5.

Triangular Element (Biaxial Element)

5.1 Introduction
A triangular element can be used for the analysis of plane strain and stress of continua, or similar
other problems. The triangular element is characterized by constant strain and stress throughout as
observed previously in the case of bar or uniaxial element.

Figure 1: Local Node Numbering Convention

Fig. (1) shows the local node numbering convention. The nodes are always numbered in succession
moving counter clockwise as shown. Each node of triangle as well as all points in the interior of
element can undergo an x-displacement u, and a y-displacement v.
The force-displacement analysis to determine the equation [𝐾]{𝑞} = {𝑅} involves again the
familiar five steps,
1. Assume a displacement function for the element using Rayleigh-Ritx form.
2. Derive the strain ε and the strain energy U.
3. Derive the P.E due to applied loads.
4. Sum the P.E terms to get total P.E π.
5. Minimize the total P.E π w.r.t the nodal displacements to get [𝐾]{𝑞} = {𝑅}.
Finally, the simultaneous linear algebraic equations obtained after assemblage are solved using a
numerical algorithm such as that based on Gaussian elimination.
5.2 Details of the five steps of FEM for the Triangular Element in the global coordination
Step 1: Assume a displacement function
We assume displacement functions u and v linear in X and Y as the simplest functions that yield
piecewise continuous first derivatives over the entire domain of the continuum.
𝑢(𝑋, 𝑌) = 𝑑1 + 𝑑2 𝑋 + 𝑑3 𝑌 (1)
𝑣(𝑋, 𝑌) = 𝐵1 + 𝐵2 𝑋 + 𝐵3 𝑌 (2)
The constant coefficients di and Bi often referred to as generalized coordinates, can be determined
as follows.
The X-component of displacement of each node may be written by using Eq. (1). Thus
 𝑢1 = 𝑑1 + 𝑑2 𝑋1 + 𝑑3 𝑌1
𝑢2 = 𝑑1 + 𝑑2 𝑋2 + 𝑑3 𝑌2
𝑢3 = 𝑑1 + 𝑑2 𝑋3 + 𝑑3 𝑌3
Or in matrix form,
𝑢1 1 𝑋1 𝑌1 𝑑1
𝑢
{ 2 } = [ 1 𝑋2 𝑌2 ] {𝑑2 } (3)
𝑢3 1 𝑋3 𝑌3 𝑑3
Or
{ 𝑞1 } = [𝐴]{𝑑} (4)
{ 𝑑} = [𝐴]−1 { 𝑞1 } (5)
Similarly
𝑣1 = 𝐵1 + 𝐵2 𝑋1 + 𝐵3 𝑌1
𝑣2 = 𝐵1 + 𝐵2 𝑋2 + 𝐵3 𝑌2
𝑣3 = 𝐵1 + 𝐵2 𝑋3 + 𝐵3 𝑌3
Or in matrix form,
𝑣1 1 𝑋1 𝑌1 𝐵1
𝑣
{ 2 } = [ 1 𝑋2 𝑌2 ] {𝐵2 } (6)
𝑣3 1 𝑋3 𝑌3 𝐵3
Or
{ 𝑞2 } = [𝐴]{𝐵} (7)
{ 𝐵} = [𝐴]−1 { 𝑞2 } (8)
Thus, the generalized coordinates di and Bi can be determined by pre-multiplying both sides of
equations (4) and (7) by the inverse of square matrix of nodal coordinates. Now the inverse of a
square non-singular matrix A is,
𝑎𝑑𝑗 𝐴 𝐴𝑐 𝑇
[𝐴]−1 = = (9)
|𝐴| |𝐴|

For the given matrix A

1 𝑋1 𝑌1
𝐴 = [ 1 𝑋2 𝑌2 ] (10)
1 𝑋3 𝑌3
The cofactor matrix is
 𝑋2 𝑌3 − 𝑋3 𝑌2 𝑌2 − 𝑌3 𝑋3 − 𝑋2
𝐴𝑐 = [ 𝑋3 𝑌1 − 𝑋1 𝑌3 𝑌3 − 𝑌1 𝑋1 − 𝑋3 ] (11)
𝑋1 𝑌2 − 𝑋2 𝑌1 𝑌1 − 𝑌2 𝑋2 − 𝑋1
Thus, for the given matrix A
𝑋2 𝑌3 − 𝑋3 𝑌2 𝑋3 𝑌1 − 𝑋1 𝑌3 𝑋1 𝑌2 − 𝑋2 𝑌1 𝑎1 𝑎2 𝑎3
𝐴𝑐 𝑇 1 1
[𝐴]−1 = = [ 𝑌2 − 𝑌3 𝑌3 − 𝑌1 𝑋1 − 𝑋3 ] = [𝑏1 𝑏2 𝑏3 ] (12)
|𝐴| |𝐴| |𝐴|
𝑋3 − 𝑋2 𝑌1 − 𝑌2 𝑋2 − 𝑋1 𝑐1 𝑐2 𝑐3
Where,
𝑎1 = 𝑋2 𝑌3 − 𝑋3 𝑌2 , 𝑎2 = 𝑋3 𝑌1 − 𝑋1 𝑌3 , 𝑎3 = 𝑋1 𝑌2 − 𝑋2 𝑌1 (13)
𝑏1 = 𝑌2 − 𝑌3 , 𝑏2 = 𝑌3 − 𝑌1 , 𝑏3 = 𝑋1 − 𝑋3
𝑐1 = 𝑋3 − 𝑋2 , 𝑐2 = 𝑌1 − 𝑌2 , 𝑐3 = 𝑋2 − 𝑋1
|𝐴| = det[𝐴] = (𝑋2 𝑌3 − 𝑋3 𝑌2 ) + 𝑋1 (𝑌2 − 𝑌3 ) + 𝑌1 ( 𝑋3 − 𝑋2 ) (14)
= 2(𝑎𝑟𝑒𝑎 𝑜𝑓 𝑡𝑟𝑖𝑎𝑛𝑔𝑙𝑒 )
Now from eq. (1)
𝑑1
𝑢 = 𝑑1 + 𝑑2 𝑋 + 𝑑3 𝑌 = [1 𝑋 𝑌] {𝑑2 } = [1 𝑋 𝑌]{𝑑} = [1 𝑋 𝑌][𝐴]−1 { 𝑞1 }
𝑑3
Thus,
𝑢1
𝑢 = [1 𝑋 𝑌][𝐴] −1 {
𝑞1 } 𝑤ℎ𝑒𝑟𝑒 𝑞1 = {𝑢2 } (15)
𝑢3
𝑣1
𝑣 = [1 𝑋 𝑌][𝐴] −1 {
𝑞2 } 𝑣
𝑤ℎ𝑒𝑟𝑒 𝑞2 = { 2 } (16)
𝑣3
Also
1
[1 𝑋 𝑌][𝐴]−1 = [𝑎 + 𝑋𝑏1 + 𝑌𝑐1 𝑎2 + 𝑋𝑏2 + 𝑌𝑐2 𝑎3 + 𝑋𝑏3 + 𝑌𝑐3 ]
|𝐴| 1
= [𝑁1 𝑁2 𝑁3 ]
𝑢1 𝑢1
𝑢 𝑢
𝑢 = [𝑁1 𝑁2 𝑁3 ] { 2 } and 𝑣 = [𝑁1 𝑁2 𝑁3 ] { 2 } (17)
𝑢3 𝑢3
Thus,
 𝑢1
𝑣1
𝑢 𝑁 0 𝑁2 0 𝑁3 0 𝑢2
{ }=[ 1 ] = [𝑁]{𝑞} (18)
𝑣 0 𝑁1 0 𝑁2 0 𝑁3 𝑣2
𝑢3
{𝑣3 }
Where [N] is the shape function matrix and {q} are the row vector of displacements.
Note:
It may be noted that the displacement of any point in the interior of triangle has been expressed in
terms of nodal displacements.
Before explaining step 2(derivation of strain and strain energy) we need a review of theory of
elasticity.
A Review of theory of Elasticity
Let u, v and w represent the X, Y and Z components of the displacement of a point in the
continuum. Then, the longitudinal strain components in the X, Y and Z directions are:
𝜕𝑢
𝜀𝑥 =
𝜕𝑥
𝜕𝑣
𝜀𝑦 =
𝜕𝑦
𝜕𝑤
𝜀𝑧 =
𝜕𝑧
The shearing strains are,
𝜕𝑢 𝜕𝑣
𝛾𝑥𝑦 = +
𝜕𝑦 𝜕𝑥
𝜕𝑣 𝜕𝑤
𝛾𝑦𝑧 = +
𝜕𝑧 𝜕𝑦
𝜕𝑢 𝜕𝑤
𝛾𝑧𝑥 = +
𝜕𝑧 𝜕𝑥
So,
 𝜕
0 0
𝜕𝑥
𝜕
𝜀𝑥 0 𝜕𝑦 0
𝜀𝑦 𝜕
𝜀𝑧 0 0 𝑢
𝜕𝑧
𝛾𝑥𝑦 = 𝜕 𝜕 {𝑣} (19)
0 𝑤
𝛾𝑦𝑧 𝜕𝑦 𝜕𝑥
𝜕 𝜕
{ 𝛾𝑧𝑥 } 0 𝜕𝑧 𝜕𝑦
𝜕 𝜕
[ 𝜕𝑧 0 ]
𝜕𝑥

The constitutive Equations:

The constitutive equations relate stress to strain. We shall consider a linearly elastic, homogeneous,
and isotropic material. The simplest stress-strain relationship is based on Hooke’s law which states
that stress is directly proportional to strain (a straight line passing through the origin in a stress,
strain plot).

For uniaxial deformations in X-direction,

𝜎 = 𝐸𝜀𝑥
Figure 2:Tensile Stress on X-face in x-direction

Where E is the elastic modulus of elasticity. It is to be noted that the strain in the Y and Z directions
is not zero, but rather
𝜀𝑦 = 𝜀𝑧 = −𝜈𝜀𝑥 (20)

Where 𝜈 is poisson’s ratio. Thus,

𝑇𝑟𝑎𝑛𝑠𝑣𝑒𝑟𝑠𝑒 𝑠𝑡𝑟𝑎𝑖𝑛 𝜀
𝜈 = − 𝐿𝑜𝑛𝑔𝑖𝑡𝑢𝑑𝑖𝑛𝑎𝑙 𝑠𝑡𝑟𝑎𝑖𝑛 = − 𝜀𝑇 (21)
𝐿

For a bar in tension, the longitudinal strain is positive and the transverse strain is negative (because
the width of the bar decreases). For compression we have the opposite situation with the bar
becoming shorter (negative longitudinal strain) and wider (positive transverse strain). Therefore,
for ordinary materials, poison’s ratio always has a positive value.
Plane Stress:
For this case the normal and shear stress components act in
a plane that is in two coordinate directions only. Let us
consider the X and Y directions only. Thus,

𝜎𝑧 = 0 = 𝑇𝑧𝑥 = 𝑇𝑧𝑦 (22) Figure 3: Normal and shear stresses on X and Y faces
The relations between stress and strain are
𝐸
𝜎𝑥 = (𝜀 + 𝜈𝜀𝑦 )
1 − 𝜈2 𝑥
𝐸
𝜎𝑦 = 1−𝜈2 (𝜀𝑦 + 𝜈𝜀𝑥 ) (23)

𝑇𝑥𝑦 = 𝐺𝛾𝑥𝑦

Where,
𝐸
𝐺=
2(1 + 𝜈)
Is the shear modulus.
𝜎𝑥 1 𝜈 0 𝜀𝑥
𝐸
{ 𝜎𝑦 } = (1+𝜈) [ 𝜈 1 0 ] { 𝜀𝑦 }
1−𝜈
(24)
𝑇𝑥𝑦 0 0 𝛾𝑥𝑦
2

{𝜎} = [𝑐] {𝜀} (25)

Where {𝜎} and {𝜀} are the stress and strain vectors repectively and [c] is the constitutive matrix
for plane stress. The strain energy stored in a differential element of volume dΩ of a continuum
subject to a uniaxial stress state is given as:
1
𝑑𝑈 = 2 𝜎𝑥 𝜀𝑥 dΩ (26)

For a biaxial stress state,

1
𝑑𝑈 = 2 (𝜎𝑥 𝜀𝑥 + 𝜎𝑦 𝜀𝑦 )dΩ (27)

For the case of pure shear,

1
𝑑𝑈 = 2 𝑇𝑥𝑦 𝛾𝑥𝑦 dΩ (28)

Where 𝛾𝑥𝑦 is the shear strain. The strain energy stored in the element dΩ under a general 3-
dimentional stress state is
1
𝑑𝑈 = (𝜎 𝜀 + 𝜎𝑦 𝜀𝑦 + 𝜎𝑧 𝜀𝑧 + 𝑇𝑥𝑦 𝛾𝑥𝑦 + 𝑇𝑦𝑧 𝛾𝑦𝑧 + 𝑇𝑧𝑥 𝛾𝑧𝑥 )dΩ
2 𝑥 𝑥
This equation, written in matrix form is
1
𝑑𝑈 = 2 {𝜎} {𝜀}𝑇 dΩ (29)

{𝜀} = [𝜀𝑥 𝜀𝑦 𝜀𝑧 𝛾𝑥𝑦 𝛾𝑦𝑧 𝛾𝑧𝑥 ]𝑇 , {𝜎} = [𝜎𝑥 𝜎𝑦 𝜎𝑧 𝑇𝑥𝑦 𝑇𝑦𝑧 𝑇𝑧𝑥 ]𝑇 (30)

From (29) we get,

 1
𝑈=∫ {𝜀}𝑇 {𝜎} dΩ (31)
2
𝛺

Where Ω is the region of integration. For plane stress calculations we have

{𝜀} = [𝜀𝑥 𝜀𝑦 𝛾𝑥𝑦 ]𝑇 , {𝜎} = [𝜎𝑥 𝜎𝑦 𝑇𝑥𝑦 ]𝑇 (32)

Substituting the constitutive relation for plane stress from (25) that is
{𝜎} = [𝑐] {𝜀}
Into (31) we get,
1
𝑈=∫ {𝜀}𝑇 [𝑐] {𝜀} dΩ (33)
2
𝛺

Eq (33) is preferably to (31) because it is in terms of 𝜀 only as compared to (31) which contains
both 𝜀 and 𝜎.
Step 2: Derive Strain and Strain Energy:
We can now develop our model of the plane stress state thickness of the continuum being small in
the z-direction from eq (18) and (19) leading to
𝜕 𝜕
0 0
𝜀𝑥 𝜕𝑥 𝜕𝑥
𝜕 𝑢 𝜕
{ 𝜀𝑦 } = 0 { }= 0 [𝑁]{𝑞} (34)
𝛾𝑥𝑦
𝜕𝑦 𝑣 𝜕𝑦
𝜕 𝜕 𝜕 𝜕
[ 𝜕𝑦 𝜕𝑥 ] [ 𝜕𝑦 𝜕𝑥 ]

𝜕
𝑢1
0 𝑣1
𝜀𝑥 𝜕𝑥
𝜕 𝑁 0 𝑁2 0 𝑁3 0 𝑢2
{ 𝜀𝑦 } = 0 [ 1 ] (35)
𝛾𝑥𝑦
𝜕𝑦 0 𝑁1 0 𝑁2 0 𝑁3 𝑣2
𝜕 𝜕 𝑢3
[ 𝜕𝑦 𝜕𝑥 ] {𝑣3 }

𝑢1
𝑣1
𝜀𝑥 𝑌2 − 𝑌3 0 𝑌3 − 𝑌1 0 𝑌1 − 𝑌2 0
1 𝑢
𝜀
{ 𝑦 } = |𝐴| [0 𝑋3 − 𝑋2 0 𝑋1 − 𝑋3 0 𝑋2 − 𝑋1 ] 𝑣2 (36)
𝛾𝑥𝑦 2
𝑋3 − 𝑋2 𝑌2 − 𝑌3 𝑋1 − 𝑋3 𝑌3 − 𝑌1 𝑋2 − 𝑋1 𝑌1 − 𝑌2 𝑢3
{𝑣3 }

So,
{𝜀} = [𝐵] {𝑞} (37)
Where {𝜀} is strain vector, [B] is strain displacement matrix and {q} are the nodal displacement.
This gives us strain vector. Now, to determine strain energy we substitute from eq (37) into (33),
yielding,
1
𝑈=∫ {𝑞}𝑇 [𝐵]𝑇 [𝑐][𝐵]{𝑞} dΩ
𝛺 2
1
𝑈 = 2 {𝑞}𝑇 [𝐵]𝑇 [𝑐][𝐵]{𝑞} ℎ𝐴 (38)

Where dΩ = hA here h is thickness of element and A is differential area

Note:
Note that matrices {q}, [B] and [C] are constants. This is obvious because matrix [B] contains only
the nodal coordinate values and [C] contains material constants. Thus, the triangular element used
here is a constant strain element.
Example 1: Strain in a triangular element
A steel triangular element of thickness h is located in the global XY plane
as shown in fig (4). Solve for the strain in the triangle in terms of nodal
displacements.

Solution: Figure 4:Triangular Element

We want to determine the strain vector

{𝜀} = [𝐵] {𝑞}
Now,

1 𝑌2 − 𝑌3 0 𝑌3 − 𝑌1 0 𝑌1 − 𝑌2 0
[𝐵] = [0 𝑋3 − 𝑋 2 0 𝑋1 − 𝑋3 0 𝑋2 − 𝑋1 ]
|𝐴|
𝑋3 − 𝑋2 𝑌2 − 𝑌3 𝑋1 − 𝑋3 𝑌3 − 𝑌1 𝑋2 − 𝑋1 𝑌1 − 𝑌2
Also,
|𝐴| = det[𝐴] = (𝑋2 𝑌3 − 𝑋3 𝑌2 ) + 𝑋1 (𝑌2 − 𝑌3 ) + 𝑌1 ( 𝑋3 − 𝑋2 )
= (6𝑥5 − 6𝑥2)2 + (2 − 5) + (6 − 6) = 12
Substituting for |A| and nodal coordinates in
{𝜀} = [𝐵] {𝑞}
We get,
 𝑢1
𝑣1
𝜀𝑥 −3 0 4 0 −1 0
1 𝑢2
𝜀
{ 𝑦}= [ 0 0 0 −4 0 4 ] 𝑣
𝛾𝑥𝑦 |𝐴| 2
0 −3 −4 4 4 −1 𝑢
3
{𝑣3 }
Example 2: Strain Energy in a triangular element
The strain energy in the triangular element is determined using Eq (38). This equation requires
the matrix product
[𝐵]𝑇 [𝑐][𝐵]
Evaluate this product for the element in Example 1 and then determine the strain energy in terms
of displacement of nodes.
Solution:
1. According to eq (38) the strain energy is
1
𝑈 = 2 {𝑞}𝑇 [𝐵]𝑇 [𝑐][𝐵]{𝑞} ℎ𝐴 (1)
2. From Example 1 the strain displacement matrix is,
𝑌2 − 𝑌3 0 𝑌3 − 𝑌1 0 𝑌1 − 𝑌2 0
1
[𝐵] = [0
|𝐴|
𝑋3 − 𝑋2 0 𝑋1 − 𝑋3 0 𝑋2 − 𝑋1 ] (2)
𝑋3 − 𝑋2 𝑌2 − 𝑌3 𝑋1 − 𝑋3 𝑌3 − 𝑌1 𝑋2 − 𝑋1 𝑌1 − 𝑌2
3. From eq (24) the constitutive matrix is
1 𝜈 0
𝐸
[𝐶] = (1+𝜈)
[𝜈 1 0]
1−𝜈
(3)
0 0 2
1−𝜈
4. Setting = ∅ we get,
2
−3 0 4 − 4𝜈 −1 − 4𝜈
𝐸
[𝐵][𝐶] = [ −3𝜈 0 4𝜈 −4 −𝜈 4 ] (4)
12(1−𝜈^2)
0 − 3∅ − 4∅ 4∅ 4∅ − 1∅

5. For Steel: E=30x 106 psi and 𝜈 =0.3

6. 2(area)= determinant >> 2A=12 >> A =6 in2
So, area = 1/2hA = 3h
7. The strain energy:
1
𝑈 = 𝑞 𝑇 [𝐵]𝑇 [𝑐][𝐵]𝑞ℎ𝐴
2
9 0 − 12 3.6 3 − 3.6
3.15 4.2 − 4.2 − 4.2 1.05
= {𝑞}𝑇 21.6 − 10.4 − 9.6 6.2 {𝑞}(6.87 𝑥 105 )(ℎ) 𝑖𝑛. 𝑙𝑏
6.6 − 2.6
[ 16.35 ]
Step 3: Derive Applied load Potential Energy Functions
We consider three types of element loading
1) Concentrated Forces applied to nodes (nodal forces):

As before the nodal P.E function Figure 5: Nodal Forces

𝑉𝑁 = −𝐹𝑥 𝑢 − 𝐹𝑦 𝑣 (39)

Where Fx and Fy are the global components of the force F, thus for node 1 in fig.5,
𝑉1 = −𝐹1𝑥 𝑢1 − 𝐹1𝑦 𝑣1

Extending this development to all force components acting at all nodes of element the complete
scaler P.E function
𝑉𝑁 = −{𝑞}𝑇 {𝐹𝑁 } (40)
Where,
{𝐹}𝑇𝑁 = {𝐹1𝑥 𝐹1𝑌 𝐹2𝑥 𝐹2𝑌 𝐹3𝑥 𝐹3𝑌 }
(2) Forces distributed along the edge of element
The distributed edge loading intensities are specified as force per
unit length 1-D or force per unit area 2-D. The force per unit area
at a particular point is usually called stress traction vector. The
force produced by the stress traction vector acting in the direction
defined by the unit vector on the area dA along one side of element
is given by
Figure 6:The distributed stress traction

𝑑𝐹 = 𝑇𝑑𝐴 𝑛̂ (41)
The differential potential energy function dVT for the distributed differential force can be expressed
as
d𝑉𝑇 = −(𝑇𝑥 𝑑𝐴)𝑢 − (𝑇𝑦 𝑑𝐴)𝑣

Where u and v are the displacements for the arbitrary point on the element boundary. Letting the
thickness of element be h, this can be written as
𝑇
𝑉𝑇 = − ∫[𝑢 𝑣] { 𝑥 } ℎ𝑑𝑆
𝑇𝑌
𝑆

Substituting from (18)

 𝑇
𝑉𝑇 = −ℎ ∫ {𝑞}𝑇 {𝑁}𝑇 { 𝑥 } 𝑑𝑆
𝑇𝑌
𝑆

Or since the vector {q} is independent of the integration variable, this can be written as

𝑇
𝑉𝑇 = −ℎ {𝑞}𝑇 ∫ {𝑁}𝑇 { 𝑥 } 𝑑𝑆 (42)
𝑆
𝑇𝑌

Example 3: Potential energy Function for distributed loading

The triangular element of example 1 is h units thick and has a uniform

stress traction on one side as shown in fig (7) Solve for the P.E function
for this distributed loading in terms of element nodes displacement
where
Figure7::Uniform stress traction on edge

𝑇
{𝐹} 𝑇 = −ℎ ∫ {𝑁}𝑇 { 𝑥 } 𝑑𝑆
𝑇𝑌
𝑆

Solution:
Using equation (42), that is
0
𝑇𝑋
𝑉𝑇 = −ℎ{𝑞} ∫ [𝑁]𝑇 {
𝑇
} 𝑑𝑠
𝑆 𝑇𝑌
𝑁1 0 ∫ 𝑁1 𝑑𝑦
0 𝑁1 0
0
0 𝑊
𝑉𝑇 = −ℎ{𝑞}𝑇 ∫ 𝑁2 { } 𝑑𝑦 = −𝑊ℎ{𝑞}𝑇 ∫ 𝑁2 𝑑𝑦
𝑆 0 𝑁2 0 0
𝑁3 0 ∫ 𝑁3 𝑑𝑦
[0 𝑁3 ] { 0 }
Where
0 5
1 1 5
∫ 𝑁1 𝑑𝑦 = ∫ (𝑎1 + 𝑥𝑏1 + 𝑦𝑐1 )𝑑𝑦 = ∫ (18 + (6)(−3) + 𝑌(0))𝑑𝑦 = 0
𝑆 |𝐴| 2 12 2
0 5
1 1 5 3
∫ 𝑁2 𝑑𝑦 = ∫ (𝑎2 + 𝑥𝑏2 + 𝑦𝑐2 )𝑑𝑦 = ∫ (−4 + (6)(4) + 𝑌(−4))𝑑𝑦 =
𝑆 |𝐴| 2 12 2 2
0 5
1 1 5 3
∫ 𝑁3 𝑑𝑦 = ∫ (𝑎3 + 𝑥𝑏3 + 𝑦𝑐3 )𝑑𝑦 = ∫ (−2 + (6)(−1) + 𝑌(4))𝑑𝑦 =
𝑆 |𝐴| 2 12 2 2
Thus,
0
0
3 1
𝑉𝑇 = − 𝑊ℎ{𝑞}𝑇
2 0
1
{0}
It is very important to note by comparing with equation (40) that that the result due to specified
distributed loading is equivalent to two concentrated forces each equal to half the resultant of the
distributed loading and applied to notes 2 and 3 in the X direction.
(3) Body forces:
A course acting per unit volume in an element or continuum is called
a body force gravitational loading is a common example of a body
force. Defining Bx as the body force in the X direction and by s the
force in y direction the differential potential function for these forces
may be written as
Figure 8: Differential Body forces

𝐵𝑋
𝑑𝑉𝐵 = −(𝐵𝑋 𝑑Ω)𝑢 − (𝐵𝑦 𝑑Ω)𝜗 = −[𝑢 𝜗 ]{ } ℎ𝑑𝐴 (43)
𝐵𝑌
Where 𝑢 and 𝜗 are the displacement of an arbitrary point in the interior of the element, and the
arbitrary element volume.
𝑑Ω= ℎ𝑑𝐴 h=thickness of the element. (44)
dA= arbitrary differential area
𝑢
Substituting from (18):{ } =[N]{q}, into (43), we get:
𝜗
0 𝐵 0 𝐵
𝑉𝐵 = −ℎ{𝑞}𝑇 ∫𝑆 [𝑁]𝑇 {𝐵𝑋 } 𝑑𝐴 = −{𝑞}𝑇 {𝐹}𝐵 , 𝑤ℎ𝑒𝑟𝑒 {𝐹𝐵 } = ℎ ∫𝑆 [𝑁]𝑇 {𝐵𝑋 } 𝑑𝐴 (45)
𝑌 𝑌

Example 4: Energy potential of distributed weight (body force)

The triangle shown in figure (8) is of uniform thickness and uniform
mass density equation (45) solve for distributed weight body humans it
to be acted in negative y –direction
Figure 9: Triangular element of uniform density

Solution: Using Eq. (45), we get

 𝑁1 0 0
0 𝑁1 ∫ 𝑁1 𝑑𝑦
0
0 𝑊 0
𝑉𝐵 = −ℎ{𝑞}𝑇 ∫ 𝑁2 { } 𝑑𝐴 = −ℎ𝜌𝑔{𝑞}𝑇
𝐴 0 𝑁2 −𝜌𝑔 ∫ 𝑁2 𝑑𝑦
𝑁3 0 0
[0 𝑁3 ] {∫ 𝑁3 𝑑𝑦}
Where, according to Eq (13),
𝑎1=𝑋2𝑌3 −𝑋3𝑌2 =𝑋2𝑌3 =6

𝑏1=𝑌2−𝑌3 =0

𝑐1=𝑋1−𝑋2=−𝑋2=−3

det=2(area of triangle) =6

⇒ ∫ 𝑁1 𝑑𝐴 = 1, ∫ 𝑁2 𝑑𝐴 = 1, ∫ 𝑁3 𝑑𝐴 = 1
Note that the total weight of the triangle
𝑊
W= (𝜌𝑔)(ℎ ∗ 𝑎𝑟𝑒𝑎) = (𝜌𝑔)(ℎ ∗ 3) = 𝜌𝑔ℎ = 3

Triangular element of uniform density:

5
1
∫ 𝑁1 𝑑𝐴 = ∫ (𝑎1 + 𝑥𝑏1 + 𝑦𝑐1 )𝑑𝑦
𝑑𝑒𝑡 2
3 2
1
∫ ∫ (𝑎1 + 𝑥𝑏1 + 𝑦𝑐1 )𝑑𝑦
𝑑𝑒𝑡 0 2
3
3 2
1 𝑦2
∫ ∫ (𝑎1 𝑦 + 𝑥𝑦𝑏1 + 𝑐1 ) 𝑑𝑦
𝑑𝑒𝑡 0 2 2
3
3 2
1
∫ ∫ (3𝑎1 + 3𝑏1 + 4𝑐1 )𝑑𝑦
𝑑𝑒𝑡 0 2
3

Hence the body force P.E is

0
1
𝑇 0 𝑊
𝑉𝐵 = {𝑞}
1 3
0
{1}
Comparing this with
 𝑉𝑁 = −{𝑞}𝑇 {𝐹𝑁 },
We not that the body force is equivalent to the nodal forces in the negative by direction and at
every node equal to one third of the total weight.
Step 4: Sum the energy terms to get total P.E:
𝜋 = 𝑢 + 𝑉𝑁 + 𝑉𝐵 + 𝑉𝑇
𝐹1𝑥
𝐹1𝑦
1 0 𝑇 𝑇 [𝐶][𝐵]{𝑞}𝑑Ω 𝑇 𝐹2𝑥 𝑇 0[𝑁]𝑇 𝑇𝑋 0 𝐵
∫ {𝑞} {𝐵} − {𝑞}
𝐹2𝑦 − ℎ{𝑞} ∫𝑆 { } 𝑑𝑠 − ℎ{𝑞}𝑇 ∫𝑆 [𝑁]𝑇 { 𝑋 } 𝑑𝐴 (46)
2 Ω 𝑇𝑌 𝐵𝑌
𝐹3𝑥
𝐹
{ 3𝑦 }
Step5: Apply the principle of minimum P.E

𝑢1 𝐹1𝑥
𝜗1 𝐹1𝑦
0 𝑢2 𝐹2𝑥 0 𝑇 0 𝐵
∫Ω {𝐵}𝑇 [𝐶][𝐵]𝑑Ω − {𝑞}𝑇 𝜗 = 𝐹 ℎ ∫𝑆 [𝑁]𝑇 {𝑇𝑋} 𝑑𝑠 + ℎ ∫𝑆 [𝑁]𝑇 {𝐵𝑋 } 𝑑𝐴 (47)
2 2𝑦 𝑌 𝑌
𝑢3 𝐹3𝑥
{𝜗3 } {𝐹 }
3𝑦

⇒ [𝐾]{q}={𝐹𝑁 } + {𝐹𝑇 } + {𝐹𝐵 } (48)

Where all the quantities referred to the element these quantities are than assemble to get the
representation for the whole continuums.
⇒ [𝐾]{r}={𝑅}, where:
⇒{R}={𝐹𝑁 } + {𝐹𝑇 } + {𝐹𝐵 } (49)

Example 5: Nodal displacement due to distributed loading

Solve for the nodal displacement of the triangle shown in the (figure 10) subjected to a uniform
distributive loading the same triangular element as used in example 1 with the loading of example
3 the material is steel and it is h unit thick.
Solution: Using Eq (48);
⇒ [𝐾]{q}={𝐹𝑁 } + {𝐹𝑇 } + {𝐹𝐵 }
Now, {𝐹𝐵 } = 0
 Figure 10:Triangular element with zero traction on one side

Using the equation (47), and substituting for 𝑉𝑇 from Example 3, we have
𝐹1𝑥 0
𝐹1𝑦 0
0
𝐹 3 1
{𝐵}𝑇 [𝐶][𝐵]ℎ ∫ 𝑑𝐴{𝑞} = 2𝑥 + 𝑊ℎ
𝐴 𝐹2𝑦 2 0
0 1
{ 0 } {0}
0 1 1
We note that ∫𝐴 𝑑𝐴{𝑞} = 𝐴 = 2 (𝑑𝑒𝑡) = 2 (12) = 6

9.00 000 12.0 3.60 3.00 −3.60

000 3.15 4.20 −4.20 −4.20 1.05
000 21.6 −10.4 −9.60 6.20 =
For example 2 : {𝐵}𝑇 [𝐶][𝐵] = (2.289 × 105 ) 000
000 000 000 21.6 6.80 −17.40
000 000 000 000 6.60 −2.60
[ 000 000 000 000 000 16.35 ]
𝐾𝑥𝑦 = 1.37 × 106

Now we substitute the displacement boundary conditions (row and column) reduce the equation
and solve for the unknown displacement
𝐾55 𝐾56 𝑢3 3 1 6.6 −2.6 3 1
1.37 × 106 ℎ [ ] {𝜗 } = 𝑊ℎ { } ⇒ 1.37 × 106 ℎ [ ] = 𝑊ℎ { }
𝐾65 𝐾66 3 2 0 −2.60 16.4 2 0
𝑢3 = 1.76 × 10−7 𝑊 𝑖𝑛
The resulting displacement are: { }
𝜗3 = 2.80 × 10−8 𝑊 𝑖𝑛
Using

−1 1 −1
𝑢1
0 0 0 𝜗1
𝐸𝑋 4 3 12
𝑢2
−1 1
{𝐸} = [𝑞][𝐵], Example 1 gives { 𝐸𝑌 } = 0 0 0 0 3 𝜗
3 2
𝛾𝑋𝑌 −1 −1 1 1 −1 𝑢3
[0 4 3 3 3 12 ] {𝜗 }
3

Substituting the displacements and then reducing the matrix gives:

 𝐸𝑋 = 1.46 × 10−8 𝑊
{𝐸𝑌 = 0.93 × 10−8 𝑊 }
𝛾𝑋𝑌 = 5.6 × 10−8 𝑊
Then {𝜎}=[C]{E}: gives,
{𝜎𝑥 = 0.39 𝑊𝑝𝑠𝑖}
{𝜎𝑦 = 0.16 𝑊𝑝𝑠𝑖}

{𝜏𝑥𝑦 = 0.65 𝑊𝑝𝑠𝑖}

Note:
The triangular element yields constant strain and constant stress this is similar to the situation from
a uniaxial element and is a sequence of using a line displacement model.