FINITE ELEMENT DYNAMIC

ANALYSIS
Introductionn
● From introduction to finite element analysis the problems are time invariant ie
external excitation as well as the response of the system
● However in many practical situations such steady state conditions are
reached after a period of time in which the external disturbances cause the
system response to fluctuate with time
● When studying unstead and vibrating systems we are interested in finding the
response as a function of time given the disturbances
Vibration problems
● When structures are excited by time-varying forces, their responses are also time-dependent.

● Inertia and dissipation (damping) effects become important in such cases.

● Dynamic response analysis is more complex than static (time-invariant) analysis.

● In slowly varying forces, response may be treated as quasi-static — allowing static analysis.

● If excitation frequency is < 1/3 of the lowest natural frequency, static analysis is typically sufficient.

● The Magnification Factor (MF) quantifies how much larger the dynamic response is compared to the static one.
Types of dynamic analysis problems
1. Wave Propagation Problems
● Caused by sudden, high-intensity events (e.g., impacts, blasts).

● Structure doesn’t "feel" the load instantaneously—stress waves must travel

through the material.

● Short time scales comparable to wave travel time (microseconds to milliseconds).

2. Structural Dynamics Problems
● Caused by lower frequency, longer duration forces.

● The entire structure participates in the response at once.

● Time scales are longer (seconds or more).

1. Wave Propagation Problems

Require explicit time integration due to:

● High-frequency content.

● Need for small time steps and efficiency in handling rapid events.

Examples:

● Car crash simulations.

● Missile impact on a target.

Structural Dynamics Problems

Suitable for implicit time integration due to:

● Larger stable time steps.

● Quasi-static or slowly varying loads.

Examples:

● Automotive crankshaft vibrations.

● Earthquake response (≤ 25 Hz).

● Offshore structures under wave loading.

Structural Dynamics in Finite Element Analysis

● Focus is on finite element solutions for structural dynamics problems.

● Two types of vibration problems:

Free vibration: System vibrates due to initial disturbance.

Forced vibration: System vibrates under time-dependent loads.

Free Vibration Problems
● Treated as eigenvalue problems.

● Determine:

○ Natural (resonance) frequencies

○ Mode shapes

● Important for design and also useful in forced vibration

analysis (e.g., using mode superposition).
Damped Natural Frequency
● For a single degree of freedom (SDOF) system:

○ ωd— Damped natural frequency

○ ωn— Undamped natural frequency

○ ξ— Damping factor

● If

● Damping can often be neglected in natural frequency evaluation, but is

critical near resonance.
Finite Element Formulation

● Derived similarly to static problems:

Use weak form of the governing PDE.

Apply either direct weak form or energy-based methods.

● Leads to dynamic finite element equations for various structural systems.

Equations of motion based on the weak from
● Governing equation for free axial vibration of a rod

● Using the technique of separation of variables and assuming harmonic

vibration we have
● After substitution will have
● The weighted residual statement can be written as

● After integrating by parts

● Observing the equations we now have an additional mass density effects

● Equations for a single element
● For a typical bar element we have 2 nodes and axial deformation d.o.f at each
node

Following the Galerkin formulation the weight functions are the same as the shape
functions so
Writing the weak form equations with respect to W1 and W2

Where P is AE(dU/dx)
The sum of all coefficients in the mass matrix equals pAl, mass of the element

Equation 6,14 can be written as

Where sigma contains the nodal d.o.f

● The element mass matrix can be assembled following exactly the same
procedure as for the element stiffness matrix
● The forces column reduces to zero since free vibration doesn't have external
forces
● This will be the typical free vibration equation
Transverse vibration of a beam
This is how we would obtain the mass matrix of a Euler bernoulli beam element

Refer to 238 for the derivation

The euler bernoulli mass matrix will be

Equations of motion using Lagrange’s approach
The Lagrangian equation of motion in the independent generalised coordinates q is
given by

Where F is the generalised force in the coodinate q

Example

Consider a 2 d.o.f spring mass system shown below, x1(t) and x2(t) are the
independent generalised coordinates. The kinetic energy of the system is
The potential energy of the system is given by

For further use in lagrange’s equation we will derive T and pi with respect to the 4
different x (coordinates)
The required equation of motion

And on matrix form

 Consistent mass matrix for beam and bar elements
Bar element — obtained from the weak form
Beam element
Mass matrices
Represents how mass is distributed in a structure

Affects inertia forces in dynamic problems

Types:

● Consistent Mass Matrix: Derived from shape functions

● Lumped Mass Matrix: Diagonal approximation

Consistent Mass Matrix
● Derived using the same shape functions used for stiffness matrix

● Fully populated matrix

● More accurate for dynamic behavior

● Requires numerical integration

Lumped Mass Matrix
● Approximates mass as concentrated at the nodes

● Diagonal matrix

● Simpler, computationally efficient

● May be less accurate

Consistent and lumped matrices for 1D bar element
HRZ lumping scheme
● The idea is to use only the diagonal elements of the consistent mass matrix but
to scale them in such a way that the total mass of the element is preserved
● For example bar element HRZ lumped is

● Calculation of the HRZ lumped matrix for a beam element

 Beam element mass matrix from HRZ scheme
We write the diagonal elements of the beam element

The total mass is obtained by summing all the diagonal elements corresponding to
translational d.o.f in one direction

We get since the total mass is pAl so the scaling factor becomes
420/312
We get

Example : Consider a uniform cross sectional bar of length L made up of a material

whose Young’s modulus and density are known. Estimate the natural frequencies of the
axial vibration of the bar using both consistent and lumped mass matrices
Using one element meaning l = L and there are 2 nodes
Now let’s use a 2 element mesh meaning l=L/2
● A continuous rod has infinity d.o.f and natural frequencies
● In FEM simplifies the system either 1 element or 2 element and each having 1
natural frerquence and the other 2 respectively
● As the mesh is refined, more dof are introduced improving frequency prediction
and accuracy
Consider a simply supported beam

Using 3 types of mass matrices HRZ lumped, consistent and simple lumped find the first
3 natural frequencies and use 1 , 2 and 3 elements

Explain the findings from each mass matrix used

Form of finite equation for vibrating problems
The governing equation

Considering free vibrations there is no external force and the damping is neglected and
the equation becomes
Properties of eigenpairs
General form
Orthogonality Properties
Normalisation of mode shapes
Properties of eigenvalues and eigenvectors
Modal decoupling
Solution of eigenvalue problems
Eigenvalue problems arise in vibration and stability analyses.

General form:

Goal: Find eigenvalues λ and eigenvectors {x}.

Three main solution strategies:

● Transformation-Based Methods

● Vector Iteration Methods

● Determinant-Based Methods
Transformation based problems
Vector iteration based method
● Efficient for large sparse systems.

Power Iteration Method:

● Finds dominant eigenvalue λ

● Simple but slow convergence.

Inverse Iteration / Shifted Inverse Iteration:

● Finds eigenvalues near a shift σ

● Faster and can target specific modes.

Determinant-Based Methods
Solve characteristic equation:

Gives exact eigenvalues (in theory).

Not feasible for large systems due to:

● High computational cost

● Numerical instability

Mostly used for small models or symbolic derivation

Application in FEA
Transformation and iteration methods commonly used in:

● Modal analysis (natural frequencies and mode shapes)

● Buckling analysis (critical loads)

Software implementation:

● Abaqus/ANSYS use Lanczos, Subspace Iteration, etc.

● Solver choice affects speed and stability.

Transient vibration analysis
● Transient vibration: Time-dependent dynamic response due to
time-varying loads.

● Governed by the second-order differential equation:

● Involves inertia, damping, and stiffness effects.

Classification of Dynamic Response
● Free vibration: f(t)=0

○ Response depends on initial conditions.

● Forced vibration: f(t)≠0

○ External loading dominates the response.

● Transient vs. Steady-State:

○ Transient: short-lived, due to load changes or impacts.

○ Steady-state: continuous periodic loading.

Time Integration Methods
Direct Integration Methods
● Solve the equation step-by-step in time.

● Popular methods:

○ Newmark-β Method (implicit, stable)

○ Wilson-θ Method

○ Central Difference Method (explicit, conditionally stable)

Newmark Method (Common Choice)
● Two parameters: β,γ

● Unconditionally stable for β≥0.25 γ≥0.5

Damping in Structural Dynamics
● Damping represents energy dissipation.

● Crucial for accurate prediction of realistic dynamic response.

Common Damping Models:

1. Viscous Damping:

○ α, β determined to match modal damping ratios.

2. Modal Damping:

○ Damping applied in modal coordinates.

○ Easier for decoupled modal analysis.

3. Hysteretic / Structural Damping:

○ Energy loss modeled as complex stiffness.

○ Less common in time-domain analysis.

Rayleigh damping
FEA Implementation
Finite Element Form:

Input:

● Initial conditions:

● Time-dependent load history

● Damping parameters

Output:

● Displacement, velocity, acceleration time histories

Applications of Transient Analysis
Impact analysis (drop tests, crash)

Seismic response of structures

Blast loading and shock analysis

Switching loads in machinery

Mode superposition scheme
A reduced-order method for solving transient dynamic problems.

Based on using a limited number of mode shapes to approximate the total

response.

Efficient for large systems where full time integration is costly.

Key assumptions
System is linear (constant [M],[K],[C]).

Damping can be represented in modal coordinates.

External load is expanded in modal space.

Modal transformation
Displacement vector expressed as:

● [Φ]: Matrix of eigenvectors (mode shapes)

● {q(t)} Generalized coordinates (modal participation)

Substituting into the equation of motion:

transforms into decoupled equations in modal space.

Decoupled modal equations
Pre-multiply by [Φ]T

where:

● ωi : Natural frequency

● ζi - Modal damping ratio

● Modal force

Each mode behaves like a single-degree-of-freedom oscillator.

Time intergratrion in modal space
Time integration (e.g., Newmark method) performed on each modal equation:

● Only r significant modes used (r≪n).

Highly efficient, especially for systems where low-frequency modes

dominate.
Advantages of mode superposition
Reduced computational cost – fewer DOFs.

Decoupled system – each mode integrated separately.

Suitable for:

● Linear transient analysis

● Seismic and harmonic excitation

Effective when:

● Load is not high frequency

● First few modes dominate response

Limitations
Not suitable for nonlinear systems.

Requires accurate modal damping information.

May miss high-frequency content if insufficient modes are used.

Load must be expandable in modal space.

FEA software implimentation
Abaqus, ANSYS allow:

● Modal extraction (Lanczos, Subspace)

● Transient integration in modal space

● Selection of number of modes based on modal participation factors

Optionally combine with Rayleigh damping or modal damping ratios

question
Consider the undamped mass spring system below. Find the response of the
system when the first mass alone is given an initial displacement and left to vibrate
freely
REFERENCE
Textbook of Finite Element Analysis