Lecture 9

Transient Analysis
16.0 Release

ANSYS Mechanical
Linear and Nonlinear Dynamics
 … 1. Definition and Purpose
• Transient dynamic analysis is a technique used to determine the dynamic
response of a structure under the action of any general time-dependent
loads.
– also known as time-history analysis or transient structural analysis
– can include inertia and/or damping effects
– can include nonlinear effects

• Typically more involved than a static analysis

• generally requires more computer resources and more "engineering" time

2 © 2015 ANSYS, Inc. March 1, 2016

 … Definition and Purpose
Hints for a preliminary work to understand the physics of the problem:

• Simplify: Analyze a simpler model first to provide good insight into the
behavior at minimal cost.
• Nonlinearities: If nonlinearities exist, first assess their affect using a static
analysis
– It may be possible to eliminate nonlinearities in the dynamic analysis.
• Modal Analysis: assess dynamics behavior using modal analysis to determine
the natural frequencies and mode shapes.
– natural frequencies are also useful for calculating the correct integration time step.

3 © 2015 ANSYS, Inc. March 1, 2016

 … Definition and Purpose
• The non-linear governing equation for the Transient Dynamic Analysis is:
u
Finertia F Fstiffness F
     
damping
    
applied

M u C u K (u )u  F t 

[M]: is structural mass matrix uሷ : is nodal acceleration vector

[C]: is structural damping matrix uሶ : is nodal velocity vector
[K]: is structural stiffness matrix {u}: is nodal displacement vector
{F}: is the load vector (t): is time

4 © 2015 ANSYS, Inc. March 1, 2016

 … Definition and Purpose
Fin ertia F Fstiffn ess F
     
d amp in g
   
ap p lied

M u C u K (u )u  F t 

• At any given time, t, these equations are thought of as a set of "static"
equilibrium equations that take into account
• inertia forces M uሷ and
• damping forces C uሶ .
• To solve these equations ANSYS uses:
• the Newmark time integration method or
• an improved method called HHT
• Integration time step: time increment between successive time points.
t  t n  t n 1
5 © 2015 ANSYS, Inc. March 1, 2016
 … Definition and Purpose
• The Newmark method uses finite difference expansions in the time interval Δt, with
a primary aim of computing displacements 𝒖𝒏+𝟏 :
un 1  un  1   un   un 1t
un 1  un  un t  0.5   un   un 1t 2
• 𝒖𝒏+𝟏 can then be obtained such that,

a0 M   a1 C   K  un 1  F a  M a0 un  a2 un  a3 un 

 C a1 u n   a4 u n   a5 u
n 

where integration constants a0 through a5 are functions of  and t. ( represents

numerical damping and can be directly input in the analysis settings)

6 © 2015 ANSYS, Inc. March 1, 2016

 … Definition and Purpose
• In ANSYS Mechanical, the newmark parameters
, and  are calculated using:

1
1   2 ,   1  
4 2

•  is a numerical damping value (amplitude decay

factor).

7 © 2015 ANSYS, Inc. March 1, 2016

 … 2. Solution Techniques in Mechanical
Technique
The Full Method  Relatively simple to set-up
 Uses full matrices [K, M, and C].
 Allows all types of nonlinearities.
 Calculates displacements and stresses in a single pass.
 Accepts most load types (e.g., nodal forces, non-zero displacements, element loads, tabular
boundary condition, etc.)
 Allows effective use of solid-model loads.
x Typically computationally more expensive than the mode-superposition method

The Mode-  Faster and less expensive than full method

Superposition  Allows damping as a function of frequency
Method x Time step must remain constant (i.e., automatic time stepping is not allowed).
x The only nonlinearity allowed is simple node-to-node contact (gap condition).
x Does not accept imposed (nonzero) displacements.

8 © 2015 ANSYS, Inc. March 1, 2016

Transient Analysis
Full Method
16.0 Release

ANSYS Mechanical
Linear and Nonlinear Dynamics
 … Full Method
• Allows all types of nonlinearities.
• Accepts most load types (e.g., nodal forces, non-zero
displacements, element loads, tabular boundary condition, etc.)
• Uses full matrices [K, M, and C]
• The mesh should be fine enough to resolve the highest mode of
interest.

10 © 2015 ANSYS, Inc. March 1, 2016

 … 3. Nonlinearities in Transient Analysis
Linear analysis:
F
• Force and displacement are linearly related.
• structural stiffness (K) is constant K
u
Nonlinear analysis:
• Force and displacement are not linearly related KT
• stiffness KT is not constant F
• it changes through the load path.
• Newton-Raphson method is used to solve nonlinear
analysis u
11 © 2015 ANSYS, Inc. March 1, 2016
 … Nonlinearities in Transient Analysis
Three sources of structural nonlinearity:

1. Changing status:
Contact pair either in or out of contact status,
tension-only cable is either slack or taut, frictional
contact, etc.

12 © 2015 ANSYS, Inc. March 1, 2016

 … Nonlinearities in Transient Analysis
Three sources of structural nonlinearity:
2. Geometric nonlinearities:
Changing geometric configuration (large deformation)
causes structure to respond nonlinearly (classic fishing
pole behavior)

13 © 2015 ANSYS, Inc. March 1, 2016

 … Nonlinearities in Transient Analysis
Three sources of structural nonlinearity:
3. Material nonlinearities:
Nonlinear stress-strain relationships (metal
plasticity, creep, hyperelasticity, etc.)

14 © 2015 ANSYS, Inc. March 1, 2016

 … Available Contact Types
All five contact types are allowed:

Contact Type Iterations Normal Behavior (Separation) Tangential Behavior (Sliding)

Bonded 1 No Gaps No Sliding
No Separation 1 No Gaps Sliding Allowed
Frictionless Multiple Gaps Allowed Sliding Allowed
Rough Multiple Gaps Allowed No Sliding
Frictional Multiple Gaps Allowed Sliding Allowed

• Bonded and No Separation contact are linear and require only 1

iteration.
• Frictionless, Rough and Frictional contact are nonlinear and require
multiple iterations.
15 © 2015 ANSYS, Inc. March 1, 2016
 … Available Contact Types
• Interface Treatment
– Nonlinear contact types allow an “interface treatment” option:
– “Adjusted to Touch”: ANSYS closes any gap to a just touching position.
– “Add Offset”: input zero or non-zero value for initial adjustment.

16 © 2015 ANSYS, Inc. March 1, 2016

 … 4. Newton-Raphson Technique
• In a nonlinear analysis, relationship between load and displacement
cannot be determined with a single solution based on initial stiffness.
• Newton-Raphson method uses a series of linear approximations with
corrections

17 © 2015 ANSYS, Inc. March 1, 2016

 … 4. Newton-Raphson Technique
• Total external load Fa is applied in iteration 1 and
displacements (x1) are calculated
• Using x1, internal forces F1 at iteration 1 are calculated.
• If Fa  F1, the system is not in equilibrium.
• Difference of between applied external and calculated
internal forces (Fa - F1) are the out-of-balance or
residual forces.
• If residual forces are within an acceptable tolerance,
the solution is converged
• If residual forces are outside an acceptable tolerance,
the solution is not converged, so a new stiffness matrix
is assembled and the process is repeated
• In this example, the system achieves convergence
after iteration 4.
18 © 2015 ANSYS, Inc. March 1, 2016
 ... Load Steps and Substeps
Each solution point is defined in terms of a unique monotonically increasing
time and a unique load step and substep combination
• Load steps are typically used to differentiate changes in general loading.
– Fa and Fb are loadsteps
• Substeps are typically used to increment loading within load steps
– Because of the complex response, it desirable to
incrementally apply the load. Fb
Fb2
– For example, Fa1 may be near 50% of the Fa load. Fb1
Fa
– After Fa1 is converged, full Fa load is applied. Fa1

– Fa has 2 substeps while Fb has 3 substeps in this example

ua ub

19 © 2015 ANSYS, Inc. March 1, 2016

 ... Equilibrium Iterations
In a nonlinear solution, equilibrium iterations are corrective solutions needed
for convergence using the Newton-Raphson method
– equilibrium iterations occur at the same time point (and same load step and
substep)
– In this example, the iterations between the dotted lines indicate equilibrium
iterations.
F
F
b
b2
Fb1
Fa
Fa1

ua u
b

20 © 2015 ANSYS, Inc. March 1, 2016

 5. Analysis Settings in Full Transient Analysis
• In a full transient analysis, the control
options are set under “Analysis Settings”
– Step Controls
– Solver Controls
– Restart Controls
– Nonlinear Controls
– Output Controls
– Damping Controls
– Analysis Data Management

21 © 2015 ANSYS, Inc. March 1, 2016

 … 5.1 Time Step Size
Integration time step Δt: time increment between successive
time points.
• It is one of the most important parameters in a transient
structural analysis
• It must be small enough to:
• correctly describe the time-varying loads
• capture the dynamic response
• Running a preliminary modal analysis is suggested.
• Time step size controls the accuracy and convergence
behavior of nonlinear systems (Newton-Raphson
method)
22 © 2015 ANSYS, Inc. March 1, 2016
 … Automatic Time Stepping
• By default, transient structural analysis uses automatic time-stepping.
– Proper selection of the initial, minimum, and maximum time steps is
important.

• Auto Time Stepping automatically adjusts the time step size (hence the load
increment) throughout the solution.
– Smaller increments when convergence is difficult, larger increments when
convergence is easy.

• Transient structural analysis uses implicit time integration (time step is

usually large).

23 © 2015 ANSYS, Inc. March 1, 2016

 … Automatic Time Stepping
• It is recommended to use automatic time-
stepping (default):
– The maximum time step can be chosen based on
accuracy concerns.
– The minimum time step is input to prevent
Mechanical from solving indefinitely.
Load
(1/100 or 1/1000 of the initial time step)
– A general suggestion for selection of the initial
time step is to use the following equation:
1
tinitial 
20 f response
– fresponse is the frequency of the highest mode of Time
interest (obtained from Modal Analysis)

24 © 2015 ANSYS, Inc. March 1, 2016

tinitial tmin tmax
 … Automatic Time Stepping
• For Auto Time Stepping = Program Controlled (Default), Mechanical will
automatically set specifications depending on the nature of the nonlinearity
in the model.
- User should always verify that these values are adequate by checking the Solution
Information folder at the beginning of the run and watching for bisections.

25 © 2015 ANSYS, Inc. March 1, 2016

 … Time Integration
• Time Integration: indicates whether a solution
step should include transient effects (e.g.,
structural inertia).
• Transient effects can be turned “off” to set up the
Initial Conditions for a transient analysis.
- On: Default for transient analyses.
- Off: Do not include structural inertia or thermal
capacitance in solving this step.
• Note: with Time Integration “Off”, Mechanical
does not compute velocity results. Therefore,
damping forces, which are derived from velocity
will equal zero.
27 © 2015 ANSYS, Inc. March 1, 2016
 … 5.2 Solver Controls – Solver Type
• Solver Type: a reference to the way ANSYS builds
the stiffness matrix for each Newton-Raphson
equilibrium iteration:
- Direct (Sparse):
- more robust
- recommended for challenging nonlinear models
with non-continuum elements (shells and beams).
- Iterative (PCG):
- more efficient,
- recommended for large bulk solid models
dominated by linear elastic behavior.
- The default “Program Controlled”: automatically
selects a solver based on the problem.
28 © 2015 ANSYS, Inc. March 1, 2016
 … Weak springs
• Weak Springs: to prevent numerical instability,
while not having an effect on real world
engineering loads.

29 © 2015 ANSYS, Inc. March 1, 2016

 … Large deflection
• Large Deflection: If set to “ON”
- Stiffness matrix is adjusted over multiple
iterations to account for changes such as
- large deflection
- large rotation
- large strain.

- Stress stiffening effect is included.

- Spin softening effect is included.

30 © 2015 ANSYS, Inc. March 1, 2016

 … 5.3 Nonlinear Control
Convergence Criteria
• Tolerances on Convergence are calculated
automatically.
• Tolerances are used during the Newton-Raphson
process to dictate when a model is Converged or
“balanced”
– The default convergence criterion works very well for
most engineering applications.
– For special situations, users can override these
defaults to tighten or loosen the convergence
tolerance.
– A tighter tolerance gives better accuracy, but can
make convergence more challenging
31 © 2015 ANSYS, Inc. March 1, 2016
 … Convergence Criteria

A moment balance is included if

rotational degrees of freedom are
present.

32 © 2015 ANSYS, Inc. March 1, 2016

 … Line Search
• It can be useful for enhancing convergence.
• When active, line search multiplies the displacement increment by a
program-calculated scale factor between 0 and 1, whenever a stiffening
response is detected.

33 © 2015 ANSYS, Inc. March 1, 2016

 … Line Search
• You might consider setting Line Search on in the following cases:
– If the structure is force-loaded.
– If the structure is "flimsy" structure which exhibits increasing stiffness (such
as a fishing pole).
– If the convergence pattern is oscillatory.

34 © 2015 ANSYS, Inc. March 1, 2016

 … Stabilization
• Nonlinear stabilization technique can help
achieve convergence.
• Adds artificial dampers to all of the DOFs in the
system.
• Keys for controlling nonlinear stabilization:
– Off - Deactivate stabilization (Default).
– Constant - The energy dissipation ratio or damping
factor remains constant during the load step.
– Reduce - The energy dissipation ratio or damping
factor is reduced linearly to zero at the end of the
load step from the specified or calculated value.

35 © 2015 ANSYS, Inc. March 1, 2016

 … 5.4 Damping Control
 i
C    M    K  Damping Ratio i 
2i

2

• Alpha and Beta damping are used to define

Rayleigh damping constants α and β. 1
• [C] is calculated using α and β to multiply the
mass matrix [M] and stiffness matrix [K].
• They can be input via: β
α
– [1] directly as global damping value (Details
section of Analysis Settings).
– [2] material-dependent damping value (Mass-
Matrix Damping Multiplier, and k-Matrix
Damping Multiplier) 2
α
β
36 © 2015 ANSYS, Inc. March 1, 2016
 … Numerical Damping
• Also referred to as amplitude decay factor ()
• Controls numerical noise produced by the higher frequencies of a structure
• A default value of 0.1 is used for Transient Structural analysis
• A default value of 0.005 is used for Transient Structural analysis using a linked
Modal analysis system).

37 © 2015 ANSYS, Inc. March 1, 2016

 … 6. Initial Conditions
2
• Initial conditions  conditions at Time = 0
• The default initial condition is that the
structure is “at rest”, that is, both initial
displacement and initial velocity are zero. 1
3
• Initial velocity can be scoped to one or more
parts of the structure.

• The remaining parts of the structure which 4

are not part of the scoping will retain the
“at rest” initial condition. 5

38 © 2015 ANSYS, Inc. March 1, 2016

 … Initial conditions using steps
• Initial conditions can also be specified using
two steps:
1. Initial Displacement = 0, Initial Velocity ≠ 0
Initial displacement of 0.005 mm
The first step end time = 0.001 sec.
Initial velocity of (0.005/0.001) = 5 mm/sec.
– Deactivate the specified displacement load in
the second step so that the part is free to
move with the specified initial velocity.

Make sure that time integration effects are

turned on for the second step.
39 © 2015 ANSYS, Inc. March 1, 2016
 … Initial conditions using steps
2. Initial Displacement ≠ 0, Initial Velocity ≠ 0
Initial velocity of 0.5 mm/sec.
Initial displacement of 0.1 mm
The first step end time =
(0.1/0.5) = 0.2 sec.

3. Initial Displacement ≠ 0, Initial Velocity = 0

Initial displacement of 0.1 mm
The first step end time = 0.001 sec.
Note the step application
of the displacement.
40 © 2015 ANSYS, Inc. March 1, 2016
 … 7. Loads and Supports
• All inertial and structural loads, and all
structural supports are allowed.
• Joint Loads are used to kinematically drive
joints.
• Magnitude could be:
• Constant,
• Tabular (time Varying), or
• Function

41 © 2015 ANSYS, Inc. March 1, 2016

 … Joint Loads
• Joints define the allowed motion (kinematic
constraint) on
surface(s).
• Various types of joints can be defined for flexible or
rigid bodies
• Absolute DOF are specified.
• A Joint Load object is used to apply a kinematic
driving condition to a single DOF on a Joint object.
• For translation DOF
– displacement, velocity, acceleration, or force is
applied.
• For rotation DOF
– rotation, angular velocity, angular acceleration, or
moment is applied.

42 © 2015 ANSYS, Inc. March 1, 2016

Workshop 9 A
Full Transient Analysis
16.0 Release

ANSYS Mechanical
Linear and Nonlinear Dynamics
Transient Analysis
Mode Superposition Method
16.0 Release

ANSYS Mechanical
Linear and Nonlinear Dynamics
 … Mode Superposition Method

• Faster and less expensive than full method

• Allows damping as a function of frequency

45 © 2015 ANSYS, Inc. March 1, 2016

 … 8. Mode Superposition Method
• Mode superposition method use the natural frequencies
and mode shapes from the modal analysis to characterize
the transient dynamic response of a structure.

• It scales the mode shapes obtained from a modal analysis

and sums them to calculate the dynamic response.

46 © 2015 ANSYS, Inc. March 1, 2016

 … Mode Superposition Method
• Example:

y1 + y2 =

1 2

– Here, the sum of mode shape 1 and mode shape 2 approximates the final
response. Since mode shapes are relative, the coefficients y1 and y2 are
required.
– Mode shapes (eigenvectors) are also known as generalized coordinates, and
in this case, coefficients y1 and y2 are the DOF.

47 © 2015 ANSYS, Inc. March 1, 2016

 … Mode Superposition Method
• The equations of motion are copied below (linear only):
M u C u K u  F t 
• Instead of using nodal coordinates, generalized coordinates will be used.
Assume that the deformation {u} can be constructed from a linear
combination of mode shapes fi:
n
u   yi i      n is the number of modes
i 1

• From this, the equations of motion can be written in generalized coordinates

as the following (with some additional substitutions not shown here for
brevity):
yi  2ii y
 i  i2 yi  f i

48 © 2015 ANSYS, Inc. March 1, 2016

 … Mode Superposition Method
• The modal equations of motion are repeated below:
yi  2ii y
 i  i2 yi  f i
• Advantages of this approach:
– If a model consists of m DOF, instead of solving m equations, the modal equations only solve n DOF,
where n represents the number of modes calculated in the modal analysis
– If 200 modes are extracted for a 1 million DOF model, instead of solving 1 million equations for
the dynamic analysis, a user only solves for 200 DOF!
– Because of the reduced number of DOF, the solution is very fast
• Points to remember:
– Linear combination means that only linear behavior is allowed
– Time step is fixed – use same guidelines as before.
– The results are based on a truncated set of modes.
– A 1 million DOF model has 1 million modes, but one typically solves for a fewer number of
modes. This means that a large enough number of modes is required to obtain good accuracy.
49 © 2015 ANSYS, Inc. March 1, 2016
 … Mode Superposition Method
• Setup a mode-sup transient analysis in the schematic by linking a modal system to a
transient structural system at the solution level.

• Notice in the transient branch, the modal analysis result

becomes an initial condition.

50 © 2015 ANSYS, Inc. March 1, 2016

 … 9. Analysis Settings for Mode-Sup
• Under “Analysis Settings” in mode-sup transient
dynamics, there are many control options that
need to be considered:
– Step Controls
– Time step must remain constant (i.e., automatic time
stepping is not allowed).
– Auto time stepping is turned off
– Substeps or time step value is defined and is applicable to all
the load steps
– The time integration is turned on by default. g
– Damping Control β
α
– The damping matrix is not explicitly computed, but rather the
damping is defined directly in terms of a damping ratio d.
– The damping ratio id for mode i is the combination of:
 
 id  g   i
2i 2
51 © 2015 ANSYS, Inc. March 1, 2016
 … 10. Restriction in Mode-sup Transient
• Time step must remain constant (i.e., automatic time stepping is not
allowed).
• The only nonlinearity allowed is simple node-to-node contact (gap
condition).
• Does not accept imposed (nonzero) displacements.
• All contact will behave as bonded or no separation in a modal analysis:
• If a gap is present:
– Nonlinear contacts will be free (no contact).
– Bonded and no separation contact will depend on the pinball size.
Modal Analysis
Contact Type
Initially Touching Inside Pinball Region Outside Pinball Region
Bonded Bonded Bonded Free
No Separation No Separation No Separation Free
Rough Bonded Free Free
Frictionless No Separation Free Free
Frictional Bonded Free Free

52 © 2015 ANSYS, Inc. March 1, 2016

 … 11. Pre-stress MSUP Transient Analysis
1. Pre-Stress MSUP Transient Analysis.
You can now perform a Mode Superposition Transient Structural
analysis that is linked to a pre-stressed Modal analysis.

53 © 2015 ANSYS, Inc. March 1, 2016

 … Pre-stress MSUP Transient Analysis
Pre-Stress MSUP Transient Analysis.

54 © 2015 ANSYS, Inc. March 1, 2016

 … Pre-stress MSUP Transient Analysis
2. Output Controls (Analysis Settings>Output Control> Nodal forces)
– Calculate Reactions on Constrained Nodes.
– Only constrained nodes are used in the calculation of reaction forces and
moments.
– The result file size and processing time are reduced.

55 © 2015 ANSYS, Inc. March 1, 2016

 … Pre-stress MSUP Transient Analysis
3. Include Residual Vector (Analysis Settings>Options>Include Residual Vector)
It allows including residual vectors for MSUP Harmonic and Transient Structural
Analyses.
In MSUP analysis, the dynamic response will be approximate when the applied
loading excites the higher frequency modes of a structure.
The residual vector method employs additional modal transformation vectors in
addition to the eigenvectors in the modal transformation .
This feature accounts for high frequency
dynamic responses with fewer eigen-modes.
The default setting is Off (RESVEC,off)

56 © 2015 ANSYS, Inc. March 1, 2016

Workshop 9B and 9C
16.0 Release

ANSYS Mechanical
Linear and Nonlinear Dynamics