Rev.

00

Alandi (D), Pune,

An Autonomous Institute Affiliated to SPPU India

School of Civil & Mechanical Engineering

UNIT 4 : 1 D Elements: Beam & Truss

By
Prof. A.B. Belvekar-Patil

Prof. A.B. Belvekar-Patil

 SEM003 TYPES OF 1D ELEMENTS Rev. 00

DESIGN LAND Accuracy Matters

 SEM003 TYPES OF 1D ELEMENTS Rev. 00

Line elements – Springs, Bars, Trusses, beams

1. LINEAR ONE DIMENSIONAL ELEMENT TYPE

• Line element with two nodes one at each side

2. HIGHER ORDER ONE DIMENSIONAL ELEMENT TYPE

• Line element with three nodes one at each side and one is at center.
• Line element with four nodes where two nodes are internal nodes to the element.

Quadratic Element Cubic Element

DESIGN LAND Accuracy Matters

 SEM003 TYPES OF 1D ELEMENTS Rev. 00

Sr. No. Element Name Schematic Representation DOF

1 1D Bar Element or Linear Bar The element has two
Element Nodes with 1 DOF at each
node Total DOF= 2
2. 3 Noded bar element or quadratic The element has three
bar element nodes with 1 DOF at each
node Total DOF= 3
3. 4 Noded bar element or cubic bar The element has 4 nodes
element with 1 DOF at each node
Total DOF= 4
4 Beam Element or Hermite Element The element has 2 nodes
with 2 DOF at each node
i.e. Linear displacement
(V1 or V2) and rotation
(1 or 2) at each node
Total DOF= 4

DESIGN LAND Accuracy Matters

 SEM003 DISPLACEMENT FUNCTION Rev. 00
In FEA the first step is to discretize the given continuum into smaller number of parts called elements.
The displacement variations for these element are unknown hence trial function is assumed for the displacement of an
element. There are two functions available for assuming the displacement variation of an element.
1. Polynomial Displacement Function
2. Trignometric Displacement Function

Polynomial function is best suited for the displacement model since mathematical calculations are simpler. A polynomial
function of N th order will give an exact solution. But for all practical purpose the function is truncated to a finite order so as
to simplify the calculations involved. Hence the solution obtained is an approximate one.

For 1 D element polynomial displacement model is

U= a0 + a1X + a2X2 + a3X3 + …….+ anXn
The number of terms in polynomial indicate the shape of displacement for the element.
More number of terms used solution
converge to exact one.

DESIGN LAND Accuracy Matters

 SEM003 Convergence Criteria Rev. 00

• Convergence is a process in which the results obtained by FEA of physical problem reaches to the exact solution.
• Due to use of finite polynomial function in the calculation there is a difference between actual result and obtained results
in FEA.
• If the difference between these two keeps reducing the results of analysis are said to be converging.

DESIGN LAND Accuracy Matters

 SEM003 Convergence Criteria Rev. 00

• The conditions for convergence are as follows.

• The selected displacement model should be continuous:

• The displacement model selected should satisfy the rigid body displacement:

• The displacement model selected should satisfy the constant strain state.

• The displacement function selected should be independent of natural coordinate system.

DESIGN LAND Accuracy Matters

 SEM003 Global and local & Natural coordinate systems Rev. 00

• Global Coordinate system is used to reference all the elements of the body.

• Local coordinate system is one which is defined for a particular element. This can be used to reference any point within
that particular element . This system is related to only for an element.

• Natural coordinate system is a type of coordinate system in which local coordinate of dimensionless number is used. It
makes the integration & differentiation involved in FE formulation of stiffness matrix and load vector.

DESIGN LAND Accuracy Matters

 SEM003 Mesh Refinements Methods: h- Element vs. p- Element Rev. 00

h- Element p- Element
1. Element length is reduced for achieving 1. No change in Element length
convergence (increase no. of elements)
2. No change in Element order 2. Element order is increasing for achieving convergence.
3. Applicable for all types of element 3. Applicable for limited type of elements
4. It requires more time for convergence
5. Re meshing with smaller element length in critical 4. It requires less time for convergence
areas and rerunning the analysis is required. 5. No need to mesh again and again.
6. Applicable for all types of analysis
6. Limited to linear static and normal mode dynamics

DESIGN LAND Accuracy Matters

 SEM003 SHAPE FUNCTION Rev. 00

DESIGN LAND Accuracy Matters

 SEM003 TYPES OF 2D & 3D ELEMENTS Rev. 00

DESIGN LAND Accuracy Matters

 SEM003 SHAPE FUNCTION Rev. 00

DESIGN LAND Accuracy Matters

 SEM003 SHAPE FUNCTION Rev. 00

DESIGN LAND Accuracy Matters

 SEM003 SHAPE FUNCTION Rev. 00

DESIGN LAND Accuracy Matters

 SHAPE FUNCTION Rev. 00

Why shape functions?

Discretization leads to solution in the nodes, but no information concerning the space in between
Shape functions required to approximate quantities between nodes.
Underlying assumption of how quantities are distributed in an element (stiffness, mass, element loads; displacements,
strains, stress, internal forces, etc.)

Amit Belvekar-Patil
 SHAPE FUNCTION Rev. 00

Properties of shape functions

1. Number of Nodes = Number of Shape functions
2. Value of shape function at respective node = 1
3. Sum of All the shape functions equals to one.

Amit Belvekar-Patil
 SHAPE FUNCTION Rev. 00

Derivation of Shape function for Linear Bar element in Natural Coordinate System

Let,
N1 = Shape Function corresponding to Node 1
N2 = Shape Function corresponding to Node 2 N1= 1 N1= 0
N2=0 N2=1
Let us assume polynomial function in natural coordinate system is

N=α0 + α1 𝝃 ______ (A)

For Node 1, N=N1

At Node 1, 𝝃 = -1 , N1= 1 At Node 2, 𝝃 = +1 , N1= 0

1= α0 - α1-------------------------(a) 0 = α0 + α1--------------------(b)
α0 = 1/2 α1 = −1/2
Amit Belvekar-Patil
 SHAPE FUNCTION Rev. 00

Substituting the values of α0 & α1

𝟏 𝟏 𝟏−𝝃
N1 = − 𝝃 N1 =
𝟐 𝟐 𝟐

For Node 2, N=N2

At Node 2, 𝝃 = -1 , N2= 0 At Node 2, 𝝃 = +1 , N2= 1

0= α0 - α1-------------------------(c) 1 = α0 + α1--------------------(d)
Substituting the values of α0 & α1

α0 = 1/2 α1 = 1/2

𝟏 𝟏 𝟏+𝝃
N2 = + 𝝃 N2 =
𝟐 𝟐 𝟐

Amit Belvekar-Patil
 SHAPE FUNCTION Rev. 00

𝟏−𝝃 𝟏+𝝃
N1 = N2 =
𝟐 𝟐
N1= 1 N2= 1

Variation of Shape function for linear bar in natural coordinates

Amit Belvekar-Patil