol{B}^T\boldsymbol{D}\boldsymbol{B}\,d\bar x

can be calculated analytically, in the MATLAB subprogram beam3d we prefer to compute it using the Gauss method with two integration points.

The result is exact because B\boldsymbol{B} is linear in xˉ\bar x, and therefore the integrand

BTDB \boldsymbol{B}^T\boldsymbol{D}\boldsymbol{B}

contains polynomials of at most second degree.

This approach is very convenient for nonlinear problems, for instance problems with large displacements.

Therefore,

LBTDBdxˉ=L2[B(xˉG1)TDB(xˉG1)+B(xˉG2)TDB(xˉG2)], \int_L \boldsymbol{B}^T\boldsymbol{D}\boldsymbol{B}\,d\bar x = \frac L2 \left[ \boldsymbol{B}(\bar x_{G1})^T \boldsymbol{D} \boldsymbol{B}(\bar x_{G1}) + \boldsymbol{B}(\bar x_{G2})^T \boldsymbol{D} \boldsymbol{B}(\bar x_{G2}) \right],

where xˉG1\bar x_{G1} and xˉG2\bar x_{G2} are the Gauss points:

xˉG1=L2(133), \bar x_{G1} = \frac L2 \left( 1-\frac{\sqrt3}{3} \right),

and

xˉG2=L2(1+33). \bar x_{G2} = \frac L2 \left( 1+\frac{\sqrt3}{3} \right).

The origin of the local xˉ\bar x-axis is at node 1.

07 Example

A clamped, curved helicoidal beam is analysed:

R=50 mm,H=80 mm, R=50\ {\rm mm}, \qquad H=80\ {\rm mm},

with a rectangular cross-section

h×b=9×6 mm. h\times b=9\times6\ {\rm mm}.

Refer to the MATLAB subprogram gen for all input data.

Figure 2
Figure 2.

The undeformed and deformed shapes of the beam are shown in the figure below; the displacement scale is 1.

Figure 3
Figure 3.

08 References

1. Andersen, L., Nielsen, S. R. K., Elastic Beams in Three Dimensions, Aalborg University, Department of Civil Engineering, 2008.