MD Nastran - Dynamic Analysis

MSC.
Software Corporation
Europe Corporate Asia Pacific
MSC.Software GmbH MSC.Software Corporation MSC.Software Japan Ltd.
Am Moosfeld 13 2 MacArthur Place Shinjuku First West 8F
81829 Munich, Germany Santa Ana, CA 92707 USA 23-7 Nishi Shinjuku
Telephone: (49) (89) 43 19 87 0 Telephone: (800) 345-2078 1-Chome, Shinjuku-Ku
Fax: (49) (89) 43 61 71 6 Fax: (714) 784-4056 Tokyo 160-0023, JAPAN
Telephone: (81) (3)-6911-1200
Fax: (81) (3)-6911-1201
MD Nastran - Dynamic Analysis
NAS102 Seminar
April 2007
Part Number: MDNA*R2*Z*Z*Z*SM-NAS102-NT1 Copyright 2007 MSC.Software Corporation
NA*V2004*Z*Z*Z*SM-NAS102-NT1
DISCLAIMER
MSC.Software Corporation reserves the right to make changes in specifications and other information contained in this
document without prior notice.
The concepts, methods, and examples presented in this text are for illustrative and educational purposes only, and are not
intended to be exhaustive or to apply to any particular engineering problem or design. MSC.Software Corporation assumes
no liability or responsibility to any person or company for direct or indirect damages resulting from the use of any
information contained herein.
User Documentation: Copyright 2007 MSC.Software Corporation. Printed in U.S.A. All Rights Reserved.
This notice shall be marked on any reproduction of this documentation, in whole or in part. Any reproduction or distribution
of this document, in whole or in part, without the prior written consent of MSC.Software Corporation is prohibited.
MSC, MSC., MSC.Actran, MSC.ADAMS, MSC.Dytran, MSC.EASY5, MSC.Fatigue, MSC.FlightLoads, MSC.Laminate
Modeler, MSC.Marc, MSC.Marc Mentat, MSC.Mvision, MSC.Nastran, MSC.Patran, MSC.Robust Design, SimDesigner,
SimManager, MSC.SOFY, the MSC.Software corporate logo, and Simulating Reality are trademarks or registered
trademarks of the MSC.Software Corporation in the United States and/or other countries.
All other brand names, product names or trademarks belong to their respective owners.
Copyright 2007 MSC.Software Corporation

TABLE OF CONTENTS
1. Review of Fundamentals
2. Dynamic Modeling Input
3. Normal Mode Analysis
4. Reduction in Dynamic Analysis
5. Rigid Body Modes
6. Damping
7. Transient Response Analysis
8. Frequency Response Analysis
9. Direct Matrix Input
10. Dynamic Equations of Motion
11. Residual Vector Methods
12. Enforced Motion
13. Shock and Response Spectrum Analysis
14. Random Response Analysis
15. Complex Eigenvalue Analysis
16. Normal Mode Analysis Using Parts Superelement
17. Extra Points, Transfer Functions, and NOLINs
18. Normal Modes of Preloaded Structures
19. Dynamic Design Optimization
20. Test-Analysis Correlation Copyright 2007 MSC.Software Corporation
BLANK
Copyright 2007 MSC.Software Corporation

Slide 1
SECTION 1
REVIEW OF FUNDAMENTALS
NAS102, Section 1, March 2007

Copyright 2007 MSC.Software Corporation S1-1
Slide 2
OVERVIEW OF DYNAMIC ANALYSIS

PROCESS
Dynamic System
Environment Design
Finite Element Interfacing

Model Media
Modal
Analysis
Dynamic
Response
Analysis

Slide 3
SINGLE DOF SYSTEM

• Dynamic equation of motion:
mu
(t) bu
(t) ku(t) p(t) n(u, u
) p(t)
m = mass (inertia) u(t)
b = damping (energy dissipation)
m
k = stiffness (restoring force)
n = nonlinear restoring force k n
p = applied force b
u = displacement of mass
u= velocity of mass

u= acceleration of mass
u, u, u, and p are time varying (in general)
m,b, and k are constants
n is a nonlinear function of u, u
Slide 4
UNITS
●Fundamental units
●Length L (inch, meter)
●Mass M (slug, kilogram)
●Time T (second)
●Length L (meter, millimeter)
●Force F (Newton)
●Time T (second)
●Fundamental and derived units

●m M Mass m FT2/L = F/(LT –2)
●b MT-1 Damping b FT/L = F/(T/L)
●k MT -2 Stiffness k F/L
●p MLT -2 Force p F
●u L Displacement u L
●u LT -1 Velocity u LT -1
●
u LT -2 Acceleration 
u LT -2
Ns 2
1 1000kg 1t
mm
Slide 5
UNITS (Cont.)
• Engineering Units for Common Variables.
Variable Dimensions* Common Common
English Units Metric Units
Length L in m
Mass M lb-sec 2/in kg
Time T sec sec
Area L2 in2 m2
Volume L3 in3 m3
Velocity LT -1 in/sec m/sec
Acceleration LT -2 in/sec 2 m/sec 2
Rotation - rad rad
Rotational Velocity T -1 rad/sec rad/sec *L denotes length
Rotational Acceleration T -2 rad/sec2 rad/sec2 M denotes mass
Circular Frequency T -1 rad/sec rad/sec T denotes time
Frequency T -1 cps; Hz cps; Hz - denotes dimensionless
Eigenvalue T -2 rad2 /sec 2 rad2/sec 2
Phase Angle - deg deg
Force MLT -2 lb N
Weight MLT -2 lb N
Moment ML2T -2 in-lb N-m
Mass Density ML -3 lb-sec 2/in4 kg/m3
-1 -2 2 2
Young’s Modulus ML T lb/in Pa; N/m
Poisson’s Ratio - - -
Shear Modulus ML -1T -2 lb/in2 Pa; N/m2
Area Moment of Inertia L4 in4 m4
Torsional Constant L4 in4 m4
Mass Moment of Inertia ML 2 in-lb-sec 2 kg-m 2
Stiffness MT -2 lb/in N/m
Viscous Damping Coefficient MT-1 lb-sec/in N-sec/m
Stress ML -1T -2 lb/in2 Pa; N/m2
Strain - - -

Slide 6
UNITS (Cont.)
● Use consistent units!
● Error in units is the number one cause of modeling

errors in dynamic analysis!
● Most common errors are in mass and damping units.
● MD Nastran contains no built-in set of units. The

analyst must verify consistency.
● Consistent Units: e.g. N, t, mm, s or N, kg, m, s

Slide 7
SINGLE DOF SYSTEM - UNDAMPED FREE

VIBRATIONS
• Dynamic
..
equation
mu(t) + ku(t) = 0
u( t ) A sin n t B cos nt
• Solution
k
n  natural frequency (rad / sec)
m

fn  n natural frequency (cycles / sec)
2
• Initial conditions
. u(t) Asinωnt Bcosωnt
u(0) and u(0) are known.
(t) Aωncosωnt Bωnsinωnt
u
B u( t 0)
.
u(t=0) u(t 0) u(0) B
A= n
(t 0) u
u (0) Aωn
• Finally
.
u(0)
u(t) = sin nt + u(0) cos nt
n
Slide 8
SINGLE DOF SYSTEM - UNDAMPED FREE

VIBRATIONS
0. 0 .2 0.4 0 .6 0 .8 1 .0
0 .2
0 .2 E1 E 1
E1
0 .1 0 .1
Disp lace m ent

0. 0 .
-0 .1 -0. 1
-0 .2 0. 0 .2 0.4 0 .6 0 .8 -0. 2
TI ME
S DO F Osci lla to r- -
-- Non ze ro In i ti a l Co n di t
io ns

Slide 9
SINGLE DOF SYSTEM - DAMPED FREE

VIBRATIONS
●Dynamic Equation
.. .
mu( t) b u( t) Ku(t) 0
●Critical Damping
b c 2 km 2mn
●Fraction of critical damping

b

bc
●The amount of damping determines the form of the solution.
●Underdamped
b b c b = 
* bc = *2mn
*2m
ζ
u(t) e bt/ 2m(A sin ωd t B cos ωd t) e
ωn t
(Asin ωdt B cos ωdt)
●Damped natural frequency

d n 1 2

Slide 10
SINGLE DOF SYSTEM - DAMPED FREE

VIBRATIONS (Cont.)
● Critically damped
• b = bc
● No oscillation occurs.
• u(t) = (A + Bt) e-bt/2m
● Overdamped
• b > bc
• No oscillation occurs. The system gradually returns to equilibrium (at
rest, undisplaced) position.
● The usual analysis case is underdamped.
● Structures typically have damping in the 0-10% range.

DAMPED FREE VIBRATIONS—
Slide 11
UNDERDAMPED CASE
0 .2
E1
0 .1
D isp l a ce m en t 0 .
-0 .1
-0 .2
0. 0 .5 1 .0 1 .5 2 .0 2 .5 3 .0 E0
C D A M P E le m e n t T im e ( S e c o n d s )
 - 0 .0 5

Slide 12
SINGLE DOF SYSTEM - UNDAMPED

SINUSOIDAL VIBRATIONS
Dynamic equation
..
mu( t) ku(t ) P sin t
where = forcing frequency

Solution
P/k
u(t ) A sin n t B cos n t  sin t
12 / n2
where
Initial Condition Steady-State
B u( t 0)
.
. u (t 0) P / k
A 
n (1 2 / 2n )n

Slide 13
SINGLE DOF SYSTEM - UNDAMPED

• Steady-state solution
• P/k is the static response.

1
• 1 2 / 2n
is the dynamic magnification factor.

Slide 14
MAGNIFICATION FACTOR
2.0
E1
1.5
Dis p l aceme nt
1.0
Ma gnit ude
0.5
P/k
0.
0. 0. 2 0 .4 nn 0. 6 0.8 1.0 E1
SD OF Syste m Freq ue n cy (Hz)

No B

Slide 15
SINGLE DOF SYSTEM - DAMPED

● Dynamic equation
..
• mu(t) .
+ bu(t) + ku(t) = P sin t
● The transient solution decays rapidly and is of no interest.

● Steady-state solution
sin(t )
• u(t ) P / k
(1 2 / 2n )2 (2 / n )2
2/ n
• tan1
1 2 / 2n
● is phase lead => 180o  360o – for Nastran output
Slide 16
SINGLE DOF SYSTEM - DAMPED


1
● For n
• Magnification factor 1 (static solution)
• Phase angle 360° (response is in phase with the force)

● For 1
n
• Magnification factor 0 (no response)

• Phase angle 180° (response has opposite sign of force)

1
● For n
1
• Magnification factor 
2
• Phase angle » 270°

Slide 17
MAGNIFICATION FACTOR
1. 2
E1
1. 0
0. 8
D is p l ac e m e nt 0. 6
M a g n i tu d e
0. 4
0. 2
0.
0. 0. 2 0 .4 0 .6 0. 8 1. 0 E1
F r e q u e nc y ( H z )

Slide 18
MULTI-DEGREE OF FREEDOM SYSTEM
● Now the equation of motion becomes

.. .
[M] {u} + [B] {u} + [K] {u} = {P} + {N}
● where
● {u} = displacement vector
● {M}= mass matrix
● {B}= damping matrix
● {K}= stiffness matrix
● {P}= forcing function
● {N}= nonlinear vector of forces

Slide 19
CLASSIFICATION OF DYNAMIC
ENVIRONMENTS
Deterministic
Transient
Periodic
Simple Shock
Harmonic Spectra
Random
Nonstationary
Stationary
MD Nastran
Ergodic

Slide 20
DYNAMIC EXCITATIONS

Slide 21
FINITE ELEMENT DYNAMIC MODELING

CONSIDERATIONS
● Frequency range
● Grid points/constraints/elements
● Linear versus nonlinear behavior
● Whole system versus superelement models
● Interaction with adjacent media
● Test/measured data integration
● Damping

Slide 22
MD NASTRAN DOCUMENTATION
● MD Nastran Documentation
• Quick Reference Guide

• Release Guide
• Installation and Operation Release Guide
● User’s Guides
• Linear Static Analysis
• Basic Dynamic Analysis
• Advanced Dynamic Analysis
• Design Sensitivity and Optimization
• DMAP Programmer’s Guide
• Numerical Methods
• Aeroelastic Analysis
• Thermal Analysis
• Superelement
• Implicit Nonlinear (SOL 600)
• Explicit Nonlinear (SOL 700)
• SOL 400 (MD Nastran)
Slide 23
MD NASTRAN DOCUMENTATION (Cont.)
● Other Documentation
• Reference Manual
• Common Questions and Answers
• Bibliography
http://www.mscsoftware.com/support/prod_support/nastran/biblio/index.cfm
● Documentation also available in online form

Slide 24
TEXT REFERENCES ON DYNAMIC ANALYSIS
• W. C. Hurty and M. F. Rubinstein, Dynamics of Structures, Prentice-

Hall, 1964.
• R. W. Clough and J. Penzien, Dynamics of Structures, McGraw-Hill,
1975.
• S. Timoshenko, D. H. Young, and W. Weaver, Jr., Vibration Problems
in Engineering, 4th Ed., John Wiley & Sons, 1974.
• K. J. Bathe and E. L. Wilson, Numerical Methods in Finite Element
Analysis, Prentice-Hall, 1976.
• J. S. Przemieniecki, Theory of Matrix Structural Analysis, McGraw-Hill,
1968.
• C. M. Harris and C. E. Crede, Shock and Vibration Handbook, 2nd
Ed., McGraw-Hill, 1976.
• L. Meirovitch, Analytical Methods in Vibrations, MacMillan, 1967.
• L. Meirovitch, Elements of Vibration Analysis, McGraw-Hill, 1975.
• M. Paz, Structural Dynamics Theory and Computation, Prentice-Hall,
1981.
Slide 25
TEXT REFERENCES ON DYNAMIC ANALYSIS

(Cont.)
• W. T. Thomson, Theory of Vibrations with Applications, Prentice-
Hall, 1981.
• R. R. Craig, Structural Dynamics: An Introduction to Computer
Methods, John Wiley & Sons, 1981.
• S. H. Crandall and W. D. Mark, Random Vibration in Mechanical
Systems, Academic Press, 1963.
• J. S. Bendat and A. G. Piersol, Random Data Analysis and
Measurement Techniques, 2nd Ed., John Wiley & Sons, 1986.
• R. Gasch, K. Knothe: Strukturdynamik, Band 1+2. Springer, 1987

Slide 26
BLANK

SECTION 2
DYNAMIC MODELING INPUT
NAS102,Section 2, March 2007

MD NASTRAN INPUT FILE SETUP
● FMS and NASTRAN Statements - File allocations and system cell
● Executive Control Section - Solution type, time allowed, system diagnostics
● CEND - Required Delimiter
● Case Control Section - Output requests, selects certain Bulk Data items
● BEGIN BULK - Required Delimiter
● Bulk Data Section - Structural model definition, solution conditions
● ENDDATA - Required Delimiter

MD NASTRAN BULK DATA ENTRY FORMAT
● Fixed field format
● GRID^^^^2^^^^^^^3^^^^^^^1.0^^^^^-2.0^^^^3.0^^^^^^^^^^^^^316
● Free format, same entry
● GRID,2,3,1.0,-2.0,3.0,,316
● Replicators for repetitive input
● This: GRID,1,,0.,0.,0.,,126
=, *(5),=,=,*(1.),==
=(3)
● Produces: GRID
GRID
1
6
0
0
0
1
0
0
126
126
GRID 11 0 2 0 126
GRID 16 0 3 0 126
GRID 21 0 4 0 126
● Examples in this course will use free format and replicators.

FINITE ELEMENT ANALYSIS
● The real world is not comprised exclusively of SDOF systems.
● Finite elements are used to model the mass, damping, and stiffness of
complex systems and structures.
● Degrees of freedom (DOF) are independent coordinates that describe
the motion of the structure at any instant in time.
● GRIDs are used to model the continuous structure as a discrete entity.
● Each GRID may have six DOFs: translation in the X, Y, and Z
directions and rotations about the X, Y, and Z axes.
y
Time 2
Time 1
x
2
z 1
● Bookkeeping is done via the matrices that define the relationships

between the DOFs.

COMMONLY USED ELASTIC ELEMENTS
One-Dimensional Geometry Number of DOFs
ROD Pin-ended rod 4 2 Grids x 2 DOFs

BAR Prismatic beam 12 2 Grids x 6 DOFs
BEAM Straight beam with warping 14 2 Grids x(6+1)DOFs
BEND Curved beam, pipe, or elbow 12 2 Grids x 6 DOFs
Two-Dimensional Geometry
TRIA3 / TRIAR Triangular plate 18 3 Grids x "6" DOFs

QUAD4/ QUADR Quadrilateral plate 24 4 Grids x "6" DOFs
SHEAR 4-sided shear panel 8 4 Grids x 2 DOFs
TRIA6 Triangular plate with midside nodes 30 6 Grids x 5 DOFs
QUAD8 Quadrilateral plate with midside nodes 40 8 Grids x 5 DOFs
Three-Dimensional
Geometry
HEXA Solid with six quadrilateral faces 24-60 8-20 Grids x 3 DOFs
TETRA Solid with four triangular faces 12-30 4-10 Grids x 3 DOFs
PENTA Solid with two triangular faces and three quadrilateral faces 18-45 6-15 Grids x 3 DOFs
Zero-Dimensional
Geometry
ELAS Simple spring connecting two degrees of freedom 2 1-2 Grids x 1DOF
BUSH Frequency-dependent spring/damper connecting up to six
degrees of freedom 6 1-2 Grid x 6 DOFs

COUPLED VERSUS LUMPED MASS
● Coupled mass is generally more accurate than lumped mass.
● Lumped mass is preferred for computational speed in dynamic analysis.
● User-selectable coupled mass matrix for elements
● PARAM, COUPMASS, 1 to select coupled mass
● The default is lumped mass.
● Elements which have either lumped or coupled mass:
● BAR, BEAM, CONROD, HEXA, PENTA, QUAD4/QUADR, QUAD8, ROD, TETRA,
TRIA3/TRIAR, TRIA6, TRIAX6, TUBE
● Elements which have lumped mass only:
● CONEAX, SHEAR – CONEAX obsolete (no longer documented)
● Elements which have coupled mass only:
● BEND

COUPLED VERSUS LUMPED MASS (Cont.)
Lumped mass contains only diagonal, translational components (no rotational
ones).
● Coupled mass contains off-diagonal translational components as well as
rotations for BAR (though no torsion for BAR, by default), BEAM, and BEND
elements.
● Setting system cell 398 to 1 generates torsional mass for BAR
● Setting system cell 398 to 1 and param,coupmass,1 generates consistent axial
mass for BAR
NASTRAN SYSTEM(398) = 1 or NASTRAN BARMASS = 1
● Neglected inertia may result in massless mechanisms.

ROD FINITE ELEMENT
2 4 Length L
1 3 Area A
Torsional constant J
Young’s modulus E
L Shear modulus G
Mass density 
Polar moment of inertia I
● Stiffness matrix:
AE AE
L 0 0 
L
 GJ GJ 
0 0 
k  L L 
AE AE
0 0 
L L 
 GJ GJ 
 0 0
 L L 

1
0
1
0 
● Classical Consistent mass: 
3 6 
 I I 

0 0 
m ρAL  3A
1
 1 
6 A  where I  r 2 dA
0 0 

6 3 
 I I 
0
 0
 6A 3 A
ROD FINITE ELEMENT (Cont.)
● Classical and MD Nastran lumped mass:
12 0 0 0
0 0 0 0
m AL  
0 0 12 0
 
0 0 0 0
5/12 = ½(1/2 + 1/3)
● MD Nastran coupled mass: 1/12 = ½( 0 + 1/6)

5 12 0 1 12 0
0 0 0 0
m AL  

1 12 0 5 12 0
 
0 0 0 0
● The axial translational terms represent the average of lumped mass

and classical consistent mass. This average is found to be best for
ROD and BAR elements.

JUSTIFICATION FOR MD NASTRAN
COUPLED MASS CONVENTION
● Consider a fixed-free rod
u(t)
1
L
Single Element Model
● Exact quarter-wave natural frequency

 E  E 
1 4  1.5708
2L L

JUSTIFICATION FOR MD NASTRAN
COUPLED MASS CONVENTION (Cont.)
● Different approximations
● Lumped mass
E E
LM  2 1.414
L L
● Classical consistent mass (10 %)
E E
C  3 1.732
● MD Nastran L L
(10 %)
● Coupled mass
E E
N  12 5 1.549
L L
(1.4%)

MASS UNITS
● MD Nastran assumes you are providing a consistent set of units.
● Weight units may be input instead of mass units if this is more
convenient. You must then convert them to mass units using
PARAM,WTMASS.
● Weight-to-mass conversion:
Mass = (1/G) Weight (G = Gravity Acceleration)
Mass Density = (1/G) Weight Density
● PARAM,WTMASS, factor performs conversion with factor = 1/G. The
default value for factor is 1.0.
● Example:
● Input RHO = 0.3 for steel weight density.
● Use PARAM, WTMASS, 0.00259 to multiply 0.3 for G = 386.4 in/sec2.
● PARAM,WTMASS is used once per run and multiplies all weight/mass
input (including MASSi, CONMi, and nonstructural mass input). Do not
mix input types. Use all mass or all weight inputs.
MASS INPUT
● Material density
● MATi entries
1 2 3 4 5 6 7 8 9 10
MAT1 MID E G NU RHO A TREF GE
MAT1 2 30.0E6 0.3 7.76E-4
● Scalar mass
● CMASSi, PMASS
● Grid point mass
● CONM1 (6x6 mass matrix) - The user defines half of the
terms, symmetry is assumed.
● CONM2 (concentrated mass) 
M 
 M SYM . 
 
 M 
 
 I11 
 I 21 I22 
 

 I31 I32 I33

MASS INPUT (Cont.)
● Structural mass (e.g., CONM2, etc.) are always included in
the model
● Nonstructural mass
● Mass input on element property entry which is not associated with
geometric properties of element. Input as mass/length for line
elements and mass/area for elements with 2-D geometry.
● Non structural mass can be specified on many property entries
(e.g., NSM on the PSHELL entry)
● This type of nonstructural mass is always included in the model
● Some nonstructural mass are Case Control selectable
● NSM Case Control selects NSM, NSM1, NSML, NSML1 entries
Bulk Data Entries

NONSTRUCTURAL MASS ENHANCEMENTS
● Five nonstructural mass Bulk Data entries—NSM,
NSM1, NSML, NSML1, and NSMADD—are available
● Distribute nonstructural mass by element lists or specific
property lists associated with property entries.
● Case Control selectable with NSM Case Control command
● NSM callout can be different between

superelements—but only one per superelement

NONSTRUCTURAL MASS ENHANCEMENTS (Cont.)
● NSM and its alternate form NSM1 allows the user to
allocate an NSM_value to selected sets of elements.
TYPE = PSHELL, PCOMP, PSHEAR
PBAR, PBARL, PBEAM, PBEAML, PBEND
Example: PROD, CONROD,
PCONEAX, PRAC2D
PTUPE
SID TYPE ID VALUE ID VALUE ID VALUE

NSM 3 PSHELL 15 .022
SID TYPE VALUE ID ID ID ID ID

NSM1 3 PSHELL .022 15
● Either one of the above examples assigns a 0.022

non-structural mass per unit area to PSHELL ID 15.
Additive if both are present

● The NSML and NSML1 entries compute a
nonstructural mass coefficient value for “area”
elements (e.g., CQUAD4)
● The NSML and NSML1 entries compute a

nonstructural mass coefficient value for“line” elements
(e.g., CBAR)
● NSML and NSML1 cannot mix “area” and “line”

element on the same entry
● NSML and NSML1 are then converted to NSM and
NSM1
Example:
SID TYPE ID VALUE ID VALUE ID VALUE
NSML 3 PSHELL 100 .95
The above example assigns a 0.95 lump mass distributed evenly

to the area elements referenced by PSHELL 100. It can also be
represented by the following entry:
SID TYPE VALUE ID ID ID ID ID
NSML1 3 PSHELL .95 100
MSMADD SID S1 S2 S3 S4 S5 S6 S7
S8 S9 S10 etc.
● NSMADD--combines NSM, NSM1, NSML, and NSML1 sets and

sum their results to selected sets of elements.

Concentrated Mass Element Connection, Rigid Body Form CONM2
Defines a concentrated mass at a grid point.
Format:
1 2 3 4 5 6 7 8 9 10
CONM2 EID G CID M X1 X2 X3
I11 I21 I22 I31 I32 I33
Example:
CONM2 2 15 6 49.7
16.2 16.2 7.8
Field Contents
EID Element identification number. (Integer > 0)
G Grid point identification number. (Integer > 0)
CID Coordinate system identification number. For CID of -1; see X1, X2, X3 below. (Integer > -1; Default = 0)
M Mass value. (Real)
X1, X2, X3 Offset distances from the grid point to the center of gravity of the mass in the coordinate system defined in field
4, unless CID = -1, in which case X1, X2, X3 are the coordinates, not offsets, of the center of gravity of the mass
in the basic coordinate
Iij Mass moments of inertia measured at the mass center of gravity in the coordinate system defined by field 4. If
CID=-1, the basic coordinate system is implied. (For I11, I22, and I33; Real > 0.0; for I21, I31, and I32; Real)
Remarks:
1. Element identification numbers should be unique with respect to all other element identification numbers
2. For a more general means of defining concentrated mass at grid points, see the CONM1 entry description.
3. The continuation is optional.
4. If CID = -1, offsets are internally computed as the difference between the grid point location and X1, X2, X3. The
grid point locations may be defined in a nonbasic coordinate system. In this case, the values of Iij must be in a
coordinate system that parallels the basic coordinate system.
(Continued)
Bulk Data Entry

Concentrated Mass Element Connection, Rigid Body Form CONM2
5. The form of the inertia matrix about its center of gravity is taken as:
éM ù
ê M symmetric ú
ê ú
ê M ú
ê ú
ê I11 ú
ê - I21 I22 ú
ê ú
ë - I31 - I32 I33 û
where M = ò r dV
= ò r (x + x 3 ) dV
2 2
I11 2
= ò r( x + x 3 ) dV
2 2
I22 1
= ò r( x + x 2 ) dV
2 2
I33 1
I21 = ò rx x 1 2
dV
I31 = ò rx x 1 3
dV
I32 = ò rx x 2 3
dV
and x 1, x2, x 3 are components of distance from the center of gravity in the coordinate system defined in field 4.
The negative signs for the off-diagonal terms are supplied automatically. A warning message is issued if the
inertia matrix is nonpositive definite, since this may cause fatal errors in dynamic analysis modules.
6. If CID > 0, then X1, X2, and X3 are defined by a local Cartesian system, even if CID references a spherical or
cylindrical coordinate system. This is similar to the manner in which displacement coordinate systems are
defined.
7. See the MSC.Nastran Reference Manual, Section 4.2 for a definition of coordinate system terminology.
Bulk Data Entry

BASIC MD NASTRAN SET OPERATIONS
● See the MD Nastran Quick Reference Guide.
Grid Set (G) = N + M
Multipoint Constraints
Independent DOF (N) = F + S
Constraints
Unconstrained DOF (F) = A + O O - omitted
Static Condensation
L – Left over
Analysis Set A = L + R R – Rigid Body
Free-Body Partitioning
Solve A-Set Modes
Reverse Process for Data Recovery to G-Set

BASIC MD NASTRAN SET DEFINITION
Each degree of freedom is a member of one mutually exclusive "set". Set names have the
following definitions:
Set
Name Definition
m degree of freedom eliminated by multipoint constraints
sb* degree of freedom eliminated by single-point constraints that are included in boundary
condition changes and by the AUTOSPC feature
sg* degrees of freedom eliminated by single-point constraints that are specified on the PS
field on GRID Bulk Data entries
o degrees of freedom omitted by structural matrix partitioning
q generalized degrees of freedom for dynamic reduction or component mode synthesis
r reference degrees of freedom used to determine free body motion
c degrees of freedom which are free during component mode synthesis or dynamic
reduction
b degree of freedom fixed during component mode analysis or dynamic reduction
e extra degrees of freedom introduced in dynamic analysis
sa permanently constrained aerodynamic degrees of freedom
k aerodynamic degrees of freedom
Each degree of freedom is also a Each degree of freedom is also a member of one or more
combined sets called "supersets." Supersets have the following definitions:
set supersets
m
sb
sg s
o
q n g p
ne ps
r f
a fe pa
c t d
l
b
e
sa
k
* Strictly speaking, sb and sg are not exclusive w ith respect to one another. Degrees of f reedom may exist in both
sets simultaneously. Since these sets are not used explicitly in the solution sequences, this need not concern the
user. How ever, those w ho use these sets in their ow n DMA Ps should avoid redundant specif ications w hen using
these sets f or partitioning or merging operations. That is, a degree of f reedom should not be specif ied on both a PS
f ield and a GRID entry (sg set) and on a selected SPC entry (sb set). Redundant specif ications w ill cause UFM 2120
in the V EC module and behavior listed in MSC.Nastran DMA P Module Dictionary f or the UPA RTN module. These sets
are exclusive, how ever, f rom the other mutually exclusive sets.
BASIC MD NASTRAN SET DEFINITION (Cont.)
Se t Na m e Me a ning (+ indicates union of two sets)
s = sb + sg all degrees of freedom eliminated by single point constraints
the structural degrees of freedom remaining after the reference degrees of freedom are
+ c
= b removed (degree of freedom left over)
t = +
r the total set of physical boundary degrees of freedom for superelements
a = t + q the set assembled in superelement analysis
d = a+ e the set used in dynamic analysis by the direct method
f= a + o unconstrained (free) structural degrees of freedom
fe = f +e free structural degrees of freedom plus extra degrees of freedom
n = f+ s all structural degrees of freedom not constrained by multipoint constraints
ne = n + e all structural degrees of freedom not constrained by multipoint constraints plus extra
degrees of freedom
g = n+ m all structural (grid) degrees of freedom including scalar degrees of freedom

p = g+ e all physical degrees of freedom
ps = p + sa physical and constrained aerodynamic degrees of freedom
pa = ps + k physical set for aerodynamics
fr = o +  statically independent set minus the statically determinate supports (fr = f - q - r)
v= o + c + r the set free to vibrate in dynamic reduction and component mode synthesis
The a-set and o-set created in the following ways:
1. If only OMITi entries are present then the o-set consists of degrees of freedom listed
explicitly on OMITi entries. The remaining f-set degrees of freedom are placed in the b-
set which is a subset of the a-set.
2. If ASETi or QSETi are present, then the a-set consists of all degrees of freedom listed
on ASETi entries and any entries listing its subsets, such as QSETi, SUPORTi, CSETi,
and BSETi entries. Any OMITi entries are redundant. The remaining f-set degrees
3. If there are no ASETi, QSETi, or OMITi entries present but there are SUPORTi, BSETi,
or CSETi entries present then the entire f-set is placed in the a-set and the o-set is not
created.
4. There must be at least one explicitly ASETi, QSETi, or OMITi entry for the o-set to
exist, even if the ASETi, QSETi, or OMITi entry is redundant.
In dynamic analysis, additional vector sets are obtained by a modal transformation derived from real
eigenvalue analysis of the a-set. These sets are as follows:

O = rigid body (zero frequency) modal degrees of freedom

f = finite frequency modal degrees of freedom

i = O + f , the set of all modal degrees of freedom
One Vector set is defined that combines physical and modal degrees of freedom:
u h = i + u e , the set of all modal degrees of freedom
The membership of each degree of freedom can be printed by use of the Bulk Data entries PARAM,
USETPRT and PARAM, USETSEL.
BASIC MD NASTRAN SET DEFINITION (Cont.)
Degre e of Fre edom Set Bulk Data Entrie s
Degrees of freedom are placed in sets as specified by the user on the following Bulk Data entries:
Name Bulk Data Entry Name

m MPC, MPCADD, MPCAX, POINTAX, RBAR, RBE1, RBE2, RBE3, RROD, RSPLINE,
RTRPLT, GMBC, GMSPC*
sb SPC, SPC1, SPCADD, SPCAX, FLSYM, GMSPC*, BNDGRID, (PARAM,AUTOSPC,YES)
sg GRID, GRIDB, GRDSET (PS field)
o OMIT, OMIT1, OMITAX, GRID (SEID field), SESET
q QSET, QSET1
r SUPORT, SUPORT1, SUPAX
c CSET, CSET1
b BSET, BSET1
e EPOINT
sa CAEROi
k CAEROi
a ASET, ASET1, Superelement exterior degrees of freedom, CSUPEXT
*Placed in set only if constraints are not specified in the basic coordinate system.
In superelement analysis, the appropriate entry names are preceded by the letter SE, and have a field
reserved for the superelement identification number. This identification is used because a boundary
(exterior) grid point may be in one mutually exclusive set in one superelement and in a different set in the
adjoining superelement. The SE-type entries are internally translated to the following types of entry for
the referenced superelement:
Entry Type Equivalent Type

SEQSETi QSETi
SESUP SUPORT
SECSETi CSETi
SEBSETi BSETi

TIPS ON MODEL VERIFICATION
V1 : ID of reference grid point
● PARAM, GRDPNT, V1 (V1 > 0) V1=0 : The location [0 0 0] in the
basic coordinate system will
● Grid point weight generator be used as reference point
See example on next (additional) page
● PARAM, USETPRT, V1 (V1 = 0, 1, 2, 10, 11, 12)

● MD Nastran set tables The membership of each degree-
printed.
degree-of
of--freedom can be
● Various rigid body and equilibrium checks (e.g.,

GROUNDCHECK)
● Engineering judgment

WEIGHT GENERATOR TABLE IN .F06 FILE
PARAM,GRDPNT,1
MO - Rigid Body Mass Properties

S - Identity Matrix when the mass is the same in each coordinate direction
I(S) - Inertia Matrix of structure for C.G. with respect to the basic coordinate system
I(Q) - Corresponding Principal Moments of Inertia Matrix
Q - Transformation from the principal direction to the basic coordinate system

SECTION 3
NORMAL MODE ANALYSIS

REASONS TO COMPUTE NATURAL
FREQUENCIES AND NORMAL MODES
● Assess the dynamic characteristics of the structure. For example, if rotating
machinery is going to be installed on a certain structure, to avoid excessive
vibrations, it might be necessary to see if the frequency of the rotating mass is
close to one of the natural frequencies of the structure.
● Assess the possible dynamic amplification of the loads.
● Use the natural frequencies and normal modes to guide subsequent dynamic
analysis (transient response, response spectrum analysis), i.e., what should
the appropriate time step be for integrating the equations of motion in transient
analysis?
● Use the natural frequencies and mode shapes for subsequent dynamic
analysis, i.e., transient analysis of the structure using modal expansion.
● Guide the experimental analysis of structures, i.e., location of accelerometers,
etc.
● Evaluate the design changes.

THEORETICAL RESULTS
● Consider ··

M x
 + 
K x
 = 0 (1)
● Assume a harmonic solution of the form
it
x  = e (2)
● (Physically, this means that all the coordinates perform synchronous
motions. The system configuration does not change its shape during
motion, only its amplitude.) 2 i t
··
● From Equation 2 x = – e (3)
● Substituting Equations 2 and 3 into Equation 1, we obtain

2 it it
– M
 e +
K
 e = 0
2
● which (after dividing by eit) simplifies to 
K –  M
 = 0
● This is then an eigenvalue problem.

THEORETICAL RESULTS (Cont.)
● Therefore, there are two cases:
1. If det ([K ] 2 [M]) 0 , the only possibility (from Equation 4) is
{} 0
 which is the so-called trivial solution and is not interesting from a
physical point of view.
2. Otherwise, we need det ([K ] 2 [M]) 0 in order to have a
nontrivial solution for {} .
● The eigenvalue problem reduces to solve the following:
det ([K ] 2 [M]) 0
 or
 where
det ([K ] [M]) 0
2

● If the structure has N degrees of freedom with attached
mass, there will be N ' s that are solutions of the
eigenvalue problem. These s (1,2,…,n) are natural
frequencies, characteristic frequencies, fundamental
frequencies, or resonance frequencies.
● The eigenvector { }j associated with the natural frequency
{} j is called the normal mode or mode shape. The
normal mode corresponds to the deflected shape patterns
of the structure.
● When a structure is vibrating, its shape at any given time
is a linear combination of its normal modes.

Simply Supported Beam
 Example
1
Mode 1
Mode 2
2
Mode 3 3
etc.

IMPORTANT FACTS AND RESULTS REGARDING
NORMAL MODES AND NATURAL FREQUENCIES
● When [K] and [M] are symmetric and real (this is true for all the
standard structural finite elements), the following orthogonality
property holds:
{} [M]{} 0
i
T
If i j
j
● and
{} [K ]{ } 0
i
T
j
If i j
● also {  } T
[K ]{ j }
2j  j T Rayleigh Quotient
{j } [M]{ j }
● The natural frequencies (1,2,…) are expressed in

radians/seconds. They can also be expressed in hertz
(cycles/seconds), using
(radian / sec ond)
fj (hertz)  j
2
IMPORTANT FACTS AND RESULTS REGARDING NORMAL
MODES AND NATURAL FREQUENCIES (Cont.)
● Example: The following unconstrained structure has a
rigid body mode.
x1 x2
m m
● If a structure is not totally constrained, i.e., if it admits a

rigid body mode (stress-free mode) or a mechanism, then
at least one natural frequency will be zero.
1 
 = 0 1=  
1
1 
● The scaling of normal modes is arbitrary. For example,
m x1
m x2
 represent the same “mode of vibration.”
1  300  0.66

1 =   , 1 =   , 1 =  
0.5  150  0.33

● For practical considerations, modes should be normalized
by a chosen convention. In MSC.Nastran there are three
normalization choices (except when using Lanczos):
● The unit value of generalized mass (default)
T

i  
M 
 i = 1.0
● The unit value of the largest A-set component in each mode

● The unit value of a specific component (not recommended)
● In the Lanczos method, normalization is to a unit value of
generalized mass and to a unit value of the largest
component.

ADDITIONAL MODAL PROPERTIES
● Since strains, internal loads and stresses develop when a
structure deforms, we may recover additional useful modal
information utilizing
● Strain-displacement relationships

= 
K u


u
● Stress-strain relationships
 = 

K

● Static force - displacement relationships
P st=  K u

● Element strain energy relationships
T
Ve = 1 2 ue  
K ee u
 e

ADDITIONAL MODAL PROPERTIES (Cont.)
● Thus, for a given modal displacement

u = 
i
i
● we have

 = 
 u 
K i 

● Modal strain i i

 = 
K K
 u 
 i
i
i
● Modal stress
P = 
K 
 i 
i
● Modal force i

1   T  2
● Modal strain energy --
-
V e =  e  
K ee 
 e i
i 2  i  i

METHODS OF COMPUTATION
MD Nastran provides 3 types of methods for eigenvalue extraction:
● Tracking methods (see Appendix A)
Eigenvalues (or natural frequencies) are determined one at a time using an
iterative technique. Two variations of the inverse power method are provided,
INV and SINV. This approach is more convenient when few natural
frequencies are to be determined. In general, SINV is more reliable than INV.
● Transformation methods (see Appendix A)
The original eigenvalue problem is transformed to the form
where

K – M

 = 0

A 
 = 

–1

A = 
M  K

METHODS OF COMPUTATION (Cont.)
Then the matrix A is transformed into a tridiagonal matrix using
either the Givens technique or the Householder technique.
Finally, all the eigenvalues are extracted at once using the QR
algorithm. Two variations of the Givens technique and two
variations of the Householder technique are provided: GIV, MGIV,
HOU, and MHOU. These methods are more efficient when a
large proportion of eigenvalues are needed.
● Lanczos Method (recommended method)

This method is a combined tracking-transformation method.

STURM SEQUENCE THEORY
● Choose 
.
T
● Factor 
K – M  into
i 
L 
D  
L .
● The number of negative terms on the factor diagonal is the
number of eigenvalues below . 
No. Neg No. Neg
Terms=7 Terms=8
0.0 = i2
8 (must be in the range)

LANCZOS METHOD
● Block, shifted, inverted Lanczos
● Random starting vectors
● Automatic shift logic
● Partial and selective orthogonalization
● Sturm sequence diagnosis
● Givens plus QL eigensolution
● Can be used for both buckling and normal modes analysis
● Mass and largest component normalize only

USER INTERFACE FOR LANCZOS METHOD
Default = 7, can be
increased up to 15

USER INTERFACE FOR ALL METHODS

USER INTERFACE (Cont.)

SOLUTION CONTROL FOR NORMAL MODES
● Executive Control Section

● SOL 103
● Case Control Section

● METHOD (required - selects EIGRL / EIGR entry)
● Bulk Data Section

● EIGRL / EIGR (Lanczos method)

CASE CONTROL OUTPUT
● Grid output
● DISPLACEMENT (or VECTOR)
● GPFORCE
● GPSTRESS
● SPCFORCE
● GPKE
● Element output
● ELSTRESS (or STRESS)
● ESE
● EKE
● ELFORCE (or FORCE)
● STRAIN
● Special entry
● OMODES – selects output for selective modes

PROBLEM #1
MODAL ANALYSIS OF A FLAT PLATE

PROBLEM #1 - MODAL ANALYSIS OF A FLAT
PLATE
 For this problem, use Lanczos method to find the first ten natural
frequencies and mode shapes of a flat rectangular plate. Below is a finite
element representation of the rectangular plate. It also contains the
geometric dimensions and the loads and boundary constraints. Table 3A
contains the necessary parameters to construct the input file.
 Grid Coordinates and Element Connectivities

WORKSHOP #1 - MODAL ANALYSIS OF A FLAT
PLATE
 Loads and Boundary Conditions
Table 3A.
Length (a) 5 in
Height (b) 2 in
Thickness 0.100 in
Weight Density 0.282 lbs/in3
Mass/Weight Factor 2.59E-3 sec2/in
Elastic Modulus 30.0E6 lbs/in2
Poisson’s Ratio 0.3

GEOMETRIC DESCRIPTION OF PLATE
$ CQUAD4 21 1 23 24 35 34
$ plate.bdf CQUAD4 22 1 24 25 36 35
$
$ geometric input file for plate model CQUAD4 23 1 25 26 37 36
$ CQUAD4 24 1 26 27 38 37
PSHELL 1 1 .1 1 1 CQUAD4 25 1 27 28 39 38
CQUAD4 1 1 1 2 13 12 CQUAD4 26 1 28 29 40 39
CQUAD4 2 1 2 3 14 13 CQUAD4 27 1 29 30 41 40
CQUAD4 3 1 3 4 15 14
CQUAD4 28 1 30 31 42 41
CQUAD4 4 1 4 5 16 15
CQUAD4 5 1 5 6 17 16 CQUAD4 29 1 31 32 43 42
CQUAD4 6 1 6 7 18 17 CQUAD4 30 1 32 33 44 43
CQUAD4 7 1 7 8 19 18 CQUAD4 31 1 34 35 46 45
CQUAD4 8 1 8 9 20 19 CQUAD4 32 1 35 36 47 46
CQUAD4 9 1 9 10 21 20
CQUAD4 33 1 36 37 48 47
CQUAD4 10 1 10 11 22 21
CQUAD4 11 1 12 13 24 23 CQUAD4 34 1 37 38 49 48
CQUAD4 12 1 13 14 25 24 CQUAD4 35 1 38 39 50 49
CQUAD4 13 1 14 15 26 25 CQUAD4 36 1 39 40 51 50
CQUAD4 14 1 15 16 27 26 CQUAD4 37 1 40 41 52 51
CQUAD4 15 1 16 17 28 27 CQUAD4 38 1 41 42 53 52
CQUAD4 16 1 17 18 29 28
CQUAD4 17 1 18 19 30 29 CQUAD4 39 1 42 43 54 53
CQUAD4 18 1 19 20 31 30 CQUAD4 40 1 43 44 55 54
CQUAD4 19 1 20 21 32 31 $
CQUAD4 20 1 21 22 33 32 MAT1 1 3.+7 .3 .282
$
Weight Density
see PARAM,WTMASS

GEOMETRIC DESCRIPTION OF PLATE (Cont.)
$
GRID 1 0. 0. 0. GRID 31 4. 1. 0.
GRID 2 .5 0. 0. GRID 32 4.5 1. 0.
GRID 3 1. 0. 0. GRID 33 5. 1. 0.
GRID 4 1.5 0. 0. GRID 34 0. 1.5 0.
GRID 5 2. 0. 0. GRID 35 .5 1.5 0.
GRID 6 2.5 0. 0. GRID 36 1. 1.5 0.
GRID 7 3. 0. 0. GRID 37 1.5 1.5 0.
GRID 8 3.5 0. 0. GRID 38 2. 1.5 0.
GRID 9 4. 0. 0. GRID 39 2.5 1.5 0.
GRID 10 4.5 0. 0. GRID 40 3. 1.5 0.
GRID 11 5. 0. 0. GRID 41 3.5 1.5 0.
GRID 12 0. .5 0. GRID 42 4. 1.5 0.
GRID 13 .5 .5 0. GRID 43 4.5 1.5 0.
GRID 14 1. .5 0. GRID 44 5. 1.5 0.
GRID 15 1.5 .5 0. GRID 45 0. 2. 0.
GRID 16 2. .5 0. GRID 46 .5 2. 0.
GRID 17 2.5 .5 0. GRID 47 1. 2. 0.
GRID 18 3. .5 0. GRID 48 1.5 2. 0.
GRID 19 3.5 .5 0. GRID 49 2. 2. 0.
GRID 20 4. .5 0. GRID 50 2.5 2. 0.
GRID 21 4.5 .5 0. GRID 51 3. 2. 0.
GRID 22 5. .5 0. GRID 52 3.5 2. 0.
GRID 23 0. 1. 0. GRID 53 4. 2. 0.
GRID 24 .5 1. 0. GRID 54 4.5 2. 0.
GRID 25 1. 1. 0. GRID 55 5. 2. 0.
GRID 26 1.5 1. 0. $
GRID 27 2. 1. 0. SPC1 1 12345 1 12 23 34 45
GRID 28 2.5 1. 0.
GRID 29 3. 1. 0.
GRID 30 3.5 1. 0.

WORKSHOP # 1
● Start with the partial input file on the following page and
modify

● Add the solution sequence

● Add eigenvalue method callout
● Add eigenvector printout

● Add weight/mass conversion factor
● Add eigenvalue method

WORKSHOP # 1
$
$ wkshp1.dat
$
CEND
TITLE = NORMAL MODES EXAMPLE
ECHO = UNSORTED
SUBCASE 1
SUBTITLE= USING LANCZOS
SPC = 1
BEGIN BULK
param,post,-1
PARAM COUPMASS 1
$
include 'plate.bdf'
$
ENDDATA

SOLUTION FILE FOR PROBLEM #1
$
$ soln1.dat
$
ID SEMINAR, PROB1
SOL 103
CEND
TITLE = NORMAL MODES EXAMPLE
ECHO = UNSORTED
SUBCASE 1
SUBTITLE= USING LANCZOS
METHOD = 1
SPC = 1
VECTOR=ALL
BEGIN BULK
param,post,0
PARAM COUPMASS 1
PARAM WTMASS .00259
EIGRL 1 10
$
include 'plate.bdf'
$
ENDDATA

PARTIAL OUTPUT FOR PROBLEM #1
*** SYSTEM INFORMATION MESSAGE 6916 (MREORDR)
DECOMP ORDERING METHOD CHOSEN: DEFAULT, ORDERING METHOD USE D: BEND
*** USER INFORMATION MESSAGE 5010 (LNCILD)
STURM SEQUENCE DATA FOR EIGENVALUE EXTRACTION.
TRIAL EIGENVALUE = 8.507511D+07, CYCLES = 1.467984D+03 NUMB ER OF EIGENVALUES BELOW THIS VALUE = 3
*** USER INFORMATION MESSAGE 5010 (LNCILD)
STURM SEQUENCE DATA FOR EIGENVALUE EXTRACTION.
TRIAL EIGENVALUE = 1.908154D+09, CYCLES = 6.952274D+03 NUMB ER OF EIGENVALUES BELOW THIS VALUE = 10
0
E I G E N V A L U E A N A L Y S I S S U M M A R Y (READ MODULE)
BLOCK SIZE USED ........... ........... 7

NUMBER OF DECOMPOSITIONS .. ........... 2
NUMBER OF ROOTS FOUND ..... ........... 10
NUMBER OF SOLVES REQUIRED . ........... 7
1 NORMAL MODES EXAMPLE MARCH 9, 2007 MD NASTRAN 3/ 1/07 PAGE 9
0 SUBCASE 1
R E A L E I G E N V A L U E S
MODE EXTRACTION EIGENVALUE RADIANS CYCLES GENERALIZED GENERALIZED
NO. ORDER [rad/sec] [Hz] MASS STIFFNESS
1 1 7.055894E+05 8.399937E+02 1.336891E+02 1.000000E+00 7.055894E+05

2 2 1.877186E+07 4.332651E+03 6.895628E+02 1.000000E+00 1.877186E+07
3 3 2.811177E+07 5.302053E+03 8.438480E+02 1.000000E+00 2.811177E+07
4 4 1.929422E+08 1.389036E+04 2.210720E+03 1.000000E+00 1.929422E+08
5 5 2.221657E+08 1.490523E+04 2.372240E+03 1.000000E+00 2.221657E+08
6 6 2.328451E+08 1.525926E+04 2.428587E+03 1.000000E+00 2.328451E+08
7 7 6.832396E+08 2.613885E+04 4.160127E+03 1.000000E+00 6.832396E+08
8 8 9.600053E+08 3.098395E+04 4.931249E+03 1.000000E+00 9.600053E+08
9 9 1.365293E+09 3.694987E+04 5.880754E+03 1.000000E+00 1.365293E+09
10 10 1.850317E+09 4.301531E+04 6.846099E+03 1.000000E+00 1.850317E+09
i = i2 i fi = iTMi = iTKi

PARTIAL OUTPUT FOR PROBLEM #1 (Cont.)
0 SUBCASE 1
USING LANCZOS
0 SUBCASE 1
EIGENVALUE = 7.055894E+05
CYCLES = 1.336891E+02 R E A L E I G E N V E C T O R N O . 1
POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 0.0 0.0 0.0 0.0 -1.255791E-15
2 G -1.556529E-15 -1.386242E-15 -9.732725E-01 -1.099511E+00 4.003941E+00 -2.834058E-15
3 G -2.915758E-15 -3.505252E-15 -4.168619E+00 -1.599071E+00 8.686669E+00 -4.892793E-15
4 G -3.230860E-15 -6.176625E-15 -9.445514E+00 -1.530855E+00 1.230146E+01 -5.718867E-15
5 G -3.941097E-15 -9.087511E-15 -1.636256E+01 -1.365152E+00 1.522518E+01 -4.092859E-15
6 G -2.401380E-15 -1.049688E-14 -2.455516E+01 -1.080930E+00 1.740494E+01 -6.150749E-15
7 G -2.307428E-15 -1.346573E-14 -3.367780E+01 -8.040399E-01 1.895419E+01 -3.032976E-15
8 G -9.768109E-16 -1.680592E-14 -4.342896E+01 -5.477144E-01 1.993718E+01 -1.908953E-15
9 G -2.291250E-15 -1.610515E-14 -5.355173E+01 -3.472317E-01 2.046533E+01 -3.178451E-15
10 G -3.957610E-15 -2.128953E-14 -6.384777E+01 -2.192121E-01 2.066150E+01 -8.190604E-15
11 G -3.857835E-15 -2.386429E-14 -7.419273E+01 -1.687060E-01 2.069914E+01 -8.426097E-15
12 G 0.0 0.0 0.0 0.0 0.0 -1.208914E-15
13 G -3.253508E-16 -1.184481E-15 -1.263801E+00 -1.865870E-01 4.860682E+00 -2.281848E-15

Blank

Slide 1
SECTION 4
REDUCTION IN DYNAMIC ANALYSIS

Slide 2
INTRODUCTION TO DYNAMIC REDUCTION
 Definition
● Dynamic reduction means reducing a given dynamic math model to one with fewer
degrees of freedom.
 Why Reduction for Dynamics?
● The math model may be too big to be solved without reduction.

● The math model has more detail than required.
● Dynamic reduction allows the deletion of selective local modes.
● Dynamic reduction is more accurate (and probably cheaper) than constructing a
separate, smaller dynamic model.

REDUCTION METHODS FOR DYNAMICS
Slide 3
AVAILABLE WITH MD NASTRAN
● Guyan reduction (static condensation)

● Generally not recommended except for correlation with test data
● Modal reduction
● Component mode synthesis (superelement option) - see Section 16.

STATIC CONDENSATION (INTERNAL
Slide 4
CALCULATION)
● Let {uf} be the set of the unconstrained (free) structural coordinates.
● Partition
 where u a 
uf= ------- 
 ua = analysis set u o 
 uo = omitted set
OO--Set Degrees of freedom removed during Guyan reduction

AA-Set
-Set User-selected dynamic degrees of freedom

STATIC CONDENSATION (Cont.) (INTERNAL
Slide 5
CALCULATION)
● Form a static equation for uf and partition the stiffness matrix into the O-set and
the A-set.
● Assume Po is zero and solve for uo in terms of ua
● Transformation from the A-set to F-set is

uo Goa
{uf} = = {ua}
ua I

● O-set is dependent upon the A-set via Equation 2. The motion of the O-set is a
linear combination of the A-set motions. The columns of Goa are the static shape
vectors.
Slide 6
CALCULATION)
● The equations of motion for the F-set are written in terms of the A-set
.. .
TM f {u a} + TBf {u a } + T Kf{u a} = TP f
or
.. .
Maa ua + Baau a + Kaaua = Pa
● Dynamics problems are solved in terms of the reduced coordinates (A-set). O-set components are
recovered using Equation 2.
● O-set mass, damping, and stiffness is spread to the A-set.
● The largest cost is associated with the formulation of Maa and Baa, particularly for nondiagonal (coupled
mass) Mff.
● The resulting Kaa, Baa, and Maa are small and dense (i.e., matrix bandedness is destroyed).

Slide 7
CALCULATION)
 SUMMARY
● Separate free degrees of freedom (U f) into the omitted set (U 0) and
the analysis set (U A) by means of OMIT entries or ASET entries.
● Retain only a small fraction of the DOFs (typically 10% or less) in
the analysis set because the computer costs for static
condensation increase rapidly with the size of the analysis set.
Otherwise, retain all of the DOFs.
● Retain DOFs with large concentrated masses in the analysis set.
● Retain DOFs that are loaded (in transient and frequency response
analysis).
● Retain DOFs to adequately describe deflected shape or modes of
interest.

Slide 8
USER INTERFACE
 Either
1 2 3 4 5 6 7 8 9 10
ASET ID C ID C ID C ID C
ASET 1 123 2 12 4 1 5 1
 and/or
ASET1 C G G G G G G G
 or ASET1 123 1 2 3 4 5
 OMIT, OMIT1
 Specify either the A-set (with ASET entry) or the O-set (with OMIT entry). The remaining DOFs
automatically are placed in the complementary set.

SOLUTION CONTROL FOR GUYAN
Slide 9
REDUCTION
● Any SOL

● No special commands required

● ASET (optional* - specifies A-set)
● OMIT (optional* - specifies O-set)
 *Components not specified are placed in the complementary set. If both

ASET and OMIT are present, components not specified are placed in the
O-set.

Slide 10
DIFFICULTIES WITH GUYAN REDUCTION

● User effort in selecting A-set points
● Accuracy depends on the user’s skill in selecting A-set points
● Regardless of user’s skill, high accuracy requires a large number
of A-set points (cost consideration) - 2 to 5 times number of
accurate modes wanted
● Stiffness reduction is exact; mass and damping reductions are
only approximate
● No loss in accuracy of modes occurs when omitting massless
degrees of freedom
● Errors are most pronounced in higher modes
● Local modes may be missed altogether
● Not generally recommended, except when performing test-
analysis correlation (see Section 20)

Slide 11
Difficulties With Guyan Reduction (Cont.)
● The static condensation approximation may miss the

local dynamic effects.
0
uo o
= G oa Ua + Uo
Local Dynamic Effects
Physical Variables
Static Transformation
o -1
uo = K oo Po
Loads on O-set Components

Slide 12
WORKSHOP #2
NORMAL MODES ANALYSIS USING

GUYAN REDUCTION

WORKSHOP #2 - MODAL ANALYSIS OF A
Slide 13
FLAT PLATE USING STATIC REDUCTION

 For this example, use Guyan Reduction to reduce the model used in
Workshop #1. Then find the first five natural frequencies and mode
shapes using the Lanczos method. Use the points indicated in Figure
4B for the A-set.
 Figure 4A. Grid Coordinates and Element Connectivities.

Slide 14
WORKSHOP #2 - MODAL ANALYSIS OF A FLAT PLATE
USING GUYAN REDUCTION (Cont.)
 Figure 4B. Loads and Boundary Conditions.

Slide 15
WORKSHOP # 2
● Start with the following partial input file and add request for
Guyan reduction. Compare the results with workshop # 1
$
$ wkshp2.dat
$
SOL 103
CEND
TITLE = REDUCTION PROCEDURES, NORMAL MODES EXAMPLE
SUBTITLE = USING STATIC REDUCTION
ECHO = UNSORTED
SUBCASE 1
SUBTITLE=USING LANCZOS
METHOD = 1
SPC = 1
VECTOR=ALL
BEGIN BULK
param,post,-1
EIGR,1,AHOU,,,,5
PARAM, COUPMASS, 1
PARAM, WTMASS, 0.00259
INCLUDE 'plate.bdf'
$
$ SELECT A-SET, STATIC REDUCTION IS DONE AUTOMATICALLY
$
ENDDATA

Slide 16
SOLUTION FILE FOR WORKSHOP #2

$
$ soln2.dat
$
ID SEMINAR, PROB2
SOL 103
CEND
TITLE = REDUCTION PROCEDURES, NORMAL MODES EXAMPLE
SUBTITLE = USING STATIC REDUCTION
ECHO = UNSORTED
SUBCASE 1
SUBTITLE=USING LANCZOS
METHOD = 1
SPC = 1
VECTOR=ALL
BEGIN BULK
param,post,0
EIGR,1,LAN,,,,5
PARAM, COUPMASS, 1
INCLUDE 'plate.bdf'
$
$ SELECT A-SET, STATIC REDUCTION IS DONE AUTOMATICALLY
$
ASET1,345,3,5,7,9,11
ASET1,345,25,27,29,31,33
ASET1,345,47,49,51,53,55
ENDDATA

Slide 17
PARTIAL OUTPUT FILE FOR WORKSHOP #2

1 REDUCTION PROCEDURES, NORMAL MODES EXAMPLE SEPTEMBER 28, 2006 MD NASTRAN 8/22/03 PAGE 9
USING STATIC REDUCTION
0 SUBCASE 1
NO. ORDER MASS STIFFNESS
1 1 7.056352E+05 8.400210E+02 1.336935E+02 1.000000E+00 7.056352E+05
2 2 1.879631E+07 4.335472E+03 6.900117E+02 1.000000E+00 1.879631E+07
3 3 2.817725E+07 5.308225E+03 8.448301E+02 1.000000E+00 2.817725E+07
4 4 1.953805E+08 1.397786E+04 2.224645E+03 1.000000E+00 1.953805E+08
5 5 2.367517E+08 1.538674E+04 2.448875E+03 1.000000E+00 2.367517E+08
.
.
.
0 SUBCASE 1
CYCLES = 1.336935E+02 R E A L E I G E N V E C T O R N O . 1

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G 0.0 0.0 -9.732642E-01 -1.099629E+00 4.004109E+00 0.0
3 G 0.0 0.0 -4.168889E+00 -1.599175E+00 8.687232E+00 0.0
4 G 0.0 0.0 -9.445539E+00 -1.529646E+00 1.230197E+01 0.0
5 G 0.0 0.0 -1.636362E+01 -1.365241E+00 1.522617E+01 0.0
6 G 0.0 0.0 -2.455527E+01 -1.077562E+00 1.740565E+01 0.0
7 G 0.0 0.0 -3.367999E+01 -8.040918E-01 1.895542E+01 0.0

Slide 18
MODAL REDUCTION
● All MD Nastran linear dynamic solutions have two versions.
● Direct - The solution is solved in terms of A-set coordinates.
● Modal - The solution is solved in terms of modal coordinates (H-set).
● In the modal solution sequences the A-set coordinates are written in terms of
modal coordinates.
ua = 
a 
Modal Coordinates
Matrix of Mode Shapes
● Modal vectors (mode shapes) are solutions to the undamped eigenvalue problem
(A-set coordinates)
..
[M aa] {ua} + [K aa] {ua } = 0

Slide 19
MODAL REDUCTION (Cont.)

● Equations of motion for the A-set are written in terms of modal
coordinates (H-set notation, modal coordinates are handled internally.)
(Note: E-set DOFs are not shown here for clarity.)
.. .
[aT] [M aa ] {a} {} + [a] [Baa ] {a} {} + [aT] [K aa] {a} {}
T
= [a T] {Pa}
 If [f] is mass normalized and there are no K2PP, M2PP, B2PP, or TF,
then:
Generalized .. . Generalized
Mass matrix ] {} + [aT] Baa {a } {} + [ W2 ] {} = [aT] {Pa}
[ Stiffness
matrix
 Note: A-set matrices may be reduced matrices from Guyan
reduction or GDR. Transformation from modal coordinates to the F-set
would require two transformations.
{u f} = [] {ua }
{ua} = [a] {}

.
.. {u f} = [] {a } {}

SOLUTION CONTROL FOR MODAL
Slide 20
REDUCTION
● Any modal dynamic analysis SOL

● METHOD (required - selects Bulk Data EIGR or EIGRL entry)

● EIGR or EIGRL (required - selects parameters for eigenanalysis)

Slide 1
SECTION 5
RIGID BODY MODES

Slide 2
RIGID BODY MODES AND RIGID BODY VECTORS
THEORETICAL CONSIDERATIONS
• A structure has the ability to displace without developing internal loads of
stresses if it is not sufficiently grounded. Examples of this are:
No Constraints
(a)
(b) Partial Constraints
(c) Mechanism(s)
P
• In cases (a) and (b), the structure can displace as a rigid body.

Slide 3
RIGID BODY MODES AND RIGID BODY VECTORS
THEORETICAL CONSIDERATIONS (Cont.)
• The presence of rigid body and/or mechanism modes is evidenced by

zero frequency values in the solution of the eigenvalue problem.

K 
= 
M
 
• On the assumption that the mass matrix [M] is positive definite, zero
eigenvalues result from a positive semi-definite stiffness, i.e.,
 T 
 RIG M
 RIG> 0
T 
RIG 
K 
RIG 
= 0
• SUPORT does not constrain the structure. It simply defines the R-set
components. In normal modes analysis, rigid body modes are
calculated using the R-set as reference degrees of freedom.

Slide 4
CALCULATION OF RIGID BODY MODES
• If R-set is present, rigid body modes are calculated in MD Nastran

by the following method for methods other than the Lanczos
method:
• The Lanczos method calculates rigid body modes directly
Step 1: “a”-set partitioning

 ul l- set (left over set)
 ua =
 ur r- set (suport
(suport DOF)
Step 2: Solve for u l in terms of ur .
K ll K l u
r l = 0
K l K˜ u Pr
r rr r
Note: Pr is not actually applied!

CALCULATION OF RIGID BODY MODES Slide 5
(Cont.)
 ul = Dm ur
where
 [ Dm ] = - [ Kll ]-1 [ Klr ]
This may be used to construct a set of rigid body vectors.
D

 = m
RIG I
r

CALCULATION OF RIGID BODY MODES Slide 6
(Cont.)
 Step 3: Mass matrix operations
T
D D

M = m M  m

r Ir aa I
r
where [Mr] is not diagonal in general
• Using Gram-Schmidt orthogonalization (in the READ module), the matrix
[Mr] is orthogonalized by the transformation [ro], that is,

Mo= T Mr 
ro
ro
 Step 4: Rigid body mode construction

Dm ro

a RIG
=
ro
T
with the property: a K  = K rr = 0*
RIG aa a RIG
T
 M  = 
M 
a RIG aa a RIG o
___________________________________________________________________________________________
*Only if R-set DOFs truly support rigid body modes

SELECTION OF “SUPORT” DEGREES OF Slide 7
FREEDOM
• Care must be taken when selecting SUPORT DOFs.
• SUPORT DOFs must be able to displace independently without
developing internal stresses (statically determinate).
2 5 2 5
1 1
4
3 6 3 4
Bad Selection Good Selection
(The independent displacement of 1

and 4 may produce internal stress.)
Slide 8
CHECKING OF “SUPORT” DOFs

MD Nastran calculates internal strain-energy (work) for each rigid body vector.
Kll Klr D
[ X ] = [ D T Ir ]
Krl Krr Ir
[ X ] = [ Krr ] + [ Klr ]T D
• If actual rigid body modes exist, the strain-energy is 0 .
• Note that [X] is also the transformation of the stiffness matrix [Kaa] to R-set
coordinates, which by definition of rigid body (zero frequency) vector properties,
should be null.
• MD Nastran also calculates the rigid body error ratio
 X
= --------------
Krr
• where means Euclidian norm of the matrix = x 2ij

i j
• Note: Only one value of is calculated using [X], and [Krr] based on all SUPORT
DOFs.
Slide 9
CHECKING OF “SUPORT” DOFS (Cont.)

• Except for round-off errors, the rigid body error ratio and the strain energy
should be zero if a compatible set of statically determinate supports are
chosen by the user. These quantities may be nonzero for any of the following
reasons:
• Round-off error accumulation
• The ur set is over determined leading to redundant supports (high strain energy).
• The ur set is underspecified leading to a singular reduced stiffness matrix (high rigid
body error ratio).
• The multipoint constraints are incompatible (high strain energy and high rigid body
error ratio).
• There are too many single-point constraints (high strain energy and high rigid body
error ratio).
• Krr is null (unit value for rigid body error but low strain energy). This is an
acceptable condition and may occur when generalized dynamic reduction is used.

Slide 10
Rigid Body Modes and Rigid Body Vectors

• In MD Nastran, flexible body modes associated with the A-set mass
and stiffness matrices are calculated. The first N modes calculated by
the eigen analysis (where N is the number of DOFs in the R-set) are
discarded. The N rigid body modes are substituted in their place.
• Note: MD Nastran does not check that discarded modes are rigid
body modes (i.e., = 0).
• When this transformation is applied to the dynamic system and the
modes are unit mass normalized, we obtain
 ··   ·   
I
RIG 0  RIG   RIG  0 0  RIG 
 + T B  + 2  
0 I
F LEX 
··
   ·  0 FLE X FL EX 
 FL EX   FLEX   
 T  T
  P  
=  RIG +
RIG N + Q 
 T P  T
 F LEX  F LEX
RIGID BODY MODES AND RIGID BODY Slide 11
VECTORS (Cont.)
• As a result of the transformation, the following consequences
occur:
• Constraint forces are not externally active, i.e.,
T
RIG
Q= 0 
T
FL EX
• If damping elements are not connected to ground, then
T 
B = 
0
RIG
Thus,
T 0 0
RIG B  =
RIG F LEX 0 T 
B 
T F LEX FL EX
F LEX

RIGID BODY MODES AND RIGID BODY Slide 12
VECTORS (Cont.)
• If damping is “proportional,” then
T 0 0
RIG B   =
RIG F LEX 0 2
T i i
F LEX
• The modal dynamic equations are fully uncoupled.

Slide 1
SECTION 6
DAMPING

Slide 2
DAMPING
● Damping represents energy dissipation observed in structures.
● Damping is difficult to accurately model since damping results from many
mechanisms:
● Viscous effects (dashpot, shock absorber)
● External friction (slippage in structural joints)
● Internal friction (characteristic of material type)
● Structural nonlinearities (plasticity)
● Analytical conveniences used to model damping
● Viscous damping force
f v = bu·
.
·· + bu
mu · + ku = p .
.
● Structural damping force
fs = igku where i = –1
g = structural damping coefficient
mu·· + 
1 + ig
ku = p
Slide 3
STRUCTURAL DAMPING VERSUS VISCOUS

DAMPING
 Assume sinusoidial response:
u = u eit
 Then: u· = iue it u·· = –2ue it

● Viscous damping:
mu·· + bu· + ku = p(t)
–2 ue it+ b
m iueit+ kue it = p 
t
– 2 mue it + ibueit + kue it = p 
t
● Structural damping:
mu·· + 
1 + ig 
ku = p
t
– 2 ue it + 
m kueit = p
1 + ig  t
– 2 mue i t + ig kuei t + kue i t = p 

t

Slide 4

DAMPING (Cont.)
● Both equations are identical if:
gk
gk = b b = -------

 Therefore, if structural damping g is to be modeled using viscous
damping b, then the equality holds at only one frequency w3 (or w4).
 if
gk
b = -------

=  = k
 but n -----
m
gk
b = -------- = g m
 n
n
b = 2m
c n

Slide 5

DAMPING (Cont.)
 then b g
------- =  = ---
b 2
c
 = critical damping ratio (percent critical damping)
1
 g = ----
Q
= structural damping factor
 Q = quality factor or magnification factor

Slide 6
STRUCTURAL DAMPING VERSUS VISCOUS DAMPING

(CONSTANT DISPLACEMENT)
Damping Structural Damping, f s = igku

Force
Equivalent
b = gk/3 (or 4)
Viscous
f = b u·
v
3 (or 4 ) 
● Viscous and structural damping are equivalent at frequency

3 (or 4).

Slide 7
DAMPING SUMMARY
● Viscous damping force proportional to velocity

● Structural damping force proportional to displacement
● Critical damping ratio b / b cr 
● Quality factor Q inversely proportional to energy
dissipated per cycle of vibration
● At resonance n
● = g/2
● Q = 1/(2
)
● Q = 1/g

Slide 8
STRUCTURAL DAMPING
● Structural damping
● MATi entries
1 2 3 4 5 6 7 8 9 10
MAT1 MID E G NU RHO A TREF GE
MAT1 2 30.0E6 0.3 0.10
● PARAM,G, factor (Default = 0.0)

● Overall structural damping coefficient to multiply entire system
stiffness matrix
● PARAM,W3, factor (Default = 0.0)
● Converts overall structural damping to equivalent viscous damping
● PARAM,W4, factor (Default = 0.0)
● Converts element structural damping to equivalent viscous damping
● Units for W3,W4 are radians/unit time
● If PARAM,G is used, PARAM,W3 must be given a setting greater
than zero; otherwise, PARAM,G is ignored in transient response
analysis (see Section 7 for more information).

Slide 9
VISCOUS DAMPING
● Scalar viscous damping

CDAMP1 Scalar damper between two DOFs with reference to a property entry.
Scalar damper between two DOFs without reference to a property entry

CDAMP2 (PDAMP).
Scalar damper between two scalar points (SPOINT) with reference to a

CDAMP3 property entry (PDAMP).
Scalar damper between two scalar points (SPOINT) without reference to

CDAMP4 a property entry.
Element damper between two grid points; references a property entry

CVISC (PVISC).
Generalized spring and damper element that may also be frequency

CBUSH dependent.

Slide 10
VISCOUS DAMPING (Cont.)

CDAMP1 Scalar Damper Connection
Defines a scalar damper element.
Format:
1 2 3 4 5 6 7 8 9 10
CDAMP1 EID PID G1 C1 G2 C2
Example:
CDAMP1 19 6 0 23 2
Field Contents
EID Unique element identification number. (Integer > 0)
PID Property identification number of a PDAMP property entry. (Integer > 0; Default = EID)
G1, G2 Geometric grid point identification number. (Integer > 0)
C1, C2 Component number. (0 < Integer < 6; 0 or up to six unique integers, 1 through 6 may
be specified in the field with no embedded blanks. 0 applies to scalar points and 1
through 6 applies to grid points.)
Remarks:
1. Scalar points may be used for G1 and/or G2, in which case the corresponding C1 and/or C2 must
be zero or blank. Zero or blank may be used to indicate a grounded terminal G1 or G2 with a
corresponding blank or zero C1 or C2. A grounded terminal is a poin
2. Element identification numbers should be unique with respect to all other element identification
3. The two connection points (G1, C1) and (G2, C2), must be distinct.
4. For a discussion of the scalar elements, see the MSC.Nastran Reference Manual, Section 5.6.
5. When CDAMP1 is used in heat transfer analysis, it generates a lumped heat capacity.
6. A scalar point specified on this entry need not be defined on an SPOINT entry.
7. If Gi refers to a grid point then Ci refers to degrees of freedom(s) in the displacement coordinate
system specified by CD on the GRID entry.

Slide 11

PDAMP Scalar Damper Property
Specifies the damping value of a scalar damper element using defined CDAMP1 or CDAMP3 entries.
Format:
1 2 3 4 5 6 7 8 9 10
PDAMP PID1 B1 PID2 B2 PID3 B3 PID4 B4
Example:
PDAMP 14 2.3 2 6.1
Field Contents
PIDi Property identification number. (Integer > 0)
Bi Force per unit velocity. (Real)
Remarks:
1. Damping values are defined directly on the CDAMP2 and CDAMP4 entries, and therefore do not
require a PDAMP entry.
2. A structural viscous damper, CVISC, may also be used for geometric grid points.
3. Up to four damping properties may be defined on a single entry.
· For a discussion of scalar elements, see the MSC.Nastran Reference Manual, Section 5.6..

Slide 12

CDAMP2 Scalar Damper Property and Connection
Defines a scalar damper element without reference to a material or property entry.
Format:
1 2 3 4 5 6 7 8 9 10
CDAMP2 EID B G1 C1 G2 C2
Example:
CDAMP2 16 2.98 32 1
Field Contents
B Value of the scalar damper. (Real)
G1, G2 Geometric grid point identification number. (Integer > 0)
C1, C2 Component number. (0 < Integer < 6; 0 or up to six unique integers, 1 through 6 may
be specified in the field with no embedded blanks. 0 applies to scalar points and 1
through 6 applies to grid points.)
Remarks:
1. Scalar points may be used for G1 and/or G2, in which case the corresponding C1 and/or C2 must
be zero or blank. Zero or blank may be used to indicate a grounded terminal G1 or G2 with a
corresponding blank or zero C1 or C2. A grounded terminal is a poin
2. Element identification numbers should be unique with respect to all other element identification numbers.
3. The two connection points (G1, C1) and (G2, C2), must be distinct.
7. If Gi refers to a grid point then Ci refers to degrees of freedom(s) in the displacement coordinate
system specified by CD on the GRID entry.

Slide 13
CDAMP3 Scalar Damper Connection to Scalar Points Only
Defines a scalar damper element that is connected only to scalar points.
Format:
1 2 3 4 5 6 7 8 9 10
CDAMP3 EID PID S1 S2
Example:
CDAMP3 16 978 24 36
Field Contents
PID Property identification number of a PDAMP entry. (Integer > 0; Default = EID)
S1, S2 Scalar point identification numbers. (Integer > 0; S1 S2)
Remarks:
1. S1 or S2 may be blank or zero, indicating a constrained coordinate.
2. Element identification numbers should be unique with respect to all other element identification
numbers.
3. Only one scalar damper element may be defined on a single entry.

Slide 14
CDAMP4 Scalar Damper Property and Connection to Scalar Points Only
Defines a scalar damper element that connected only to scalar points and without reference to a material or
property entry.
Format:
1 2 3 4 5 6 7 8 9 10
CDAMP4 EID B S1 S2
Example:
CDAMP4 16 -2.6 4 9
Field Contents
B Scalar damper value. (Real)
S1, S2 Scalar point identification numbers. (Integer > 0; S1 S2)
Remarks:
1. S1 or S2 may be blank or zero, indicating a constrained coordinate.
2. Element identification numbers should be unique with respect to all other element identification numbers.
3. Only one scalar damper element may be defined on a single entry.
5. If this entry is used in heat transfer analysis, it generates a lumped heat capacity.

Slide 15

Slide 16

PVISC Viscous Damping Element Property
Defines properties of a one-dimensional viscous damping element (CVISC entry).
Format:
1 2 3 4 5 6 7 8 9 10
PVISC PID1 CE1 CR1 PID2 CE2 CR2
Example:
PVISC 3 6.2 3.94
Field Contents
PIDi Property identification number. (Integer > 0)
CE1, CE2 Viscous damping values for extension in units of force per unit velocity. (Real)
CR1, CR2 Viscous damping values for rotation in units of moment per unit velocity. (Real)
Remarks:
1. Viscous properties are material independent; in particular, they are temperature independent.
2. One or two viscous element properties may be defined on a single entry.

Slide 17
MODAL DAMPING
 CASE CONTROL
 SDAMP = n $ selects the modal damping table to be used.
 BULK DATA
 TABDMP1,n,CRIT
 ,x1,y1,x2,y2,..endt
 $ Lists damping values (in "G", "CRIT", or "Q") versus

 $ frequencies.

Slide 18
Rayleigh Damping
● Proportional to either the mass or stiffness matrix
● Also known as proportional damping
● Proportional to mass matrix (param,alpha1,x)
● Proportional to stiffness matrix (param,alpha2,y)
● Available in transient and frequency response analysis
● Scale factors applied to d-set (direct) and h-set (modal)
● Added to the viscous damping matrix as follows:
[B’] = [B] + alpha1 * [M] + alpha2 * [K]

Slide 19
Rayleigh Damping (cont.)
● ALPHA1 and ALPHA2 are complex parameters, e.g.

Real , Imaginary
PARAM, ALPHA2, 1.25E-4, 0.
● Modal Damping Matrix reads
φT Bφ  1φT Mφ  2φTKφ
 1I  2Ω2

Slide 20
Blank

SECTION 7
TRANSIENT RESPONSE ANALYSIS

INTRODUCTION TO TRANSIENT RESPONSE
ANALYSIS
● Compute response to time-varying excitation.
● Excitation is explicitly defined in the time domain. All of

the applied forces are known at each instant in time.
● Computed response usually includes nodal displacements

and accelerations, and element forces and stresses.
● Two categories of analysis - direct and modal.

DIRECT TRANSIENT RESPONSE
● Dynamic equation of motion
.. .
[M] {u(t)} + [B] {u(t)} +[K] {u(t)} = {P(t)}
● Response solved at discrete times with a fixed t

.
● Using central finite difference representation for {u(t)} and
..
{u(t) } at discrete times
.
{un} = 1 {u n+1 - un-1}
2t
..
{un} = 1 2 {u n+1 - 2un + u n-1}
t
● Note: These equations are also used by MSC.Nastran to
compute velocity and acceleration output.

DIRECT TRANSIENT RESPONSE (Cont.)
● Numerical integration (Central Difference type method)
(except “smear” force over 3 adjacent time points)
m b un + 1 + u n + un – 1
--------- 
u – 2u + u – + --------- 
u –u –  ------------------------------------------------------
2 n+1 n u 1 2t n + 1 n 1 3
t
p n + 1 + p n + p n – 1 Time Average “Filters”
k 1 ------------------------------------------------------
+ --- 
u + u + u – = ---
P +P +P –  3 in MD
3 n +1 n n 1 3 n +1 n n 1
Nastran
p
t
t

● Solution 
A u
1

n+1
= 
A + 
2
A u
3
 + 
n
A u
4
 – 
n 1
2
where 
A 1 = 
Mt + B 
2t + K 
3 Dynamic Matrix

A 2 = 1 
3P n + 1 + P n + P n – 1  Applied Force
2

A = 
2M 
t – K 
3
3 Initial Conditions,
from Previous
2

A4 = 
–M 
t + B 
2 t – K 
3
● Solve by decomposing A1 and applying it to the right-hand side of

the above equation.
● Similar to classical Newmark-Beta method

● M, B, and K do not change with time.
● A1 needs to be decomposed only once if t is unchanged

throughout the entire solution. If t is changed, A1 must be
redecomposed (which may be a costly operation).
● The output time interval may be greater than the solution

time interval (i.e., use solution t of 0.001 second and
output results every fifth time step or with output t of
0.005 second).

DAMPING IN DIRECT TRANSIENT RESPONSE
● Damping matrix B comprised of several matrices:
1 2 G 1 1
B = B + B + -------- K + --------
W3 W4 GE K E
where B1 = damping elements (VISC,DAMP) + B2GG
B2 = B2PP direct input matrix + transfer functions
G = overall structural damping coefficient (PARAM,G)
W3 = frequency of interest - rad/sec (PARAM,W3)
K1 = global stiffness matrix
Ge = element structural damping coefficient (GE on MATi entry)
W4 = frequency of interest - rad/sec (PARAM,W4)
KE = element stiffness matrix
● Transient analysis does not permit complex coefficients. Therefore,

structural damping is included by means of equivalent viscous damping.
● The default values for W3, W4 are 0.0. In this case, they cause associated
damping terms to be ignored.

MODAL TRANSIENT RESPONSE
● Transform from physical to modal coordinates.
(1) u =  
● Temporarily remove damping. The equation of motion becomes
 ··
(2) M u +  K u
 = P 
t 
● Substitute Equation 1 into Equation 2 to obtain
 ··
(3) M 
  + 
K 
 = P  t

● Pre-multiply by [T] to obtain
 ··
(4) T 
 M 

 T K
+  
 T P
 =   t

 where 
TM = modal mass matrix (diagonal)
 
TK = modal stiffness matrix (diagonal)
 
TP = modal force vector
MODAL TRANSIENT RESPONSE (Cont.)
● Equation 4 can be written as decoupled SDOF systems:
··
(5) m + k = p t
i i i i i
 where mi = ith modal mass

 ki = ith modal stiffness
 pi = ith modal force

DAMPING IN MODAL TRANSIENT RESPONSE
● If damping matrix B exists, then the assumption is made that it is not
diagonalized by :
T Bdiagonal
● The coupled problem is solved using modal coordinates utilizing the direct
transient response Newmark-Beta-type numerical integration.

A 1
n + 1= 
A 2 + 
A 3 
 n + 
A 4
n – 1 
M B K
● where 
A =  T --------- + --------- + ---- 
 Dynamic Matrix
1 2 2 t 3
t
1 T

A = --- 
 P
 +P +P – 
2 3 n+1 n n 1 Applied Force
2M K
 T --------- – ---- 
A =  
3
t2 3 Initial Conditions,
from Previous
M B K
 T – --------- + --------- – ---- 
A =  
4
t 2 2t 3
(Cont.)
● If modal damping is used, then each mode has
damping bi.
● The equations of motion become uncoupled

(Cont.)
● Use Duhamel’s integral to solve for modal response as
decoupled SDOF systems.
● Duhamel’s integral:

(Cont.)
● Most efficient to use modal damping ratios since equations are
decoupled
● TABDMP1 Bulk Data entry defines the modal damping ratios.
1 2 3 4 5 6 7 8 9 10
TABDMP1 ID TYPE +ABC
+ABC f1 g1 f2 g2 f3 g3 f4 g4 +DEF
+DEF f5 g5 ... ... ENDT
● Type = G (default), CRIT, or Q

= b 
bcr = G 
2
Q = 1
2 
= 1 /G
● Example: for 10% critical damping

CRIT = 0.10
Q = 5.0
G = 0.2
(Cont.)
● The TABDMP1 Bulk Data entry is selected with the SDAMPING Case
Control command.
● fi (units: Hz) and gi define pairs of frequencies and dampings. Straight-
line interpolation is used for modal frequencies between consecutive fis.
Linear extrapolation is used at the ends of the table. ENDT ends the
table input.
● Example: Assume modes at 1.0, 2.5, 3.6, and 5.5 Hz.
Entered Computed Calculated by
f g f g Extrapolation
2.0 0.10 1.0 0.02
3.0 0.18 2.5 0.14 Calculated by
4.0 0.13 3.6 0.15
6.0 0.13 5.5 0.13
Interpolation
● May add nonmodal damping (PARAM, G; VISC; DAMP; GE on MATi entry)

● Computational cost due to coupled B causing direct integration to
be used
● Recommended practice: Use only modal damping (TABDMP1) in modal
transient response analysis. If discrete damping is desired, use direct
transient response analysis.

DATA RECOVERY IN MODAL TRANSIENT
RESPONSE
● Recover physical response as the summation of the modal
responses.
u = 
 
● Not as large a computational penalty for changing t in

modal transient response as in direct. However, the
constant is still recommended.
● The output time interval may be greater than the solution

time interval.

MODE TRUNCATION
● May not need all of computed modes. Often only the
lowest few will suffice for dynamic response calculation.
● PARAM,LFREQ specifies the lower limit on the frequency
range of retained modes.
● PARAM,HFREQ specifies the upper limit on the frequency
range of retained modes.
● PARAM,LMODES specifies the number of the lowest
modes to be retained.
● Truncating high-frequency modes truncates high-
frequency response.

SELECTIVE MODES DELETION
● By default all modes calculated are included in the
response analysis
● Parameters available for excluding modes at low or high
end
● Param,lfreq,value--modes below frequency “value” are not included
Default value = 0.0
● Param,hfreq,value--modes above frequency “value” are not included
Default value = 1.E+30
● Param,lmodes,number—only the lowest “number” modes are included
Default number = 0 - > the retained modes are
determined by parameter LFREQ and HFREQ
● Similar set of parameters are available for fluid modes

DELETING HIGH/LOW MODES
● Command to delete modes below and/or above certain
frequencies FLSFEL is a CASE CONTROL command
0.0 1.+30
FLSFSEL LFREQ = , HFREQ = ,
fs1 fs2
fs1 = lower freq range for

structure (real number)
0.0 fs2 = upper freq range for
LMODES = structure (real number)
ms ms = number of lowest
modes to use for structure
Example: portion of model
FLSFSEL HFREQ = 4.
SELECTIVE MODES DELETION (Cont.)
● Previous commands remove high/low modes
● MODESELECT allows you to remove selective mode
MODESELECT is a CASE CONTROL command
● Use to either:
● Include a selective set of modes (n>0)
or
● Excludes a selective set of modes ( n < 0 )

SELECTIVE MODES DELETION (Cont.)
● Note that the modes that are deleted will not participate in
the response
● This may lead in incorrect results if the wrong modes are deleted
Examples:
1. Selects all 10 modes excluding modes 6 and 7.
SET 100 = 1 THRU 10 EXCEPT 6,7
MODESELECT = 100
or
SET 200 = 6,7
MODESELECT = -200
TRANSIENT EXCITATION
● Define force as a function of time.
● Several methods in MD Nastran:
● TLOAD1 “Brute force”; ordered time, force pairs table input
● TLOAD2 Efficient definition for analytical-type loadings
● LSEQ Generates dynamic loads from static loads (not

recommended)

TLOAD1 ENTRY
● Defines excitation in the form:
P
t = AF 
t – 
 Where A =
spatial load distribution and scale factor
(DAREA, static load, thermal load, or LSEQ)
  = DELAY entry (Integer) or time delay  
(Real),Default=0.
 F(t-
) = TABLEDi entry
● DELAY defines DOFs and time delay
● TID – specifies TABLEDi for defining time and force pairs.
● Selected by DLOAD Case Control command.
Delay SID P1 C1 T1 P2 C2 T2
Pi – GRID number
Ci – Compenent number
Ti – Time delay for designated point Pi and component Ci (real)
TLOAD1 ENTRY (Cont.)
● Excitation is defined by TYPE.
● Only loads (first row) will be discussed in this section. For

enforced motion, see section 12.

TLOAD2 ENTRY
● Defines excitation in the form of ˜
t < 0 or ˜
t >T –T t<(T1+ ) or
t<(T1+
0 2 1
t>(T2+
t>(T2+ )
where ˜
t = t–T – P(t) =
1 Be Ct˜
At˜ 2Ft˜+ P  (T1+
cos  (T1+)  t  (T2+
(T2+ )
 A Defined as a spatial load distribution and scale factor

(DAREA, static load, thermal load, or LSEQ)
  Defined on a DELAY entry
 TYPE Defined as in the TLOAD1
 T1,T2 Time constants (T2>T1)
 F Frequency (Hz)
 P Phase angle (degrees)
 C Exponential coefficient
 B Growth coefficient
● Selected by the DLOAD Case Control command

LOAD SET COMBINATION - DLOAD
● The applied load PC is constructed from a combination of
component load sets PK
P
C
= S
C S K P K
K
where SC = overall scale factor
SK = scale factor for k-th load set
PK = SID of TLOAD
1 2 3 4 5 6 7 8 9 10
DLOAD SID SC SI P1 S2 P2 -etc-
● TLOAD1s and TLOAD2s must have unique SIDs.

● Use the DLOAD entry to combine TLOADs.
● The DLOAD Bulk Data entry is selected by DLOAD Case
Control command.

DAREA ENTRY
● Defines the degree of freedom where the dynamic load is
to be applied to the scale factor.
● Can point to static load directly, e.g., FORCE, PLOAD4,
etc. (recommended) instead of DAREA
● Relationship to other input:
DLOAD
Case Control
Bulk Data
TLOADi,SID,EXCITEID
i,SID,EXCITEID,,DELAY
Dynamic Delay DAREA ,SID,P1,C1,A1,P2,C1,A2

SID,P1,C1,A1,P2,C1,A2
Load Time Lag Scale
Factor
Temporal Spatial
EXAMPLE Type = 0 => Structural
Load
TLOAD1 SID EXCITEID DELAY TYPE TID
TLOAD1 35 29 0.2 0 40
FORCE SID G CID F N1 N2 N3
FORCE 29 30 5.2 1.
TABLED1 ID XAXIS YAXIS
X1 Y1 X2 Y2 X3 Y3 X4 Y4
TABLED1 40
0. 0. 0.1 1.5 0.2 2.0 3.0 1.5
4.0 1.5 ENDT
Result is the load specified by the TLOAD1, scaled by 5.2, delayed

by 0.2 seconds, and applied to grid point 30, component T1

STATIC LOAD – INDIRECT METHOD
● Defines static loads that are being applied dynamically.
● The LSEQ Bulk Data entry is selected by the LOADSET Case Control
command.
● Contains a EXCITEID entry to identify the loadset for use with the
TLOAD entries.
● Relationship to other input
DLOAD LOADSET
Case Control
Bulk Data TLOAD LSEQ
Dynamic EXCITEID Static Load
Temporal Cross Reference Spatial

STATIC LOAD – INDIRECT METHOD (Cont.)
 DLOAD = 25 LOADSET must appear above all subcases.
Only one LOADSET may be specified per
 LOADSET = 27 superelement
 TLOAD1 25 28 5
 LSEQ 27 28 100
 PLOAD4 100 ….
TABLED1 5 ….

STATIC LOAD – DIRECT METHOD
● Defines static loads that are being applied dynamically.
● The EXCITEID references the static load (e.g., PLOAD4)

directly
DLOAD
Case Control
Bulk Data TLOAD
Dynamic EXCITEID
Temporal Static Load

STATIC LOAD – DIRECT METHOD (Cont.)
 DLOAD = 25
 TLOAD1 25 100 5
 PLOAD4 100 …..
TABLED1 5 ….

TRANSIENT EXCITATION CONSIDERATIONS
● Remember the 1/3 “smearing” of applied loads. This will smooth the force and
decreases apparent frequency content.
● Avoid discontinuous forces. These may cause different results on different computers.
A-
A
Force
B
C+
C
Time
● If NDt causes a solution at ABC, then MD Nastran should select the average force B.
● However, due to numerical roundoff, NDt on one computer may be at time A- and will
give force A. On another computer, NDt may be at time C+ and will give force C.
● The integration results will differ depending on whether the force at NDt is A, B, or C.

TRANSIENT EXCITATION CONSIDERATIONS
(Cont.)
● Smooth a discontinuous force over one t.
Force = original force

= smoothed force
Time

INITIAL CONDITIONS
● May impose initial displacement and/or velocity in transient response
via the TIC Bulk Data entry.
● The IC Case Control command selects the TIC entry.
● Be careful - initial conditions for unspecified DOFs are set to zero.
● Initial conditions may be specified only for A-set DOFs.
● Initial conditions are used to determine the values of {u0}, {u-1}, {P0},
and {P-1} used in calculating {u 1}. The acceleration for all points is
assumed to be zero for t < 0 (constant velocity).
·
u – = u – u 0t
1 0
·
P – =  Ku
 – + Bu 0
1 1
● The load specified by the user at t = 0 is replaced by:
·
P = 
K u
 + 
B u
 0
0 0

INITIAL CONDITIONS (Cont.)
● The recommended practice for any type of dynamic excitation is to use at
least one time step of zero excitation prior to applying the dynamic force.
Force
Time

Transient Initial Condition TIC
Defines values for the initial conditions of variables used in structural transient analysis. Both displacement
and velocity values may be specified at independent degrees of freedom. This entry may not be used for heat
transfer analysis.
IC = SID in Case Control
Format:
1 2 3 4 5 6 7 8 9 10
TIC SID G C U0 V0
Example:
TIC 1 3 2 .
5.0 .
-6.0
Field Contents
SID Set identification number. (Integer > 0)
G Grid, scalar, or extra point identification number. (integer > 0)
C Component numbers. (Integer zero or blank for scalar or extra points, any one of the Integers 1
through 6 for a grid point.)
U0 Initial displacement. (Real)
V0 Initial velocity. (Real)
Remarks:
1. Transient initial condition sets must be selected with the Case Control command IC = SID. Note
the use of IC in the Case Control command versus TIC on the Bulk Data entry. For heat transfer,
the IC Case Control command selects TEMP or TEMPD entries for
2. If no TIC set is selected in the Case Control Section, all initial conditions are assumed to be zero.
3. Initial conditions for coordinates not specified on TIC entries will be assumed zero.
Bulk Data Entry

TSTEP ENTRY Recommended:
T  1./(10* fmax)
● Select integration time step for direct and modal transient response.
● Integration errors increase with increasing natural frequency.
● Recommended t is to use at least eight solution time steps per period (cycle) of
response.
● The TSTEP Bulk Data entry controls solution and output t, and is selected by
the TSTEP Case Control command.
● The cost of integration is directly proportional to the number of time steps
when t is constant.
● Use an adequate length of time to properly capture long-period (low
frequency) response.
··
● User may change t during a run. u is assumed constant for t < Nt1 .
 ··
u calculates new initial conditions for the integration based on {un} and the
calculated velocity and acceleration at the transition. The assumption of
uniform acceleration assures a smooth transition when t is changed.
· 1
u 0 = --------- uN – u N – 1 
t 1
·· 1

u 0 = --------- uN – 2u N – 1 + u N – 2 
NAS102,Section 7, March 2007 t12
TSTEP ENTRY (Cont.)
● The initial conditions for the new integration are:

u 0 = u N 
·  1 ··
u – 1= u N – t2 u 0 – ---t22u 0 
  2 Uniform
P 0 = PN  Acceleration
·  ··
P –1 = 
Ku
 – 1+ B u
 – 1 + Mu – 1

 ·  ··  ··
= 
Ku – 1+ 
B u
 0 – t 2u 0 + Mu
0
 
● Note: New matrices A1 - A4 (page S7-7) are

assembled, and the new A1 must be decomposed.

Transient Time Step TSTEP
Defines time step intervals at which a solution will be generated and output in transient analysis.
TSTEP = SID in CASE Control
Format:
1 2 3 4 5 6 7 8 9 10
TSTEP SID N1 DT1 NO1
N2 DT2 NO2
-etc.-
Example:
TSTEP 2 10 0.001 5
9 0.01 1
Field Contents
Ni Number of time steps of value DTi. (integer 1)
DTi Time increment. (Real 0.0)
NOi Skip factor for output. Every NOi-th step will be saved for output. (Integer > 0; Default = 1)
Remarks:
1. TSTEP entries must be selected with the Case Control command TSTEP = SID.
2. Note that the entry permits changes in the size of the time step during the course of the solution.
Thus, in the example shown, there are 10 time steps of value .001 followed by 9 time steps of value
.01. Also, the user has requested that output be reco
3. See the MSC.Nastran Basic Dynamic Analysis User’s Guide, Chapter 10 for a discussion of
considerations leading to the selection of time steps.
4. In modal frequency response analysis (SOLs 111 and 146), this entry is required only when
TLOADi is requested; i.e., when Fourier methods are selected.
5. The maximum and minimum displacement at each time step and the SIL numbers of these
variables can be printed by altering DIAGON(30) before the transient module TRD1 and by altering
DIAGOFF(30) after the module. This is useful for runs that terminate due
6. For hear transfer analysis in SOL 159, use the TSTEPNL entry.
Bulk Data Entry

DYNAMIC DATA RECOVERY
● Three options for recovering displacements and stresses in modal solutions:
mode displacement method, matrix method, and mode acceleration method
● The mode displacement method computes physical displacements directly from
modal displacements and then computes element stresses from the physical
displacements. The number of operations is proportional to number of time steps
(T).
● The matrix method computes displacements per mode and element stresses per
mode and then computes physical element stresses as the summation of modal
element stresses. Costly operations are proportional to the number of modes (H).
Number of Modes
Cost of matrix method H

Cost of mode displaceme nt method T Number of Time Steps
● Since H is usually << T, the matrix method is cheaper.
● The matrix method is the default and is the recommended method for most cases.
The mode displacement method can be selected via PARAM, DDRMM, -1.
● The mode acceleration method automatically accounts for the quasi-static
response of all high frequency modes (See Appendix C ). (Requires mode
displacement method)

MODAL TRANSIENT VERSUS DIRECT TRANSIENT
Modal Direct
Small Model X
Large Model X
Few Time Steps X
Many Time Steps X
High Frequency Excitation X
Nonlinearities X
Initial Conditions X X
Modal Damping X

SOLUTION CONTROL FOR TRANSIENT ANALYSIS

● SOL (for required input see below)
Method Structured
Solution
Sequences
Direct 109
● Case Control Section Modal 112
● DLOAD (both - required)

● LOADSET (both - optional)
● METHOD (modal - required)
● SDAMPING (modal - optional)
● IC (Physical) (both - optional)
● IC (Modal) (modal – optional)
● TSTEP (both - required)
SOLUTION CONTROL FOR TRANSIENT
ANALYSIS (Cont.)
 ASET,OMIT (both - optional)
 EIGRL or EIGR (modal - required)
 TSTEP (both - required)
 TIC (both - optional)
 TLOADi (both - required)
 LSEQ (both - optional)
 DAREA (both – optional)
 TABLEDi (both - optional)
 DELAY (both - optional)
 DLOAD (both - optional)
 TABDMP1 (modal - optional)

CASE CONTROL OUTPUT
● Grid output
● ACCELERATION
● DISPLACEMENT (or VECTOR)
● GPSTRESS
● NLLOAD (nonlinear load output)
● OLOAD (output applied load)
● SACCELERATION (solution set output - A-set in direct
● SDISPLACEMENT solutions, modal variables in
● SVELOCITY modal solutions)
● SVECTOR (A-set eigenvector)
● SPCFORCES
● VELOCITY
● MPCFORCE
● Element output
● ELSTRESS (or STRESS)
● ELFORCE (or FORCE)
● STRAIN
● Special
● OTIME (controls solution output times)
WORKSHOP #3
DIRECT TRANSIENT RESPONSE

WORKSHOP #3 - DIRECT TRANSIENT RESPONSE
Using the direct method, determine the transient response of the flat rectangular plate,
created in Workshop 1, subject to time-varying excitation. This example structure is
excited by 1 psi pressure load over the total surface of the plate varying at 250 Hz. In
addition, a 50 lb force is applied at a corner of the tip also varying at 250 Hz, but 180o
out-of-phase with the pressure load. Both time dependent dynamic loads are applied for
the duration of 0.008 seconds only. Use structural damping of g=0.06 and convert this
damping to equivalent viscous damping at 250 Hz. Carry the analysis for 0.04 seconds.
Below is a finite element representation of the flat plate. It also contains the loads and
boundary constraints.
1 p s i o v e r t h e t o t a l s u rf a ce
5 0.0 0

Figure 7-1. Loads and Boundary Conditions.
WORKSHOP # 3
• Use the following file as the starting point
$ wkshp3.dat
$
$
SOL 109
CEND
TITLE= TRANSIENT RESPONSE WITH TIME DEPENDENT PRESSURE AND POINT LOADS
SUBTITLE= USE THE DIRECT METHOD
ECHO= PUNCH
SPC= 1
SET 1= 11, 33, 55
DISPLACEMENT= 1
SUBCASE 1
DLOAD= 700 $ SELECT TEMPORAL COMPONENT OF TRANSIENT LOADING
TSTEP= 100 $ SELECT INTEGRATION TIME STEPS
$
OUTPUT (XYPLOT)
XGRID=YES
YGRID=YES
XTITLE= TIME (SEC)
YTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER
XYPLOT DISP RESPONSE / 11 (T3)
YTITLE= DISPLACEMENT RESPONSE AT CENTER TIP
YTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER
$
BEGIN BULK
param,post,0
PARAM, COUPMASS, 1
$
INCLUDE 'plate.bdf'
$
ENDDATA

$
$ soln3.dat $
$ $ SPECIFY STRUCTURAL DAMPING
ID SEMINAR, PROB3 $ 3 PERCENT AT 250 HZ. = 1571 RAD/SEC.
SOL 109 $
CEND PARAM, G, 0.06
TITLE= TRANSIENT RESPONSE WITH TIME DEPENDENT PARAM, W3, 1571.
PRESSURE AND POINT LOADS $
SUBTITLE= USE THE DIRECT METHOD $ TIME VARYING PRESSURE LOAD (250 HZ)
ECHO= PUNCH $
SPC= 1 TLOAD2, 200, 400, , 0, 0., 8.E-3, 250., -90.
SET 1= 11, 33, 55 PLOAD2, 400, 1., 1, THRU, 40
DISPLACEMENT= 1 $
SUBCASE 1 $ APPLY POINT LOAD OUT OF PHASE WITH PRESSURE LOAD
DLOAD= 700 $ SELECT TEMPORAL COMPONENT OF $
TRANSIENT LOADING TLOAD2, 500, 600, , 0, 0., 8.E-3, 250., 90.
TSTEP= 100 $ SELECT INTEGRATION TIME STEPS $
$ DAREA, 600, 11, 3, 1.
OUTPUT (XYPLOT) $
XGRID=YES $ COMBINE LOADS
YGRID=YES $
XTITLE= TIME (SEC) DLOAD, 700, 1., 1., 200, 50., 500
YTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER $
XYPLOT DISP RESPONSE / 11 (T3) $ SPECIFY INTERGRATION TIME STEPS
YTITLE= DISPLACEMENT RESPONSE AT CENTER TIP $
XYPLOT DISP RESPONSE / 33 (T3) TSTEP, 100, 100, 4.0E-4, 1
YTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER $
XYPLOT DISP RESPONSE / 55 (T3) ENDDATA
$
BEGIN BULK
param,post,0
PARAM, COUPMASS, 1
$
INCLUDE 'plate.bdf'
$

0 SUBCASE 1ML
0
SUBCASE 1
POINT-ID = 11
D I S P L A C E M E N T V E C T O R
TIME TYPE T1 T2 T3 R1 R2 R3
0.0 G 0.0 0.0 0.0 0.0 0.0 0.0
4.000000E-04 G 0.0 0.0 -2.173954E-02 1.105127E-02 1.051570E-02 0.0
8.000000E-04 G 0.0 0.0 -7.205779E-02 2.849259E-02 2.853552E-02 0.0
1.200000E-03 G 0.0 0.0 -1.433632E-01 4.084704E-02 4.916111E-02 0.0
1.600000E-03 G 0.0 0.0 -2.060064E-01 3.055025E-02 6.223300E-02 0.0
2.000000E-03 G 0.0 0.0 -2.459063E-01 1.403890E-03 6.812792E-02 0.0
0 SUBCASE
1
POINT-ID = 33
0.0 G 0.0 0.0 0.0 0.0 0.0 0.0
4.000000E-04 G 0.0 0.0 -1.122321E-02 9.220317E-03 6.136194E-03 0.0
8.000000E-04 G 0.0 0.0 -4.424645E-02 2.577100E-02 2.014592E-02 0.0
1.200000E-03 G 0.0 0.0 -1.030766E-01 3.820221E-02 3.921979E-02 0.0
1.600000E-03 G 0.0 0.0 -1.756299E-01 2.929389E-02 5.577315E-02 0.0
2.000000E-03 G 0.0 0.0 -2.443358E-01 1.775214E-03 6.761315E-02 0.0
2.400000E-03 G 0.0 0.0 -2.882814E-01 -2.449635E-02 7.435711E-02 0.0
2.800000E-03 G 0.0 0.0 -2.843486E-01 -3.428704E-02 7.112917E-02 0.0
0 SUBCASE
1
POINT-ID = 55
0.0 G 0.0 0.0 0.0 0.0 0.0 0.0
4.000000E-04 G 0.0 0.0 -2.850083E-03 7.786633E-03 4.615276E-03 0.0
8.000000E-04 G 0.0 0.0 -1.992840E-02 2.321643E-02 1.681915E-02 0.0
1.200000E-03 G 0.0 0.0 -6.642522E-02 3.539676E-02 3.502940E-02 0.0

0
0 SUBCASE 1
0 X Y - O U T P U T S U M M A R Y ( R E S P O N S E )
0 SUBCASE CURVE FRAME CURVE ID./ XMIN-FRAME/ XMAX-FRAME/ YMIN-FRAME/ X FOR YMAX-FRAME/ X FOR
ID TYPE NO. PANEL : GRID ID ALL DATA ALL DATA ALL DATA YMIN ALL DATA YMAX
0 1 DISP 1 11( 5) 0.000000E+00 4.000000E-02 -2.623262E-01 2.400000E-03 2.823412E-01 5.200000E-03
0.000000E+00 4.000000E-02 -2.623262E-01 2.400000E-03 2.823412E-01 5.200000E-03
0 1 DISP 2 33( 5) 0.000000E+00 4.000000E-02 -2.882814E-01 2.400000E-03 3.220591E-01 5.200000E-03
0.000000E+00 4.000000E-02 -2.882814E-01 2.400000E-03 3.220591E-01 5.200000E-03
0 1 DISP 3 55( 5) 0.000000E+00 4.000000E-02 -3.166115E-01 2.800000E-03 3.570641E-01 5.200000E-03
0.000000E+00 4.000000E-02 -3.166115E-01 2.800000E-03 3.570641E-01 5.200000E-03

(Cont.)

(Cont.)

(Cont.)

WORKSHOP #4
MODAL TRANSIENT RESPONSE

WORKSHOP #4 - MODAL TRANSIENT
RESPONSE
 Using the Modal Method, determine the transient response of the flat rectangular plate, created in
Workshop 1, subject to time-varying excitation. This example structure is excited by a 1 psi pressure
load over the total surface of the plate varying at 250 Hz. In addition, a 25 lb force is applied at a
corner of the tip also varying at 250 Hz, but starting 0.004 seconds after the pressure load begins.
Both time-dependent dynamic loads are applied for a duration of 0.008 seconds. Use a modal
damping of = 0.03 for all modes. Carry out the analysis for 0.04 seconds.
 Below is a finite element representation of the flat plate. It also contains the loads and boundary
constraints.
12345
12345 1.0 psi over the total surface
12345
12345
12345
25.0
Z
Y
X

WORKSHOP # 4
● Use the following file as a starting point
$ BEGIN BULK
$ wkshp4.dat PARAM, COUPMASS, 1
$ PARAM, WTMASS, 0.00259
SOL 112 $
diag 8 $ PLATE MODEL DESCRIBED IN NORMAL MODES EXAMPLE PROBLEM
CEND $
TITLE = TRANSIENT RESPONSE WITH TIME DEPENDENT PRESSURE INCLUDE 'plate.bdf'
AND POINT LOADS $
SUBTITLE = USE THE MODAL METHOD ENDDATA
ECHO = UNSORTED
SPC = 1
SET 111 = 11, 33, 55
DISPLACEMENT(SORT2) = 111
SDAMPING = 100
SUBCASE 1
METHOD = 100
DLOAD = 700
TSTEP = 100
$
OUTPUT (XYPLOT)
$

ID SEMINAR, PROB4 BEGIN BULK
$ PARAM, COUPMASS, 1
$ soln4.dat PARAM, WTMASS, 0.00259
$ $
ID SEMINAR, PROB4 $ PLATE MODEL DESCRIBED IN NORMAL MODES EXAMPLE
SOL 112 PROBLEM
diag 8 $
CEND INCLUDE 'plate.bdf'
TITLE = TRANSIENT RESPONSE WITH TIME DEPENDENT $
PRESSURE AND POINT LOADS $ EIGENVALUE EXTRACTION PARAMETERS
SUBTITLE = USE THE MODAL METHOD $
ECHO = UNSORTED EIGRL, 100, , ,5
SPC = 1 $
SET 111 = 11, 33, 55 $ SPECIFY MODAL DAMPING
DISPLACEMENT(SORT2) = 111 $
SDAMPING = 100 TABDMP1, 100, CRIT,
SUBCASE 1 +, 0., .03, 10., .03, ENDT
METHOD = 100 $
DLOAD = 700 $ TIME VARYING PRESSURE LOAD (250 HZ)
TSTEP = 100 $
$ TLOAD2, 200, 400, , 0, 0., 8.E-3, 250., -90.
OUTPUT (XYPLOT) PLOAD2, 400, 1., 1, THRU, 40
XGRID=YES $
YGRID=YES $ APPLY POINT LOAD (250 HZ)
XTITLE= TIME (SEC) $
YTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER TLOAD2, 500, 600,610, 0, 0.0, 8.E-3, 250., -90.
XYPLOT DISP RESPONSE / 11 (T3) $
YTITLE= DISPLACEMENT RESPONSE AT TIP CENTER DAREA, 600, 11, 3, 1.
XYPLOT DISP RESPONSE / 33 (T3) DELAY, 610, 11, 3, 0.004
YTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER $
XYPLOT DISP RESPONSE / 55 (T3) $ COMBINE LOADS
$ $
DLOAD, 700, 1., 1., 200, 25., 500
$
$ SPECIFY INTERGRATION TIME STEPS
$
TSTEP, 100, 100, 4.0E-4, 1
$
NAS102,Section 7, March 2007 ENDDATA
(Cont.)
(BEFORE AUGMENTATION OF RESIDUAL VECTORS)
1 1 7.055894E+05 8.399937E+02 1.336891E+02 1.000000E+00 7.055894E+05
2 2 1.877186E+07 4.332651E+03 6.895628E+02 1.000000E+00 1.877186E+07
3 3 2.811177E+07 5.302053E+03 8.438480E+02 1.000000E+00 2.811177E+07
4 4 1.929422E+08 1.389036E+04 2.210720E+03 1.000000E+00 1.929422E+08
5 5 2.221657E+08 1.490523E+04 2.372240E+03 1.000000E+00 2.221657E+08
(AFTER AUGMENTATION OF RESIDUAL VECTORS)
1 1 7.055894E+05 8.399937E+02 1.336891E+02 1.000000E+00 7.055894E+05
2 2 1.877186E+07 4.332651E+03 6.895628E+02 1.000000E+00 1.877186E+07
3 3 2.811176E+07 5.302053E+03 8.438479E+02 1.000000E+00 2.811176E+07
4 4 1.929422E+08 1.389036E+04 2.210720E+03 1.000000E+00 1.929422E+08
5 5 2.221657E+08 1.490523E+04 2.372240E+03 1.000000E+00 2.221657E+08
6 6 2.351324E+08 1.533403E+04 2.440486E+03 1.000000E+00 2.351324E+08
7 7 7.974902E+08 2.823987E+04 4.494515E+03 1.000000E+00 7.974902E+08
8 8 1.453224E+09 3.812117E+04 6.067173E+03 1.000000E+00 1.453224E+09
9 9 2.625274E+09 5.123743E+04 8.154690E+03 1.000000E+00 2.625274E+09
10 10 4.154733E+09 6.445722E+04 1.025868E+04 1.000000E+00 4.154733E+09
11 11 4.205890E+09 6.485284E+04 1.032165E+04 1.000000E+00 4.205890E+09
12 12 3.216783E+10 1.793539E+05 2.854506E+04 1.000000E+00 3.216783E+10

PARTIAL OUTPUT FILE FOR WORKSHOP #4 (Cont.)
POINT-ID = 11
0.0 G 0.0 0.0 0.0 0.0 0.0 0.0
4.000000E-04 G -5.468071E-16 8.320540E-17 2.038250E-04 -3.199117E-06 2.623944E-05 2.230269E-16
8.000000E-04 G -2.152597E-15 6.019959E-15 1.980818E-03 -1.510274E-05 -2.074038E-04 3.010193E-15
1.200000E-03 G -4.731214E-15 1.976858E-14 6.911292E-03 4.029346E-07 -1.653521E-03 9.216824E-15
1.600000E-03 G -8.185502E-15 3.999204E-14 1.407448E-02 2.646134E-05 -3.815677E-03 1.824066E-14
2.000000E-03 G -1.174987E-14 6.113124E-14 2.121053E-02 4.361140E-05 -5.847335E-03 2.759835E-14
POINT-ID = 33
0.0 G 0.0 0.0 0.0 0.0 0.0 0.0
4.000000E-04 G -8.130047E-16 5.783684E-16 2.019039E-04 1.682611E-08 2.577562E-05 8.783539E-16
8.000000E-04 G -5.437500E-15 9.431799E-15 1.971792E-03 -1.131783E-09 -2.092973E-04 7.575778E-15
1.200000E-03 G -1.469855E-14 2.928089E-14 6.911632E-03 -3.483805E-08 -1.653002E-03 2.206479E-14
1.600000E-03 G -2.785101E-14 5.819658E-14 1.409060E-02 -1.660608E-08 -3.811871E-03 4.290151E-14
2.000000E-03 G -4.146778E-14 8.828006E-14 2.123710E-02 1.814818E-10 -5.841007E-03 6.439503E-14
2.400000E-03 G -4.989242E-14 1.075860E-13 2.615920E-02 1.753988E-08 -7.446125E-03 7.806971E-14
2.800000E-03 G -4.913322E-14 1.067912E-13 2.613206E-02 2.261718E-08 -7.588134E-03 7.724409E-14
POINT-ID = 55
0.0 G 0.0 0.0 0.0 0.0 0.0 0.0
4.000000E-04 G -2.313598E-15 1.279507E-15 2.038511E-04 3.216275E-06 2.622230E-05 1.886289E-15
8.000000E-04 G -1.738001E-14 1.431151E-14 1.980840E-03 1.514255E-05 -2.074704E-04 1.465446E-14
1.200000E-03 G -4.910303E-14 4.295687E-14 6.911171E-03 -5.569029E-07 -1.653289E-03 4.203939E-14
1.600000E-03 G -9.447002E-14 8.441534E-14 1.407446E-02 -2.646987E-05 -3.815674E-03 8.127868E-14
2.000000E-03 G -1.412618E-13 1.273955E-13 2.121054E-02 -4.360946E-05 -5.847338E-03 1.216724E-13
2.400000E-03 G -1.707585E-13 1.548346E-13 2.611754E-02 -6.863064E-05 -7.455783E-03 1.473061E-13
2.800000E-03 G -1.686479E-13 1.534346E-13 2.608514E-02 -7.741245E-05 -7.598883E-03 1.456237E-13
3.200000E-03 G -1.287358E-13 1.173798E-13 1.986141E-02 -5.680266E-05 -5.762664E-03 1.111591E-13
3.600000E-03 G -5.511781E-14 5.048424E-14 8.672326E-03 -2.997875E-05 -2.603049E-03 4.766799E-14

(Cont.)
X Y - O U T P U T S U M M A R Y ( R E S P O N S E )
0 1 DISP 1 11( 5) 0.000000E+00 4.000000E-02 -1.666260E-01 9.200000E-03 1.482653E-01 6.400000E-03
0.000000E+00 4.000000E-02 -1.666260E-01 9.200000E-03 1.482653E-01 6.400000E-03
0 1 DISP 2 33( 5) 0.000000E+00 4.000000E-02 -1.846309E-01 9.200000E-03 1.602995E-01 6.800000E-03
0.000000E+00 4.000000E-02 -1.846309E-01 9.200000E-03 1.602995E-01 6.800000E-03
0 1 DISP 3 55( 5) 0.000000E+00 4.000000E-02 -2.001817E-01 9.200000E-03 1.730418E-01 6.800000E-03
0.000000E+00 4.000000E-02 -2.001817E-01 9.200000E-03 1.730418E-01 6.800000E-03

(Cont.)

(Cont.)

(Cont.)

Slide 1
SECTION 8
FREQUENCY RESPONSE ANALYSIS
NAS102,Seminar 8, March 2007

Copyright 2007 MSC.Software Corporation Sem 8-1
Slide 2
INTRODUCTION TO FREQUENCY RESPONSE

ANALYSIS
• Compute the response to oscillatory excitation.

• Excitation explicitly defined in the frequency domain -
all of the applied forces are known at each forcing
frequency.
• Computed response usually includes nodal
displacements and element forces and stresses.
• The computed responses are the complex numbers
defined as magnitude and phase (with respect to the
forcing) or as real and imaginary components.
• Two categories of analysis - direct and modal.
Slide 3
DIRECT FREQUENCY RESPONSE

• Dynamic equation of motion:
2
(1) –  M + iB + K u
 
= P 

• PARAM,G and GE on MATi entry do not form a damping matrix. They form a
complex stiffness matrix
1
K = 
1 + iG 
K +i
G EkE
where K1 = global stiffness matrix
G = overall structural damping coefficient (PARAM,G)
KE = element stiffness matrix
GE = element structural damping coefficient (GE on MATi entry)
• Contrast this with transient response analysis
1 2 G 1 1
B
T RANS
= B + B + --------- K + ---------
W
3
W
4
G EkE
• Solve the equation by inserting ω to form a complex left-hand side, and then
solve it similar to a static problem (using complex arithmetic).

Slide 4
MODAL FREQUENCY RESPONSE

• Convert to modal coordinates and solve as decoupled
SDOF systems
P
i
 = --------------------------------------------------
i 2
– m  + ib + k
i i i
• Much quicker to solve this equation than in direct

method
• Decoupled procedure can be used only if either no
damping is present or if modal damping alone (via
TABDMP1) is used. Otherwise, use the less efficient
direct approach (on smaller modal coordinate
matrices) if non-modal damping (VISC,DAMP) is
present.
Slide 5
EXCITATION DEFINITION
• Define force as a function of frequency.
• Several methods in MD Nastran:
• RLOAD1 (defines frequency-dependent load in real and imaginary
forms)
• RLOAD2 (defines frequency-dependent load in magnitude and
phase forms)
• LSEQ (generates dynamic loads from static loads)
• DLOAD Bulk Data entries are used to combine
frequency-dependent forces.
• RLOADi entries are selected by DLOAD Case Control
commands.

Slide 6

Slide 7

Slide 8
FREQUENCY RESPONSE CONSIDERATIONS

• Exciting an undamped (or modal damped) system at 0.0 Hz gives the same results
as a static analysis. Therefore, if the maximum excitation frequency is much less
than the lowest resonant frequency of the system, a static analysis is sufficient.
• Very lightly-damped structures exhibit large dynamic responses for excitation
frequencies near resonant frequencies. A small change in the model (or running
it on another computer) may give large changes in such response.
• Use a fine-enough frequency step size (f) to adequately predict peak response.
Use at least 5 points per half-power bandwidth.
Peak Response
Response Peak / = Half-

Power Point
f1 f2 Frequency
Half-Power
• For maximum efficiency, use an uneven frequency step size. Use smaller Df in
regions around resonant frequencies and larger Df in regions far away from
resonant frequencies.
Slide 9
FREQi ENTRIES
Applicable for Direct and Modal Method
• Select frequency step size.

• The FREQ entry defines discrete excitation frequencies.
• The FREQ1 entry defines fSTART, frequency increment, and the number of
increments.
• The FREQ2 entry defines fSTART, fEND, and the number of logarithmic
intervals.
Applicable for Modal Method
• The FREQ3 entry defines F1, F2, and the number of frequencies in
between using either a linear or log interpolation. Biased towards end
points or center.
• The FREQ4 entry specifies a frequency at each resonant frequency and
the number of equally spaced excitation frequencies within the spread.
• The FREQ5 entry specifies a frequency range and fractions of the natural
frequencies within that range.
• The FREQ3, FREQ4, and FREQ5 entries are available only for the modal
method.
Slide 10
FREQi ENTRIES (CONT.)
• FREQi Bulk Data entries are selected by the

FREQUENCY Case Control commands.
• All FREQi Bulk Data entries with the same set ID are
used. Therefore, FREQ, FREQ1, FREQ2, FREQ3,
FREQ4, and FREQ5 entries may all be used in an
analysis.

Slide 11
Frequency List FREQ

Defines a set of frequencies to be used in the solution of frequency response problems.
Format:
1 2 3 4 5 6 7 8 9 10
FREQ SID F1 F2 F3 F4 F5 F6 F7
F8 F9 F10 -etc.-
Example:
FREQ 3 2.98 3.05 17.9 21.3 25.6 28.8 31.2
29.2 22.4 19.3
Field Contents
Fi Frequency value in units of cycles per unit time. (Real 0.0)
Remarks:
1. Frequency sets must be selected with the Case Control command FREQUENCY = SID.
2. All FREQi entries with the same frequency set identification numbers will be used. Duplicate
frequencies will be ignored. f N and f N-1 are considered duplicated if |f N - f N-1 |<DFREQ*|f MAX - f MIN|,
where DFREQ is a user parameter, with default of 10-5. f
3. In modal analysis, solutions for modal DOFs from rigid body modes at zero excitation frequencies
may be discarded. Solutions for nonzero modes are retained.
Bulk Data Entry

Slide 12
Frequency List, Alternate Form 1 FREQ1

Defines a set of frequencies to be used in the solution of frequency response problems by specification of a
starting frequency, frequency increment, and the number of increments desired..
Format:
1 2 3 4 5 6 7 8 9 10
FREQ1 SID F1 DF NDF
Example:
FREQ1 6 2.9 0.5 13
Field Contents
F1 First frequency set. (Real 0.0)
DF Frequency increment. (Real > 0.0)
NDF Number of frequency increments. (Integer > 0; Default = 1)
Remarks:
1. FREQ1 entries must be selected with the Case Control command FREQUENCY = SID.
2. The units for F1 and DF are cycles per unit time.
3. The frequencies defined by this entry are given by
fi = F1 + DF * (i - 1)
where i = 1 to (NDF + 1)
frequencies will be ignored. fN and fN-1 are considered duplicated if |f N - f N-1|<DFREQ*|fMAX - f MIN|,
where DFREQ is a user parameter, with default of 10-5. fMAX and fMIN are the maximum and
minimum excitation frequencies of the combined FREQi entries.
(Continued)
Bulk Data Entry

Slide 13

Defines a set of frequencies to be used in the solution of frequency response problems by specification of a
starting frequency, final frequency, and the number of logarithmic increments desired..
Format:
1 2 3 4 5 6 7 8 9 10
FREQ2 SID F1 F2 NF
Example:
FREQ2 6 1 8 6
Field Contents
F1 First frequency. (Real > 0.0)
F2 Last frequency. (Real > 0.0, F2 > F1)
NF Number of logarithmic intervals. (Integer > 0; Default = 1)
Remarks:
2. The units for F1 and F2 are cycles per unit time.
3. The frequencies defined by this entry are given by 1
fi = F1 * e (i - 1)d f i1 F NF
where d = (1/NF) * ln (F2/F1) and i = 1, 2, . . . , (NF + 1)

e d  2 
In the example above, the list of frequencies will be 1.0, 1.4142, 2.0, 2.8284, 4.0, 5.6569, and 8.0
fi F1 
cycles per unit time
frequencies will be ignored. f N and f N-1 are considered duplicated if |f N - fN-1|<DFREQ*|f MAX - fMIN|,
where DFREQ is a user parameter, with default of 10-5. f
(Continued)
Bulk Data Entry

Slide 14

Defines a set of excitation frequencies for modal frequency response solutions by speciying a number of
excitation frequencies between two modal frequencies.
Format:
1 2 3 4 5 6 7 8 9 10
FREQ3 SID F1 F2 TYPE NEF CLUSTER
Example:
FREQ3 6 20 2000 LINEAR 10 2
Field Contents
F1 Lower bound of modal frequency range in cycles per unit time. (Real > 0.0)
F2 Upper bound of modal frequency range in cycles per unit time. (Real>0.0, F2F1, Default = F1)
TYPE LINEAR or LOG. Specifies linear or logarithmic interpolation between frequencies.
(Character; Default = "LINEAR")
NEF Number of excitation frequencies within each subrange including the end points. The
first subrange is between F1 and the first modal frequency within the bounds. The
second subrange is between first and second modal frequencies between the bounds.
The
CLUSTER Specifies clustering of the excitation frequency near the end points of the range.See
Remark3. (Real > 0.0; Default=1.0)
Remarks:
1. FREQ3 applies only to modal frequency-response solutions (SOLs 11, 111, 146, and 200) and is
ignored in direct frequency response solutions.
3. In the example above, there will be 10 frequencies in the interval between each set of modes
within the bounds 20 and 2000, plus 10 frequencies between 20 and the lowest mode in the range,
plus 10 frequencies between the highest mode in the range and 2000
4. Since the forcing frequencies are near structural resonances, it is important that some amount of
damping be specified.
(Continued)
Bulk Data Entry
Slide 15

Defines a set of frequencies used in the solution of modal frequency-response problems by specifying the
amount of "spread" around each natural frequency and the number of equally spaced excitation frequencies
within the spread.
Format:
1 2 3 4 5 6 7 8 9 10
FREQ4 SID F1 F2 FSPD NFM
Example:
FREQ4 6 20 200 0.3 21
Field Contents
F1 Lower bound of frequency range in cycles per unit time. (Real 0.0, Default=0.0)
F2 Upper bound of frequency range in cycles per unit time. (Real>0.0, F2 F1, Default = 1.0E20)
FSPD Frequency spread, +/- the fractional amount specified for each mode which occurs un
the frequency range F1 to F2. (1.0 > Real > 0.0; Default = 0.10)
NFM Number of evenly spaced frequencies per "spread"mode.. (Integer > 0; Default = 3; If
NFM is even, NFM + 1 will be used.)
Remarks:
3. There will be NFM excitation frequencies between (1-FSPD) * f N and (1+FSPD) * fN , for each
natural frequency in the range F1 to F2.
4. In the example above there will be 21 equally spaced frequencies across a frequency band of 0.7 *
f N to 1.3 * fN for each natural frequency that occurs between 20 and 2000. See Figure 1 for
defintion of frequency spread.
(Continued)
Bulk Data Entry

Slide 16
(1-FSPD) * fN fN (1+FSPD )* fN
Figure 1. Frequency Spread Definition

Excitation frequencies may be based on natural frequencies that are not within the range (F1 and
F2) as long as the calculated excitation frequencies are within the calculated range. Similarly, an
excitation frequency calculated based on the natural frequencies within the range (F1through F2)
may be excluded if it falls outside the range.
5. The frequency spread can also be used to define the half-power bandwidth. The half-poer
bandwidth is given by 2 * x * fN, where x is the damping ratio. Therefore, if FSPD is specified
equal to the damping ratio for the mode, NFM specifies the number of excitation frequency within
the half-power bandwidth. See Figure 2 for the definition of half-power bandwidth.
(Continued)
Bulk Data Entry

Slide 17
Peak Response
2 * * f N
Amplitude
Half-Power Point (.707 Peak)
fN Frequency
Half-Power
Bandwidth
Figure 2. Half-Power Bandwidth Definition
frequencies will be ignored. f N and fN-1 are considered duplicated if
|f N - fN-1 |<DFREQ*|f MAX - fMIN|,
where DFREQ is a user parameter, with default of 10-5. f MAX and f MIN are the maximum and
8. In design optimization, (SOL 200), the excitation frequencies generated from this entry are derived
from the natural frequencies computed in the first design cycle and the ecitation frequencies
remain fixed through subsequent design cycles. In other word
(Continued)
Bulk Data Entry

Slide 18

Defines a set of frequencies used in the solution of modal frequency-response problems by speciification of a
frequency range and fractions of the natural frequencies within that range.
Format:
1 2 3 4 5 6 7 8 9 10
FREQ5 SID F1 F2 FR1 FR2 FR3 FR4 FR5
FR6 FR7 -etc.-
Example:
FREQ5 6 20 200 1. 0.6 0.8 0.9 0.95
1.05 1.1 1.2
Field Contents 
SID Set identification number. (Integer > 0) 
F1 Lower bound of frequency range in cycles per unit time. (Real 0.0, Default=0.0)
F2 Upper bound of frequency range in cycles per unit time. (Real>0.0, F2 F1, Default = 1.0E20)
FRi Fractions of the natural frequencies in the range F1 to F2 (Real > 0.0)
Remarks:
3. The frequencies defined by this entry are given by
fi = FRi * f Ni
where fNi are the natural frequencies in the range F1 through F2.
4. In the example above, the list of frequencies will be 0.6, 0.8, 0.9, 0.95, 1.0, 1.05, 1.1, 1.2 times
each natural frequency band between 20 and 2000. If this computation results in excitation
frequencies less than F1 and greater than F2, those computed frequencies are ignored.
(Continued)
Bulk Data Entry
Slide 19

frequencies will be ignored. f N and f N-1 are considered duplicated if
|fN - f N-1|<DFREQ*|fMAX - f MIN|,
where DFREQ is a user parameter, with default of 10-5 . fMAX and f MIN are the maximum and
7. In design optimization, (SOL 200), the excitation frequencies generated from this entry are derived
from the natural frequencies computed in the first design cycle and the ecitation frequencies
remain fixed through subsequent design cycles. In other words, the excitation frequencies will not
be readjusted even if the natural frequencies change.
(Continued)
NAS102,Seminar 8, March 2007 Bulk Data Entry
Slide 20
DYNAMIC DATA RECOVERY

• The matrix method and the mode displacement method are used
to recover data from modal frequency analysis.
Cost of matrix method H
= ----
Cost of mode displacement method F
where H = number of modes

F = number of excitation frequencies
• The matrix method is the default and is cheaper for H < F and is
the recommended method for most cases.
• PARAM,DDRMM,0 (default)
• The mode displacement method may be selected by

PARAM,DDRMM,-1.
Slide 21
MODAL VERSUS DIRECT FREQUENCY

RESPONSE
Modal Direct*
Small Model X
Large Model X
Few Excitation Frequencies X
Many Excitation Frequencies X
Modal Damping X

Slide 22
SORT1 VERSUS SORT2 OUTPUT

• SORT1 is the listing of the output for each excitation frequency.
• SORT2 is the listing of the output for each requested grid point or
element.
• Primarily useful for frequency response analysis
Transient Response Frequency Response

Direct Modal Direct Modal
Default 2 2 1 1
Deformed Plot Requests 1 1 1 1
XY Plot Requests 2 2 2 2
• If SORT1 and SORT2 are mixed in a run, all output will default to SORT1
for frequency response and SORT2 for transient response.

Slide 23
SOLUTION CONTROL FOR FREQUENCY

RESPONSE
• Executive Control Section
SOL (for required input see below)
Method Structured
Soultion
Sequences
Direct 108
Modal 111
• Case Control Section
DLOAD (both - required)
LOADSET (both - optional)
METHOD (modal - required)
SDAMPING (modal - optional)
FREQUENCY (both - required)

Slide 24
SOLUTION CONTROL FOR FREQUENCY

RESPONSE (Cont.)
• Bulk Data Section
• ASET,OMIT (both - optional)

• EIGRL or EIGR (modal - required)
• FREQ (both - required)
• RLOADi (both – required)
• LSEQ (both - optional)
• DAREA (both - optional)
• DELAY (both - optional)
• DPHASE (both - optional)
• TABDMP1 (modal - optional)
• DLOAD (both - optional)
• TLOADi (both - optional)
Slide 25
CASE CONTROL OUTPUT

• Grid Point
• ACCELERATION
• DISPLACEMENT (or VECTOR)
• OLOAD
• SACCELERATION
• SDISPLACEMENT
• SVELOCITY
• SVECTOR
• SPCFORCES
• VELOCITY
• MPCFORCE
• Element Output
• ELSTRESS (or STRESS)
• ELFORCE (or FORCE)
• STRAIN
• ESE
• EKE
• EDE
• Special
• OFREQUENCY (control solution output frequencies)
Slide 26
WORKSHOP #5
DIRECT FREQUENCY RESPONSE

Slide 27
WORKSHOP #5 - DIRECT FREQUENCY

RESPONSE
Using the direct method, determine the frequency response of the flat
rectangular plate, created in Workshop 1, subject to frequency-varying
excitation. The structure is excited by a unit load at a corner of the tip.
Use a frequency step of 20 Hz between a range of 20 and 1000 Hz. Use
structural damping of g=0.06. Below is a finite element representation of
the flat plate. It also contains the loads and boundary constraints.
1.0

Slide 28
WORKSHOP # 5
• Start with following input file
$ OUTPUT (XYPLOT)
$ wkshp5.dat $
$ XTGRID= YES
SOL 108
CEND YTGRID= YES
TITLE = FREQUENCY RESPONSE DUE TO UNIT FORCE AT TIP XBGRID= YES
ECHO = UNSORTED YBGRID= YES
SPC = 1 YTLOG= YES
SET 111 = 11, 33, 55
DISPLACEMENT(SORT2, PHASE) = 111 YBLOG= NO
SUBCASE 1 XTITLE= FREQUENCY (HZ)
DLOAD = 500 YTTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER,
FREQUENCY = 100 MAGNITUDE
$ YBTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, PHASE
XYPLOT DISP RESPONSE / 11 (T3RM, T3IP)
YTTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, MAGNITUDE
YBTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, PHASE
YTTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER,
MAGNITUDE
YBTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER, PHASE
$
BEGIN BULK
param,post,0
PARAM, COUPMASS, 1
$
$ PLATE MODEL DESCRIBED IN NORMAL MODES EXAMPLE
$
INCLUDE 'plate.bdf'
$
ENDDATA

Slide 29

ID SEMINAR, PROB5 BEGIN BULK
SOL 108 param,post,0
TIME30 PARAM, COUPMASS, 1
CEND PARAM, WTMASS, 0.00259
TITLE = FREQUENCY RESPONSE DUE TO UNIT FORCE AT TIP $
ECHO = UNSORTED $ PLATE MODEL DESCRIBED IN NORMAL MODES
SPC = 1 EXAMPLE
SET 111 = 11, 33, 55 $
DISPLACEMENT(SORT2, PHASE) = 111 INCLUDE ’plate.bdf’
SUBCASE 1 $
DLOAD = 500 $ SPECIFY STRUCTURAL DAMPING
FREQUENCY = 100 $
$ PARAM, G, 0.06
OUTPUT (XYPLOT) $
$ $ APPLY UNIT FORCE AT TIP POINT
XTGRID= YES $
YTGRID= YES RLOAD2, 500, 600, , ,310
XBGRID= YES $
YBGRID= YES DAREA, 600, 11, 3, 1.0
YTLOG= YES $
YBLOG= NO TABLED1, 310,
XTITLE= FREQUENCY (HZ) , 0., 1., 1000., 1., ENDT
YTTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, MAGNITUDE $
YBTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, PHASE $ SPECIFY FREQUENCY STEPS
XYPLOT DISP RESPONSE / 11 (T3RM, T3IP) $
YTTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, MAGNITUDE FREQ1, 100, 20., 20., 49
YBTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, PHASE $
XYPLOT DISP RESPONSE / 33 (T3RM, T3IP) ENDDATA
MAGNITUDE
$

Slide 30

POINT-ID = 11
C O M P L E X D I S P L A C E M E N T V E C T O R
(MAGNITUDE/PHASE)
FREQUENCY TYPE T1 T2 T3 R1 R2 R3
0 2.000000E+01 G .0 .0 8.817999E-03 6.435859E-04 2.632016E-03 .0
.0 .0 356.4954 176.5664 176.5000 .0
0 4.000000E+01 G .0 .0 9.404316E-03 6.434992E-04 2.795561E-03 .0
.0 .0 356.2596 176.5677 176.2785 .0
.
.
.
0 9.799999E+02 G .0 .0 9.965085E-04 2.691742E-04 4.097779E-04 .0
.0 .0 187.6832 7.8008 15.1581 .0
0 1.000000E+03 G .0 .0 8.803169E-04 2.354655E-04 3.317750E-04 .0
.0 .0 186.9298 8.2146 14.6645 .0
.
.
.
POINT-ID = 33
(MAGNITUDE/PHASE)
0 2.000000E+01 G .0 .0 8.183126E-03 5.993296E-04 2.443290E-03 .0
.0 .0 356.4899 176.5639 176.4950 .0
0 4.000000E+01 G .0 .0 8.768992E-03 6.006201E-04 2.606561E-03 .0
.0 .0 356.2376 176.5565 176.2581 .0
.
.
.
0 9.799999E+02 G .0 .0 6.867234E-04 3.836353E-04 5.393046E-04 .0

.0 .0 188.0180 5.5597 10.0794 .0
0 1.000000E+03 G .0 .0 6.062436E-04 3.454143E-04 4.648783E-04 .0
.0 .0 186.8358 5.4959 8.8514 .0
.
.
.

Slide 31

(Cont.)
1 FREQUENCY RESPONSE DUE TO UNIT FORCE AT TIP MARCH 12, 2007 MD NASTRAN 3/ 1/07 PAGE 24
0 SUBCASE 1
0 SUBCASE CURVE FRAME CURVE ID./ XMIN -FRAME/ XMAX-FRAME/ YMIN-FRAME/ X FOR YMAX-FRAME/ X FOR
0 1 DISP 1 11( 5, --) 2.000000E+01 1.000000E+03 1.002672E-04 3.800000E+02 6.786681E-02 1.400000E+02
2.000000E+01 1.000000E+03 1.002672E-04 3.800000E+02 6.786681E-02 1.400000E+02
0 1 DISP 1 11( --, 11) 2.000000E+01 1.000000E+03 1.834657E+02 2.200000E+02 3.564954E+02 2.000000E+01
2.000000E+01 1.000000E+03 1.834657E+02 2.200000E+02 3.564954E+02 2.000000E+01
0 1 DISP 2 33( 5, --) 2.000000E+01 1.000000E+03 4.982490E-05 6.000000E+02 6.850390E-02 1.400000E+02
2.000000E+01 1.000000E+03 4.982490E-05 6.000000E+02 6.850390E-02 1.400000E+02
0 1 DISP 2 33( --, 11) 2.000000E+01 1.000000E+03 1.815141E+02 3.000000E+02 3.564899E+02 2.000000E+01
2.000000E+01 1.000000E+03 1.815141E+02 3.000000E+02 3.564899E+02 2.000000E+01
0 1 DISP 3 55( 5, --) 2.000000E+01 1.000000E+03 2.148861E-04 1.000000E+03 6.888805E-02 1.400000E+02
2.000000E+01 1.000000E+03 2.148861E-04 1.000000E+03 6.888805E-02 1.400000E+02
0 1 DISP 3 55( --, 11) 2.000000E+01 1.000000E+03 6.490854E+00 7.599999E+02 3.579576E+02 7.800000E+02
2.000000E+01 1.000000E+03 6.490854E+00 7.599999E+02 3.579576E+02 7.800000E+02

Slide 32

(Cont.)
The display of Phase

will be different in Phase Lead
MD Patran!
Patran! calculated by
-180o < Phase < 180o MD Nastran:

0oo < Phase < 360o
Any Phase Lead larger
180oo will be displayed
as ( Phase - 360o )

Slide 33

(Cont.)

Slide 34

(Cont.)

Slide 35
WORKSHOP #6
MODAL FREQUENCY RESPONSE

Slide 36
WORKSHOP #6 - MODAL FREQUENCY

RESPONSE
Using the modal method, determine the frequency response of the flat
rectangular plate, created in Workshop 1, excited by a 0.1 psi pressure load over
the total surface of the plate and a 1.0 lb. force at a corner of the tip lagging 45 o.
Use a modal damping of  = 0.03. Use a frequency step of 20 hz between a range
of 20 and 1000 hz; in addition, specify five evenly spaced excitation frequencies
between the half power points of each resonant frequency between the range of
20-1000 hz. Below is a finite element representation of the flat plate. It also
contains the loads and boundary constraints.
0.1 psi over the total surface
1.0

Slide 37
WORKSHOP # 6
• Start with the following input file
$ $
$ wkshp6.dat
$ BEGIN BULK
$ executive control add : modal frequency response solution PARAM,COUPMASS,1
sequence PARAM,WTMASS,0.00259
$ $
$ case control, add : eigenvalue callout $ PLATE MODEL DESCRIBED IN NORMAL MODES EXAMPLE
$ frequency selection $
$ modal damping selection INCLUDE 'plate.bdf'
$ xyplot commands $
$ ENDDATA
$ bulk data, add : eigenvalue method
$ modal damping
$ loading
$ frequencies of load application
$
CEND
TITLE = FREQUENCY RESPONSE WITH PRESSURE AND POINT LOADS
SUBTITLE = USING THE MODAL METHOD WITH LANCZOS
ECHO = UNSORTED
SEALL = ALL
SPC = 1
SET 111 = 11, 33, 55
DISPLACEMENT(PHASE, PLOT) = 111
SUBCASE 1
DLOAD = 100
$
OUTPUT (XYPLOT)
$

Slide 38
SOLUTION FOR WORKSHOP #6

BEGIN BULK
PARAM,COUPMASS,1
ID SEMINAR, PROB6 PARAM,WTMASS,0.00259
$ $
$ soln6.dat $ PLATE MODEL DESCRIBED IN NORMAL MODES EXAMPLE
$ $
ID SEMINAR, PROB6 INCLUDE 'plate.bdf'
SOL 111 $
CEND $ EIGENVALUE EXTRACTION PARAMETERS
TITLE = FREQUENCY RESPONSE WITH PRESSURE AND POINT LOADS $
SUBTITLE = USING THE MODAL METHOD WITH LANCZOS EIGRL, 100, 10., 2000.
ECHO = UNSORTED $
SEALL = ALL $ SPECIFY MODAL DAMPING
SPC = 1 $
SET 111 = 11, 33, 55 TABDMP1, 100, CRIT,
DISPLACEMENT(PHASE ) = 111 +, 0., .03, 10., .03, ENDT
METHOD = 100 $
FREQUENCY = 100 $ APPLY PRESSURE LOAD
SDAMPING = 100 $
SUBCASE 1 RLOAD2, 400, 300, , ,310
DLOAD = 100 PLOAD2, 300, 1., 1, THRU, 40
$ $
OUTPUT (XYPLOT) TABLED1, 310,
$ , 10., 1., 1000., 1., ENDT
XTGRID= YES $
YTGRID= YES $ POINT LOAD
XBGRID= YES $
YBGRID= YES RLOAD2, 500, 600, , -45., 310
YTLOG= YES $
YBLOG= NO FORCE,600,11,,1.0,,,1.0
XTITLE= FREQUENCY (HZ) $
YTTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, MAGNITUDE $ COMBINE LOADS
YBTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, PHASE $
XYPLOT DISP RESPONSE / 11 (T3RM, T3IP) DLOAD, 100, 1., .1, 400, 1.0, 500
YTTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, MAGNITUDE $
YBTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, PHASE $ SPECIFY FREQUENCY STEPS
YTTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER, MAGNITUDE FREQ1, 100, 20., 20., 49
YBTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER, PHASE FREQ4, 100, 20., 1000., .03, 5
$ ENDDATA

Slide 39

POINT-ID = 11
(MAGNITUDE/PHASE)
0 2.000000E+01 G 1.647094E-12 2.530238E- 12 1.122426E-02 6.413405E-04 3.278660E-03 2.135828E-12
146.8949 146.8692 325.678 3 134.5891 144.8534 146.8698
0 4.000000E+01 G 1.770526E-12 2.719501E- 12 1.199385E-02 6.408464E-04 3.493634E-03 2.295604E-12
146.2678 146.2447 325.172 5 134.4988 144.4191 146.2453
0 6.000000E+01 G 2.022193E-12 3.105379E- 12 1.356089E-02 6.394051E-04 3.931127E-03 2.621361E-12
145.4318 145.4127 324.515 9 134.3886 143.8702 145.4131
0 8.000000E+01 G 2.520848E-12 3.869944E- 12 1.666288E-02 6.353444E-04 4.797074E-03 3.266807E-12
144.1163 144.1025 323.442 7 134.2561 142.9458 144.1028
0 1.000000E+02 G 3.674813E-12 5.639226E- 12 2.383472E-02 6.228435E-04 6.799188E-03 4.760436E-12
141.4430 141.4358 321.081 4 134.1330 140.7818 141.4359
0 1.200000E+02 G 8.109207E-12 1.243793E- 11 5.136045E-02 5.649115E-04 1.448202E-02 1.049992E-11
131.6994 131.7002 311.728 4 135.1967 131.6851 131.7000
POINT-ID = 55
(MAGNITUDE/PHASE)
0 2.000000E+01 G 2.299811E-12 2.628319E- 12 1.003391E-02 5.635861E-04 3.021529E-03 2.074211E-12
326.8539 146.8715 326.973 5 135.2839 145.7045 146.8706
0 4.000000E+01 G 2.471663E-12 2.824950E- 12 1.079900E-02 5.666142E-04 3.235993E-03 2.229377E-12
326.2311 146.2468 326.332 0 135.1997 145.1825 146.2461
0 6.000000E+01 G 2.822033E-12 3.225849E- 12 1.235921E-02 5.723091E-04 3.672785E-03 2.545738E-12
325.4014 145.4144 325.477 3 135.1333 144.5074 145.4138
0 8.000000E+01 G 3.516238E-12 4.020179E- 12 1.545119E-02 5.823869E-04 4.537677E-03 3.172566E-12
324.0942 144.1037 324.138 4 135.0829 143.4117 144.1033

Slide 40
OUTPUT FOR WORKSHOP #6 (Cont.)

0 SUBCASE 1
0 1 DISP 1 11( 5,--) 2.000000E+01 1.000000E+03 3.651020E-04 4.200000E+02 1.699705E-01 1.336891E+02
2.000000E+01 1.000000E+03 3.651020E-04 4.200000E+02 1.699705E-01 1.336891E+02
0 1 DISP 1 11(--, 11) 2.000000E+01 1.000000E+03 1.400466E+02 1.000000E+03 3.256783E+02 2.000000E+01
2.000000E+01 1.000000E+03 1.400466E+02 1.000000E+03 3.256783E+02 2.000000E+01
0 1 DISP 2 33( 5,--) 2.000000E+01 1.000000E+03 2.236997E-04 6.400000E+02 1.700564E-01 1.336891E+02
2.000000E+01 1.000000E+03 2.236997E-04 6.400000E+02 1.700564E-01 1.336891E+02
0 1 DISP 2 33(--, 11) 2.000000E+01 1.000000E+03 1.384945E+02 1.000000E+03 3.263299E+02 2.000000E+01
2.000000E+01 1.000000E+03 1.384945E+02 1.000000E+03 3.263299E+02 2.000000E+01
0 1 DISP 3 55( 5,--) 2.000000E+01 1.000000E+03 1.824597E-04 1.000000E+03 1.696912E-01 1.336891E+02
2.000000E+01 1.000000E+03 1.824597E-04 1.000000E+03 1.696912E-01 1.336891E+02
0 1 DISP 3 55(--, 11) 2.000000E+01 1.000000E+03 1.702332E+01 6.999999E+02 3.577236E+02 7.102496E+02
2.000000E+01 1.000000E+03 1.702332E+01 6.999999E+02 3.577236E+02 7.102496E+02

Slide 41

Slide 42

Slide 43

Slide 44
FREQUENCY-DEPENDENT SPRINGS AND

DAMPERS
• Specifies stiffness as a function of forcing frequency
• Specifies damping as a function of forcing frequency
• Different impedance in different direction
• CBUSH
• Defines generalized spring, damper connectivity
• PBUSH
• Defines nominal spring and damper values
• PBUSHT
• Defines frequency-dependent spring and damper values

Slide 45
FREQUENCY-DEPENDENT SPRINGS AND

DAMPERS
• Specifies stiffness as a function of forcing frequency

• Specifies damping as a function of forcing frequency
• Different impedance in different direction
• CBUSH
• Defines generalized spring, damper connectivity
• PBUSH
• Defines nominal spring and damper values
• PBUSHT
• Defines frequency-dependent spring and damper values

Slide 46

Slide 47

Slide 48
Generalized Spring-and-Damper Connection CBUSH

z elem
GA v
y elem
S*
(1 - S ) *
GB
xelem
Figure 1. CBUSH Element.
zelem
y elem
(S1, S2, S3)OCID
GB
GA
Note : 1. The material stiffness and damping properties of the elastomer are located at (S1, S2, S3).
2. The elastomer itself has zero length; i.e., GA and GB are coincident. It is shown here in an exploded view
Figure 2. Definition of Offset S1, S2, S3
Bulk Data Entry

Slide 49

Slide 50

Slide 51

Slide 52

Slide 53
FREQUENCY DEPENDENT IMPEDANCE

SAMPLE
P() = 2.0 P(f) sin t PARAM.WTMASS
12 mass 12 = 1.0 * (1./2)2
spring(f) damper(f)
CBUSH 1000
11
Table of Impedance
Forcing
Frequency K(f) B(f) P(f)
(Hz)
0.9 0.81 0.286479 0.81
1 1 0.31831 1
1.1 1.21 0.350141 1.21
fi f i22 f i/  fi2
Slide 54
SAMPLE USING CBUSH ELEMENT

$
$ cbush1.dat
$
TIME 10
SOL 108
CEND
TITLE = VERIFICATION PROBLEM, FREQ. DEP. IMPEDANCE BUSHVER
SUBTITLE = SINGLE DOF, CRITICAL DAMPING, 3 EXCITATION FREQUENCIES
ECHO = BOTH
SPC = 1002
DLOAD = 1
DISP = ALL
FREQ = 10
ELFO = ALL
BEGIN BULK
$ CONVENTIONAL INPUT FOR MOUNT
GRDSET, , , , , , , 23456 $ PS
$ TIE DOWN EVERYTHING BUT THE 1 DOF
GRID, 11, , 0., 0., 0.0 $ GROUND
=, 12, =, =, =, , $ ISOLATED DOF
SPC1, 1002 123456 11 $ GROUND
CONM2, 12, 12, , 1.0 $ THE ISOLATED MASS
$
$ EID PID GA GB GO/X1 X2 X3 CID
$
CBUSH 1000 2000 11 12 0
$
PBUSH 2000 K 1.0
B 0.0
$
PBUSHT 2000 K 2001
B 2002
$
TABLED1, 2001 $ STIFFNESS TABLE
, 0.9 0.81, 1.0, 1.0, 1.1, 1.21 ENDT
TABLED1 2002 $ DAMPING TABLE
, 0.9 .2864789, 1.0, .318309, 1.1 , .3501409 ENDT
$CONVENTIONAL INPUT FOR FREQUENCY RESPONSE
PARAM, WTMASS , .0253303 $ 1/(2*PI)**2. GIVES FN=1.0
DAREA, 1, 12, 1, 2. $CAUSES UNIT DEFLECTION
FREQ, 10, 0.9, 1.0, 1.1 $ BRACKET THE NATURAL FREQUENCY
RLOAD1 , 1, 1, , , 3
TABLED1,3 $ TABLE FOR FORCE VS. FREQUENCY
, 0.9, 0.81, 1., 1., 1.1, 1.21,ENDT $ P = K
ENDDATA

Slide 55
DISPLACEMENT OUTPUT FOR CBUSH

ELEMENT
FREQUENCY = 9.000000E-01
(REAL/IMAGINARY)

0 11 G 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
0 12 G -6.682744E-08 0.0 0.0 0.0 0.0 0.0
-1.000000E+00 0.0 0.0 0.0 0.0 0.0
1 VERIFICATION PROBLEM, FREQ. DEP. IMPEDANCE BUSHVER MARCH 12, 2007 MD NASTRAN 3/ 1/07 PAGE 10
SINGLE DOF, CRITICAL DAMPING, 3 EXCITATION FREQUENCIES
0
FREQUENCY = 1.000000E+00
(REAL/IMAGINARY)

0 11 G 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
0 12 G -1.046835E-07 0.0 0.0 0.0 0.0 0.0
-9.999999E-01 0.0 0.0 0.0 0.0 0.0
0
(REAL/IMAGINARY)

0 11 G 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
0 12 G -6.855670E-08 0.0 0.0 0.0 0.0 0.0
-9.999998E-01 0.0 0.0 0.0 0.0 0.0

Slide 56
FORCE OUTPUT FOR CBUSH ELEMENT

FREQUENCY = 9.000000E-01
C O M P L E X F O R C E S I N B U S H E L E M E N T S ( C B U S H )
(REAL/IMAGINARY)
ELEMENT-ID FORCE-X FORCE-Y FORCE-Z MOMENT-X MOMENT-Y MOMENT-Z

0 1000 1.620000E+00 0.0 0.0 0.0 0.0 0.0
-8.100001E-01 0.0 0.0 0.0 0.0 0.0
0
(REAL/IMAGINARY)

0 1000 2.000000E+00 0.0 0.0 0.0 0.0 0.0
-1.000000E+00 0.0 0.0 0.0 0.0 0.0
0
(REAL/IMAGINARY)

0 1000 2.419999E+00 0.0 0.0 0.0 0.0 0.0
-1.210000E+00 0.0 0.0 0.0 0.0 0.0
0

Slide 57
BLANK PAGE

SECTION 9
DIRECT MATRIX INPUT

DIRECT MATRIX INPUT
● In addition to generating the mass, stiffness, and damping matrix automatically
from the FE model, you can add mass, stiffness, and damping to specific
DOFs.
● The DMIG Bulk Data entry is used to input a mass, damping, or stiffness
matrix connecting specified DOFs.
● Two types of DMIGs
● Type 1 G-Type Matrices

● G-type matrices are the size of the G-set.
● G-type matrices are applied at the system level prior to any constraints operations.
● G-type matrices are real, symmetric. They are selected in the Case Control Section via:
● M2GG = name of mass matrix addition
● B2GG = name of damping matrix addition
● K2GG = name of stiffness matrix addition
● G-type matrices can be added to the superelement and the residual structure.
DIRECT MATRIX INPUT (Cont.)
● Type 2 P-Type Matrices
● P-type matrices are the same size as the P-set (G-set plus E-set
for extra points).
● P-type matrices are not applied at the system level prior to
constraint operations.
● The P-type direct input matrices are processed through the
constraints and reduction procedures in parallel with the G-type
matrices, then added to the reduced model (A- or H-set) before
the analysis operation.
● Note that modal reduction (modal solutions) operations do not
include the effects for P-type matrices.
● The load reduction operations to the analysis sets does not
include the effects of P-type matrices.
DIRECT MATRIX INPUT (Cont.)
● P-type matrices need not be real or symmetric. They are
selected in the Case Control Section via:
● M2PP = name of mass matrix addition
● B2PP = name of damping matrix addition
● K2PP = name of stiffness matrix addition
● P-type matrices can only be added to the residual

structure. They cannot be added to superelements.
● PARAM,WTMASS does not affect direct input matrices M2GG or
M2PP. PARAM,CM2 can be used to scale M2GG.

DMIG Direct Matrix Input at Points
Defines direct input matrices related to grid, extra, and/or scalar points. The matrix is defined by a single header
entry and one or more column entries. A column entry is required for each column with nonzero elements.
Header Entry Format:

1 2 3 4 5 6 7 8 9 10
DMIG NAME “0" IFO TIN TOUT POLAR NCOL
Column Entry Format:

DMIG NAME GJ CJ G1 C1 A1 B1
G2 C2 A2 B2 -etc.-
Example:
DMIG STIF 0 1 3 4
DMIG STIF 27 1 2 3 3.+5 3.+3
2 4 2.5+10 0 50 1 0
Field Contents
NAME Name of the matrix. See Remark 1. (One to eight alphanumeric characters, the first of which is
alphabetic.)
IFO Form of matrix input. IFO = 6 must be specified for matrices selected by the K2GG, M2GG, and
B2GG Case Control commands. (Integer)
1 = Square
9 or 2 = Rectangular
6 = Symmetric
TIN Type of matrix being input: (Integer)

1 = Real, single precision (One field is used per element.)
2 = Real, double precision (One field is used per element.)
3 = Complex, single precision (Two fields are used per element.)
4 = Complex, double precision (Two fields are used per element.)
TOUT Type of matrix that will be created: (Integer)

0 = Set by precision system cell (Default)
1 = Real, single precision
2 = Real, double precision
3 = Complex, single precision
4 = Complex, double precision
Field Contents
POLAR Input format of Ai, Bi. (Integer=blank or 0 indicates real, imaginary format; Integer > 0 indicates
amplitude, phase format.)
NCOL Number of columns in a rectangular matrix. Used only for IFO = 9. (Integer > 0)
GJ Grid, scalar or extra point identification number for column index. (Integer > 0)
CJ Component number for grid point GJ. (0 < Integer £ 6; blank or zero if GJ is a scalar or extra point.)
Gi Grid, scalar, or extra point identification number for row index. (Integer > 0)
Ci Component number for Gi for a grid point. (0 < CJ £ 6; blank or zero if Gi is a scalar or extra point.)
Ai, Bi Real and imaginary (or amplitude and phase) parts of a matrix element. If the matrix is real (TIN =
1 or 2), then Bi must be blank. (Real)
Remarks:
1. Matrices defined on this entry may be used in dynamics by selection in the Case Control with
K2PP = NAME, B2PP = NAME, M2PP = NAME for [K pp], [B pp ], or [M pp ], respectively.
Matrices may also be selected for all solution sequences by K2GG = NAME, B2GG = NAME, and
M2GG = NAME. The g-set matrices are added to the structural matrices before constraints are
applied, while p-set matrices are added in dynamics after constraints are applied. Load matrices
may be selected by P2G = NAME for dynamic and superelement analyses.
2. The header entry containing IFO, TIN and TOUT is required. Each nonnull column is started with
a GJ, CJ pair. The entries for each row of that column follows. Only nonzero terms need be
entered. The terms may be input in arbitrary order. A GJ, CJ pai
3. Field 3 of the header entry must contain an integer 0.

4. For symmetric matrices (IFO = 6), a given off-diagonal element may be input either below or
above the diagonal. While upper and lower triangle terms may be mixed, a fatal message will be
issued if an element is input both below and above the diagonal.
5. The recommended format for rectangular matrices requires the use of NCOL and IFO = 9. The
number of columns in the matrix is NCOL. (The number of rows in all DMIG matrices is always
either p-set or g-set size, depending on the context.) The GJ term i
6. If NCOL is not used for rectangular matrices, two different conventions are available:
· If IFO = 9, GJ and CJ will determine the sorted sequence, but will otherwise be
ignored; a rectangular matrix will be generated with the columns submitted being in the
1 to N positions, where N is the number of logical entries submitted (not counting th
· If IFO = 2, the number of columns of the rectangular matrix will be equal to the index
of the highest numbered non-null column (in internal sort). Trailing null columns of the
g- or p-size matrix will be truncated.
7. The matrix names must be unique among all DMIGs.
8. TIN should be set consistent with the number of decimal digits required to read the input data
adequately. For a single-precision specification on a short-word machine, the input will be
truncated after about eight decimal digits, even when more digits are needed, a double precision
specification should be used instead. However, note that a double precision specification requires
a “D” type exponent even for terms that do not need an exponent. For example, unity may be
input as 1.0 in single precision, but the longer form 1.0D0 is required for double precision.
9. On long-word machines, almost all matrix calculations are performed in single precision and on
short-word machines, in double precision. It is recommended that DMIG matrices also follow
these conventions for a balance of efficiency and reliability. The recommended value for TOUT is
0, which instructs the program to inspect the system cell that measures the machine precision at
run time and sets the precision of the matrix to the same value. TOUT = 0 allows the same DMIG
input to be used on any machine. If TOUT is contrary to the machine type specified (for example,
a TOUT of 1 on a short-word machine), unreliable results may occur.
10. If any DMIG entry is changed or added on restart then a complete re-analysis is performed.
Therefore, DMIG entry changes or additions are not recommended on restart.

SCALE FACTOR FOR X2GG AND X2PP
MATRICES
● The K2GG, M2GG, B2GG, and P2G commands accept
the specifications of multiple matrices and real scale
factor(s) Example:
K2GG = 1.25*KDMIG
M2gg = 2.8*MDMIG1,3.9*MDMIG2
● K2PP, M2PP, and B2PP commands accept the

specifications of multiple matrices and complex or real
scale factor(s) Example:
K2PP = KDMIG
K2PP = KDMIG1, KDMIG2,KDMIG3
K2PP = 2.04*KDMIG1,0.82*KDMIG2
K2PP = (2.04,.5)*KDMIG1,(0.82,0.)*KDMIG2
DIRECT INPUT OF STRUCTURAL ELEMENT
DAMPING MATRICES
● Can also add structural element damping by using the
K42GG command and DMIG entries.
● Can also specify multiple matrices
Example:
K42GG = KDMIG
K42GG = KDMIG1, KDMIG2,KDMIG3
K42GG = 2.04*KDMIG1,0.82*KDMIG2

DMIG EXAMPLE
1 2 3 4 5 6
CBAR,1 CBAR,2 CBAR,3 CBAR,4 DMIG = EL5
The stiffness and mass matrices from a contractor which connect the
grid points 5 and 6. They contain two DOFs (R2 and T3) per grid point.
5,3 5,5 6,3 6,5

5,3 500039. SYMMETRIC
5,5 -250019 166680.
K55 =
6,3 -500039. 250019. 500039.
6,5 -250019. 83340. 250019. 166680.
5,3 5,5 6,3 6,5

5,3 3.5829 0. 0. 0.
5,5 0. 0. 0. 0.
M5 =
6,3 0. 0. 3.5829 0.
6,5 0. 0. 0. 0.

INPUT FILE FOR THE DMIG EXAMPLE
$
DMIG,EXSTIF,5,3,,5,3,500039.
$ DMIG EXAMPLE
+,5,5,-250019.,,6,3,-500039.
$ NORMAL MODES
+,6,5,-250019.
$
$
SOL 103 $ NORMAL MODES ANALYSIS
DMIG,EXSTIF,5,5,,5,5,166680.
TIME 10
+,6,3,250019.,,6,5,83340.
CEND
$
TITLE = DMIG TO READ STIFFNESS AND MASS FOR ELEM 5
DMIG,EXSTIF,6,3,,6,3,500039.
SUBTITLE = PLANAR PROBLEM
+,6,5,250019.
$
$
SPC = 10
DMIG,EXSTIF,6,5,,6,5,166680.
$
$
$ SPECIFY K2GG AND M2GG
$ HEADER ENTRY FOR MASS
K2GG = EXSTIF
DMIG,EXMASS,0,6,1
M2GG = EXMASS
$
$
$ DATA ENTRIES FOR MASS
METHOD = 10
$
$
DMIG,EXMASS,5,3,,5,3,3.5829
BEGIN BULK
DMIG,EXMASS,6,3,,6,3,3.5829
$
$
$
GRID,1,,0.,0.,0.,,1246
$EIGRL SID V1 V2 ND
GRID,2,,1.,0.,0.,,1246
EIGRL 10 2
GRID,3,,2.,0.,0.,,1246
$
GRID,4,,3.,0.,0.,,1246
CBAR,1,1,1,2,10
GRID,5,,4.,0.,0.,,1246
CBAR,2,1,2,3,10
GRID,6,,5.,0.,0.,,1246
CBAR,3,1,3,4,10
GRID,10,,0.,0.,10.,,123456
CBAR,4,1,4,5,10
MAT1,1,7.1+10,,.33 2700.
$
PBAR,1,1,2.654-3,5.869-7
$ HEADER ENTRY FOR STIFFNESS
SPC1,10,123456,1
DMIG,EXSTIF,0,6,1
$
$
ENDDATA

BLANK

SECTION 10
DYNAMIC EQUATIONS OF MOTION

DYNAMIC MATRIX ASSEMBLY
● MD Nastran provides direct and modal methods for
performing transient and frequency response and complex
mode analysis.
● The dynamic matrices are assembled differently

depending on the analysis and method.

DIRECT METHODS
● The general dynamic equation used in the direct methods is:
Mdd p 2 + Bd dp + K d du
  d = P d 
where p = a derivative operator

ud = the union of the analysis set ua and extra points ue
● For frequency response and complex eigenvalue analysis, the dynamic
matrices are: 1 + 

Kd d= 1 + ig 
K dd K2dd + i 
4 
K dd
 B1
B dd=  dd + 
2 
Bd d
Mdd =  1
Mdd +  2
Mdd 
● For transient response, the dynamic matrices are:

K K 1 + 
=  K2 
dd dd dd
1 g 1
B =  B +  B 2 + ------ 
K 1 + ------ 
K4 
dd dd dd 3 dd 4 d d

M =  M 1 +  2
M 
dd dd dd

DYNAMIC MATRIX DEFINITIONS
[K1dd] is the reduced structural stiffness matrix plus the reduced direct input K2GG
(symmetric).
[K2dd] is the reduced direct input matrix K2PP plus the reduced transfer function
input (symmetric or unsymmetric).
[K4dd] is the reduced structural damping matrix obtained by multiplying the

stiffness matrix [Ke] of an individual structural element by an element damping
factor ge and combining the results for all structural elements (symmetric).
[B1dd] is the reduced viscous damping matrix plus the reduced direct input B2GG
(symmetric).

DYNAMIC MATRIX DEFINITIONS (Cont.)
● [B2dd] is the reduced direct input matrix B2PP plus the reduced
transfer function input (symmetric or unsymmetric).
● [M1dd ] is the reduced mass matrix plus the reduced direct input M2GG
(symmetric).
● [M2dd] is the reduced direct input matrix M2PP plus the reduced
transfer function input (symmetric or unsymmetric).
● g, 3, 4 are the constants specified by the user.

DYNAMIC MATRIX DEFINITIONS (Cont.)
● The structural matrices K aa ,Ka4a , M a a
, and[Baa] are expanded by the
addition of zeroes in the rows and columns corresponding to extra
points to form Kdd
1  4 
, Kdd 1 
, Mdd , and B dd 1 
.
● Only the Kdd
2  2 
, B dd , and Mdd 2 
matrices can reference extra points.
● The direct input matrices K pp 2 
, B 2p p, and M 2pp are processed through
the multipoint and single-point constraint elimination and any reduction
procedure.
● Note: The extra points are unaffected by any constraint or reduction
procedures. The constraint and reduction procedures can only
eliminate grid or scalar point DOFs but not extra points.
● The matrices K dd , B d d, and Mdd are examined to identify rows and
columns which are null in all three matrices. For transient and
frequency response, K dd is augmented by placing unity in each null
row and column. In complex eigenvalue analysis, null rows and
columns are discarded from K dd , B d d, and Mdd .
MODAL METHODS
● The general dynamic equation used in the modal methods is:
M h hp 2 + Bhh p + K hh u
  h = ph 
where p = a derivative operator
uh = the union of the modal coordinates i and extra points ue.
● The transformation between i and ua is:

u a= ai 
 i
where [ai] is the matrix of eigenvectors obtained in real eigenvalue analysis.
● The transformation from uh to ud is obtained by augmenting [ai] to include the extra

points.
u = 
 u  
d dh h
 0

d h= ai
where 0 I ee
i
uh =
ue
MODAL METHODS (Cont.)
● For frequency response and complex eigenvalue analysis the dynamic
matrices are: 
Khh = 
ki +  T
dh  ig 1 + 
K2 i K4 
Kdd d d+  dd 
dh

B h h= bi +  T
dh   1 + 
B2
B dd d d
dh

M h h = 
m i +  T
dh  2 
Md d dh 
where [mi] = a diagonal matrix with terms m ii = aiT M aaai 

[bi] = a diagonal matrix with terms bii = ig(i)m ii ,
where i = is the radian frequency of the i-th normal mode and
g(i) is a damping factor obtained from interpolation of
a user-supplied table (TABDMP1)
[ki] = a diagonal matrix with terms kii = 2imii
● If parameter KDAMP = -1, then ( Default: PARAM,KDAMP,1 )
mii = m ii
bii = 0
k =  1 + ig 
 k
ii i ii

MODAL METHODS (Cont.)
● g(wi) is a damping factor obtained from the interpolation of a user-
supplied table (TABDMP1).
● [mi], [bi], and [ki] are expanded by the addition of zeros to the rows and
columns corresponding to the extra points (ue).
● For transient response the dynamic matrices are:


K hh = 
ki+  T
dh  2 
K dd 
dh

B hh = 
b i + dhTB
g
1 + B 2 + ------- 
K
1
1 + ------
- 1 
Kd 
d h
 dd dd  dd  d 
3 4

Mhh =  m i + dhT 2 
Mdd dh
● If only [mi], [bi], and [ki] are present in any modal dynamic analysis,
then the modal dynamic equations are uncoupled.
BLANK

SECTION 12
ENFORCED MOTION

ENFORCED MOTION IN DYNAMIC ANALYSIS
● Used to analyze constrained structures with base input

acceleration, displacement, and velocity.
● Common examples are earthquakes (for transient

analysis), swept-sine shaker test simulation (for frequency
response analysis) and automobile suspensions.

ANALYSIS METHODS
● Four available methods:
● 1. Direct Specification of enforced displacement, velocity, or

acceleration (recommended method).
● 2. Large Mass (See Appendix B)
● 3. Large Stiffness (Enforced displacement only, see appendix B)
● 4. Lagrange Multiplier (See Appendix B)

METHOD 1
● Method 1 is the recommended method and it is the
method discussed in this seminar.
● This method of enforced motion can be specified by the

direct specification of displacements, velocities or
accelerations via SPC / SPC1 and SPCD Bulk Data
entries.
● Interface very similar to enforced motion in static analysis.

DEGREE OF FREEDOM SETS
● In the direct method, enforced motion is applied to the
degrees of freedom in the S-set.
G-Set: 6 Degrees of Freedom per Grid

N-Set M-Set:
F-Set S-Set: Eliminated by
multipoint
Prescribed by single-point constraints
constraints (SPCs, (MPCs, R-type
AUTOSPC, PS) elements)

METHODS FOR DIRECT ENFORCED MOTION
● Two approaches are available for performing this type of

enforced motion
● Absolute motion (v2001)
● Relative motion (v2004+)

BASIC EQUATIONS FOR USING ABSOLUTE
MOTION
● For the N-set, the equation of motion is as follows:
.. .

M ff M fs 
uf  B ff B fs 
uf  K ff K fs 
u f   p f 1
 .. 
 . 
  
 

M sf u s  
M ss  B sf u s  
B ss  K sf u s  
K ss  
ps qs 
where
p f , ps Applied Loads (known)
us Prescribed Displacements (known)
qs Constraint Forces (unknown)

EQUATIONS FOR TRANSIENT RESPONSE
● Equation 1 can be solved for the F-set displacements:
M ff u f B ff u f K ff u f p f M fsu s B fsu s K fsu s  2


..  .  .. .
  
● Subsequently, the constraint forces are obtained from the

second matrix equation:
.. .
u f  uf  uf 3
q s p s 
M sf M ss .. 
 B sf B ss . 
 K sf K ss
 
u s  u s  u s 

EQUATIONS FOR FREQUENCY RESPONSE
● In frequency response analysis, the F-set displacements

are obtained from
 M 2
ff  
iB ff K ff U f P f 2 M fs iB fs K fs U s  4
● The constraint forces read
U f 
 
Q s Ps  M sf 2
M ss 
i
B sf B ss 
K sf K ss 
U  5
 s

RELATIVE MOTION APPROACH
● { Uf } = { Yf } + { X } 6
KffUf + KfsUs = Pf
KffYf + KffX + KfsUs = Pf
where { Uf } = absolute motion Rel.
Rel. Motion: KffYf = Pf
{ Yf } = relative motion  K X + K U = 0
ff fs s
{ X } = base motion X = -K -1 K U
ff fs s
● { X } - motion of f-set DOFs due to specified enforced

motion of the s-set DOFs (computed purely from static
considerations)
● { X } = -[ Kff ]–1 [ Kfs ] { Us } = [ Z ] { Us } 7

RELATIVE MOTION APPROACH
● Substituting equations 6 and 7 into equation 2, we get:
.
   
.. .. .
    
M ff Yf B ff Yf K ff Yf p f M ff Z M fs s B ff Z B fs s
u u
.. .
● Solve for Yf, Y f, and Y f
.. .
● Uf, Uf, and U f are obtained using equation 6
● This is the default output
.. .
● To get Y f, Y f, and Yf, set PARAM,ENFMOTN,REL
PARAM,ENFMOTN,ABS - absolute motion of the model (default)
PARAM,ENFMOTN,REL - relative motion to the enforced motion
of the base
USER INTERFACE
● SPC / SPC1 Bulk Data entries are used to identify the degrees of
freedom with enforced motion. These entries are activated by an SPC
Case Control command.
● SPCD Bulk Data entries are used to define the enforced motion.
These entries are referenced by the EXCITEID field of TLOADi or
RLOADi Bulk Data entries.
● The TYPE field of the TLOADi or RLOADi Bulk Data entries indicates
the type of enforced motion.
● For modal method, residual vectors should always be included which

is the default

THE TYPE FIELD
● The type of excitation is defined in field 5 of TLOADi Bulk

Data entries or in field 8 of RLOADi Bulk Data entries:
● The character fields may be shortened to as little as a

single character
EXAMPLE: EXECUTIVE AND CASE CONTROL
SOL 111
111
CEND
$
TITLE =Example for Direct Enforced Motion
SUBTITLE=Modal Frequency Response Analysis
$
SPC =1
METHOD =10
FREQUENCY=20
$
SET 1 = 1000,1001
ACCELERATION(SORT2,PRINT,PHAS)=1
$
SUBCASE 1
LABEL=Unit Acceleration in x-
x-Direction
DLOAD=100
$
SUBCASE 2
LABEL=Unit Acceleration in y-
y-Direction
DLOAD=200
$

EXAMPLE: BULK DATA DECK
BEGIN BULK
$
PARAM, G, 0.02 $ 2% Structural Damping
SPC1, 1, 3456, 1000 $ z -Displ. and Rotations are fixed
SPC1, 1, 12, 1000 $ x- and y- Accelerations are prescribed
$
$ Modal Reduction
EIGRL, 10,, 150. $ Modes up to 150Hz
$
$ Base Motion Excitation
$
RLOAD1, 100
100,
, 1001
1001,,,
,,, 10
10,,
,, A $ Load of Subcase 1:
SPCD,
SPCD, 1001
1001,
, 1000
1000,
, 1, 1. $ Unit x-
x-Acceleration
$
RLOAD1, 200
200,
, 1002
1002,,,
,,, 10
10,,
,, A $ Load of Subcase 2:
SPCD, 1002, 1000, 2, 1. $ Unit y -Acceleration
$
TABLED1, 10 $ Constant for all Frequencies
, 0., 1., 100., 1., ENDT
FREQ1, 20, 1., 1., 49 $ Frequency Range from 1Hz to 50Hz
$
INCLUDE 'tower.bdf
'tower.bdf'
' $ Structural Model
$
ENDDATA

INITIAL CONDITION SPECIFICATION FOR
ENFORCED MOTION (SPC/SPCD)
● Enforced acceleration or velocity in transient analysis

(using SPC/SPCD) requires integration to compute
corresponding enforced velocities and/or displacements
● Above integration involves the use of initial conditions
● The US0 (initial displacement) and VS0 (initial velocity)

fields on the TLOADi entries accomplish this.

INITIAL CONDITION SPECIFICATION FOR
ENFORCED MOTION (SPC/SPCD)
● Format of TLOAD1 is shown below
TLOAD1 SID EXCITEID DELAY TYPE TID US0 VS0
● Format of TLOAD2 is shown below

TLOAD2 SID EXCITEID DELAY TYPE T1 T2 F P
C B US0 UV0
● US0—initial displacement
● VS0—initial velocity
● US0 and VS0 can be used in conjunction with other initial
condition (TIC) at independent DOFs
EXAMPLE SPECIFYING INITIAL DISPLACEMENT
● A simple 2 dof system is used to illustrate the use of initial

displacement.
● The following enforced displacement is applied to grid 1, x
direction
● d(t) = cos (2* * 5 * t)
● Note that d(0) = 1.0
● The following enforced velocity is applied to grid 1, y
direction
● v(t) = -10 * * sin (2 * *5 * t)
● Note that v(0) = 0
● The following enforced acceleration is applied to grid 1, z
direction
● a(t) = -100 * 2 * cos ( 2 * * 5 * t )

EXAMPLE SPECIFYING INITIAL
DISPLACEMENT (Cont.)
● Applied enforced displacement at grid 1 in x direction
d(t) = cos (2* * 5 * t)

● Applied enforced velocity at grid 1 in y direction
v(t) = -10 * * sin (2 * *5 * t)

● Applied enforced acceleration at grid 1 in z direction
a(t) = -100 * 2 * cos ( 2 * * 5 * t )

● Without US0, the enforced displacement is off when using
enforced velocity and acceleration
● This affects the subsequent displacement, velocity, and
acceleration output
● With the appropriate US0 added to the enforced velocity

and enforced acceleration input, the correct enforced
displacement is applied
● The product of US0 and the value specified on the SPCD should
equal to the enforced value

begin bulk
param,post,0
tstep,40,500,.001,1
grid,1,,,,,,123
grid,2
dload,5000,1.0,1.0,30,1.0,300,1.0,3000
$
$ Enforced displacement at Grid 1 - Component 1
$ Corresponding response is at Grid 2 - Component 1
$
celas2,100,986.9604,1,1,2,1
cmass2,200,1.,2,1
spcd,20,1,1,1.
tload2,30,20,,d,0.,1000.,5.
$
$ Enforced velocity at Grid 1 - Component 2
$
celas2,1000,986.9604,1,2,2,2
cmass2,2000,1.,2,2
spcd,200,1,2,-31.4159
tload2,300,200,,v,0.,1000.,5.,-90.
,,,-3.183-2
$
$ Enforced acceleration at Grid 1 - Component 3
$
celas2,10000,986.9604,1,3,2,3
cmass2,20000,1.,2,3
spcd,2000,1,3,-986.960
tload2,3000,2000,,a,0.,1000.,5.
,,,-1.013-3
enddata

● Displacement output at grid 1

● Velocity output at grid 1

● Acceleration output at grid 1

● Displacement output at grid 2

● Velocity output at grid 2

● Acceleration output at grid 2

WORKSHOP # 7A
DIRECT TRANSIENT RESPONSE WITH
ENFORCED ACCELERATION

WORKSHOP # 7A - DIRECT TRANSIENT
RESPONSE WITH ENFORCED ACCELERATION
Using the direct method, determine the transient response due to a
unit acceleration sine pulse of 250 Hz applied at the base in the z-
direction. Use a structural damping coefficient of g = 0.06 and convert
this damping to equivalent viscous damping at 250 Hz.
Below is a finite element representation of the flat plate. It also
Large Mass
Drive Point
1000 lbs
1245
1245
1245
1245
1245
Z
Y
X

WORKSHOP
● Start with the following input file
7A
$ $
$ wkshp7a.dat OUTPUT (XYPLOT)
$ XGRID=YES
$ add : boundary conditions at left end YGRID=YES
$ rbe2 for left edge in the vertical direction XTITLE= TIME (SEC)
$ loading and enforced motion YTITLE= BASE ACCELERATION
$ XYPLOT ACCELERATION RESPONSE / 23 (T3)
SOL 109 YTITLE= BASE DISPLACEMENT
CEND XYPLOT DISP RESPONSE / 23 (T3)
TITLE = TRANSIENT RESPONSE WITH ENFORCED MOTION YTITLE= TIP CENTER DISPLACEMENT RESPONSE
SUBTITLE = USING DIRECT TRANSIENT METHOD XYPLOT DISP RESPONSE / 33 (T3)
ECHO = UNSORTED $
SPC = 200 BEGIN BULK
SET 111 = 23, 33 $
set 1000 = 23 $ PLATE MODEL DESCRIBED IN NORMAL MODES EXAMPLE
spcf = 1000 $
DISPLACEMENT (SORT2) = 111 param,post,0
VELOCITY (SORT2) = 111 INCLUDE 'plate.bdf'
ACCELERATION (SORT2) = 111 PARAM, COUPMASS, 1
SUBCASE 1 PARAM, WTMASS, 0.00259
DLOAD = 500 $
TSTEP = 100 $ SPECIFY STRUCTURAL DAMPING
$ $
PARAM, G, 0.06
PARAM, W3, 1571.
$
TSTEP, 100, 200, 2.0E-4, 1
$
ENDDATA

SOLUTION FOR WORKSHOP # 7A
ID SEMINAR, PROB7A YTITLE= TIP CENTER DISPLACEMENT RESPONSE
SOL 109 XYPLOT DISP RESPONSE / 33 (T3)
CEND $
TITLE = TRANSIENT RESPONSE WITH BASE EXCITATION BEGIN BULK
SUBTITLE = USING DIRECT TRANSIENT METHOD, NO
$
REDUCTION
ECHO = UNSORTED $ PLATE MODEL DESCRIBED IN NORMAL MODES
SPC = 200 EXAMPLE
$
SET 111 = 23, 33
DISPLACEMENT (SORT2) = 111 INCLUDE ’plate.bdf’
PARAM, COUPMASS, 1
VELOCITY (SORT2) = 111
ACCELERATION (SORT2) = 111 PARAM, WTMASS, 0.00259
SUBCASE 1 $
$ SPECIFY STRUCTURAL DAMPING
DLOAD = 500
TSTEP = 100 $
$ PARAM, G, 0.06
PARAM, W3, 1571.
OUTPUT (XYPLOT)
XGRID=YES $
$ APPLY EDGE CONSTRAINTS
YGRID=YES
XTITLE= TIME (SEC) $
YTITLE= BASE ACCELERATION SPC1, 200, 12456, 1, 12, 23, 34, 45
XYPLOT ACCELERATION RESPONSE / 23 (T3)
YTITLE= BASE DISPLACEMENT

SOLUTION FOR WORKSHOP # 7A (Cont.)
$
$ APPLY ACCELERATION TO THE BASE
$
SPC1, 200, 3, 23
SPCD, 600, 23, 3, 1.0
TLOAD2, 500, 600, , ACCE, 0.0, 0.004, 250., -90.
$
$ RBE MASS TO REMAINING BASE POINTS
$
RBE2, 101, 23, 3, 1, 12, 34, 45
$
$ SPECIFY INTEGRATION TIME STEPS
$
TSTEP, 100, 200, 2.0E-4, 1
$
ENDDATA

SOLUTION FOR WORKSHOP # 7A (Cont.)
0 SUBCASE 1
0 1 ACCE 1 23( 5) 0.000000E+00 4.000000E-02 -1.000000E+00 3.000000E-03 1.000000E+00 9.999999E-04
0.000000E+00 4.000000E-02 -1.000000E+00 3.000000E-03 1.000000E+00 9.999999E-04
0 1 DISP 2 23( 5) 0.000000E+00 4.000000E-02 6.123031E-25 0.000000E+00 2.525501E-06 4.000000E-03
0.000000E+00 4.000000E-02 6.123031E-25 0.000000E+00 2.525501E-06 4.000000E-03
0 1 DISP 3 33( 5) 0.000000E+00 4.000000E-02 -4.217794E-07 9.400000E-03 5.628117E-06 5.800000E-03
0.000000E+00 4.000000E-02 -4.217794E-07 9.400000E-03 5.628117E-06 5.800000E-03

PARTIAL OUTPUT FILE FOR WORKSHOP # 7A
(Cont.)
Base
Acceleration:
Grid Point 23

PARTIAL OUTPUT FILE FOR WORKSHOP #
7A (Cont.)
Base
Displacement:
Grid Point 23

7A (Cont.)
Tip Center
Displacement:
Grid Point 33

WORKSHOP # 7B
MODAL TRANSIENT RESPONSE WITH

ENFORCED ACCELERATION

WORKSHOP # 7B - MODAL TRANSIENT
RESPONSE WITH ENFORCED ACCELERATION
Using the modal method, determine the transient response due to a
unit acceleration sine pulse of 250 Hz applied at the base in the z-
direction. Use a structural damping coefficient of g = 0.06 and convert
this damping to equivalent viscous damping at 250 Hz. Be sure to
include residual vector.
Large Mass
Drive Point
1000 lbs
1245
1245
1245
1245
1245
Z
Y
NAS102,Section 12, March 2007 X
WORKSHOP 7B
● Use solution from workshop 7A as a starting point and

make appropriate changes.

SOLUTION FOR WORKSHOP # 7B (Cont.)
ID SEMINAR, PROB7B
SOL 112
CEND
TITLE = TRANSIENT RESPONSE WITH BASE EXCITATION
SUBTITLE = USING MODAL TRANSIENT METHOD, NO REDUCTION
BEGIN BULK
ECHO = UNSORTED
$
SPC = 200 $ PLATE MODEL DESCRIBED IN NORMAL MODES $EXAMPLE
METHOD = 1000 $
SET 111 = 23, 33 INCLUDE ’plate.bdf’
DISPLACEMENT (SORT2) = 111 PARAM, COUPMASS, 1
VELOCITY (SORT2) = 111 PARAM, WTMASS, 0.00259
ACCELERATION (SORT2) = 111 $
SUBCASE 1 $ SPECIFY STRUCTURAL DAMPING
DLOAD = 500 $
TSTEP = 100 PARAM, G, 0.06
$ PARAM, W3, 1571.
OUTPUT (XYPLOT) $
XGRID=YES $ APPLY EDGE CONSTRAINTS
YGRID=YES $
XTITLE= TIME (SEC) SPC1, 200, 12456, 1, 12, 23, 34, 45
YTITLE= BASE ACCELERATION
XYPLOT ACCELERATION RESPONSE / 23 (T3)
YTITLE= BASE DISPLACEMENT
YTITLE= TIP CENTER DISPLACEMENT RESPONSE
$

EIGRL, 1000, , , 10
$
$ APPLY ACCELERATION TO THE BASE
$
SPC1, 200, 3, 23
SPCD, 600, 23, 3, 1.0
TLOAD2, 500, 600, , ACCE, 0.0, 0.004, 250., -90.
$
$
RBE2, 101, 23, 3, 1, 12, 34, 45
$
$
TSTEP, 100, 200, 2.0E-4, 1
$
ENDDATA

SOLUTION FOR WORKSHOP # 7B
0 SUBCASE 1
0 1 ACCE 1 23( 5) 0.000000E+00 4.000000E-02 -1.000000E+00 3.000000E-03 1.000000E+00 9.999999E-04
0.000000E+00 4.000000E-02 -1.000000E+00 3.000000E-03 1.000000E+00 9.999999E-04
0 1 DISP 2 23( 5) 0.000000E+00 4.000000E-02 6.123031E-25 0.000000E+00 2.525501E-06 4.000000E-03
0.000000E+00 4.000000E-02 6.123031E-25 0.000000E+00 2.525501E-06 4.000000E-03
0 1 DISP 3 33( 5) 0.000000E+00 4.000000E-02 -4.332059E-07 9.400000E-03 5.631920E-06 5.800000E-03
0.000000E+00 4.000000E-02 -4.332059E-07 9.400000E-03 5.631920E-06 5.800000E-03

7B (Cont.)
Base
Acceleration:
Grid Point 23

7B (Cont.)
Base
Displacement:
Grid Point 23

7B (Cont.)
Tip Center
Displacement:
Grid Point 33

WORKSHOP # 8A
DIRECT FREQUENCY RESPONSE WITH

ENFORCED DISPLACEMENT

WORKSHOP # 8A – DIRECT FREQUENCY
RESPONSE WITH ENFORCED DISPLACEMENT
Using the direct method, determine the frequency response of the flat
rectangular plate, created in Workshop 1, subject to a 0.1 enforced
displacement at the corner of the tip. Use a frequency step of 20 Hz in
the range of 20 to 1000 Hz. Use a structural damping of g = 0.06.

WORKSHOP 8A
● Use the following input file as a starting point
$
OUTPUT (XYPLOT)
$
XTGRID= YES
YTGRID= YES
XBGRID= YES
YBGRID= YES
YTLOG= YES
YBLOG= NO
$
$ wkshp8a.dat XTITLE= FREQUENCY (HZ)
$ YTTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER,
$ add : enforced motion input MAGNITUDE
$ frequencies of loading YBTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, PHASE
$ XYPLOT DISP RESPONSE / 11 (T3RM, T3IP)
SOL 108
CEND YTTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, MAGNITUDE
TITLE= FREQUENCY RESPONSE DUE TO .1 DISPLACEMENT YBTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, PHASE
AT TIP XYPLOT DISP RESPONSE / 33 (T3RM, T3IP)
SUBTITLE= DIRECT METHOD YTTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER,
ECHO= UNSORTED MAGNITUDE
SPC= 1
SET 111= 11, 33, 55 YBTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER, PHASE
DISPLACEMENT(PHASE, SORT2)= 111 XYPLOT DISP RESPONSE / 55 (T3RM, T3IP)
$SDISP(PHASE, SORT2)= ALL $
set 222 = 11 BEGIN BULK
OLOAD= 222
SUBCASE 1 $
DLOAD= 500 $ PLATE MODEL DESCRIBED IN NORMAL MODES EXAMPLE
FREQUENCY= 100 $
$ param,post,0
INCLUDE 'plate.bdf'
PARAM, COUPMASS, 1
$
$
PARAM, G, 0.06
$
ENDDATA
SOLUTION FOR WORKSHOP # 8A
ID SEMINAR, PROB8A
XTITLE= FREQUENCY (HZ)
SOL 108
YTTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER,
CEND
MAGNITUDE
TITLE= FREQUENCY RESPONSE DUE TO .1 DISPLACEMENT AT
YBTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, PHASE
TIP
SUBTITLE= DIRECT METHOD
YTTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, MAGNITUDE
ECHO= UNSORTED
YBTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, PHASE
SPC= 1
SET 111= 11, 33, 55
DISPLACEMENT(PHASE, SORT2)= 111
MAGNITUDE
$SDISP(PHASE, SORT2)= ALL
YBTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER,
set 222 = 11
PHASE
OLOAD= 222
SUBCASE 1
$
DLOAD= 500
BEGIN BULK
FREQUENCY= 100
$
$
OUTPUT (XYPLOT)
$
$
INCLUDE ’plate.bdf’
XTGRID= YES
PARAM, COUPMASS, 1
YTGRID= YES
XBGRID= YES
$
YBGRID= YES
YTLOG= YES
YBLOG= NO

SOLUTION FOR WORKSHOP # 8A (cont.)
$
PARAM, G, 0.06
$
$ APPLY UNIT DISPLACEMENT AT TIP POINT
$
SPC1, 1, 3, 11
SPCD, 600, 11, 3, 0.1
$
RLOAD2, 500, 600, , ,310, , DISP
$
TABLED1, 310, 0., 1., 10000., 1., ENDT
$
$
$ SPECIFY FREQUENCY STEPS
$
FREQ1, 100, 20., 20., 49
$
ENDDATA

8A (Cont.)
0 SUBCASE 1
0 1 DISP 1 11( 5,--) 2.000000E+01 1.000000E+03 1.000000E-01 2.000000E+01 1.000000E-01 2.000000E+01
2.000000E+01 1.000000E+03 1.000000E-01 2.000000E+01 1.000000E-01 2.000000E+01
0 1 DISP 1 11(--, 11) 2.000000E+01 1.000000E+03 0.000000E+00 2.000000E+01 0.000000E+00 2.000000E+01
2.000000E+01 1.000000E+03 0.000000E+00 2.000000E+01 0.000000E+00 2.000000E+01
0 1 DISP 2 33( 5,--) 2.000000E+01 1.000000E+03 2.310271E-03 6.000000E+02 8.439928E-01 3.800000E+02
2.000000E+01 1.000000E+03 2.310271E-03 6.000000E+02 8.439928E-01 3.800000E+02
0 1 DISP 2 33(--, 11) 2.000000E+01 1.000000E+03 3.306744E-01 9.799999E+02 3.599946E+02 2.000000E+01
2.000000E+01 1.000000E+03 3.306744E-01 9.799999E+02 3.599946E+02 2.000000E+01
0 1 DISP 3 55( 5,--) 2.000000E+01 1.000000E+03 2.443192E-02 1.000000E+03 1.623458E+00 3.800000E+02
2.000000E+01 1.000000E+03 2.443192E-02 1.000000E+03 1.623458E+00 3.800000E+02
0 1 DISP 3 55(--, 11) 2.000000E+01 1.000000E+03 3.672138E+00 1.000000E+03 3.599891E+02 2.000000E+01
2.000000E+01 1.000000E+03 3.672138E+00 1.000000E+03 3.599891E+02 2.000000E+01

8A (Cont.)

8A (Cont.)

8A (Cont.)

WORKSHOP # 8B
MODAL FREQUENCY RESPONSE WITH

ENFORCED DISPLACEMENT

WORKSHOP # 8B – MODAL FREQUENCY
RESPONSE WITH ENFORCED DISPLACEMENT
Using the modal method, determine the frequency response of the flat
rectangular plate, created in Workshop 1, subject to a 0.1 enforced
displacement at the corner of the tip. Use a frequency step of 20 Hz in
the range of 20 to 1000 Hz. Use a structural damping of g = 0.06. Be
sure to include residual vectors.

WORKSHOP 8B
● Use solution from workshop 8A as a starting point and

make appropriate changes.

ID SEMINAR, PROB8B
SOL 111 XTITLE= FREQUENCY (HZ)
CEND YTTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, MAGNITUDE
TITLE= FREQUENCY RESPONSE DUE TO .1 DISPLACEMENT YBTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, PHASE
AT TIP XYPLOT DISP RESPONSE / 11 (T3RM, T3IP)
SUBTITLE= MODAL METHOD YTTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, MAGNITUDE
ECHO= UNSORTED YBTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, PHASE
SPC= 1 XYPLOT DISP RESPONSE / 33 (T3RM, T3IP)
SET 111= 11, 33, 55 YTTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER,
DISPLACEMENT(PHASE, SORT2)= 111 MAGNITUDE
$SDISP(PHASE, SORT2)= ALL YBTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER, PHASE
set 222 = 11 XYPLOT DISP RESPONSE / 55 (T3RM, T3IP)
OLOAD= 222 $
SUBCASE 1 BEGIN BULK
METHOD= 1000 $
DLOAD= 500 $ PLATE MODEL DESCRIBED IN NORMAL MODES EXAMPLE
FREQUENCY= 100 $
$ INCLUDE ’plate.bdf’
OUTPUT (XYPLOT) PARAM, COUPMASS, 1
$ PARAM, WTMASS, 0.00259
XTGRID= YES $
YTGRID= YES $ SPECIFY STRUCTURAL DAMPING
XBGRID= YES
YBGRID= YES
YTLOG= YES
YBLOG= NO

$
PARAM, G, 0.06
EIGRL, 1000, , , 10
$
$ APPLY UNIT DISPLACEMENT AT TIP POINT
$
SPC1, 1, 3, 11
SPCD, 600, 11, 3, 0.1
$
RLOAD2, 500, 600, , ,310, , DISP
$
TABLED1, 310, 0., 1., 10000., 1., ENDT
$
$
$
FREQ1, 100, 20., 20., 49
$
ENDDATA

8B (Cont.)
0 SUBCASE 1
0 1 DISP 1 11( 5,--) 2.000000E+01 1.000000E+03 1.000000E-01 2.000000E+01 1.000000E-01 2.000000E+01
2.000000E+01 1.000000E+03 1.000000E-01 2.000000E+01 1.000000E-01 2.000000E+01
0 1 DISP 1 11(--, 11) 2.000000E+01 1.000000E+03 0.000000E+00 2.000000E+01 0.000000E+00 2.000000E+01
2.000000E+01 1.000000E+03 0.000000E+00 2.000000E+01 0.000000E+00 2.000000E+01
0 1 DISP 2 33( 5,--) 2.000000E+01 1.000000E+03 2.307091E-03 6.000000E+02 8.439932E-01 3.800000E+02
2.000000E+01 1.000000E+03 2.307091E-03 6.000000E+02 8.439932E-01 3.800000E+02
0 1 DISP 2 33(--, 11) 2.000000E+01 1.000000E+03 3.325545E-01 9.799999E+02 3.599946E+02 2.000000E+01
2.000000E+01 1.000000E+03 3.325545E-01 9.799999E+02 3.599946E+02 2.000000E+01
0 1 DISP 3 55( 5,--) 2.000000E+01 1.000000E+03 2.447906E-02 1.000000E+03 1.623457E+00 3.800000E+02
2.000000E+01 1.000000E+03 2.447906E-02 1.000000E+03 1.623457E+00 3.800000E+02
0 1 DISP 3 55(--, 11) 2.000000E+01 1.000000E+03 3.658197E+00 1.000000E+03 3.599891E+02 2.000000E+01
2.000000E+01 1.000000E+03 3.658197E+00 1.000000E+03 3.599891E+02 2.000000E+01

8B (Cont.)

8B (Cont.)

8B (Cont.)

BLANK

SECTION 13
SHOCK AND RESPONSE SPECTRUM

ANALYSIS

RESPONSE SPECTRUM
● Response spectrum depicts the maximum response of an SDOF
system as a function of its resonant frequency for base excitation.
The following graph
Response
Resonant Frequency (Hz)
is generated from
x(t)
FN = f1 f2 f3 FMAX
uB (t)
Point on larger, vibrating structure

RESPONSE SPECTRUM (Cont.)
● The peak response of each SDOF oscillator is calculated from its X(t).
The oscillator base motion UB is derived from the force or base
excitation applied to a larger structure.
● Example: An earthquake time history is applied to a power plant.
Response spectra are calculated at the locations of the floors to be
used in the design of components (i.e., machinery and piping
systems).
● An implicit assumption is that the oscillator’s mass is very small
relative to the larger, vibrating mass. Therefore, no dynamic
interaction occurs between the two. (Consequently, the response
spectrum analysis is decoupled from the transient analysis.)

● Analysis is repeated for several damping values to generate a family of curves.
1
1 damping = 0% critical
2
Response 2 damping = 3% critical
3
Resonant Frequency (Hz)
Damping applies to each oscillator, not to the vibrating structure.

● Maximum displacement response from X(t) is calculated for each oscillator. The maximum relative
displacement between each oscillator and its base (a point on the vibrating structure) is also
computed. X = maximum intertial (absolute) displacement
X r = maximum relative displacement

● Relative velocity and absolute acceleration are approximately related to the relative displacements
by  Xr 
X r

X2X r
● 
For design, useful variables are Xr, Xr, and X. Design spectra are usually in terms of these
variables. ..
.

.
ZETA=.05 SDOF RESPONSE TO 4 HZ SINE PULSE
SINE 978.2 2 HZ SYS 2.327 4 HZ SYS
1.000
0.0 0.0
-1.000
-2.000
-1.145 -2.726
0.0 1.490 0.0 1.490 0.0 1.490
10 HZ SYS 1.039 20 HZ SYS 1.068 40 HZ SYS
0.0 0.0
-1.016 -999.7m
1.490 0.0 1.490 0.0 1.490

SHOCK SPECTRUM OF 4 HZ PULSE FOR ZETA = .05
2.009 4.025 9.928 20.41
100.0 1.00 10.0 20.0

● For very low oscillator frequencies 
 0 
X 0
Xr  UB
● For very high oscillator frequencies 
 
X  UB
Xr  0
. ..
● The approximate relationships between Xr, Xr, and X are not valid at
very high or very low frequencies or for large damping values.
● Note that only the magnitudes of response are computed with no
phase information.

● Response spectra may be generated in any transient
solution (e.g., SOLs 109, 112).
● The transient response for selected DOFs in the model is

used as the input time history for the generation of the
response spectra curves.
● Further information is in the MSC.Nastran Advanced

Dynamics User’s Guide.

RESPONSE SPECTRUM GENERATION
Required input

SOL command selecting a transient solution (e.g., SOL 109)

XYPLOT SPECTRAL Compute spectra
XYPUNCH SPECTRAL Punch spectra
Example:
XYPUNCH ACCELERATION SPECTRAL 1/1(T1RM)
This XYPLOT command creates a set of absolute (RM) acceleration

spectra based on record number 1 on the DTI,SPSEL entry using the
motion of Grid Point 1 in the x (T1) direction.

RESPONSE SPECTRUM GENERATION (Cont.)
PARAM,RSPECTRA,0 Requests calculation of spectra

DTI,SPSEL Correlates frequency and damping requests
FREQ Used to specify frequencies and damping
values (one FREQ set each)
● Example Input:

APPLYING SPECTRA
● Available in SOL 103
● “Poor man’s transient.” The input spectra are used to determine

the peak response of each mode.
● These peak modal responses are combined to obtain the system

response (the only problem is that the timing of each mode’s peak
is not known).
● Three methods of combining the modal responses are available

(ABS, SRSS, NRL).

APPLYING SPECTRA (Cont.)
Procedure
● A model of the structure to be analyzed is created with the input

points identified as ‘SUPORT’ DOFs.
● A “large mass” (usually 103 to 106 times the structural mass) is

attached to the ‘SUPORT’ DOFs.
● System modes are obtained for the model (including the 0.0 Hz
modes) with the ‘SUPORT’ DOFs unconstrained.
● This approximates the “cantilevered” modes of the model attached

to the “exciting” structure.

● The 0.0 Hz modes (Dm) approximate the ‘static’ motion the model
experiences when the supporting structure moves statically.
● “Participation Factors” (PF) are calculated using the following
expression:

TMD
m
● where = the set of “elastic” modes
● PF is used in conjunction with the spectra described as shown in the

input section to calculate the peak response for each mode.
● Data recovery quantities (displacements, stresses, forces, etc.) are
then calculated for each mode based on its peak motion.
● These quantities are then combined for the modes using the selected
method (ABS, SRSS, NRL), and the results are printed.

● Xr, response of a single DOF oscillator due to the base motion is
calculated as follows:
·· · ··
xr + gxr + 2 xr = urt
● The actual transient response at a physical point is
uk 
t=
ik ir x r ig it 
i r
● ABSOLUTE Option
uk 
i k ir xriigi 
i r
● where xri igi = max xri igi t 

● and i Represents a mode
● and r Represents a direction

● SRSS option
2
uk  i k i
i
where the average peak modal magnitude, i is

2
i   
ir xr 
ig i
r
● NRL option
2
uk jk j +
 
ik i
i j
where jk j is the peak modal magnitude

Required Input
SOL statement selecting SOL 103
SDAMP To select modal damping ratios
DLOAD To select input spectra
METHOD To select eigenvalue solver
● Example
METHOD = 1 Selects eigenvalue solution method 1 from the Bulk
Data (be sure that the range includes 0.0)
SDAMP = 1 Selects modal damping to be used for the calculated
modes. Refers to a TABDMP1 entry in the Bulk Data
DLOAD = 1 Selects DLOAD Bulk Data entry that describes which
spectra are applied at which ‘SUPORT’ DOF

Bulk Data
PARAM,SCRSPEC,0 Requests application of response

spectra
DLOAD Selects spectra and ‘SUPORT’ DOF at
which to apply them
DTI,SPECSEL Selects spectra, states associated
damping and type of spectrum
TABLED1 Provides input spectra
SUPORT Selects spectrum input locations
TABDMP1 Describes modal damping for the
calculated modes
PARAM,OPTION Selects modal combination method

● Sample Input
SUPORT 1 3
$
$ Define input dof for the spectra - in this case, dof 3 for GRID 1
$ is selected
$
CONM2 1001 3 0 1000000.
$ apply large masses in the directions of the spectra input
TABDMP1 1 CRIT +DMP1
+DMP1 0.0 .01 100. .01 100.01 .02 1000. .02 +DMP2
+DMP2 ENDT
$
$ Select damping ratios for the calculated modes - in this case, a ratio of
$ 1% of ctrtical $ is used for all modes from 0hz to 100hz and 2% of critical
$ is used for all modes above $ 100.01hz
$
$
PARAM,SCRSPEC,0
$ Tells MSC.Nastran to perform shock spectrum analysis
$ DEFINE WHERE AND HOW TO APPLY SPECTRA
$
$ NOTE THAT SPECTRA ARE APPLIED USING INTERNAL SORT...NOT ASCENDING ORDER
$
$ SID S S1 L1 S2 L2 ....
DLOAD 1 1.0 1.0 1
$
$ Define where spectra are to be applied - this entry is called from
$ Case Control by a $ ’DLOAD=1’ command - for this entry, an overall
$ scale factor of 1.0 (S) is applied, $ a factor of 1.0 (S1) is used
$ to apply spectrum 1 (L1) at ’SUPORT’ dof 1.
$
$ (It should be noted that the order of the ’SUPORT’ dof used on this
$ entry is the MSC.Nastran internal sort. If only one GRID point is used,
$ this is no problem, but if more than one GRID point is used,
$ then PARAM,USETPRT,1 should be used to obtain the internal order)
$

DTI SPECSEL 0
DTI SPECSEL 1 A 2 0.0 3 .01 +SP1
+SP1 4 .02
$
$ This table defines the relationship between the input tables
$ (from the spectrum creation run) and the spectra sets. For example,
$ record 1 defines a spectra set representing acceleration spectra,
$ containing spectrum 2 for 0% of critical damping, spectrum 3 for 1% of
$ critical damping, and spectrum 4 for 2% of critical damping.
$ The program will interpolate between the spectra if a mode has a
$ damping value other than those defined in the table.
$
$ GRID Component (XYPLOT terminology)
$ACCE 0 1 3 1
$ 0.000000E+00
TABLED1 2
.5 3.156-4 1.0 .001263 1.5 .002842 2. .005056
2.5 .007905 3. .011393 3.5 .015524 4. .020303
4.5 .025738 5. .031839 5.5 .038615 6. .046073
6.5 .054219 7. .063052 7.5 .072569 8. .082766
.
.
100.5 3.87229ENDT
$ Table representing the input spectra

WORKSHOP #9 (PART I)
Generates The Shock Spectrum Input
Generate a shock spectrum of the plate due to a 2.0 in/sec2
sine pulse applied at the clamped edge for 0, 2, and 4 percent
damping.

WORKSHOP # 9A
● Start with the following partial input file
$
$ wkshp9a.dat
$
$ case control, add : generate response spectrum curve at grid point 3000
$ using xyplot/xypunch commands
$ bulk data, add : appropriate parameters
$ create response spectrum curves for appropriate damping
$ and frequencies at grid point 3000
$
ID SEMINAR, PROB9a
SOL 109
TIME 30
CEND
TITLE= TRANSIENT RESPONSE
SUBTITLE= USING DIRECT TRANSIENT METHOD
LABEL= SHOCK SPECTRUM CALCULATION
ECHO= UNSORTED
SPC= 100
SET 111= 3000
DISPLACEMENT (SORT2)= 111 $ AT LEAST DISP AND VEL MUST APPEAR
VELOCITY (SORT2)= 111
ACCELERATION ()= 111
DLOAD= 500
TSTEP= 100
$

WORKSHOP # 9A
$
OUTPUT (XYPLOT)
$
$ SHOCK RESPONSE IS ONLY AVAILABLE IN PLOT OR PUNCH OUTPUT. THEREFORE,
$ THE `OUTPUT(XYPLOT)' SECTION OF THE CASE CONTROL MUST BE USED.
$
XGRID=YES
YGRID=YES
XYPLOT ACCE / 3000(T1)
XLOG= YES
YLOG= YES
$
$ RELATIVE SHOCK RESPONSES ARE CONTAINED IN THE IMAGINARY/PHASE
$ COMPONENTS OF THE OUTPUT
$ ABSOLUTE SHOCK RESPONSES ARE CONTAINED IN THE REAL/MAGNITUDE
$
$
BEGIN BULK
$
$ DEFINE GRID POINT
$
GRID, 3000, ,0.,0.,0., ,23456
$
$ DEFINE MASS
$
CMASS2, 100, 1.0, 3000, 1
$
$ APPLY LOADING TO MASS
$
TLOAD2, 500, 600, , 0, 0., 0.004, 250., -90.
$
DAREA, 600, 3000, 1, 1.
$
$
TSTEP, 100, 100, 4.0E-4, 1
$
ENDDATA

SOLUTION FILE FOR WORKSHOP #9 (PART I)
ID SEMINAR, PROB9a
SOL 109
TIME 30 XTITLE= FREQUENCY (CYCLES/SEC)
CEND YTITLE= RELATIVE DISPLACEMENT
TITLE= TRANSIENT RESPONSE
XYPLOT DISP SPECTRAL 1 / 3000 (T1IP)
SUBTITLE= USING DIRECT TRANSIENT METHOD
LABEL= SHOCK SPECTRUM CALCULATION YTITLE= RELATIVE VELOCITY
ECHO= UNSORTED XYPLOT VELOCITY SPECTRAL 1 / 3000 (T1IP)
SPC= 100 YTITLE= ABSOLUTE ACCELERATION
SET 111= 3000
XYPLOT ACCELERATION SPECTRAL 1 / 3000 (T1RM)
DISPLACEMENT (SORT2)= 111 $ AT LEAST DISP AND VEL MUST APPEAR
VELOCITY (SORT2)= 111 $
ACCELERATION ()= 111 $ PUNCH SHOCK SPECTRUM FOR LATER USE
DLOAD= 500 $
TSTEP= 100
XYPUNCH ACCELERATION SPECTRAL 1 / 3000(T1RM)
$
OUTPUT (XYPLOT) $
$ BEGIN BULK
$ SHOCK RESPONSE IS ONLY AVAILABLE IN PLOT OR PUNCH OUTPUT. THEREFORE, $
$ THE ‘OUTPUT(XYPLOT)’ SECTION OF THE CASE CONTROL MUST BE USED.
$ DEFINE GRID POINT
$
XGRID=YES $
YGRID=YES GRID, 3000, ,0.,0.,0., ,23456
XYPLOT ACCE / 3000(T1)
XLOG= YES
YLOG= YES
$
$ RELATIVE SHOCK RESPONSES ARE CONTAINED IN
$IMAGINARY/PHASE
$ ABSOLUTE SHOCK RESPONSES ARE CONTAINED IN THE REAL/MAGNITUDE
$

SOLUTION FILE FOR WORKSHOP #9 (PART I)
(Cont.)
$
$ DEFINE MASS
$
CMASS2, 100, 1.0, 3000, 1
$
$ APPLY LOADING TO MASS
$
TLOAD2, 500, 600, , 0, 0., 0.004, 250., -90.
$
DAREA, 600, 3000, 1, 1.
$
$
TSTEP, 100, 100, 4.0E-4, 1
$
$ PARAMETER TO CALCULATE SHOCK SPECTRUM
$
PARAM, RSPECTRA, 0
$
$ SPECIFY FREQUENCY AND DAMPING VALUES FOR
$ THE SDOF OSCILLATORS AT GRID 3000
$
DTI, SPSEL, 0
DTI, SPSEL, 1, 111, 222, 3000
$ 1= SUBCASE... 111= DAMPING... 222= FREQUENCIES... 3000= GRID NUMBER
$
$ DAMPING INFORMATION FOR OSCILLATORS
$
FREQ, 111, 0., 0.02, 0.04
$
$ NATURAL FREQUENCIES OF OSCILLATORS
$
FREQ1, 222, 20., 20., 49
$
ENDDATA

(PART I)

(PART I) (Cont.)

(PART I) (Cont.)

(PART I) (Cont.)

WORKSHOP #9 (PART II)
Applies The Shock Spectrum

Apply the shock spectrum generated in Part I and sum the response using the
SRSS option. Include modes up to 1000 Hz using 3% critical damping.

WORKSHOP 9 (PART 2)
● Use the following partial input file as a starting point
$
$ wkshp9b.dat
$
$ case control, add: appropriate solution sequence
$
$ bulk data, add: apply the response spectrum curves generated in run 9a.
$ use the srss summing method.
$
ID SEMINAR, PROB9b
TIME 30
CEND
TITLE= RESPONSE SPECTRUM ANALYSIS
SUBTITLE= USING CALCULATED SHOCK RESPONSE
LABEL= SHOCK WILL BE INPUT IN Z DIRECTION
ECHO= UNSORTED
SET 111= ALL
DISPLACEMENT= 111
SPC= 200
SUBCASE 1
METHOD= 100
SDAMP= 200
DLOAD= 500
$

WORKSHOP 9 (PART 2)
BEGIN BULK
$
$
INCLUDE 'plate.bdf'
PARAM,COUPMASS,1
PARAM,WTMASS,0.00259
$
$ BOUNDRY CONDITIONS FOR `CLAMPED' MODES
$
SPC1, 200, 1245, 1, 12, 23, 34, 45
$
$ PLACE BIG FOUNDATION MASS (BFM) AT BASE
$ TO STIMULATE `CLAMPED' MODES
$
CMASS2, 110, 1000., 23, 3
$
$
RBE2, 101, 23, 3, 1, 12, 34, 45
$
$ SUPORT CARD TO IDENTIFY EXCITATION DOFS
$
SUPORT, 23, 3
$
$ EIGENVALUE EXTRACTION
$ MUST BE MASS NORMALIZED (DEFAULT)
$
eigrl,100,0.,1000.
$
$ TABLE TO SPECIFY DAMPING FOR USE IN THE ANALYSIS
$
TABDMP1, 200, CRIT,
, 0., 0.03, 1000., 0.03, ENDT
$
$ SPECIFICATION OF SHOCK SPECTRUM TO BE USED
$
$ MODAL FREQUENCY RANGE CAN BE SELECTED USING
PARAM, LFREQ, 0.1
PARAM, HFREQ, 1000.
$
ENDDATA

SOLUTION FILE FOR WORKSHOP #9 (PART II)
$
ID SEMINAR, PROB9b $ PLACE BIG FOUNDATION MASS (BFM) AT BASE
SOL 103 $ TO STIMULATE ‘CLAMPED’ MODES
TIME 30 $
CEND CMASS2, 110, 1000., 23, 3
TITLE= RESPONSE SPECTRUM ANALYSIS $
SUBTITLE= USING CALCULATED SHOCK RESPONSE $ RBE MASS TO REMAINING BASE POINTS
LABEL= SHOCK WILL BE INPUT IN Z DIRECTION $
ECHO= UNSORTED RBE2, 101, 23, 3, 1, 12, 34, 45
SET 111= ALL $
DISPLACEMENT= 111 $ SUPORT CARD TO IDENTIFY EXCITATION DOFS
SPC= 200 $
SUBCASE 1 SUPORT, 23, 3
METHOD= 100 $
SDAMP= 200 $ EIGENVALUE EXTRACTION
DLOAD= 500 $ MUST BE MASS NORMALIZED (DEFAULT)
$ $
BEGIN BULK eigrl,100,0.,1000.
$ $
$ PLATE MODEL DESCRIBED IN NORMAL MODES $ TABLE TO SPECIFY DAMPING FOR USE IN THE ANALYSIS
EXAMPLE
$ $
INCLUDE ’plate.bdf’ TABDMP1, 200, CRIT,
PARAM,COUPMASS,1 , 0., 0.03, 1000., 0.03, ENDT
PARAM,WTMASS,0.00259 $
$ $ SPECIFICATION OF SHOCK SPECTRUM TO BE USED
$ BOUNDRY CONDITIONS FOR ‘CLAMPED’ MODES $
$ DLOAD, 500, 1.0, 2.0, 1
SPC1, 200, 1245, 1, 12, 23, 34, 45 $
$ $ DLOAD, ID, OVERALL SCALE, SCALE FOR R-SET DOF# 1, SHOCK TABLE FOR
$DOF# 1,
$ SCALE FOR R-SET DOF# 2, SHOCK TABLE FOR DOF# 2, ETC.
$

(Cont.)
$ SELECT SHOCK RESPONSE CALCULATION
1 RESPONSE SPECTRUM ANALYSIS MARCH 12, 2007 MD NASTRAN 3/ 1/07 PAGE 7
USING CALCULATED SHOCK RESPONSE
0 SHOCK WILL BE INPUT IN Z DIRECTION
0 I N P U T B U L K D A T A E C H O
. 1 .. 2 .. 3 .. 4 .. 5 .. 6 .. 7 .. 8 .. 9 .. 10 .
$
PARAM, SCRSPEC, 0
$
$ SELECT SUMMATION OPTION
$
PARAM, OPTION, SRSS
$
$ MODAL FREQUENCY RANGE CAN BE SELECTED USING
PARAM, LFREQ, 0.1
PARAM, HFREQ, 1000.
$
$ SPECIFICATION FOR SHOCK TABLES
$
DTI, SPECSEL, 0
DTI, SPECSEL, 1, , A, 2, 0., 3, 0.02,
, 4, 0.04, ENDREC
$
$ DTI, SPECSEL, SHOCK TABLE NUMBER, , [(A)CCELERATION, (V)ELOCITY, OR (D)ISP],
$ TABLED1 POINTER, DAMPING FOR TABLE, ETC.
$
$ PUNCH OUTPUT FOR SHOCK SPECTRUM CALCULATION
$
$ ACCE 4 3000 3 1
$ 0.000000E+00
$
TABLED1 2
20. .038683 40. .152539 60. .33511 80. .576059
100. .862049 120. 1.17619 140. 1.50169 160. 1.82018
180. 2.11404 200. 2.36801 220. 2.56617 240. 2.70027
260. 2.76275 280. 2.75073 300. 2.74632 320. 2.61887
340. 2.4218 360. 2.39068 380. 2.24931 400. 2.02296
420. 1.78538 440. 1.70355 460. 1.57056 480. 1.40493
500. 1.22608 520. 1.20483 540. 1.17631 560. 1.14097
580. 1.10048 600. 1.05582 620. 1.00818 640. .958761
660. .908725 680. .859158 700. .827667 720. .782127
740. .728996 760. .694088 780. .668602 800. .635044
820. .598496 840. .571831 860. .563072 880. .550499

(Cont.)
900. .528854 920. .509281 940. .500534 960. .498016
980. .488793 1000. .468321 ENDT
$ACCE 4 3000 3 52
$ 2.0000000E-02
TABLED1 3
20. .037708 40. .143365 60. .314936 80. .541342
100. .80976 120. 1.10506 140. 1.40671 160. 1.69567
180. 1.98167 200. 2.22217 220. 2.35249 240. 2.53055
260. 2.56231 280. 2.55577 300. 2.58668 320. 2.45921
340. 2.29411 360. 2.25956 380. 2.12901 400. 1.92605
420. 1.68656 440. 1.61355 460. 1.4968 480. 1.35263
500. 1.19796 520. 1.17707 540. 1.14947 560. 1.11613
580. 1.07807 600. 1.03637 620. .992124 640. .946383
0 SHOCK WILL BE INPUT IN Z DIRECTION
0 I N P U T B U L K D A T A E C H O
. 1 .. 2 .. 3 .. 4 .. 5 .. 6 .. 7 .. 8 .. 9 .. 10 .
660. .900171 680. .854434 700. .810016 720. .767647
740. .727923 760. .691288 780. .658039 800. .628311
820. .602091 840. .579207 860. .559362 880. .542128
900. .526973 920. .51329 940. .500403 960. .487602
980. .474171 1000. .459408 ENDT
$ACCE 4 3000 3 103
$ 4.0000000E-02
TABLED1 4
20. .039336 40. .137673 60. .297382 80. .511244
100. .764891 120. 1.04406 140. 1.31588 160. 1.58461
180. 1.85678 200. 2.10175 220. 2.19165 240. 2.3921
260. 2.39929 280. 2.42782 300. 2.44263 320. 2.317
340. 2.17923 360. 2.14283 380. 2.0227 400. 1.8407
420. 1.62279 440. 1.53417 460. 1.43168 480. 1.30597
500. 1.17212 520. 1.15165 540. 1.12513 560. 1.09349
580. 1.05768 600. 1.01868 620. .977462 640. .934986
660. .892143 680. .849752 700. .808538 720. .769114
740. .731968 760. .69746 780. .665814 800. .637115
820. .611319 840. .588261 860. .567655 880. .549125
900. .532205 920. .516369 940. .501047 960. .485644
980. .469568 1000. .452243 ENDT
$
ENDDATA

(PART II)
0 SHOCK WILL BE INPUT IN Z DIRECTION SUBCASE 1
1 1 0.0 0.0 0.0 1.000000E+00 0.0
2 2 7.057114E+05 8.400663E+02 1.337007E+02 1.000000E+00 7.057114E+05
3 3 1.877186E+07 4.332651E+03 6.895628E+02 1.000000E+00 1.877186E+07
4 4 2.811329E+07 5.302197E+03 8.438708E+02 1.000000E+00 2.811329E+07
^^^ USER INFORMATION MESSAGE 9047 (POSTREIG)
^^^ SCALED RESPONSE SPECTRA FOR RESIDUAL STRUCTURE ONLY
*** USER INFORMATION MESSAGE 7588 (GKAM)
BASED ON THE USER PARAMETERS LMODES, LFREQ OR HFREQ, ONLY 3 OF THE 4 COMPUTED STRUCTURE MODES
(MODES 2 THROUGH 4) WILL BE USED IN THIS MODAL TRANSIENT RESPONSE ANALYSIS
(DETAILS OF THE EIGENVALUE DATA FOR THE MODES USED ARE GIVEN BELOW)
(ACTUAL MODES USED IN THE DYNAMIC ANALYSIS)
2 2 7.057114E+05 8.400663E+02 1.337007E+02 1.000000E+00 7.057114E+05
3 3 1.877186E+07 4.332651E+03 6.895628E+02 1.000000E+00 1.877186E+07
4 4 2.811329E+07 5.302197E+03 8.438708E+02 1.000000E+00 2.811329E+07
0 MATRIX FN (GINO NAME 101 ) IS A DB PREC 1 COLUMN X 3 ROW RECTANG MATRIX.
0COLUMN 1 ROWS 1 THRU 3 --------------------------------------------------
ROW
1) 1.3370D+02 6.8956D+02 8.4387D+02
0THE NUMBER OF NON-ZERO TERMS IN THE DENSEST COLUMN = 3
0THE DENSITY OF THIS MATRIX IS 100.00 PERCENT.
0 PSIT
POINT VALUE POINT VALUE POINT VALUE POINT VALUE POINT VALUE
COLUMN 1
23 T3 -2.11561E-02
COLUMN 2
23 T3 9.14315E-13
COLUMN 3
23 T3 1.18597E-02
U S E T D E F I N I T I O N T A B L E ( I N T E R N A L S E Q U E N C E , R O W S O R T )
R DISPLACEMENT SET
0 -1- -2- -3- -4- -5- -6- -7- -8- -9- -10-
1= 23-3
0 SCALED SPECTRAL RESPONSE, SRSS OPTION, DLOAD = 500 CLOSE = 1.00

(PART II) (Cont.)
PARTIAL OUTPUT FILE FOR PROBLEM #9 (PART II) (Cont.)
0 MATRIX UHVR (GINO NAME 101 ) IS A REAL 3 COLUMN X 3 ROW SQUARE MATRIX.
0COLUMN 1 ROWS 1 THRU 3 --------------------------------------------------
ROW
1) 7.6204E-08 8.1011E-20 4.8920E-10
0COLUMN 2 ROWS 1 THRU 3 --------------------------------------------------
ROW
1) 6.4017E-05 3.5099E-16 2.5938E-06
0COLUMN 3 ROWS 1 THRU 3 --------------------------------------------------
ROW
1) 5.3778E-02 1.5207E-12 1.3753E-02
0THE NUMBER OF NON-ZERO TERMS IN THE DENSEST COLUMN = 3
0THE DENSITY OF THIS MATRIX IS 100.00 PERCENT.
SCALED SPECTRAL RESPONSE, SRSS OPTION

1 G 0.0 0.0 6.222940E-10 0.0 0.0 7.935161E-19
2 G 1.656899E-18 7.935161E-19 7.362168E-08 8.380406E-08 3.053683E-07 1.051286E-18
3 G 3.038017E-18 1.358321E-18 3.172597E-07 1.218669E-07 6.622535E-07 1.255057E-18
4 G 4.072633E-18 2.361632E-18 7.194942E-07 1.166965E-07 9.376346E-07 1.666253E-18
5 G 4.328228E-18 3.551627E-18 1.246647E-06 1.041301E-07 1.160351E-06 1.080113E-18
6 G 3.627554E-18 4.017149E-18 1.870941E-06 8.255398E-08 1.326449E-06 9.153620E-19
49 G 4.083018E-18 2.280663E-18 1.246647E-06 1.041301E-07 1.160351E-06 1.558670E-18

50 G 6.021022E-18 3.094015E-18 1.870941E-06 8.255398E-08 1.326449E-06 3.331682E-19
51 G 8.089630E-18 3.015453E-18 2.566087E-06 6.151257E-08 1.444593E-06 1.727660E-18
52 G 9.842749E-18 1.590179E-18 3.309156E-06 4.199305E-08 1.519643E-06 4.018735E-18
53 G 1.094405E-17 9.800442E-19 4.080598E-06 2.667464E-08 1.560018E-06 5.598891E-18
54 G 1.141588E-17 4.045981E-18 4.865306E-06 1.686023E-08 1.575035E-06 6.242042E-18
55 G 1.153212E-17 7.227172E-18 5.653790E-06 1.297877E-08 1.577920E-06 6.353067E-18

SECTION 14
RANDOM RESPONSE ANALYSIS

CLASSIFICATION OF DYNAMIC
ENVIRONMENTS
Deterministic
Transient
Periodic
Simple Shock
Harmonic Spectra
Random
Nonstationary
Stationary
MD Nastran
Ergodic

● Random vibration is vibration that can be described only in a statistical
sense. Its instantaneous magnitude at any time is not known; rather,
the probability of its magnitude exceeding a certain value is given.
● Examples include earthquake ground motion, ocean wave heights and
frequencies, wind pressure fluctuations on aircraft and tall buildings,
and acoustic excitation due to rocket and jet engine noise.
● MD Nastran performs random response analysis as post processing to
frequency response. Inputs include the output from frequency
response analysis as well as user-supplied loading conditions in the
form of auto- and cross spectral densities. Outputs are response
power spectral densities, autocorrelation functions, number of zero
crossings with positive slope per unit time, and the RMS values of
response.
● The theory is described in Random Vibration in Mechanical Systems,
by S. H. Crandall and W. D. Mark, Academic Press, 1963.
● Further information is in the MSC.Nastran Advanced Dynamics User’s
Guide.

RANDOM ANALYSIS THEORY
● There are several conventions used to define random
analysis quantities. Care must be taken to use MD
Nastran random capability properly (see the MSC.Nastran
Advanced Dynamics User’s Guide for details and Bendat
and Piersol (Reference 13) to understand the
conventions).
● MD Nastran random analysis assumes ergodic processes.
● The concepts of autocorrelation, autospectrum (power
spectrum), cross-correlation, and cross-spectrum must be
defined.
● The mean square value and apparent frequency are the
key statistical quantities.

EXAMPLES OF RANDOM DYNAMIC
ENVIRONMENT
STATIONARY RANDOM
3.363
-3.242 0.0 10.23

1.751 NON-STATIONARY RANDOM
-1.449
0.0 10.23

EXAMPLE OF ENSEMBLE OF ERGODIC
RANDOM DATA

AUTOCORRELATION AND AUTOSPECTRUM
● Autocorrelation function:
● Note: Rj(o) is the mean-square value of uj(t).

● Autospectrum function:
2 T –it 2
S = lim --- u 
t e dt
j T  T 0 j
● Fourier transform pairs:

1 
Rj 
 = ------ S j 
cos 

d
2 0
● Mean square value:

2 1 
u j 
t = R j
o = ------ S j

d
20
AUTOCORRELATION AND AUTOSPECTRUM
● Apparent frequency N0 (zero crossings):

 2
    2  Sj  d
0
N 2 = -----------------------------------------------------
0 
 S   d
0 j

CALCULATION OF LINEAR SYSTEM RESPONSE
TO ERGODIC RANDOM EXCITATION
Uj ()
Fa ()
● From frequency response analysis

u
j 
= H 
ja 
F 
a
● where Hja(w) is the frequency response or transfer function

relating output uj to input Fa.
● If we have several inputs, then
u = H  
F + H  
F  + 
j ja a jb b
DEFINITION OF MULTIPLE INPUT-OUTPUT SPECTRAL
RELATIONSHIP FOR A LINEAR SYSTEM
● In matrix form we have
Fa ()
uj() = [ Hja() Hjb() … ] Fb ()
.
.
● The output autospectrum is
Fa () H*ja
Sujuj = T[ Hja Hjb … ] Fb () [F*a()F*b()…] H*jb
. .
. .
RELATIONSHIP FOR A LINEAR SYSTEM (Cont.)
● The individual input spectra are

TF a
F
a= S aa

TF a 
F
b= S ab

TFb 
F
b= S bb

● The multiple input-output spectral relationship is therefore:

S S 
aa ab
T S ba S bb 
S uju j
= 
Hj Hj 

. . .
. . .
. . .

  
H ja 
T  
where 
Hj  = 
Hja Hjb   H 
H j=  jb
 
 . 
 . 
 . 
 
The input cross-spectral matrix is
Saa() Sab() …
[ S ]in = S ba() S bb() …
. .
. .
and it has the special properties

= Sba
Sab 
Saa () , S bb 
= rea l 0

● Commonly used special cases
● Single input analysis (fully correlated inputs)
2
Sujuj 
= H ja
 S aa 

● Uncorrelated multiple inputs
2 2
S ujuj 
 = H ja 
 S aa
 + H jb 
 Sbb
+

RANDOM ANALYSIS AS IMPLEMENTED IN
MD NASTRAN
● It is assumed that the output from the frequency response

calculations is Hja(). It does not calculate
H ja 
= u j F a 

● If Hja() is desired, use F() = 1.0.

INPUT REQUIRED FOR RANDOM RESPONSE
ANALYSIS
SOL (required)
Structured
Solution
Direct 108
Modal 111

RANDOM (selects Bulk Data RANDPS, RANDT entries and entries
for frequency response, and must be above the
subcases)

RANDPS (PSD specification)
RANDT1 (autocorrelation time lag entries for frequency response)

RANDPS ENTRY

TABRND1 ENTRY

RANDOM RESPONSE OUTPUT
● Output option 1:
● Auto-PSDF, auto correlation functions, N0, and CRMS (cumulative root
mean square) print and punch output are available using “normal” Case
Control command.
● Available for acce, disp, velocity, force, oload, spcf, mpcf, stress, and
strain output, using the RPRINT and RPUNCH options
Format for Disp:
● Output option 2:
● Auto-psdf, auto correlation functions, and N0 (# of positive crossings) are
available using the xyplot/xypeak/xypunch options

RANDOM RESPONSE OUTPUT (CONT.)
● Format for Disp (cont.):

where
PSDF—request output for auto power spectral density function

ATOC—request output for auto correlation function
CRMS—request output for cumulative root mean square
RALL—request output for psdf, atoc, and crms
RPRINT—request printed output in the f06 file
RPUNCH—request punch output
NORPRINT—none of the above output
● Log-Log option available when computing RMS, N0, and CRMS

● Param,rmsint,log-log
RANDOM RESPONSE OUTPUT (Cont.)
● Additional Output:
● cross power spectral
● cross-correlation functions
● These output requires
● RCROSS (and RANDOM) Case Control commands
PSDF—request output for cross power spectral density function

CROF—request output for cross correlation function
RALL—request output for both psdf and crof
● RCROSS (and RANDPS) Bulk Data Entries

Example:
● The following simplified car model is subjected to random loadings that are fully
correlated at the front and back wheels
● Output request:
● The auto psdf disp, RMS, CRMS, N0 at grid points 1,2, and 5
● The cross psdf displacement output between grid 1 (t2 ) and grid 2 (t2)

RANDOM RESPONSE ANALYSIS (Cont.)
Example (cont):
SOL 108 $ $
CEND $ 2 3 4 5 6 7 8 9 10
TITLE = SIMPLE CAR WITH RANDOM INPUT RCROSS 100 DISP 1 2 DISP 2 2
SPC = 100 $
FREQUENCY = 1000 FREQ1 1000 0.1 .05 40
set 50 = 1,2,5 $
disp(phase,psdf,crms) = 50 $ DEFINE THE INPUT PSD
rcross(phase,psdf) = 100 $ 2 3 4 5 6 7 8 9 10
$ RANDPS 1000 1 1 1. 0. 145
random = 1000 RANDPS 1000 2 2 1. 0. 145
SUBCASE 1 RANDPS 1000 1 2 1. 0. 146
DLOAD = 111 RANDPS 1000 1 2 0. 1. 147
SUBCASE 2 TABRND1 145
DLOAD = 112 .1 .1 5. 1. 10. .05 ENDT
$ $
output (xyplot) .
xtitle = frequency (hz) .
ytitle = disp psd at grid pt 5 ENDDATA
xypunch disp psdf /5(t2)
$
BEGIN BULK

Example (cont):

EXAMPLE(cont.):

EXAMPLE(cont.):

RANDOM ANALYSIS RECOMMENDATIONS
● Most spectra are given as a log function. Use the log features on the
TABRND1 entry if PSD is given in log scale.
● Always generate the output PSD at the input location if possible.
● Plot the output PSD. Do not use the summary output blindly.
● Use several frequencies in the vicinity of each mode. For the modal
method, a combination of FREQ1 (or FREQ2) and FREQ4 usually
works best.
● For low frequencies (<20 Hz), use many frequencies since the
displacement spectra is changing rapidly for a constant input
acceleration.

WORKSHOP #10
RANDOM RESPONSE WITH SINGLE INPUT

WORKSHOP #10 - RANDOM RESPONSE
WITH SINGLE INPUT
● For the plate model below, apply a base motion in the z-
direction using the following power spectral density,
(PSD). 55 Autospectra of the
Base Excitation
Frequency (Hz) G2/Hz
9999 33 20 0.1
30 1
100 1
y 500 0.1
1000 0.1
x
● Connect the left edge with an RBE2 to grid point 9999 and
apply the enforced motion at grid point 9999
● Use modal solution

WITH SINGLE INPUT (Cont.)
● Assume a constant critical damping ratio of 3% across the

whole frequency range.
● Use a log-log input for the PSD.
● Determine the acceleration PSD response at the drive
point (grid point 9999) and at the corner and center of the
free edge (grid points 33 and 55)
● Request output in both print and xyplot format

WORKSHOP # 10
$
$ wkshp10.dat
$
$ case control, add : plot command for psd output
$
$ bulk data, add : enforced motion at grid point 9999
$ add forcing frequencies
$ add random input
$ residual vector
$
ID SEMINAR, PROB10
SOL 111
CEND
TITLE= RANDOM ANALYSIS - BASE EXCITATION
SUBTITLE= USING THE MODAL METHOD WITH LANCZOS
ECHO= UNSORTED
SPC= 101
SET 111= 33, 55, 9999
ACCELERATION(SORT2, PHASE)= 111
METHOD= 100
FREQUENCY= 100
SDAMPING= 100
RANDOM= 100
DLOAD= 100
$

WORKSHOP # 10
$
OUTPUT(XYPLOT)
XTGRID= YES
YTGRID= YES
XBGRID= YES
YBGRID= YES
YTLOG= YES
XTITLE= FREQUENCY
YTTITLE= ACCEL RESPONSE BASE, MAGNITUDE
YBTITLE= ACCEL RESPONSE AT BASE, PHASE
XYPLOT ACCEL RESPONSE / 9999 (T3RM, T3IP)
YTTITLE= ACCEL RESPONSE AT TIP CENTER, MAGNITUDE
YBTITLE= ACCEL RESPONSE AT TIP CENTER, PHASE
YTTITLE= ACCEL RESPONSE AT OPPOSITE CORNER, MAGNITUDE
YBTITLE= ACCEL RESPONSE AT OPPOSETE CORNER, PHASE
$
$ PLOT OUTPUT IS ONLY MEANS OF VIEWING PSD DATA
$
BEGIN BULK
param,post,0
PARAM,COUPMASS,1
$
INCLUDE 'plate.bdf'
$
GRID, 9999, , 0., 1., 0.
$
RBE2, 101, 9999, 12345, 1, 12, 23, 34, 45
$
SPC1, 101, 12456, 9999
$
$ EIGENVALUE EXTRACTION PARAMETERS
$
EIGRL, 100 , , 2000.
$
$ SPECIFY MODAL DAMPING
$
TABDMP1, 100, CRIT,
+, 0., .03, 10., .03, ENDT
$
ENDDATA

SOLUTION FILE FOR WORKSHOP # 10
BEGIN BULK
PARAM,COUPMASS,1
ID SEMINAR, PROB10 $
SOL 111 INCLUDE 'plate.bdf'
CEND $
TITLE= RANDOM ANALYSIS - BASE EXCITATION GRID, 9999, , 0., 1., 0.
SUBTITLE= USING THE MODAL METHOD WITH LANCZOS $
ECHO= UNSORTED RBE2, 101, 9999, 12345, 1, 12, 23, 34, 45
SPC= 101 $
SET 111= 33, 55, 9999 SPC1, 101, 12456, 9999
ACCELERATION (rall, PHASE)= 111 $
METHOD= 100 $ EIGENVALUE EXTRACTION PARAMETERS
FREQUENCY= 100 $
SDAMPING= 100 EIGRL, 100 , , 2000.
RANDOM= 100 $
DLOAD= 100 $ SPECIFY MODAL DAMPING
$ $
OUTPUT(XYPLOT) TABDMP1, 100, CRIT,
XTGRID= YES +, 0., .03, 10., .03, ENDT
YTGRID= YES $
XBGRID= YES $ POINT LOADING AT TIP CENTER
YBGRID= YES $
YTLOG= YES RLOAD2, 100, 600, , , 310,,A
XTITLE= FREQUENCY spcd,600,9999,3,1.0
YTTITLE= ACCEL RESPONSE BASE, MAGNITUDE spc1,101,3,9999
YBTITLE= ACCEL RESPONSE AT BASE, PHASE $
XYPLOT ACCEL RESPONSE / 9999 (T3RM, T3IP) TABLED1, 310,
YTTITLE= ACCEL RESPONSE AT TIP CENTER, MAGNITUDE +, 10., 1., 1000., 1., ENDT
YBTITLE= ACCEL RESPONSE AT TIP CENTER, PHASE $
XYPLOT ACCEL RESPONSE / 33 (T3RM, T3IP) $ SPECIFY FREQUENCY STEPS
YTTITLE= ACCEL RESPONSE AT OPPOSITE CORNER, MAGNITUDE $
YBTITLE= ACCEL RESPONSE AT OPPOSETE CORNER, PHASE FREQ,100,30.
XYPLOT ACCEL RESPONSE / 55 (T3RM, T3IP) FREQ1,100,20.,20.,50
$ FREQ4,100,20.,1000.,.03,5
$ PLOT OUTPUT IS ONLY MEANS OF VIEWING PSD DATA $
$ $ SPECIFY SPECTRAL DENSITY
XGRID= YES $
YGRID= YES RANDPS, 100, 1, 1, 1., 0., 111
XLOG= YES $
YLOG= YES TABRND1, 111,LOG,LOG
YTITLE= ACCEL P S D AT LOADED CORNER +, 20., 0.1, 30., 1., 100., 1., 500., .1,
XYPLOT ACCEL PSDF / 9999(T3) +, 1000., .1, ENDT
YTITLE= ACCEL P S D AT TIP CENTER $
XYPLOT ACCEL PSDF / 33(T3) ENDDATA
YTITLE= ACCEL P S D AT OPPOSITE CORNER
XYPLOT ACCEL PSDF / 55(T3)
$

POINT- ID = 33
A C C E L E R A T I O N V E C T O R
( POWER SPECTRAL DENSITY FUNCTION )
2.000000E+01 G 8.869005E-23 3.857560E-21 1.072273E-01 4.070925E- 20 9.165858E-06 2.384971E-21
3.000000E+01 G 4.759690E-21 2.069922E-19 1.171296E+00 2.184254E- 18 4.929607E-04 1.279754E-19
4.000000E+01 G 1.637099E-20 7.118072E-19 1.328467E+00 7.510322E- 18 1.700663E-03 4.400866E-19
6.000000E+01 G 1.078228E-19 4.685393E-18 1.937051E+00 4.941785E- 17 1.129788E-02 2.896886E-18
( ROOT MEAN SQUARE )

33 G 7.193928E-08 4.649358E-07 9.228086E+01 1.510409E- 06 4.589051E+01 3.657763E-07
55 G 4.272782E-07 4.836801E-07 9.166651E+01 1.969691E+00 4.610466E+01 3.796657E-07
9999 G 0.0 0.0 1.561982E+01 0.0 0.0 0.0
( NUMBER OF ZERO CROSSINGS )

33 G 2.561972E+02 2.110519E+02 3.924331E+02 2.173109E+02 7.319445E+02 2.122896E+02
55 G 2.176310E+02 2.103150E+02 3.846914E+02 8.502588E+02 7.328254E+02 2.098791E+02
9999 G 0.0 0.0 3.991722E+02 0.0 0.0 0.0

XYPLOT OUTPUT FOR WORKSHOP #10
0 1 ACCE 1 9999( 5,-- ) 2.000000E+01 1.020000E+03 9.999999E-01 1.356945E+02 1.000000E+00 1.296785E+02
2.000000E+01 1.020000E+03 9.999999E-01 1.356945E+02 1.000000E+00 1.296785E+02
0 1 ACCE 1 9999(--, 11) 2.000000E+01 1.020000E+03 0.000000E+00 2.000000E+01 0.000000E+00 2.000000E+01
2.000000E+01 1.020000E+03 0.000000E+00 2.000000E+01 0.000000E+00 2.000000E+01
0 1 ACCE 2 33( 5,-- ) 2.000000E+01 1.020000E+03 1.008510E+00 3.800000E+02 2.621251E+01 1.336891E+02
2.000000E+01 1.020000E+03 1.008510E+00 3.800000E+02 2.621251E+01 1.336891E+02
0 1 ACCE 2 33(--, 11) 2.000000E+01 1.020000E+03 1.044886E+01 1.020000E+03 3.599818E+02 2.000000E+01
2.000000E+01 1.020000E+03 1.044886E+01 1.020000E+03 3.599818E+02 2.000000E+01
0 1 ACCE 3 55( 5,-- ) 2.000000E+01 1.020000E+03 1.000853E+00 3.800000E+02 2.617639E+01 1.336891E+02
2.000000E+01 1.020000E+03 1.000853E+00 3.800000E+02 2.617639E+01 1.336891E+02
0 1 ACCE 3 55(--, 11) 2.000000E+01 1.020000E+03 1.055883E+01 1.020000E+03 3.599818E+02 2.000000E+01
2.000000E+01 1.020000E+03 1.055883E+01 1.020000E+03 3.599818E+02 2.000000E+01
0 X Y - O U T P U T S U M M A R Y ( A U T O O R P S D F )
0 PLOT CURVE FRAME CURVE ID./ RMS NO. POSITIVE XMIN FOR XMAX FOR YMIN FOR X FOR YMAX FOR X FOR*
TYPE TYPE NO. PANEL : GRID ID VALUE CROSSINGS ALL DATA ALL DATA ALL DATA YMIN ALL DATA YMAX
0
PSDF ACCE 4 9999( 5) 1.561982E+01 3.991722E+02 2.000E+01 1.020E+03 1.000E-01 9.800E+02 1.000E+00 3.000E+01
0
0

OUTPUT FILE FOR WORKSHOP #10 (Cont.)






WORKSHOP #11
RANDOM RESPONSE WITH MULTIPLE
INPUTS

WITH MULTIPLE INPUTS
 Using the modal method, determine the displacement response
spectrum of the tip center point due to the input spectrum of the
pressure and point loads listed below. Use the complex matrix
representation (SAB) for the cross spectrum.
Autospectra of Pressure Load Auto Spectra of Corner Load
Frequency (Hz) psi/Hz Frequency (Hz) lb/Hz
20 0.1 20 0.5
30 1 30 2.5
100 1 500 2.5
500 0.1 1000 0
1000 0.1
Cross-Spectrum of Pressure and Corner Loads Real/Imaginary

Frequency (Hz) Real Part Imaginary Part
20 -0.099619 0.007816
100 -0.498097 0.043579
500 0.070711 -0.070711
1000 0 0

WITH MULTIPLE INPUTS
● Request Auto psdf and CRMS displacement output at

grid points 11, 33, and 55
● Request cross spectrum displacement output between

grid point 11 direction 3 and grid point 55 direction 3

WORKSHOP # 11
$
$ wkshp11.dat
$
$ add: random input plus corresponding output
$
SOL 111
CEND
TITLE= FREQUENCY RESPONSE WITH PRESSURE AND POINT LOADS
SUBTITLE= USING THE MODAL METHOD WITH LANCZOS
ECHO= UNSORTED
SPC= 1
SET 111= 11, 33, 55
DISPLACEMENT(PLOT, PHASE)= 111
ACCELERATION(PLOT,PHASE) = 111
METHOD= 100
FREQUENCY= 100
SDAMPING= 100
DISP(PSDF,CRMS,PHASE)=111
SUBCASE 1
LABEL= PRESSURE LOAD
DLOAD= 100
SUBCASE 2
LABEL = CORNER LOAD
DLOAD= 200
$

WORKSHOP # 11
OUTPUT (XYPLOT) PARAM,COUPMASS,1
$ PARAM,WTMASS,0.00259
XTGRID= YES $
YTGRID= YES $ MODEL DESCRIBED IN NORMAL MODES EXAMPLE
XBGRID= YES $
YBGRID= YES INCLUDE 'plate.bdf'
YTLOG= YES $
YBLOG= NO $ EIGENVALUE EXTRACTION PARAMETERS
XTITLE= FREQUENCY (HZ) $
YTTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, MAGNITUDE EIGRL, 100, 10., 2000.
YBTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, PHASE $
XYPLOT DISP RESPONSE / 11 (T3RM, T3IP) $ SPECIFY MODAL DAMPING
YTTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, MAGNITUDE $
YBTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, PHASE TABDMP1, 100, CRIT,
XYPLOT DISP RESPONSE / 33 (T3RM, T3IP) +, 0., .03, 10., .03, ENDT
YTTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER, MAGNITUDE $
YBTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER, PHASE $ FIRST LOADING
$ RLOAD2, 100, 400, , , 310
$ PLOT OUTPUT IS ONLY MEANS OF VIEWING PSD DATA $
$ TABLED1, 310,
XGRID= YES +, 10., 1., 1000., 1., ENDT
YGRID= YES $
XLOG= YES $ UNIT PRESSURE LOAD TO PLATE
YLOG= YES $
YTITLE= DISP P S D AT LOADED CORNER PLOAD2, 400, 1., 1, THRU, 40
XYPLOT DISP PSDF / 11(T3) $
YTITLE= DISP P S D AT TIP CENTER $ SECOND LOADING
YTITLE= DISP P S D AT OPPOSITE CORNER RLOAD2, 200, 600, , , 310
$ $ POINT LOAD AT TIP CENTER
BEGIN BULK $
FORCE,600,11,,1.,0.,0.,1.
$
$
FREQ1, 100, 20., 20., 49
$
ENDDATA

SOLUTION FILE FOR WORKSHOP11
$
$
$ soln11.dat
OUTPUT (XYPLOT)
$ $
ID SEMINAR, PROB11 XTGRID= YES
SOL 111 YTGRID= YES
CEND XBGRID= YES
YBGRID= YES
TITLE= FREQUENCY RESPONSE WITH PRESSURE AND POINT LOADS YTLOG= YES
SUBTITLE= USING THE MODAL METHOD WITH LANCZOS YBLOG= NO
ECHO= UNSORTED XTITLE= FREQUENCY (HZ)
SPC= 1 YTTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, MAGNITUDE
YBTITLE= DISPLACEMENT RESPONSE AT LOADED CORNER, PHASE
SET 111= 11, 33, 55
DISPLACEMENT(PLOT, PHASE)= 111 YTTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, MAGNITUDE
ACCELERATION(PLOT,PHASE) = 111 YBTITLE= DISPLACEMENT RESPONSE AT TIP CENTER, PHASE
METHOD= 100 XYPLOT DISP RESPONSE / 33 (T3RM, T3IP)
FREQUENCY= 100 YTTITLE= DISPLACEMENT RESPONSE AT OPPOSITE CORNER, MAGNITUDE
SDAMPING= 100
RANDOM= 100 $
DISP(PSDF,CRMS,PHASE)=111 $ PLOT OUTPUT IS ONLY MEANS OF VIEWING PSD DATA
RCROSS(PSDF,PHASE)=1000 $
SUBCASE 1 XGRID= YES
YGRID= YES
LABEL= PRESSURE LOAD XLOG= YES
DLOAD= 100 YLOG= YES
SUBCASE 2 YTITLE= DISP P S D AT LOADED CORNER
LABEL = CORNER LOAD XYPLOT DISP PSDF / 11(T3)
YTITLE= DISP P S D AT TIP CENTER
DLOAD= 200
XYPLOT DISP PSDF / 33(T3)
$
YTITLE= DISP P S D AT OPPOSITE CORNER
XYPLOT DISP PSDF / 55(T3)

SOLUTION FOR WORKSHOP11 (Cont.)
$ $
BEGIN BULK $ SPECIFY FREQUENCY STEPS
PARAM,COUPMASS,1
PARAM,WTMASS,0.00259 $
$ FREQ1, 100, 20., 20., 49
$ MODEL DESCRIBED IN NORMAL MODES EXAMPLE
$ $
INCLUDE 'plate.bdf' RCROSS,1000,DISP,11,3,DISP,55,3
$
$ EIGENVALUE EXTRACTION PARAMETERS $
$ $ SPECIFY SPECTRAL DENSITY
EIGRL, 100, 10., 2000.
$
$
$ SPECIFY MODAL DAMPING RANDPS, 100, 1, 1, 1., 0., 100
$
TABDMP1, 100, CRIT,
RANDPS, 100, 2, 2, 1., 0., 200
+, 0., .03, 10., .03, ENDT RANDPS, 100, 1, 2, 1., 0., 300
$
$ FIRST LOADING
RANDPS, 100, 1, 2, 0., 1., 400
$ $
RLOAD2, 100, 400, , , 310
$
TABRND1, 100,
TABLED1, 310, +, 20., 0.1, 30., 1., 100., 1., 500., .1,
+, 10., 1., 1000., 1., ENDT +, 1000., .1, ENDT
$
$ UNIT PRESSURE LOAD TO PLATE $
$ TABRND1, 200,
PLOAD2, 400, 1., 1, THRU, 40
$ +, 20., 0.5, 30., 2.5, 500., 2.5, 1000., 0.,
$ SECOND LOADING +, ENDT
$
RLOAD2, 200, 600, , , 310 $
$ TABRND1, 300,
$ POINT LOAD AT TIP CENTER
$ +, 20., -.099619, 100., -.498097, 500., .070711, 1000., 0.,
FORCE,600,11,,1.,0.,0.,1. +, ENDT
$
TABRND1, 400,
+, 20., .0078158, 100., .0435791, 500., -.70711, 1000., 0.,
+, ENDT
$
ENDDATA

PARTIAL OUTPUT FOR WORKSHOP11
0 1 DISP 1 11( 5,--) 2.000000E+01 1.000000E+03 5.214993E-04 1.000000E+03 2.645494E-01 1.400000E+02
2.000000E+01 1.000000E+03 5.214993E-04 1.000000E+03 2.645494E-01 1.400000E+02
0 1 DISP 1 11(--, 11) 2.000000E+01 1.000000E+03 1.937994E+01 1.000000E+03 3.594682E+02 2.000000E+01
2.000000E+01 1.000000E+03 1.937994E+01 1.000000E+03 3.594682E+02 2.000000E+01
0 2 DISP 2 11( 5,--) 2.000000E+01 1.000000E+03 7.341043E-05 3.800000E+02 6.694620E-02 1.400000E+02
2.000000E+01 1.000000E+03 7.341043E-05 3.800000E+02 6.694620E-02 1.400000E+02
0 2 DISP 2 11(--, 11) 2.000000E+01 1.000000E+03 1.841973E+02 2.400000E+02 3.595171E+02 2.000000E+01
2.000000E+01 1.000000E+03 1.841973E+02 2.400000E+02 3.595171E+02 2.000000E+01
0 1 DISP 3 33( 5,--) 2.000000E+01 1.000000E+03 5.543213E-04 1.000000E+03 2.649228E-01 1.400000E+02
2.000000E+01 1.000000E+03 5.543213E-04 1.000000E+03 2.649228E-01 1.400000E+02
0 1 DISP 3 33(--, 11) 2.000000E+01 1.000000E+03 1.875292E+01 1.000000E+03 3.594681E+02 2.000000E+01
2.000000E+01 1.000000E+03 1.875292E+01 1.000000E+03 3.594681E+02 2.000000E+01
0 2 DISP 4 33( 5,--) 2.000000E+01 1.000000E+03 4.009693E-05 6.000000E+02 6.759480E-02 1.400000E+02
2.000000E+01 1.000000E+03 4.009693E-05 6.000000E+02 6.759480E-02 1.400000E+02
0 2 DISP 4 33(--, 11) 2.000000E+01 1.000000E+03 1.822676E+02 3.400000E+02 3.594853E+02 2.000000E+01
2.000000E+01 1.000000E+03 1.822676E+02 3.400000E+02 3.594853E+02 2.000000E+01
0 1 DISP 5 55( 5,--) 2.000000E+01 1.000000E+03 5.213259E-04 1.000000E+03 2.645494E-01 1.400000E+02
2.000000E+01 1.000000E+03 5.213259E-04 1.000000E+03 2.645494E-01 1.400000E+02
0 1 DISP 5 55(--, 11) 2.000000E+01 1.000000E+03 1.938930E+01 1.000000E+03 3.594682E+02 2.000000E+01
2.000000E+01 1.000000E+03 1.938930E+01 1.000000E+03 3.594682E+02 2.000000E+01
0 2 DISP 6 55( 5,--) 2.000000E+01 1.000000E+03 2.097658E-04 1.000000E+03 6.799091E-02 1.400000E+02
2.000000E+01 1.000000E+03 2.097658E-04 1.000000E+03 6.799091E-02 1.400000E+02
0 2 DISP 6 55(--, 11) 2.000000E+01 1.000000E+03 7.973937E+00 7.599999E+02 3.594539E+02 2.000000E+01
2.000000E+01 1.000000E+03 7.973937E+00 7.599999E+02 3.594539E+02 2.000000E+01

POINT-ID = 11
( POWER SPECTRAL DENSITY FUNCTION )
2.000000E+01 G 2.047303E-24 4.813135E-24 7.947328E-05 2.155795E-07 6.016012E-06 3.429885E-24
4.000000E+01 G 3.063225E-23 7.202310E-23 1.189875E-03 1.059814E-06 8.969508E-05 5.132451E-23
6.000000E+01 G 3.794013E-23 8.919452E-23 1.465137E-03 1.070977E-06 1.104442E-04 6.356191E-23
8.000000E+01 G 5.574504E-23 1.310384E-22 2.142037E-03 1.086198E-06 1.618793E-04 9.338235E-23
( ROOT MEAN SQUARE )
11 G 2.217665E-10 3.399741E-10 1.376854E+00 8.629130E-02 3.851211E-01 2.870088E-10
33 G 4.747619E-11 3.394890E-10 1.377021E+00 8.664112E-02 3.844410E-01 2.681251E-10
55 G 3.086634E-10 3.532137E-10 1.378706E+00 8.410328E-02 3.849473E-01 2.787290E-10
1 MARCH 13, 2007 MD NASTRAN 3/ 1/07 PAGE 47
0 RANDOM 100
( NUMBER OF ZERO CROSSINGS )
11 G 1.364916E+02 1.362757E+02 1.401700E+02 6.491689E+02 1.503498E+02 1.362929E+02
33 G 1.355889E+02 1.362910E+02 1.360215E+02 6.552164E+02 1.492609E+02 1.363054E+02
55 G 1.361721E+02 1.362927E+02 1.432664E+02 6.581000E+02 1.498830E+02 1.362823E+02
1 MARCH 13, 2007 MD NASTRAN 3/ 1/07 PAGE 48
0 RANDOM 100
POINT-ID = 11
( CUMULATIVE ROOT MEAN SQUARE )
2.000000E+01 G 0.0 0.0 0.0 0.0 0.0 0.0
4.000000E+01 G 1.807749E-11 2.771935E-11 1.126653E-01 3.571265E-03 3.093721E-02 2.339966E-11
6.000000E+01 G 3.182011E-11 4.879076E-11 1.981000E-01 5.836253E-03 5.439213E-02 4.118748E-11
8.000000E+01 G 4.415168E-11 6.769688E-11 2.744364E-01 7.458793E-03 7.537731E-02 5.714762E-11
1.000000E+02 G 6.018009E-11 9.226939E-11 3.734311E-01 8.809525E-03 1.027079E-01 7.789153E-11
1.200000E+02 G 9.969239E-11 1.528460E-10 6.180981E-01 1.001494E-02 1.708699E-01 1.290305E-10

1 FREQUENCY RESPONSE WITH PRESSURE AND POINT LOADS MARCH 13, 2007 MD NASTRAN 3/ 1/07 PAGE 52
USING THE MODAL METHOD WITH LANCZOS
0 PRESSURE LOAD RANDOM 100
SEQUENTIAL CURVE-ID = 1
C O M P L E X C R O S S - P O W E R S P E C T R A L D E N S I T Y F U N C T I O N
(MAGNITUDE/PHASE)
0 RCROSS RTYPE1 ID1 COMP1 RTYPE2 ID2 COMP2 CURID
0 1000 DISP 11 3 DISP 55 3 0
FREQUENCY CPSDF FREQUENCY CPSDF
2.000000E+01 7.782711E-05 / 359.7912 4.000000E+01 1.169207E -03 / 359.9816
6.000000E+01 1.446440E-03 / 359.9715 8.000000E+01 2.125043E -03 / 359.9651
1.000000E+02 4.256192E-03 / 359.9659 1.200000E+02 1.995366E -02 / 0.0206
1.400000E+02 5.931299E-02 / 0.0270 1.600000E+02 3.981690E -03 / 359.9641
1.800000E+02 1.110766E-03 / 359.8067 2.000000E+02 4.651507E -04 / 359.5229
2.200000E+02 2.361836E-04 / 359.0721 2.400000E+02 1.342108E -04 / 358.4012
2.600000E+02 8.205519E-05 / 357.4377 2.800000E+02 5.276163E -05 / 356.0812
3.000000E+02 3.515652E-05 / 354.1868 3.200000E+02 2.402327E -05 / 351.5385
3.400000E+02 1.670592E-05 / 347.8049 3.600000E+02 1.176429E -05 / 342.4655
3.800000E+02 8.387207E-06 / 334.7123 4.000000E+02 6.119115E -06 / 323.4214
4.200000E+02 4.722277E-06 / 307.6403 4.400000E+02 4.072001E -06 / 288.2907
4.600000E+02 4.034620E-06 / 269.2190 4.800000E+02 4.427422E -06 / 253.8799
5.000000E+02 5.106796E-06 / 242.7013 5.200000E+02 5.390143E -06 / 236.1571
5.400000E+02 5.879032E-06 / 229.9111 5.600000E+02 6.667265E -06 / 223.8539
5.800000E+02 7.941911E-06 / 217.8872 6.000000E+02 1.009844E -05 / 211.9292
6.200000E+02 1.407105E-05 / 205.9172 6.400000E+02 2.244283E -05 / 199.8112
6.600000E+02 4.367885E-05 / 193.5947 6.799999E+02 9.592211E -05 / 187.2714
6.999999E+02 7.964308E-05 / 180.8491 7.200000E+02 2.743963E -05 / 174.2928
7.400000E+02 1.057217E-05 / 167.3723 7.599999E+02 4.433168E -06 / 158.9403
7.800000E+02 1.351509E-06 / 138.3460 8.000000E+02 1.558493E -06 / 1.1614
8.200000E+02 5.536476E-06 / 343.4609 8.399999E+02 1.063914E -05 / 337.7312
8.600000E+02 6.950112E-06 / 334.6017 8.800000E+02 2.748527E -06 / 332.9402
8.999999E+02 1.156800E-06 / 332.4832 9.199999E+02 5.375278E -07 / 333.2076
9.400000E+02 2.648299E-07 / 335.2509 9.600000E+02 1.325349E -07 / 339.0211
9.799999E+02 6.395399E-08 / 345.7172 1.000000E+03 2.718711E -08 / 359.9906

X Y - O U T P U T S U M M A R Y ( A U T O O R P S D F )
0 PLOT CURVE FRAME CURVE ID./ RMS NO. POSITIVE XMIN FOR XMAX FOR YMIN FOR X FOR YMAX FOR X FOR*
TYPE TYPE NO. PANEL : GRID ID VALUE CROSSINGS ALL DATA ALL DATA ALL DATA YMIN ALL DATA YMAX
0
PSDF DISP 7 11( 5) 1.376854E+00 1.401700E+02 2.000E+01 1.000E+03 2.720E-08 1.000E+03 5.926E-02 1.400E+02
0
0

PARTIAL OUTPUT FOR WORKSHOP #11
(Cont.)

(Cont.)

(Cont.)

(Cont.)

(Cont.)

(Cont.)

(Cont.)

Partial Output For Workshop #11 (Cont.)

(Cont.)

SECTION 15
COMPLEX EIGENVALUE ANALYSIS

COMPLEX EIGENVALUE ANALYSIS
● Used to assess the stability of systems modeled with
transfer functions (including servomechanisms and
rotating systems)
● Also used to compute damped modes of systems
● Mass and stiffness matrices may be unsymmetric and may

contain complex numbers.
● See the MSC.Nastran Advanced Dynamics User’s Guide.

COMPLEX EIGENSOLUTIONS - THEORY
● Equation of motion
2

Mp + Bp + K u
 = 0 
● where p = + i
● and = real part of solution
● = imaginary part of solution
● For stable systems, a < 0
Imaginary
● Damping coefficient Real
g 
/





COMPLEX EIGENSOLUTIONS IN MD Nastran
● B matrix is the same as that used in frequency response analysis.
● The direct method solves the equation with M,B,K matrices of D-size
(physical variables plus extra points).
● Modal solves the equation with M,B,K matrices of H-size (modal
coordinates plus extra points). Undamped modes are first computed
to transform matrices from D-size to H-size.
● Four methods of complex eigenanalysis: HESS, INV, DET, and CLAN
● HESS (upper Hessenberg) is related to GIV. This method requires
nonsingular M and may be quite expensive for large problems.
Therefore, the modal method is usually recommended except for small
problems.
● HESS: solves canonical equation forms. There are two categories:
● Systems with [B] = 0
● Systems with [B] 0

COMPLEX EIGENSOLUTIONS IN MD Nastran
(Cont.)
● [B] = 0 case solves 
A – I

 = 0
● where
[B] 0 case studies:
[A] = -[M-1 K], = p 2 
A – p 
I
 = 0
 
where = u 
v 
0 I

A =
–1 –1
–M K –M B
● CLAN - Similar to real Lanczos - hybrid of a transformation and a tracking method.

● INV is related to INV for undamped eigenvalue analysis. The user must select the search
region in the complex plane. This method may be used for larger problems and a singular M is
allowed. However, the method is expensive and less reliable than HESS.
● DET is not recommended because it is awkward and inefficient.
● These methods with search regions are specified on the EIGC Bulk Data entry, which is
selected by the CMETHOD Case Control command.
● Mode acceleration is available and is selected by PARAM,MODACC,0 and PARAM,DDRMM,-
1. Mode acceleration has no effect on calculated eigenvalues, it only affects the data recovery.
INPUT REQUIRED FOR COMPLEX
EIGENANALYSIS
SOL (for required input see below)
Structured
Solution
Direct 107
Modal 110
CMETHOD (required for both)
METHOD (required for modal)
EIGC (required for both)
EIGR or EIGRL (required for modal)

WORKSHOP #12
COMPLEX EIGENANALYSIS
Compute the complex modes of the following pile driver
Exciter
Pile
Ground
WORKSHOP #12 - COMPLEX EIGENVALUE
1 m1
K1
2 m2
K2 C2
m1 3.0 lb-sec2/in
m2 1.5 lb-sec2/in
K1 50,000 lb/in
K2 12,500 lb/in
C2 30 lb-sec/in
WORKSHOP # 12
• Use the following partial input file as a starting point
$
$ wkshp12.dat
$
$ case control, add : direct complex eigenvalue callout
$
$ bulk data, add : Hess method
$ ask for eigenvalue and eigenvector output
$
CEND
TITLE= TWO-DOF MODEL (IMAC 8, PG 891)
SUBTITLE= COMPLEX MODES
SPC= 100
$
BEGIN BULK
$
$ COMPLEX EIGENVALUE EXTRACTION PARAMETERS
$
$
$ DEFINE GRIDS, MASSES, AND STIFFNESSES
$ GRID 1 = EXCITER (X=2, MASS=3) 50K STIFFNESS BETWEEN GRIDS 1 AND 2
$ GRID 2 = PILE (X=1, MASS=3) 12.5K STIFFNESS BETWEEN GRIDS 2 AND 3
$ GRID 3 = BASE (X=0, FIX BASE)
$
GRID, 1, , 2., 0., 0.
GRID, 2, , 1., 0., 0.
GRID, 3, , 0., 0., 0.
GRDSET, , , , , , , 23456
CELAS2, 1, 50000., 1, 1, 2, 1
CELAS2, 2, 12500., 2, 1, 3, 1
CONM2, 201, 1, , 3.0
CONM2, 202, 2, , 1.5
SPC, 100, 3, 1
$
$ DEFINE DAMPER OF 30 BETWEEN GRIDS 2 AND 3
$
CVISC, 101, 1, 2, 3
PVISC, 1, 30.
$
ENDDATA

SOL 107
CEND
TITLE= TWO-DOF MODEL (IMAC 8, PG 891)
SUBTITLE= COMPLEX MODES
DISPLACEMENT= ALL $ DEFAULT= REAL, IMAGINARY
SPC= 100
CMETHOD= 99
$
BEGIN BULK
$
$ COMPLEX EIGENVALUE EXTRACTION PARAMETERS
$
EIGC, 99, HESS, , , , , 4
$
$ DEFINE GRIDS, MASSES, AND STIFFNESSES
$ GRID 1 = EXCITER (X=2, MASS=3) 50K STIFFNESS BETWEEN GRIDS 1 AND 2
$ GRID 2 = PILE (X=1, MASS=3) 12.5K STIFFNESS BETWEEN GRIDS 2 AND 3
$ GRID 3 = BASE (X=0, FIX BASE)
$
GRID, 1, , 2., 0., 0.
GRID, 2, , 1., 0., 0.
GRID, 3, , 0., 0., 0.
GRDSET, , , , , , , 23456
CELAS2, 1, 50000., 1, 1, 2, 1
CELAS2, 2, 12500., 2, 1, 3, 1
CONM2, 201, 1, , 3.0
CONM2, 202, 2, , 1.5
SPC, 100, 3, 1
$
$ DEFINE DAMPER OF 30 BETWEEN GRIDS 2 AND 3
$
CVISC, 101, 1, 2, 3
PVISC, 1, 30.
$
ENDDATA

1 TWO-DOF MODEL (IMAC 8, PG 891) MARCH 13, 2007 MD NASTRAN 3/ 1/07 PAGE 9
COMPLEX MODES
0
C O M P L E X E I G E N V A L U E S U M M A R Y
0 ROOT EXTRACTION EIGENVALUE FREQUENCY DAMPING
NO. ORDER (REAL) (IMAG) (CYCLES) COEFFICIENT
1 2 -2.660969E+00 -4.983521E+01 7.931520E+00 1.067907E-01
2 1 -2.660969E+00 4.983521E+01 7.931520E+00 1.067907E-01
3 4 -7.339031E+00 -2.360312E+02 3.756553E+01 6.218695E-02
4 3 -7.339031E+00 2.360312E+02 3.756553E+01 6.218695E-02
COMPLEX MODES
0
COMPLEX MODES
0
COMPLEX EIGENVALUE = -2.660969E+00, -4.983521E+01
C O M P L E X E I G E N V E C T O R NO. 1
(REAL/IMAGINARY)

0 1 G 1.000000E+00 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
0 2 G 8.514119E-01 0.0 0.0 0.0 0.0 0.0
1.591320E-02 0.0 0.0 0.0 0.0 0.0
0 3 G 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
COMPLEX MODES
0
COMPLEX EIGENVALUE = -2.660969E+00, 4.983521E+01
(REAL/IMAGINARY)

0 1 G 1.000000E+00 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
0 2 G 8.514119E-01 0.0 0.0 0.0 0.0 0.0
-1.591320E-02 0.0 0.0 0.0 0.0 0.0
0 3 G 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0

(Cont.)
1 TWO-DOF MODEL (IMAC 8, PG 891) MARCH 13, 2007 MD NASTRAN 3/ 1/07 PAGE
13
COMPLEX MODES
0
COMPLEX EIGENVALUE = -7.339031E+00, -2.360312E+02
(REAL/IMAGINARY)

0 1 G -4.241094E-01 0.0 0.0 0.0 0.0 0.0
-3.768431E-02 0.0 0.0 0.0 0.0 0.0
0 2 G 1.000000E+00 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
0 3 G 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
1 TWO-DOF MODEL (IMAC 8, PG 891) MARCH 13, 2007 MD NASTRAN 3/ 1/07 PAGE
14
COMPLEX MODES
0
COMPLEX EIGENVALUE = -7.339031E+00, 2.360312E+02
(REAL/IMAGINARY)

0 1 G -4.241094E-01 0.0 0.0 0.0 0.0 0.0
3.768431E-02 0.0 0.0 0.0 0.0 0.0
0 2 G 1.000000E+00 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
0 3 G 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0

Section 16
Normal Mode Analysis Using Parts
Superelement

WHAT IS A SUPERELEMENT?
● Physical and mathematical representation
● Physical - substructure: a finite element model of a portion of a
structure
● Mathematical - boundary matrices: loads, mass, damping, and
stiffness reduced from the interior points to the exterior or
boundary points
● Other types of substructuring analysis
● Cyclic symmetry analysis
● GENEL and DMIG input
● There are two ways to define superelements
● PARTS superelement (discussed in this section)
● Main Bulk Data superelement (See Superelement seminar)

HOW ARE PARTS SUPERELEMENT DEFINED
IN MD NASTRAN
● The description for each superelement is completely self-contained
within its own Bulk Data Section.
● This type of superelement is often referred to as parts in this section.
● Each superelement begins with a
BEGIN [BULK] SUPER=m
delimiter and terminates with another
BEGIN [BULK] SUPER=n
or ENDATA delimiter
● The Main Bulk Data Section contains the full description of the
residual structure and other superelements not described by parts.
These superelements can be defined by either the SESET or SEELT
entries. See Superelement Seminar for further description of SESET.

HOW ARE PARTS SUPERELEMENT DEFINED
IN MD NASTRAN (Cont.)
● By default, if the location of a grid point in one part falls
within a specified tolerance of a grid point in another part,
then these parts will be connected together at that
location.
● You can manually overwrite the connection tolerance.
● Duplicate IDs (e.g., grid points, elements, properties, etc.)
are allowed across different parts.
● Each part can have its own loadings and component
modes calculation.
● Plotting of complete system modes is supported.
● A superelement may also be defined as a copy of a
superelement or obtained from outside MD Nastran.

SAMPLE PROBLEM - STEEL STAMPING
4” 4”
93 94 95 96 97 98 99 100 101 102 103 104
81 38 39 40 41 42 86 87 63 64 65 66 67 92
33 34 35 36 37 58 59 60 61 62
69 74 75 80
4” 28 29 30 31 32 53 54 55 56 57
57 23 24 25 26 27 62 63 48 49 50 51 52 68
18 19 20 21 22 43 44 45 46 47
33 34 35 36 37 38 39 40 41 42 43 44
15 17
1.6” 29 30 31 32
19 14 20 21 22 23 24 25 26 27 16
28
.8” 6 7 8 9 5 10 11 12 13
9 10 11 12 13 14 15 16 17 18
4 1.6”
1.6” 7 8
.8”
3
3.6” 5 6
2
3 4
1
3.2” 3.2”
.8”
(Cont.)
● Grid Points 1 and 2 fixed
● Material properties:
Steel t = .05”
E = 29 x 106 psi
n = .3
r = .283 lb/in3 (weight density)
● Applied loads
● 1 psi pressure on square portions
● Normal force of 2 lb on Grids 93 and 104
● Opposing normal force of 2 lb on Grids 93 and 104

MODEL DEFINITION FOR SAMPLE PROBLEM
$
$ file - fs1.dat
$
$ all 7 s.e. brought in using begin super
$ duplicate boundary grids id
$ each s.e. contains its own property description with
the same id $
$ condensed subcase setup include ’plot.blk’
$ $
$ ------------------------------------------------------- BEGIN BULK
$ $
id allsep1 dat
CQUAD4 5 1 13 14 24 23
SOL 101
TIME 15 $
CEND GRDSET 6
TITLE = S.E. SAMPLE PROBLEM 1 GRID 13 -.4 3.6 0.
SUBTITLE = S.E. STATICS - RUN 1 - MULTIPLE LOADS GRID 14 .4 3.6 0.
DISP = ALL GRID 23 -.4 4.4 0.
stress = all GRID 24 .4 4.4 0.
oload = all $
SET 999 = 0,1,2,3,4,5,6,7
include ’prop1.blk’
SUPER = 999 $ ALL CASE CONTROL IS FOR ALL SUPERELEMENTS
$
PARAM,GRDPNT,1
SUBCASE 101 begin super=1
LABEL = PRESSURE LOAD $
LOAD = 101 include ’loadse1.blk’
$ include ’prop1.blk’
SUBCASE 201 include ’se1.blk’
LABEL = 2# NORMAL LOADS
LOAD = 201
$
SUBCASE 301
LABEL = OPPOSING LOADS
LOAD = 301
$

$
begin super=2
$
(Cont.)
include ’loadse2.blk’
include ’se2.blk’
$
begin super=3
$
$
begin super=4
$
$
begin super=5
$
$
begin super=6
$
$
begin super=7
$
$
enddata

(Cont.)
$
$ plot.blk
$
output(plot) seplot 5
$ ptitle = superelement 5
set 1 = all find scale, origin 1, set 1
axes z,x,y plot static deformation set 1 origin 1 label both
view 0.,0.,0. $
seupplot 0 seplot 6
ptitle = full structure ptitle = superelement 6
find scale, origin 1, set 1 find scale, origin 1, set 1
plot set 1 origin 1 label both plot static deformation set 1 origin 1 label both
$ $
$ deform plots seplot 7
$ ptitle = superelement 7
seplot 1 find scale, origin 1, set 1
ptitle = superelement 1 plot static deformation set 1 origin 1 label both
find scale, origin 1, set 1
plot static deformation set 1 origin 1 label both $
$ seplot 0
seplot 2 ptitle = superelement 0
ptitle = superelement 2 find scale, origin 1, set 1
find scale, origin 1, set 1 plot static deformation set 1 origin 1 label both
plot static deformation set 1 origin 1 label both $
$ seupplot 0
seplot 3 ptitle = full structure
axes x,mz,y
ptitle = superelement 3
find scale, origin 1, set 1 find scale, origin 1, set 1
plot static deformation set 1 origin 1 label both plot static deformation set 1 origin 1
$ $
seplot 4
ptitle = superelement 4
find scale, origin 1, set 1
plot static deformation set 1 origin 1 label both
$

(Cont.)
$ GRID 33 -5.2 6. 0.
$ se1.blk GRID 34 -4.4 6. 0.
$ GRID 37 -2. 6. 0.
CQUAD4 18 1 33 34 46 45 GRID 38 -1.2 6. 0.
CQUAD4 19 1 34 35 47 46 GRID 45 -5.2 6.8 0.
CQUAD4 20 1 35 36 48 47 GRID 46 -4.4 6.8 0.
CQUAD4 21 1 36 37 49 48 GRID 47 -3.6 6.8 0.
CQUAD4 22 1 37 38 50 49 GRID 48 -2.8 6.8 0.
CQUAD4 23 1 45 46 58 57 GRID 49 -2. 6.8 0.
CQUAD4 24 1 46 47 59 58 GRID 50 -1.2 6.8 0.
CQUAD4 25 1 47 48 60 59 GRID 57 -5.2 7.6 0.
CQUAD4 26 1 48 49 61 60 GRID 58 -4.4 7.6 0.
CQUAD4 27 1 49 50 62 61 GRID 59 -3.6 7.6 0.
CQUAD4 28 1 57 58 70 69 GRID 60 -2.8 7.6 0.
CQUAD4 29 1 58 59 71 70 GRID 61 -2. 7.6 0.
CQUAD4 30 1 59 60 72 71 GRID 62 -1.2 7.6 0.
CQUAD4 31 1 60 61 73 72 GRID 69 -5.2 8.4 0.
CQUAD4 32 1 61 62 74 73
CQUAD4 33 1 69 70 82 81 GRID 70 -4.4 8.4 0.
CQUAD4 34 1 70 71 83 82 GRID 71 -3.6 8.4 0.
CQUAD4 35 1 71 72 84 83 GRID 72 -2.8 8.4 0.
CQUAD4 36 1 72 73 85 84 GRID 73 -2. 8.4 0.
CQUAD4 37 1 73 74 86 85 GRID 74 -1.2 8.4 0.
CQUAD4 38 1 81 82 94 93 GRID 81 -5.2 9.2 0.
CQUAD4 39 1 82 83 95 94 GRID 82 -4.4 9.2 0.
CQUAD4 40 1 83 84 96 95 GRID 83 -3.6 9.2 0.
CQUAD4 41 1 84 85 97 96 GRID 84 -2.8 9.2 0.
CQUAD4 42 1 85 86 98 97 GRID 85 -2. 9.2 0.
$ GRID 86 -1.2 9.2 0.
GRDSET 6 GRID 93 -5.2 10. 0.
$ GRID 94 -4.4 10. 0.
6 GRID 95 -3.6 10. 0.
$ boundary grids GRID 96 -2.8 10. 0.
$ GRID 97 -2. 10. 0.
GRID 35 -3.6 6. 0.
GRID 98 -1.2 10. 0.
GRID 36 -2.8 6. 0.
$

(Cont.)
$
$ GRID 39 1.2 6. 0.
$ se2.blk GRID 40 2. 6. 0.
$ GRID 43 4.4 6. 0.
CQUAD4 43 1 39 40 52 51 GRID 44 5.2 6. 0.
CQUAD4 44 1 40 41 53 52 $
CQUAD4 45 1 41 42 54 53 GRID 51 1.2 6.8 0.
CQUAD4 46 1 42 43 55 54 GRID 52 2. 6.8 0.
CQUAD4 47 1 43 44 56 55 GRID 53 2.8 6.8 0.
CQUAD4 48 1 51 52 64 63 GRID 54 3.6 6.8 0.
CQUAD4 49 1 52 53 65 64 GRID 55 4.4 6.8 0.
CQUAD4 50 1 53 54 66 65 GRID 56 5.2 6.8 0.
CQUAD4 51 1 54 55 67 66 GRID 63 1.2 7.6 0.
CQUAD4 52 1 55 56 68 67 GRID 64 2. 7.6 0.
CQUAD4 53 1 63 64 76 75 GRID 65 2.8 7.6 0.
CQUAD4 54 1 64 65 77 76 GRID 66 3.6 7.6 0.
CQUAD4 55 1 65 66 78 77 GRID 67 4.4 7.6 0.
CQUAD4 56 1 66 67 79 78 GRID 68 5.2 7.6 0.
CQUAD4 57 1 67 68 80 79 GRID 75 1.2 8.4 0.
CQUAD4 58 1 75 76 88 87
CQUAD4 59 1 76 77 89 88 GRID 76 2. 8.4 0.
CQUAD4 60 1 77 78 90 89 GRID 77 2.8 8.4 0.
CQUAD4 61 1 78 79 91 90 GRID 78 3.6 8.4 0.
CQUAD4 62 1 79 80 92 91 GRID 79 4.4 8.4 0.
CQUAD4 63 1 87 88 100 99 GRID 80 5.2 8.4 0.
CQUAD4 64 1 88 89 101 100 GRID 87 1.2 9.2 0.
CQUAD4 65 1 89 90 102 101 GRID 88 2. 9.2 0.
CQUAD4 66 1 90 91 103 102 GRID 89 2.8 9.2 0.
CQUAD4 67 1 91 92 104 103 GRID 90 3.6 9.2 0.
$ GRID 91 4.4 9.2 0.
GRDSET 6 GRID 92 5.2 9.2 0.
$ GRID 99 1.2 10. 0.
$ boundary grids GRID 100 2. 10. 0.
$ GRID 101 2.8 10. 0.
GRID 41 2.8 6. 0. GRID 102 3.6 10. 0.
GRID 42 3.6 6. 0.
GRID 103 4.4 10. 0.
$ GRID 104 5.2 10. 0.
$
(Cont.)
$
$ se3.blk
$
CQUAD4 14 1 19 20 30 29
CQUAD4 15 1 29 30 36 35
$
GRDSET 6
$
$ boundary grids
$
GRID 19 -3.6 4.4 0.
GRID 20 -2.8 4.4 0.
GRID 35 -3.6 6. 0.
GRID 36 -2.8 6. 0.
$
GRID 29 -3.6 5.2 0.
GRID 30 -2.8 5.2 0.
$

(Cont.)
$
$ se4.blk
$
CQUAD4 16 1 27 28 $
32 31 $ se5.blk
CQUAD4 17 1 31 32 $
42 41 CQUAD4 6 1 9 10 20 19
$ CQUAD4 7 1 10 11 21 20
GRDSET CQUAD4 8 1 11 12 22 21
6 CQUAD4 9 1 12 13 23 22
$ $
$ boundary grids GRDSET 6
$ $
GRID 27 2.8 4.4 $ boundary grids
0. $
GRID 28 3.6 4.4 GRID 19 -3.6 4.4 0.
0. GRID 20 -2.8 4.4 0.
GRID 41 2.8 6. GRID 13 -.4 3.6 0.
0. GRID 23 -.4 4.4 0.
GRID 42 3.6 6. $
0. GRID 9 -3.6 3.6 0.
$ GRID 10 -2.8 3.6 0.
GRID 31 2.8 5.2 GRID 11 -2. 3.6 0.
0. GRID 12 -1.2 3.6 0.
GRID 32 3.6 5.2 GRID 21 -2. 4.4 0.
0. GRID 22 -1.2 4.4 0.
$ $

(Cont.)
$
$ se6.blk
$ $
CQUAD4 10 1 14 15 25 24 $ se7.blk
CQUAD4 11 1 15 16 26 25 $
CQUAD4 12 1 16 17 27 26 CQUAD4 1 1 1 2 4 3
CQUAD4 13 1 17 18 28 27 CQUAD4 2 1 3 4 6 5
$ CQUAD4 3 1 5 6 8 7
GRDSET CQUAD4 4 1 7 8 14 13
6 $
$ GRDSET
$ boundary grids 6
$ $
GRID 27 2.8 4.4 0. GRID 1 -.4 0. 0.
GRID 28 3.6 4.4 0. 123456
GRID 14 .4 3.6 0. GRID 2 .4 0. 0.
GRID 24 .4 4.4 0. 123456
$ GRID 3 -.4 0.9 0.
GRID 15 1.2 3.6 0. GRID 4 .4 0.9 0.
GRID 16 2. 3.6 0. GRID 5 -.4 1.8 0.
GRID 17 2.8 3.6 0. GRID 6 .4 1.8 0.
GRID 18 3.6 3.6 0. GRID 7 -.4 2.7 0.
$ GRID 8 .4 2.7 0.
GRID 25 1.2 4.4 0. $
GRID 26 2. 4.4 0. $ boundary grids
$ $
GRID 13 -.4 3.6 0.
GRID 14 .4 3.6 0.
$

(Cont.)
$
$ prop1.blk
$
MAT1,1,30.+6,,.3,.283
PARAM,WTMASS,.00259
PARAM,AUTOSPC,YES
PSHELL,1,1,.05,1,,1 $
$ $ file - loadse2.blk
$ loads on s.e. 2
$
$ LOAD CASE 1 - PRESSURE LOAD
$
PLOAD2,101,-1.,43,THRU,67
$
$ $ LOAD CASE 2 - 2 POINT LOADS AT
$ file - loadse1.blk CORNERS
$ loads on s.e. 1 $
$ FORCE,201,104,,2.,0.,0.,1.
$ LOAD CASE 1 - PRESSURE LOAD $
$ $ LOAD CASE 3 - OPPOSING POINT LOADS
PLOAD2,101,-1.,18,THRU,42 AT CORNERS
$ $
$ LOAD CASE 2 - 2 POINT LOADS AT CORNERS FORCE,301,104,,2.,0.,0.,-1.
$ $
FORCE,201,93,,2.,0.,0.,1.
$
$ LOAD CASE 3 - OPPOSING POINT LOADS AT
CORNERS
$
FORCE,301,93,,2.,0.,0.,1.
$

SAMPLE SUPERELEMENT 1
Part 1
4” 4”
93 94 95 96 97 98 99 100 101 102 103 104
38 39 40 41 42 63 64 65 66 67
81 86 87 92
33 34 35 36 37 58 59 60 61 62
69 74 75 80
4” 28 29 30 31 32 53 54 55 56 57
57 62 63 68
23 24 25 26 27 48 49 50 51 52
18 19 20 21 22 43 44 45 46 47
33 34 35 36 37 38 39 40 41 42 43 44
15 17
1.6” 29 30 31 32
19 14 20 21 22 23 24 25 26 27 16
28
.8” 7 8 9 5 10 11 12
9 10 11 12 13 14 15 16 17 18
4 1.6”
1.6” 7 8
.8”
3
3.6” 5 6
2
3 4
3.2” 3.2”
.8”

PARTITIONED SOLUTIONS
● For each superelement, its degrees-of-freedom (DOFs)
are divided into two subsets:
● Exterior DOFs (called the A-set): Designates the analysis DOFs,

which are retained for subsequent processing (for Superelement
1, Grid Points 35 and 36)
● Interior DOFs: Designates the DOFs that are reduced out during
superelement processing and are omitted in subsequent
processing (for Superelement 1 of the sample problem, Grid
Points 33, 34, 37, 38, 45-50, 57-62, 69-74, 81-86, 93-98).

PARTITIONED SOLUTIONS (Cont.)
● For each superelement, produce a description in matrix terms of its
behavior as seen at the boundary or exterior degrees of freedom.
● A set of ‘G’-sized matrices is produced for each superelement based on
the input data.
● These matrices are reduced down to matrices representing the properties of
the superelement as seen by the adjacent (attached) structure.
● At the residual structure, combine and assemble the boundary
matrices.
● The BULK DATA for the RESIDUAL structure consists of all ‘residual’
data not assigned to any superelement plus any common data.
● Solve for the residual structure displacements.
● For each superelement, expand boundary (exterior) displacements to
obtain its interior displacements.

THEORY OF STATIC CONDENSATION
After generating matrices and applying MPCs and SPCs,
Kff Uf = Pf
O-Set = interior points (to be condensed out by

the reduction)
A-Set = exterior (or boundary) points (which are
retained for further analysis)
Partition
K oo K oa 
U o



P o


 =  
K oa K aa U a
T  P a 
   

THEORY OF STATIC CONDENSATION
● Extract upper equation and premultiply by k 1

oo
oo 
k k oo U o k oa U a  1
1
k oo Po
1
Let G oa k oo k oa (Boundary Transformation)
and U oo k 1
oo Po (Fixed Boundary Displacements)
then U o U oo G oa U a (Total Interior Displacements)

THEORY OF STATIC CONDENSATION (Cont.)
● Substitute expression for Uo in the lower equation
T
K oa G o aU a + Uoo + K a aU a = P a
T G
K aa = K oa
then oa + K aa (Boundary Stiffness)
and Pa = GT
oa P o + P a (Boundary Loads)
● Solve for residual structure
Ua = Kaa-1 Pa (Boundary Displacments)

ADVANTAGES OF SUPERELEMENT
ANALYSIS
● Large problems (i.e., allows solving problems that exceed
your hardware capabilities)
● Less CPU or wall clock time per run (reduced risk since
each superelement may be processed individually)
● Partial redesign requires only partial solution (cost).
● Allows more control of resource usage
● Partitioned input desirable
● Organization
● Repeated components

ADVANTAGES OF SUPERELEMENT
ANALYSIS (Cont.)
● Partitioned output desirable
● Organization
● Comprehension
● Components may be modeled by subcontractors.
● Multi-step reduction for dynamic analysis
● Zooming (or global-local analysis)
● Allows for efficient configuration studies (“What if...”)

DISADVANTAGES OF SUPERELEMENT
ANALYSIS
● Increased overhead due to DMAP compilation and
database manipulation and storage
● Mandatory static condensation may cancel other cost
savings for small models.
● Residual structure is not resequenced and its stiffness
matrix is usually dense.
● All superelements must be linear.
● Approximations must be made in dynamics for mass and
damping through static, component mode, or generalized
dynamic reduction.
● Automatic restarts are available in SOLs 101 and above.

CONVENTIONAL ANALYSIS
 Flowchart
Generation
Solution

CONVENTIONAL ANALYSIS (Cont.)
P2 = 1 P3 = 2 P4 = 3
1 2 3 4 5
K12 K 23 K 34 K 45
Kxx = unit stiffness
 Generation
K12 –K 12 0 0 0
– K12 K 12 + K 23 – K23 0 0
K GG = 0 –K 23 K23 + K 34 – K34 0
0 0 – K 34 K34 + K45 –K 45
0 0 0 – K45 K 45
1.0 –1.0 0 0 0
– 1.0 2.0 – 1.0 0 0
K GG = 0 – 1.0 2.0 –1.0 0
0 0 – 1.0 2.0 – 1.0
0 0 0 – 1.0 1.0
CONVENTIONAL ANALYSIS (Cont.)
 Apply Constraints and Solve
  –1  
U 2  K12 + K 23 – K23 P 2 
   
U 3 = –K 23 K23 + K34 –K 34 P 3 
   
U 4  – K34 K 34 + K 45 P 4 
   
 
U2  2.0 – 1.0 0
–1 
1.0 
   
U3 = –1.0 2.0 –1.0 2.0 
  3.0 
U4  0 – 1.0 2.0  
 
 
U 2  2.5 
   
U 3 = 4.0 
   
U 4  3.5 
 

SUPERELEMENT ANALYSIS
 Flowchart
DO LABELA
Phase I
Generation
Assembly
LABELA
Phase II
Solution
DO LABELB
Phase III
LABELB

SUPERELEMENT ANALYSIS (Cont.)
SEID = 1 SEID = 2
1 2 3 4 5
Residual Structure
P2 = 1
 Generation - SEID = 1 1 2 3
K12 K 23
u2 u3
K12 –K 12 0
1
K gg = –K 12 K12 + K23 – K23
0 – K23 K 23

 
 P1 
1   0 
     
P g  =  P2 = 1 
     
  0 
 P31 
 
 Reduction - SEID = 1
 Eliminate constraints:
1 K +K – K23
Kgg = 12 23
– K23 K23
Koo K oa
 Compute boundary transformation: =
Kao Kaa
1 –
Goa = – Koo 1 Koa
K23
= --------------------------
- = 0.5
K12 + K23
 Compute boundary stiffness:
1
K aa = Kaa + K T G
o a oa
K12K
1 23
Kaa = --------------------------
- = 0.5
K12 + K 23
 Compute boundary loading:
1  
  P2  1 
P f  =  =  
  P 13  0 
 
 
P 
=  o 
P a 
 
1
   
Pa  = P a + Goa
T P 
o
   
0 K23
P13 = P 31 + ---------------------------P2 = 0.5
K12 + K23
 Similarly - SEID = 2
K34 –K 34 0
2
K gg = –K 34 K 34 + K 45 –K 45
0 – K45 K 45
 
2 P 23  0 
     
P g  = P 4 = 3 
     
P 5  0 
 
2 K34
Goa = --------------------------- = 0.5
K 34 + K 45
2 K 34 K 45
Ka a = --------------------------- = 0.5
K 34 + K 45
0
K 34
P 23 = P 32 + --------------------------- P 2 = 1.5
K34 + K 45
 Residual Structure
P32
P31 P
3 3
1
K K2 K
P30
 Assembly
0
Kaa = K 1 + K 2 + K 0
aa aa gg
K = K 1 + K2 = 1
   
Pa = P 1a + P a2 + Pg0 
   
 Solution
P = P31 + P 23 + P 30 = 4
  –1  
Ua = K aa P a 
   
P
U 03 = ---- = 4
K
● Data Recovery - SEID = 1
● Enforce (transform) boundary motion.
 a   
U o = G oa U a 
   
K 23 1 2 3
U 32 = --------------------------
- U = 2.0
K 12 + K 23 3
U 23 = ? U3 = 4
● Compute fixed-boundary motion.

 o  –1  
 Uo = Koo  Po 
   
1 1 2 3
Uo2 = --------------------------
-P = 0.5
K 12 + K 23 2
U 02 = ?
● Compute total motion.

   
Uo = U oo + U oa 
   
1 2 3
K 23 U3 + P 2
U 2 = ------------------------------------ = 2.5
K12 + K 23 U2 = 2.5 U3 = 4

SUPERPOSITION OF PARTIAL SOLUTIONS
Applied Loads
A. Assembled Structure
Superelement 1 Superelement 2 Superelement 3
Residual Structure
B. Partitioned Structure
C. Partial Solution Due to External Loads
D. Partial Solution Due to Boundary Motion
E. Assembled Solution

SUPERELEMENT REDUCTION METHODS
AVAILABLE IN DYNAMIC ANALYSIS
● Static reduction
● Static condensation of stiffness and Guyan reduction of mass
● Automatically done
● Dynamic reduction
● Component modal synthesis (CMS)
● Analytical (SOL 103)

DEGREES OF REDUCTION
● Static reduction (default)
● Interior masses relumped to boundary (Guyan)
● Rigid body properties preserved
● Important masses must be made exterior (boundary)
● Component mode reduction - in addition to static reduction

● Interior masses represented by exact eigenvectors of the
component
● Eigensolutions for each superelement may be output

GUYAN OR STATIC REDUCTION
● Based on stiffness
● Perform static condensation on stiffness
K oo K oa 
U o



Po


 =  
KT K aa U a  P a 
oa    
If Po=0, then {uo} = [Goa ] {ua}, where Goa = -[Koo-1] [Koa]
Use this transformation to go from the F-set

T
to the A-set
Kaa = K + Koa G oa
aa
T T T
M aa = M aa + Moa Goa + G oa M oa + G oa M oo G oa
However, U oo (internal dynamic effects) is ignored
● No approximation if no masses, damping, or applied loads are specified on O-
set (interior DOFs)
● Good approximation if component frequencies are much higher than residual
structure frequencies and the excitation frequencies

COMPARISON OF REDUCTION METHODS
X X X X X X X
X X
uo Points
X X X
X X X X
X X ut Points
X X X X X X X
● Static reduction Local dynamic

0
      effects are ignored.
Uo = Got U t + Uoo 
     
● Generalized dynamic reduction
 
   Ut 
 Uo  = Got G oq  
  U q 
 
Approximate eigenvectors are used to represent the interior motion.
● Component mode reduction
 
   Ut 
 Uo = Got Goq  
   Uq 
 
● Exact eigenvectors are used to represent the interior motion.

ADVANTAGES OF EACH REDUCTION METHOD
● Advantages of Component Mode Reduction over Static

Reduction
● Can use experimental results
● More accurate for the same number of dynamic DOFs
● Ideal for highly coupled and uncoupled structures
● Advantages of Static Reduction over Component Mode

Reduction
● Cheaper
● Less sophisticated
● Fewer problems

CALCULATION OF NORMAL MODES USING
STATIC REDUCTION ONLY
● No generalized coordinates are needed for the superelements.
(exception: residual structure, if component reduction at the residual
structure Is desired)
● Superelement mass, damping, and stiffness are reduced statically to
exterior DOFs.
● ASETi and QSETi entries may be specified for residual structure
DOFs only.
● If no ASETi entries are present, then all DOFs interior to the residual
structure are retained for the eigensolution.
● If ASETi entries are present, then only those DOFs specified on the
ASETi entries are retained for the eigensolution.
● Case Control is similar to static analysis with the addition of a
METHOD command under the residual structure subcase.

DYNAMIC REDUCTION FOR SUPERELEMENTS
● The behavior of a superelement is represented by its component modes in addition to
the static shapes.
● The superelement stiffness, mass, and damping are transformed using both physical
and modal variables.
● The superelement modes are computed if a METHOD command appears under the
superelement subcase.
● The number of superelement modes computed (modal truncation) is controlled by the
EIGR or EIGRL entry.
● The number of superelement modes sent downstream directly to the residual is done
one of three ways:
● Param,autoqset,yes
● SENQSET DOFs
● QSETi and SPOINT DOFs.
● If the superelement modes are to be sent downstream to another superelement, QSETi
and SPOINTs must be used. These SPOINTs must also be “connected” to the
SPOINTs in the downstream superelement using the SECONCT entry.

(Cont.)
● By default, superelement modes are computed with all
exterior degrees of freedom fixed (in the B-set). This is
better known as the Craig-Bampton method.
● Superelement modes are computed in Phase I under the
SEMR operation.
● Copied superelements need to have the same number of
exterior DOFs as their primary. If the primary has an
SENQSET, then the image must have equivalent DOFs to
represent the modes.

Superelement Internal Generalized Degree of Freedom SENQSET
Defines number of internally generated scalar points for superelement dynamic reduction.
Format:
1 2 3 4 5 6 7 8 9 10
SENQSET SEID N
Example:
SENQSET 110 45
Field Contents
SEID Superelement identification number. See Remark 3.. (Integeru0 or Character=“ALL”)
Number of internally generated scalar points for dynamic reduction generalized
N coordinates. (Integer>0; Default=0)
Remarks:
1. SENQSET can only be specified in the main Bulk Data Section and is ignored after the BEGIN
SUPER=n command.
2. SENQSET is only required if the user wants to internally generated scalar points used for dynamic
reduction.
3. SEID=“ALL” will automatically generate N q-set degrees of freedom for all superelements, except
the residual structure (SEID=0). Specifying additional SENQSET entries for specific
superelements will override the value of N specified on this entry.
4. If the user manually specifies q-set degrees of freedom using a SEQSETi or QSETi entries, then
the internally generated scalar points will not be generated.LQ
Bulk Data Entry

Scalar Point Definition SPOINT
Defines scalar points.
Format:
1 2 3 4 5 6 7 8 9 10
SPOINT ID1 ID2 ID3 ID4 ID5 ID6 ID7 ID8
Example:
SPOINT 3 18 1 4 16 2
Alternate Format and Example:

SPOINT ID1 “THRU” ID2
SPOINT 5 THRU 649
Field Contents
IDi Scalar point identification number. (0<Integer<1000000; For “THRU” option, ID1<ID2)
Remarks:
1. A scalar point defined by its appearance on the connection entry for a scalar element (see the
CELASi, CMASSi, and CDAMPi entries) need not appear on an SPOINT entry.
2. All scalar point identification numbers must be unique with respect to all other structural, scalar,
and fluid points. However, duplicate scalar point identification numbers are allowed in the input.
3. This entry is used primarily to define scalar points appearing in single-point or multipoint constraint
equations to which no scalar elements are connected.
4. If the alternate format is used, all scalar points ID1 through ID2 are defined.
5. For a discussion of scalar points, see the MSC.Nastran Reference Manual , Section 5.6.
Bulk Data Entry

Generalized Degree of Freedom QSET
Defines generalized degrees of freedom (q-set) to be used for dynamic reduction or component mode synthesis.
Format:
1 2 3 4 5 6 7 8 9 10
QSET ID1 C1 ID2 C2 ID3 C3 ID4 C4
Example:
QSET 15 123456 1 7 9 2 105 6
Field Contents
IDi Grid or scalar point identification number. (Integer>0)
Component number. (Integer zero or blank for scalar points or any unique combination
Ci of the Integers 1 through 6 for grid points with no embedded blanks.)
Remarks:
1. Degrees of freedom specified on this entry form members of the mutually exclusive q-set. They
may not be specified on other entries that define mutually exclusive sets. See the MSC.Nastran
Quick Reference Guide, Appendix B for a list of these entries.
2. Degrees of freedom specified on QSET and QSET1 entries are automatically placed in the a-set.
3. When ASET, ASET1, QSET, and/or QSET1 entries are present, all degrees of freedom not
otherwise constrained (e.g., SPCi or MPC entries) will be placed in the omitted set (o-set).
Bulk Data Entry

Generalized Degree of Freedom (Alternate Form of QSET Entry QSET1
Defines generalized degrees of freedom (q-set) to be used for generalized dynamic reduction or component
mode synthesis.
Format:
1 2 3 4 5 6 7 8 9 10
QSET1 C ID1 ID2 ID3 ID4 ID5 ID6 ID7
ID8 ID9 -etc.-
Example:
QSET1 123456 1 7 9 22 105 6 22
52 53

QSET1 C ID1 “THRU” ID2
QSET1 0 101 THRU 110
Field Contents
C Component number. (Integer zero or blank for scalar points or any unique combination
of the Integers 1 through 6 for grid points with no embedded blanks.)
IDi Grid or scalar point identification number. (Integer>0; For THRU option, ID1<ID2.)
Remarks:
1. Degrees of freedom specified on this entry form members of the mutually exclusive q-set. They
2. Degrees of freedom specified on QSET and QSET1 entries are automatically placed in the a-set.
3. When ASET, ASET1, QSET, and/or QSET1 entries are present, all degrees of freedom not
otherwise constrained (e.g., SPCi or MPC entries) will be placed in the omitted set (o-set).
Bulk Data Entry

(Cont.)
● For free-free superelement component modes, all exterior DOFs should be
specified on CSETi entries.
● The rigid-body modes (f=0.0 Hz) are a linear combination of the static vectors and
should not be included in the reduction.
● Either:
● Do not calculate them (F1>0.0 on the EIGR or EIGRL entry).
● Calculate them and hope that the program will remove them (see PARAM,EPSRC in the
MD Nastran Reference Manual).
● Mixed-boundary modes may be calculated by using the CSETi and BSETi
entries to describe the exterior DOFs to be unconstrained and constrained
during CMS.
● If 0.0 Hz mixed boundary modes exist, they must be handled in a similar manner to
those in the free-free case.

Fixed Analysis Degrees of Freedom BSET
Defines analysis set (a-set) degrees of freedom to be fixed (b-set) during generalized dynamic reduction or
component mode synthesis calculations.
Format:
1 2 3 4 5 6 7 8 9 10
BSET ID1 C1 ID2 C2 ID3 C3 ID4 C4
Example:
BSET 2 135 14 6
Field Contents
Ci Component number. (Integer zero or blank for scalar points, or any unique
combinations of the Integers 1 through 6 for grid points. No embedded blanks.)
Remarks:
1. If there are no CSETi or BSETi entries present, all a-set points are considered fixed during
component mode analysis. If there are only BSETi entries present, any a-set degrees of freedom
not listed are placed in the free boundary set (c-set). If there
2. Degrees of freedom specified on this entry form members of the mutually exclusive b-set. They
3. If PARAM,AUTOSPC is YES, then singular b-set and c-set degrees of freedom will be reassigned
as follows:
a. If there are no o-set (omitted) degrees of freedom, then singular b-set and c-set
degrees of freedom are reassigned to the s-set.
b. If there are o-set (omitted) degrees of freedom, then singular c-set degrees of freedom
are reassigned to the b-set. Singular b-set degrees of freedom are not reassigned.
Bulk Data Entry

Fixed Analysis Degrees of Freedom, Alternate Form of BEST Entry BSET1
Defines analysis set (a-set) degrees of freedom to be fixed (b-set) during generalized dynamic reduction or
component mode synthesis calculations.
Format:
1 2 3 4 5 6 7 8 9 10
BSET1 C ID1 ID2 ID3 ID4 ID5 ID6 ID7
ID8 ID9 ID10 -etc.-
Example:
BSET1 2 135 14 6 23 24 25 26
122 127

BSET1 C ID1 “THRU” ID2
BSET1 3 6 THRU 32
Field Contents
C Component numbers. (Integer zero or blank for scalar points, or any unique
combinations of the Integers 1 through 6 for grid points with no embedded blanks.)
IDi Grid or scalar point identification numbers. (Integer>0; For “THRU” option, ID1<ID2)
Remarks:
1. If there are no CSETi or BSETi entries present, all a-set points are considered fixed during
component mode analysis. If there are only BSETi entries present, any a-set degrees of freedom
not listed are placed in the free boundary set (c-set). If there
2. Degrees of freedom specified on this entry form members of the mutually exclusive b-set. They
3. If PARAM,AUTOSPC is YES, then singular b-set and c-set degrees of freedom will be reassigned
as follows:
Bulk Data Entry

Free Boundry Degree of Freedom CSET
Defines analysis set (a-set) degrees of freedom to be free (c-set) during generalized dynamic reduction or
component modes calculations.
Format:
1 2 3 4 5 6 7 8 9 10
CSET ID1 C1 ID2 C2 ID3 C3 ID4 C4
Example:
CSET 124 1 5 23 6 16
Field Contents
Component numbers. (Integer zero or blank for scalar points, or any unique
Ci combination of the Integers 1 through 6 for grid points with no embedded blanks.)
Remarks:
1. If there are no CSETi or BSETi entries present, all a-set degrees of freedom are considered fixed
during component modes analysis. If there are only BSETi entries present, any a-set degrees of
freedom not listed are placed in the free boundary set (c-set
2. Degrees of freedom specified on this entry form members of the mutually exclusive c-set. They
3. If PARAM,AUTOSPC is YES then singular b-set and c-set degrees of freedom will be reassigned
as follows:
Bulk Data Entry

Free Boundry Degree of Freedom, Alternate Form of CSET Entry CSET1
Defines analysis set (a-set) degrees of freedom to be free (c-set) during generalized dynamic reduction or
component modes calculations.
Format:
1 2 3 4 5 6 7 8 9 10
CSET1 C ID1 ID2 ID3 ID4 ID5 ID6 ID7
ID8 ID9 -etc.-
Example:
CSET1 124 1 5 7 6 9 12 122
127
Alternate Formats and Examples:

CSET1 C ID1 “THRU” ID2
CSET1 3 6 THRU 32
CSET1 “ALL”
CSET1 ALL
Field Contents
C Component number. (Integer zero or blank for scalar points, or any unique combination
of the Integers 1 through 6 for grid points with no embedded blanks.)
IDi Grid or scalar point identification number. (Integer>0; For THRU option, ID1<ID2)
Remarks:
1. If there are no CSETi or BSETi entries present, all a-set degrees of freedom are considered fixed
during component modes analysis. If there are only BSETi entries present, any a-set degrees of
freedom not listed are placed in the free boundary set (c-set
(Continued)
Bulk Data Entry

Free Boundry Degree of Freedom, Alternate Form of CSET Entry CSET1
2. Degrees of freedom specified on this entry form members of the mutually exclusive c-set. They
3. If PARAM,AUTOSPC is YES then singular b-set and c-set degrees of freedom will be reassigned
as follows:
Bulk Data Entry

DEFAULT CMS METHOD “FIXED
BOUNDARY” CMS
 Description of Methodology (better known as Craig-Bampton CMS)
● The superelement matrices are partitioned into two sets of degrees of freedom (DOFs). The first set
(the B-set) represents the boundary points. The second set is the interior DOFs (the O-set).
● A set of “constraint” modes is generated. Each “constraint” mode represents the motion of the model
resulting from moving one boundary DOF 1.0 unit, while holding the other boundary DOF fixed.
Therefore, there is one “constraint” mode for each boundary DOF (these vectors are known as GOAT
in MD Nastran)
● In matrix form,
 
K oo K ob ob   0 
 =  
K bo K bb  I bb  P b 
 
 (Pb is not actually applied.)
● The first line gives
  –1  
ob = – Koo Kob Ibb  (GOAT )
   

BOUNDARY” CMS (Cont.)
● giving the following “constraint” modes:
 ob
b = -------
Ibb
● Now the O-set equations are solved for the “fixed-boundary” modes {
oo} (known as GOAQ in MD
Nastran).
   
–k2 M oo oo + Koo oo = 0
   
As many fixed-boundary modes as are desired are found. Then they are concatenated with the
“constraint” modes to form the generalized coordinates.
 
 = ob oo 
 G   
   Ibb 0 
 
● The mass and stiffness matrices are pre- and postmultiplied by these modes to obtain the
“generalized” mass and stiffness
KG = G 
T K  
ff G
M G = G 
T M  
ff G
where the F-set is the union of the B- and O-sets.

BOUNDARY” CMS (Cont.)
● These “generalized” matrices contain physical DOFs
representing the boundaries and “modal” coordinates
representing the “fixed-boundary” component modes.
● At this point, these matrices can be treated like any other
structural matrices, and data recovery can be performed
for the component in a manner similar to using modal
coordinates. That is, the displacements of the generalized
coordinates are multiplied by the associated vectors and
added together to obtain the component displacements.
● The calculated modes for each superelement are internally
scaled to have a maximum displacement = 1.0 in MD
Nastran (regardless of the scaling requested by the user).
SOLUTION BY HAND
 Component Modal Synthesis Sample
Part 2 Part 1
1 2 3 4 5
k1 k2 k3 k4
 k1 = k2 = k3 = k4 = 1.0
 m1 = m2 = m4 =m5 = 1.0 ; m3 = 1.0
Residual
Part 2 Part 1
 Theoretical solution for frequencies
1 2 3 3 3 4 5
i 1 2 3 4
fi 0.0553 0.1592 0.2438 0.2991
li = w
2
0.1206 1 2.3473 3.5321
SOLUTION BY HAND (Cont.)
● Superelement 1  
0 0 0  
1.0 –1.0 0 U3
 
K gg = – 1.0 2.0 – 1.0 Mgg = 0 1.0 0 U 
4
 
0 – 1.0 1.0 0 0 1.0 U 
5
 
● Grid Point 3 is the “boundary” point; solve for “constraint”
modes.  
1.0 –1.0 0  
1.0 

P
 b


– 1.0 2.0 –1.0  U 4 = 0 
   
0 –1.0 1.0  U 5  0 
 
Where Koo = 2.0 –1.0
– 1.0 1.0
– 1.0
Kob = 0
Koo
-1 =
1.0 1.0
1.0 2.0

1.0 1.0 
1.0  1 .0
● Where ob 
 

0.0  

1.0 2.0 
    1 .0

1 .0 
 
b 
1 .0 

1 .0 
 
● Solve for “fixed-boundary” modes.
● Note: Internally MD Nastran uses component modes scaled to a maximum deformation
of 1.0. Output for the component modes is based on the normalization performed by
the eigenvalue solution.
 
 2 
– 2 Moo + Koo oo = 0. = –  0 + 2.0 – 1.0  o o
 0 – 2 – 1.0 1.0 
 

det 2 –  – 1
2
= 0 2 = 0.3819, 2.618
– 1 1 – 2 f = 0.098 Hz, 0.2575 Hz
   0.5257 
1 = 0.6180  1 =   Normalized to
1.0000   0.8506  unit mass
– 1.0000  – 0.8506 
2 =  2 =  
 0.6180   0.5257 
oo = 0.6180 –1.0000

1.0000 0.6180
1.000 0 0
G = 1.000 0.618 – 1.000
1.000 1.000 0.618
T 0 0 0 u3
G K gg G = 0 0.5279 0 u1001
0 0 3.6180 u 1002
T 2.0000 1.6180 – 0.3820 u3

G M gg G = 1.6180 1.3820 0 u1001
–0.3820 0 1.3820 u1002
 where 1001 and 1002 are scalar points used to represent Superelement 1’s
modes.

 Superelement 2
1.0 – 1.0 0 1.0 0 0
Kgg = –1.0 2.0 –1.0 M gg = 0 1.0 0
0 – 1.0 1.0 0 0 0
• Apply constraint to grid point 1.
 0.5 

b =  
 1.0 
2 = 2.0 0.5 1.0


oo = 1  G =
f = 0.2251 1.0 0
T
 K gg G = 0.5 0 u3
G
0 2.0 u 1005
T u3
 M gg  = 0.25 0.50
G G u1005
0.50 1.00
 where 1005 is a scalar point used to represent Superelement 2’s

mode.
Residual Structure
● Before adding superelements:
0 0 0 0
Kgg = 0 0 0 0
0 0 0 0
0 0 0 0
1 0 0 0 u3
0 0 0 0 u 1001
M gg =
0 0 0 0 u 1002
0 0 0 0 u 1005

● Add Superelement 1
0 0 0 0
Kgg = 0 0.5279 0 0
0 0 3.618 0
0 0 0 0
3.0000 1.6180 – 0.3820 0 u3

u1001
M gg = 1.6180 1.3820 0 0
– 0.3820 0 1.3820 0 u1002
0 0 0 0 u1005

● Add Superelement 2
0.5 0 0 0
K gg = 0 0.5279 0 0
0 0 3.618 0
0 0 0 2.0
3.2500 1.6180 – 0.3820 0.5 u3

M gg = 1.6180 1.3820 0 0 u1001
– 0.3820 0 1.3820 0 u1002
0 0 0 1.0 u1005

0.5 0 0 0
K ff = 0 0.5279 0 0
0 0 3.618 0
0 0 0 2.0
3.2500 1.6180 –0.3820 0 u3

u 1001
Mff = 1.6180 1.3820 0 0
– 0.3820 0 1.3820 0 u 1002
0 0 0 1.0 u 1005
● Solve {KFF - 2Mff}{f} which gives  = 0.1206, 1.0000, 2.3473, 3.5321.

2
 
 0.42850 – 0.5773 – 0.2280 0.6565 
 
f =  0.23150 1.0937 0.3188 – 0.8619

–0.00572 0.0986 0.5464 0.7012 
 0.01370 – 0.2887 0.7705 –0.7568 
 
● Data recovery (grid point displacement for mode 1)
 u1
 Residual Structure =  0 
0.4285 u3

● Superelement 2
 0  u1
for exterior points 
2G
= 
0.4285

 u3
 u
0.0137  1005
1.0 0 0   0


 0

u
 1
21 = G2
 2 G = 0.5 0.5 1.0 0.4285 = 0.2280 u 2
   u
0 1.0 0 0.0137  0.4285  3
● Superelement 1
 0.4285 u
 1
for exterior points 1G=  0.2315 u1001
 u
–0.00572  1002
1.0 0 0  0.4285  0.4285 u3

   
11 = G11 G = 1.0 0.6180 –1.0  0.2315 = 0.5773 u
   4
1.0 1.0 0.6180 – 0.00572  0.6565 u5

SOLUTION USING MD NASTRAN SOL 103
 The following input data was created and run in MD Nastran:
$
$ sesp2.dat
$ BEGIN SUPER = 1
SOL 103 $
CEND EIGRL,1,,,2
TITLE = SAMPLE PROBLEM FOR CMS USING grid,3,,20.
PARTS grid,4,,30.
SPC = 1 grid,5,,40.
DISP = ALL $
PARAM,GRDPNT,0 CELAS2,3,1.,3,1,4,1
PARAM,USETPRT,0 CELAS2,4,1.,4,1,5,1
$ CONM2,14,4,,1.
SUBCASE 1
LABEL = CMS FOR PART 1 CONM2,15,5,,1.
SUPER = 1 $
METHOD=1 $ BEGIN SUPER = 2
$ $
SUBCASE 2 EIGRL,2,,,1
LABEL = CMS FOR PART 2 grid,1,,0.
SUPER = 2 grid,2,,10.
METHOD = 2 grid,3,,20.
$ $
SUBCASE 100 CELAS2,1,1.,1,1,2,1
LABEL = SYSTEM MODES CELAS2,2,1.,2,1,3,1
SUPER = 0 CONM2,11,1,,1.
METHOD = 100 CONM2,12,2,,1.
$ SPC1,1,123456,1
BEGIN BULK $
$ ENDDATA
grid,3,,20.
conm2,13,3,,1.0
$
EIGRL,100,,,4
param,autoqset,yes

OUTPUT FROM SPRING MODEL CMS RUN
1 SAMPLE PROBLEM FOR CMS USING PARTS MARCH 13, 2007 MD NASTRAN 3/ 1/07 PAGE 28
SUPERELEMENT 1
0 CMS FOR PART 1 SUBCASE 1
1 1 3.819660E-01 6.180340E-01 9.836316E-02 1.000000E+00 3.819660E-01
2 2 2.618034E+00 1.618034E+00 2.575181E-01 1.000000E+00 2.618034E+00

(Cont.)
1 SAMPLE PROBLEM FOR CMS USING PARTS MARCH 13, 2007 MD NASTRAN 3/ 1/07 PAGE
43
SUPERELEMENT 2
1 1 2.000000E+00 1.414214E+00 2.250791E-01 1.000000E+00 2.000000E+00

(Cont.)
SUPERELEMENT 0
0 SYSTEM MODES SUBCASE 100
1 1 1.206148E-01 3.472964E-01 5.527393E-02 1.000000E+00 1.206148E-01
2 2 1.000000E+00 1.000000E+00 1.591549E-01 1.000000E+00 1.000000E+00
3 3 2.347296E+00 1.532089E+00 2.438395E-01 1.000000E+00 2.347296E+00
4 4 3.532089E+00 1.879385E+00 2.991135E-01 1.000000E+00 3.532089E+00
SUPERELEMENT 0
SUPERELEMENT 0
*** USER INFORMATION MESSAGE 7321 (SEDRDR)
DATA RECOVERY FOR SUPERELEMENT 0 IS NOW INITIATED.
SUPERELEMENT 0
EIGENVALUE = 1.206148E-01
CYCLES = 5.527393E-02 R E A L E I G E N V E C T O R N O . 1

3 G 4.285251E-01 0.0 0.0 0.0 0.0 0.0

(Cont.)

3 G - 5.773503E-01 0.0 0.0 0.0 0.0 0.0
SUPERELEMENT 0

3 G 2.280134E-01 0.0 0.0 0.0 0.0 0.0
SUPERELEMENT 0

3 G - 6.565385E-01 0.0 0.0 0.0 0.0 0.0
SUPERELEMENT 0
SUPERELEMENT 1
SUPERELEMENT 1

3 G 4.285251E-01 0.0 0.0 0.0 0.0 0.0
4 G 5.773503E-01 0.0 0.0 0.0 0.0 0.0
5 G 6.565385E-01 0.0 0.0 0.0 0.0 0.0

(Cont.)
SUPERELEMENT 1

3 G -5.773503E-01 0.0 0.0 0.0 0.0 0.0
4 G -1.110223E-16 0.0 0.0 0.0 0.0 0.0
5 G 5.773503E-01 0.0 0.0 0.0 0.0 0.0
SUPERELEMENT 1

3 G 2.280134E-01 0.0 0.0 0.0 0.0 0.0
4 G 5.773503E-01 0.0 0.0 0.0 0.0 0.0
5 G -4.285251E-01 0.0 0.0 0.0 0.0 0.0
SUPERELEMENT 1

3 G -6.565385E-01 0.0 0.0 0.0 0.0 0.0
4 G 5.773503E-01 0.0 0.0 0.0 0.0 0.0
5 G -2.280134E-01 0.0 0.0 0.0 0.0 0.0
SUPERELEMENT 1
SUPERELEMENT 2

(Cont.)
SUPERELEMENT 2

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G 2.280134E-01 0.0 0.0 0.0 0.0 0.0
3 G 4.285251E-01 0.0 0.0 0.0 0.0 0.0
SUPERELEMENT 2

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G -5.773503E -01 0.0 0.0 0.0 0.0 0.0
3 G -5.773503E -01 0.0 0.0 0.0 0.0 0.0
SUPERELEMENT 2

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G -6.565385E -01 0.0 0.0 0.0 0.0 0.0
3 G 2.280134E-01 0.0 0.0 0.0 0.0 0.0
SUPERELEMENT 2

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G 4.285251E-01 0.0 0.0 0.0 0.0 0.0
3 G -6.565385E -01 0.0 0.0 0.0 0.0 0.0

Blank Page

SECTION 17
EXTRA POINTS, TRANSFER FUNCTIONS,
AND NOLINS

EXTRA POINTS
● Nonstructural DOFs are utilized to represent nonstructural variables.
● Defined by EPOINT Bulk Data entry
● Exist only in the residual structure for the dynamics solutions (E-set,
subset of D-set)
● Unaffected by any reduction procedure including modal reduction.
● Cannot be used as structural DOFs*
● Cannot be constrained by MPCs or SPCs*
● Can only be used in P-type direct input matrices or transfer functions
● Can only undergo dynamic loading with the use of the DAREA Bulk
Data entry

TRANSFER FUNCTIONS
● Transfer functions are used to define the dynamic functions of the
form: N
 2 2
b + b p+ b p u + 
 a 0 + a 1p + a2 p u = 0
0 1 2 d 
i i
i =1
where ud = dependent coordinate

ui = independent degree of freedom
p = differential operator (p = d/dt)
● Equivalent to P-type DMIG matrices (M2PP, B2PP, K2PP)
● Function is added to other P-type matrix input. ud defines the row with
b0, b 1, and b2 specifying the diagonal terms. ui defines the column for
the a0, a1, and a2 terms.
● Defined by the TF Bulk Data entry and selected by the TFL Case
Control command (residual structure only)

NONLINEAR TRANSIENT FORCE
● Grid point-oriented nonlinearity
·· ·
Mu t + Bu 
t + Ku 
t = P
t + N
t
● N(t) is the nonlinear force and is a function of the

displacement and the velocity.
● Defined on NOLINi Bulk Data entry that is selected by the

NONLINEAR Case Control command.
● Available only in transient response solution sequences

(must be in E-set for modal and D-set for direct
formulation).
NONLINEAR TRANSIENT FORCE
● Nonlinear force represents a deviation from linearity.
·· ·
Mu t + Bu 
t + Ku 
t = P
t + N 
t
Force Force Force
Displacement
= +
Displacement Displacement
Swaybrace = Elastic + Nonlinear
● Easiest to implement in direct solutions. Apply nonlinear

force directly to grid point. Modal solutions require transfer
functions and extra points since only E-set points may
have nonlinear loads in a modal formulation.
CONSIDERATIONS WHEN USING
NONLINEAR FORCES
● Nonlinear forces are based on the solution at the previous time step.
·· ·
Mu 
t + Bu 
t+ Ku 
t = P
t + N 
t – t 
● A smaller t will give a more accurate solution.

● Easier to use in direct solution sequences than in modal solution
sequences
● Nonlinear forces may be applied only to members of D-set (A-set + E-
set) for direct and H-set (modal set + E-set). DOFs cannot be
OMITted nor can they be dependent upon the motion of other omitted
DOFs.
● May need a local coordinate system for the nonlinear force applied in
a non-basic axis direction. Verify this with a pilot model.

TYPES OF NOLINS
● NOLIN1 - Nonlinear transient load as a tabular function
P
i n
= ST 
u
j n
● Function of displacement t t 
·
P t = ST 
uj 
t 
i n
● Function of velocity
u t – ut
· n n – 1
where ut n = ------------------------------------------
t
● NOLIN2 - Nonlinear transient load as the product of two

variables P t = SX t 
X 
t 
i n j n k n
where Xj and Xk can be either two displacements or

two velocities or one of each
TYPES OF NOLINS (Cont.)
● NOLIN3 - Nonlinear transient load as a positive variable raised to a
power
 A
 X 
tn = S  j tn 
 , Xj (tn) > 0
Pi 
 0 , Xj (tn) < 0

● where Xj can be a displacement or a velocity

● NOLIN4 - Nonlinear transient load as a negative variable raised to a
power
 A ,
–S 
–X t 
 Xj (tn) < 0
Pi 
tn =  j n
 0 , Xj (tn) > 0

● where Xj can be a displacement or a velocity

WORKSHOP #13
LINEAR TRANSIENT ANALYSIS USING
NOLINS

WORKSHOP #13 - CAR TRAVELING OVER A
BUMP
 Simulate the following car model travelling over a bump using the
NOLIN element. u
1 5 2
Velocity = 100 in/sec
3864 lbs
4” 3 4
50” 120”

k: u – 2.0in
197.4 lb/in
u – 2.0in 394.8 lb/in
·

c: u 0 1.88 lb/(in/sec)
u· 0 1.88 lb/(in/sec) + 0.3 lb/(in/sec)2

WORKSHOP #13 - CAR TRAVELING OVER A
BUMP
F F
.
-2.0” Urel. Urel.

WORKSHOP # 13
$
$ wkshp13.dat
$
$ case control, add : nolin callout
$ plot nonlin forces
$
$ bulk data, add : transfer function to monitor relative displacements
$ nolin for springs and dampers
$ apply appropriate enforced displacement
$
ID NAS102, WORKSHOP13
SOL 109
TIME 100
CEND
TITLE= SIMPLE CAR MODEL WITH NOLINEAR
SUBTITLE= SPRINGS AND DAMPERS RUNNING OVER A BUMP
LABEL= SOL 109, CONSTANT DELTA TIME
SPC= 100
TFL= 100
NONLINEAR = 100
DLOAD = 100
TSTEP = 100
DISPLACEMENT(PLOT)= ALL
NLLOAD(PLOT)= ALL
$

WORKSHOP # 13 (Cont.)
$ $
OUTPUT(XYPLOT) BEGIN BULK
CSCALE=1.3 PARAM,POST,-1
XAXIS= YES $
YAXIS= YES $ CARRIAGE POINTS
XGRID LINES= YES $
YGRID LINES= YES GRID, 1, , 0., 0., 0.
XTITLE= TIME (SEC) GRID, 2, , 120., 0., 0.
YTITLE= VERTICAL DISPLACEMENT OF POINT 1 GRID, 5, , 60., 0., 0.
XYPLOT DISP/1(T2) $
YTITLE= VERTICAL DISPLACEMENT OF POINT 2 $ WHEEL POINTS
XYPLOT DISP/2(T2) $
YTITLE= VERTICAL DISPLACEMENT OF POINT 3 GRID, 3, , 0., -10., 0.
XYPLOT DISP/3(T2) GRID, 4, , 120., -10., 0.
YTITLE= VERTICAL DISPLACEMENT OF POINT 4 $
XYPLOT DISP/4(T2) $ CAR CARRIAGE
YTITLE= VERTICAL DISPLACEMENT OF POINT 5 $
XYPLOT DISP/5(T2) CBAR, 5, 11, 1, 5, 0., 1., 0.
YTITLE= NONLINEAR FORCES AT POINT 1 CBAR, 6, 11, 5, 2, 0., 1., 0.
XYPLOT NONLINEAR/1(T2) PBAR, 11, 12, 10., 10., 10.
YTITLE= NONLINEAR FORCES AT POINT 2 MAT1, 12, 3.0E+7, , .33
XYPLOT NONLINEAR/2(T2) $
$ $ CONSTRAINTS TO ELIMINATE RIGID-BODY MODES
$
SPC1, 100, 1345, 1, 2, 5
SPC1, 100, 13456, 3, 4
$
$
$ INTEGRATION INFORMATION
TSTEP, 100, 200, .05, 1
$
ENDDATA

YTITLE= VERTICAL DISPLACEMENT OF POINT 4
XYPLOT DISP/4(T2)
SOL 109
XYPLOT DISP/5(T2)
CEND
YTITLE= NONLINEAR FORCES AT POINT 1
TITLE= SIMPLE CAR MODEL WITH NOLINEAR
XYPLOT NONLINEAR/1(T2)
SUBTITLE= SPRINGS AND DAMPERS RUNNING OVER A
YTITLE= NONLINEAR FORCES AT POINT 2
BUMP
XYPLOT NONLINEAR/2(T2)
LABEL= SOL 109, CONSTANT DELTA TIME
$
SEALL= ALL
BEGIN BULK
SPC= 100
PARAM,POST,-1
TFL= 100
$
NONLINEAR = 100
$ CARRIAGE POINTS
DLOAD = 100
$
TSTEP = 100
GRID, 1, , 0., 0., 0.
DISPLACEMENT(PLOT)= ALL
GRID, 2, , 120., 0., 0.
NLLOAD(PLOT)= ALL
GRID, 5, , 60., 0., 0.
$
OUTPUT(XYPLOT) $
CSCALE=1.3
$WHEEL POINTS
XAXIS= YES
$
YAXIS= YES
GRID, 3, , 0., -10., 0.
XGRID LINES= YES
GRID, 4, , 120., -10., 0.
YGRID LINES= YES
$
XTITLE= TIME (SEC)
$ CAR CARRIAGE
$
XYPLOT DISP/1(T2)
CBAR, 5, 11, 1, 5, 0., 1., 0.
CBAR, 6, 11, 5, 2, 0., 1., 0.
XYPLOT DISP/2(T2)
PBAR, 11, 12, 10., 10., 10.
MAT1, 12, 3.0E+7, , .33
XYPLOT DISP/3(T2)
$

SOLUTION FOR WORKSHOP #13 (Cont.)
$
$ CONSTRAINTS TO ELIMINATE RIGID-BODY MODES
$
SPC1, 100, 1345, 1, 2, 5
SPC1, 100, 13456, 3, 4
$ NOLIN1, 100, 1, 2, 197.4, 101, 0, 111
$ SYSTEM WILL HAVE A NATURAL FREQUENCY OF 1 HZ NOLIN1, 100, 2, 2, 197.4, 102, 0, 111
$ WITH CRITICAL DAMPING OF 1 PERCENT
$ TABLED2, 111, -2.0,
CONM2, 10, 1, ,2.5 , -1., 1., 0., 0., 1., 0.,ENDT
CONM2, 15, 2, ,2.5 $
CONM2, 20, 5, ,5. $ ADD NONLINEAR PORTION OF DAMPERS
$ $
CELAS2, 30, 197.4, 1, 2, 3, 2 NOLIN4, 100, 1, 2, -0.3, 101, 10, 2.
CELAS2, 40, 197.4, 2, 2, 4, 2 NOLIN4, 100, 2, 2, -0.3, 102, 10, 2.
$ $
CDAMP2, 50, 1.88, 1, 2, 3, 2
$ MOVE WHEELS OVER BUMP
CDAMP2, 60, 1.88, 2, 2, 4, 2
$ $
$ DEFINE EXTRA POINTS TO HOLD DIFFERENCES TLOAD2, 100, 222, 333, D, 0., 0.5, 1., -90.
$ BETWEEN WHEELS AND CARRIAGE SPCD, 222, 3, 2, 4.
$ SPCD, 222, 4, 2, 4.
EPOINT, 101, 102 SPC1,100,2,3,4
$
$ USE TRANSFER FUNCTIONS TO TRACK DIFFERENCES DELAY, 333, 4, 2, 1.2
$ 101= V1 - V3 $
$ 102= V2 - V4
$ INTEGRATION INFORMATION
$
TF, 100, 101, 0, 1., 0., 0., TSTEP, 100, 200, .05, 1
, 1, 2, -1., 0., 0., $
, 3, 2, 1., 0., 0. ENDDATA
$
TF, 100, 102, 0, 1., 0., 0.,
, 2, 2, -1., 0., 0.,
, 4, 2, 1., 0., 0.
$
$ ADD NONLINEAR PORTION OF SPRINGS
$

0 SUBCASE CURVE FRAME XMIN-FRAME/ XMAX-FRAME/ YMIN-FRAME/ X FOR YMAX-FRAME/ X FOR
ID TYPE NO. CURVE ID. ALL DATA ALL DATA ALL DATA YMIN ALL DATA YMAX
0 1 NONLIN 1 1( 4) 0.000000E+00 1.00000 2E+01 0.000000E+00 0.000000E+00 2.975151E+02 6.000001E-01
0.000000E+00 1.000002E+01 0.000000E+00 0.000000E+00 2.975151E+02 6.000001E-01
0 1 NONLIN 2 2( 4) 0.000000E+00 1.00000 2E+01 0.000000E+00 0.000000E+00 4.661460E+02 1.749999E+00
0.000000E+00 1.000002E+01 0.000000E+00 0.000000E+00 4.661460E+02 1.749999E+00
.
.
.
.
.
0 SUBCASE CURVE FRAME XMIN-FRAME/ XMAX-FRAME/ YMIN-FRAME/ X FOR YMAX-FRAME/ X FOR
ID TYPE NO. CURVE ID. ALL DATA ALL DATA ALL DATA YMIN ALL DATA YMAX
0 1 DISP 3 1( 4) 0.000000E+00 1.00000 2E+01 -2.836541E+00 2.199999E+00 5.942877E+00 4.500000E-01
0.000000E+00 1.000002E+01 -2.836541E+00 2.199999E+00 5.942877E+00 4.500000E-01
0 1 DISP 4 2( 4) 0.000000E+00 1.00000 2E+01 -2.671688E+00 1.999999E+00 7.464219E+00 1.600000E+00
0.000000E+00 1.000002E+01 -2.671688E+00 1.999999E+00 7.464219E+00 1.600000E+00
0 1 DISP 5 3( 4) 0.000000E+00 1.00000 2E+01 0.000000E+00 0.000000E+00 4.000000E+00 2.500000E-01
0.000000E+00 1.000002E+01 0.000000E+00 0.000000E+00 4.000000E+00 2.500000E-01
0 1 DISP 6 4( 4) 0.000000E+00 1.00000 2E+01 0.000000E+00 0.000000E+00 4.000000E+00 1.450000E+00
0.000000E+00 1.000002E+01 0.000000E+00 0.000000E+00 4.000000E+00 1.450000E+00
0 1 DISP 7 5( 4) 0.000000E+00 1.00000 2E+01 -2.150819E+00 2.149999E+00 3.963917E+00 1.600000E+00
0.000000E+00 1.000002E+01 -2.150819E+00 2.149999E+00 3.963917E+00 1.600000E+00

(Cont.)

(Cont.)

(Cont.)

(Cont.)

(Cont.)

(Cont.)

(Cont.)

Blank Page

SECTION 18
NORMAL MODES OF PRELOADED
STRUCTURES

NORMAL MODES WITH DIFFERENTIAL STIFFNESS
● Calculate the modes of structures with preloads, large changes in

geometry, and/or nonlinear materials.
y
x
Fx
Reaction
y
x
Differential Stiffness
Preloaded Structure

NORMAL MODES WITH DIFFERENTIAL
STIFFNESS (Cont.)
Large Geometry Changes
k1
k0
u1
K0 = k1
Nonlinear Material
NORMAL MODES WITH DIFFERENTIAL
STIFFNESS (Cont.)
Procedures for obtaining frequencies of a preloaded
structure.
● Use SOL 103.

● Material must be linear.
● Two subcases are required.
● The first subcase is a static subcase calling out the preload.
● The second subcase calculated the modes with a method = x
callout.
● The second subcase must also contain a statsub = y command,
where y = subcase ID of the first subcase.

WORKSHOP #14
NORMAL MODES WITH PRELOAD

WORKSHOP #14 - MODES OF PRELOADED
STRUCTURE
● Consider the simply supported beam as shown below.
Calculate the first bending frequency:
● Without preload
● With preload using SOL 103

WORKSHOP #14 - MODES OF PRELOADED
STRUCTURE (Cont.)
1.0 in
0.1 in
2.0 in 0.1 in
0.1 in
1.0 in
Length: 100 in
Height: 2 in
Width: 1 in
Thickness: 0.100 in
Area: 0.38 in2
I1: 0.229 in4
I2: 0.017 in4

WORKSHOP # 14a
● Run soln14a.dat as a baseline (model without preload)

SOLUTION FILE FOR WORKSHOP #14a -
MODES WITHOUT PRELOAD
SOL 103
TIME 600
CEND
TITLE = Normal Modes Example
SUBCASE 1
METHOD = 1
SPC = 1
VECTOR=ALL
BEGIN BULK
PARAM WTMASS .00259
PARAM COUPMASS 1
EIGRL 1 3 0
PBARL 1 1 I + A
+ A 2. 1. 1. .1 .1 .1
CBAR 1 1 1 2 0. 1. 0.
CBAR 2 1 2 3 0. 1. 0.
CBAR 3 1 3 4 0. 1. 0.
CBAR 4 1 4 5 0. 1. 0.
CBAR 5 1 5 6 0. 1. 0.
CBAR 6 1 6 7 0. 1. 0.
CBAR 7 1 7 8 0. 1. 0.
CBAR 8 1 8 9 0. 1. 0.
CBAR 9 1 9 10 0. 1. 0.
CBAR 10 1 10 11 0. 1. 0.
MAT1 1 1.+7 .3 .101
GRID 1 0. 0. 0. 345
GRID 2 10. 0. 0. 345
GRID 3 20. 0. 0. 345
GRID 4 30. 0. 0. 345
GRID 5 39.9999 0. 0. 345
GRID 6 49.9999 0. 0. 345
GRID 7 60. 0. 0. 345
GRID 8 70. 0. 0. 345
GRID 9 80. 0. 0. 345
GRID 10 90. 0. 0. 345
GRID 11 100. 0. 0. 345
SPC1 1 1234 1
SPC1 1 234 11
ENDDATA

PARTIAL OUTPUT FILE FOR WORKSHOP
#14a MODES WITHOUT PRELOAD
E I G E N V A L U E A N A L Y S I S S U M M A R Y (READ MODULE)
BLOCK SIZE USED ...................... 6
NUMBER OF DECOMPOSITIONS ............. 2
NUMBER OF ROOTS FOUND ................ 3
NUMBER OF SOLVES REQUIRED ............ 4
0 SUBCASE 1
1 1 2.239525E+04 1.496504E+02 2.381760E+01 1.000000E+00 2.239525E+04
2 2 3.550679E+05 5.958757E+02 9.483656E+01 1.000000E+00 3.550679E+05
3 3 1.772662E+06 1.331413E+03 2.119010E+02 1.000000E+00 1.772662E+06

WORKSHOP 14c
● Use soln14a.dat as a starting point as modify it to run in
SOL 103 with preload

SOLUTION FILE FOR WORKSHOP #14c -
MODES WITH PRELOAD USING SOL 103
SOL 103 $
DIAG 8
CEND
MAT1 1 1.+7 .3 .101
$ GRID 1 0. 0. 0. 345
TITLE = Normal Modes with Differential Stiffness GRID 2 10. 0. 0. 345
$ GRID 3 20. 0. 0. 345
SPC = 1 GRID 4 30. 0. 0. 345
DISPLACEMENT=ALL GRID 5 39.9999 0. 0. 345
$ GRID 6 49.9999 0. 0. 345
SUBCASE 1
LOAD = 2
GRID 7 60. 0. 0. 345
$ GRID 8 70. 0. 0. 345
SUBCASE 2 GRID 9 80. 0. 0. 345
METHOD = 10 GRID 10 90. 0. 0. 345
STATSUB = 1 GRID 11 100. 0. 0. 345
$ LOAD 2 1. 1. 1
BEGIN BULK SPC1 1 1234 1
PARAM COUPMASS 1
PARAM WTMASS .00259 SPC1 1 234 11
$ FORCE 1 11 0 500. 1. 0. 0.
EIGRL,10,,,3 ENDDATA
PBARL 1 1 I + B
+ B 2. 1. 1. .1 .1 .1
CBAR 1 1 1 2 0. 1. 0.
CBAR 2 1 2 3 0. 1. 0.
CBAR 3 1 3 4 0. 1. 0.
CBAR 4 1 4 5 0. 1. 0.
CBAR 5 1 5 6 0. 1. 0.
CBAR 6 1 6 7 0. 1. 0.
CBAR 7 1 7 8 0. 1. 0.
CBAR 8 1 8 9 0. 1. 0.
CBAR 9 1 9 10 0. 1. 0.
CBAR 10 1 10 11 0. 1. 0.
$

PARTIAL OUTPUT FILE FOR WORKSHOP #14c
MODES WITH PRELOAD USING SOL 103
1 NORMAL MODES EXAMPLE MARCH 13, 2007 MD NASTRAN 3/ 1/07
PAGE 11
0 SUBCASE 2
1 1 2.735964E+04 1.654075E+02 2.632542E+01 1.000000E+00 2.735964E+04
2 2 3.749262E+05 6.123122E+02 9.745251E+01 1.000000E+00 3.749262E+05
3 3 1.817352E+06 1.348092E+03 2.145555E+02 1.000000E+00 1.817352E+06

BLANK PAGE

SECTION 19
DYNAMIC DESIGN OPTIMIZATION

WHAT IS “DESIGN OPTIMIZATION”?
● Automated modifications of the analysis model parameters
to achieve a desired objective while satisfying specified
design requirements.
● WHAT ARE THE POSSIBLE APPLICATIONS?
● Structural design improvements (optimization)
● Generation of feasible designs from infeasible designs
● Model matching to produce similar structural responses
● System parameter identification
● Configuration evaluations
● Correlates analytical model with test results (see Chapter 20)
● Others - (depends on designer’s creativity)

BASIC FEATURES IMPLEMENTED IN MD
NASTRAN
● Easy access to design synthesis capabilities
● Concept of design model
● Flexibility for design model representation
● User-supplied equation interpretation capability
● Efficient solution for problems of “any” size
● Number of finite element analyses as the measure of efficiency

STRENGTHS OF MD NASTRAN STRUCTURAL
OPTIMIZATION
● Efficient performance for small- to large-scale problems
● Reliable convergence characteristics
● Flexible user interface and user-defined equations
● Full implementation of approximation concepts
● Continuous enhancements
● Results dependent on the proven reliability of MD Nastran
analysis
● High-level support as a part of MD Nastran
● Access to the familiar analysis tools in MD Nastran

OPTIMIZATION CAPABILITIES SUPPORTED
● Multi-Disciplinary Optimization
● Static Response Optimization
● Buckling Response Optimization
● Dynamic Response Optimization
● Direct Frequency
● Modal Frequency
● Modal Transient
● Acoustics
● Superelement Optimization
● Allows the design model to span superelement boundaries.
● Aeroelastic Optimization
● Static Aeroelastic
● Flutter

OPTIMIZATION CAPABILITIES SUPPORTED
(Cont.)
● Shape Optimization
● Four methods are available for Basis Vector Generation
● Manual Grid Variation
● Direct Input of Shapes
● Geometric Boundary Shapes
● Analytic Boundary Shapes

BASIC OPTIMIZATION PROBLEM
STATEMENT
● Design variables
Find X= 
X1, X 2, , Xn 
● Objective function:
 Minimize F(X)
● Subject to:
● Inequality constraints:
G
j 
● Equality constraints: X 0 j = 1, 2, ..., M
● Side constraints:
H 
k 
X = 0 k = 1, 2, ..., L
L u
X X X j = 1, 2, ..., N
i i i

DESIGN OPTIMIZATION IN MD NASTRAN
● The ANALYSIS Case Control request allows you to
specify the type of optimization analysis discipline that you
want to perform for each subcase.
● The following analysis types are allowed in the ANALYSIS
Case Control request:
● STATICS Statics
● MODES Normal Modes
● BUCK Buckling
● DFREQ Direct Frequency
● MFREQ Modal Frequency
● MTRAN Modal Transient
● SAERO Static Aeroelasticity
● FLUTTER Flutter
DESIGN OPTIMIZATION IN MD NASTRAN
(Cont.)
● In the example below you optimize two static load cases in
subcases 1 and 2, a modal response in subcase 3, and a
transient response in subcase 4. SOL 200
cend
spc = 100
DESOBJ(MIN) = 15
ANALYSIS = STATICS
subcase 1
subtitle=static load 1
DESSUB = 10
displacement = all design constraint
stress = all
load = 1
subcase 2
subtitle=static load 1
DESSUB = 20
displacement = all
strain(fiber) = all design constraint
load = 2
subcase 3
subtitle=modal analysis
ANALYSIS = MODES
DESSUB = 30
method = 3
subcase 4
subtitle=transient analysis
ANALYSIS = MTRAN
DESSUB = 40
method = 4
dload = 4
TSTEP = N
NAS102,Section 19, March 2007 begin bulk

COMMONLY USED OPTIMIZATION BULK
DATA ENTRIES
● DESVAR Defines a design variable.
● DVPREL1 Defines the relation between an analysis model property
and design variables.
● DLINK Relates one design variable to one or more other design variables.
● DRESP1 Defines a set of direct structural responses that is used in the design either
as constraints (referenced by the DCONSTR Bulk Data entry) or as an
objective (referenced by the DESOBJ Case Control Command).
● DCONSTR Defines a design constraint (referenced by the DESSUB Case Control
Command).
● DCONADD Defines the design constraints for a subcase as a union of DCONSTR
entries.
● DRESP2 Defines a synthesized response that are used in the design. This response
can be either a constraint or an objective.
● DEQATN Defines equation(s) for use in design sensitivity.
● DVCREL1 Defines the relation between a connectivity property and design variables.
● DVCREL2 Defines the relation between a connectivity property and design variables
using a user-supplied equation.
● DVMREL1 Defines the relation between a material property and design variables.
● DVMREL2 Defines the relation between a material property and design variables with a
user-supplied equation.
WORKSHOP #15
OPTIMIZATION USING NORMAL MODES

WORKSHOP #15 - OPTIMIZATION USING
NORMAL MODES
Minimize the weight of the following three bar truss problem. The first mode
must be between 1500-1550 Hz. The structure must remain symmetric.
Below is a geometric representation of the truss. It also contains the loads
and boundary constraints.
Node 1 Node 2 Node 3
Area 1 Area 2 Area 1 10”
Node 4
10”
Elastic Modulus: 10E6

Poisson’s Ratio 0.33
Density: 0.1
Wt.-Mass Conversion 0.00259
Area 1: 1.0
Area 2: 2.0

OPTIMIZATION STATEMENT
● Design variables
● The areas of the three rod elements (A1, A2, A3)
● Objective
● Minimize the weight of the truss.
● Subject to the following constraints:

● The first mode must be between 1500-1550 Hz.
● A1 = A3 to impose symmetry.

WORKSHOP # 15

$
$ wkshp15.dat
$
$ executive control, add : optimization solution sequence
$
$ case control, add : type of analysis
$ appropriate case control callout
$
$ bulk data, add : design variable
$ objective function
$ constraints
$ relate design variable to property
$
ID NAS102, WORKSHOP 15
TIME 10
SOL 200 $ OPTIMIZATION
CEND
TITLE= SYMMETRIC THREE BAR TRUSS DESIGN OPTIMIZATION - VARIATION OF D200X1
SUBTITLE= GOAL IS TO MIN WT WHILE KEEPING THE 1ST MODE BETWEEN 1500-1550 HZ
ECHO= SORT
SPC= 100
DISP(PLOT) ALL
DESOBJ(MIN)= 100 $ (DESIGN OBJECTIVE = DRESP ID)
DESSUB= 200 $ DEFINE CONSTRAINT SET FOR BOTH SUBCASES
SUBCASE 1
ANALYSIS= MODES
METHOD= 10
BEGIN BULK

WORKSHOP # 15
$ $
$---------------------------------------------------------------- ------ $----------------------------------------------------------------------
$ ANALYSIS MODEL
$---------------------------------------------------------------- ------ $ DESIGN MODEL
$ $----------------------------------------------------------------------
EIGRL, 10, , , 2 $
PARAM, POST, -1 $...DESIGN VARIABLE DEFINITION
PARAM, PATVER, 3.0
$ $
$ GRID DATA $DESVAR,ID, LABEL, XINIT, XLB, XUB, DELXV(OPTIONAL)
$ 2 3 4 5 6 7 8 9 10 DESVAR, 1, A1, 1.0, 0.1, 100.0
GRID, 1, , -10.0, 0.0, 0.0 .
GRID, 2, , 0.0, 0.0, 0.0
GRID, 3, , 10.0, 0.0, 0.0 .
GRID, 4, , 0.0, -10.0, 0.0 $
$ SUPPORT DATA $...IMPOSE X3=X1 (LEADS TO A3=A1)
SPC, 100, 1, 123456, , 2, 123456 $
SPC, 100, 3, 123456, , 4, 3456
$ ELEMENT DATA $DLINK, ID, DDVID, CO, CMULT, IDV1, C1, IDV2, C2, +
CROD, 1, 11, 1, 4 $+, IDV3, C3, ...
CROD, 2, 12, 2, 4 DLINK, 1, 3, 0.0, 1.0, 1 1.00
CROD, 3, 13, 3, 4 $
$ PROPERTY DATA
PROD, 11, 1, 1.0 $...DEFINITION OF DESIGN VARIABLE TO ANALYSIS MODEL
PROD, 12, 1, 2.0 PARAMETER RELATIONS
PROD, 13, 1, 1.0 $
MAT1, 1, 1.0E+7, , 0.33, 0.1 $DVPREL1,ID, TYPE, PID, FID, PMIN, PMAX, CO, , +
$ $+, DVID1, COEF1, DVID2, COEF2, ...
PARAM, WTMASS, .00259
DVPREL1, 10, PROD, 11, 4, , , , , +DP1
+DP1, 1, 1.0
.
.
$
DOPTPRM, DESMAX, 30
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......0
ENDDATA

ID NAS102, WORKSHOP 15
TIME 10
SOL 200 $ OPTIMIZATION
CEND
TITLE= SYMMETRIC THREE BAR TRUSS DESIGN OPTIMIZATION - VARIATION OF D200X1
SUBTITLE= GOAL IS TO MIN WT WHILE KEEPING THE 1ST MODE BETWEEN 1500-1550 HZ
ECHO= SORT
SPC= 100
DISP(PLOT) ALL
DESOBJ(MIN)= 100 $ (DESIGN OBJECTIVE = DRESP ID)
DESSUB= 200 $ DEFINE CONSTRAINT SET FOR BOTH SUBCASES
SUBCASE 1
ANALYSIS= MODES
METHOD= 10
BEGIN BULK
$
$----------------------------------------------------------------------
$ ANALYSIS MODEL
$----------------------------------------------------------------------
$
EIGRL, 10, , , 2
PARAM, POST, -1
$
$ GRID DATA
$ 2 3 4 5 6 7 8 9 10
GRID, 1, , -10.0, 0.0, 0.0
GRID, 2, , 0.0, 0.0, 0.0
GRID, 3, , 10.0, 0.0, 0.0
GRID, 4, , 0.0, -10.0, 0.0
$ SUPPORT DATA
SPC, 100, 1, 123456, , 2, 123456
SPC, 100, 3, 123456, , 4, 3456
$ ELEMENT DATA
CROD, 1, 11, 1, 4
CROD, 2, 12, 2, 4
CROD, 3, 13, 3, 4
$ PROPERTY DATA
PROD, 11, 1, 1.0
PROD, 12, 1, 2.0
PROD, 13, 1, 1.0
MAT1, 1, 1.0E+7, , 0.33, 0.1
$
PARAM, WTMASS, .00259

$
$---------------------------------------------------------------- ------
$ DESIGN MODEL
$---------------------------------------------------------------- ------
$
$...DESIGN VARIABLE DEFINITION
$
$DESVAR,ID, LABEL, XINIT, XLB, XUB, DELXV(OPTIONAL)
DESVAR, 1, A1, 1.0, 0.1, 100.0
DESVAR, 2, A2, 2.0, 0.1, 100.0
DESVAR, 3, A3, 1.0, 0.1, 100.0
$
$...IMPOSE X3=X1 (LEADS TO A3=A1)
$
$DLINK, ID, DDVID, CO, CMULT, IDV1, C1, IDV2, C2, +
$+, IDV3, C3, ...
DLINK, 1, 3, 0.0, 1.0, 1 1.00
$
$...DEFINITION OF DESIGN VARIABLE TO ANALYSIS MODEL PARAMETER RELATIONS
$
$DVPREL1,ID, TYPE, PID, FID, PMIN, PMAX, CO, , +
$+, DVID1, COEF1, DVID2, COEF2, ...
DVPREL1, 10, PROD, 11, 4, , , , , +DP1
+DP1, 1, 1.0
DVPREL1, 20, PROD, 12, 4, , , , , +DP2
+DP2, 2, 1.0
DVPREL1, 30, PROD, 13, 4, , , , , +DP3
+DP3, 3, 1.0
$
$...STRUCTURAL RESPONSE INDENTIFICATION
$
$DRESP1 ID LABEL RTYPE PTYPE REGION ATTA ATTB ATT1 +
$+ ATT2 ...
DRESP1 100 W WEIGHT
DRESP1 210 MODE1 EIGN 1
$
$...CONSTRAINTS
$
$DCONSTR,DCID, RID, LALLOW, UALLOW
DCONSTR, 200, 210, 8.883E7, 9.485E7
$
$...OPTIMIZATION CONTROL
$
DOPTPRM, DESMAX, 30
$
ENDDATA






DIRECT SPECIFICATION OF COMMONLY
USED FUNCTIONS
● Direct Support of Commonly-used Functions on the
DRESP1 entry (e.g, SUM, RSS, AVG, etc.) for transient
and frequency response analysis.
● Example:
● DRESP1,100,RSSCAL,FRDISP,,,3,RSS,100
● The above DRESP1 performs a root sum square of the

displacement at grid point 100, component 3, across the whole
forcing frequency range.

OPTIMIZATION EXAMPLE USING
FREQUENCY RESPONSE
 The automobile model shown below has the front left wheel out of balance. The amount
of mass is .3 units and the radius from the center of the wheel to the mass is 10 inches.
The frequency range of interest is from 0.5-50 Hz. The tire must not displace by more
than 0.5 inches at any frequency. The driver’s seat must not displace by more than
0.25 inches between 0.5 and 25 hz. The goal is to minimize the SRSS response of the
driver’s seat across the whole forcing frequency range. Use the modal method.

OPTIMIZATION STATEMENT
● Design Variables:
● The springs and dampers of the car
● Objective function:
● Minimize the SRSS response at the driver’s seat
● Subject to the following constraints:

● The maximum vertical displacement of the tire must be less than
0.5 inches
● The maximum vertical displacement of the driver seat must be
less than 0.25 inches

DYNAMIC LOAD INPUT
2
F = mr
Direction of rotation
= t
 Fx = F cos(t)
Fy = F sin( t)
● Use DAREA to describe magnitude of Fx and Fy (mr)
● Use one RLOAD1 entry for each loading
● Use a DPHASE to describe the phase relationship
● Use a TABLED4 to describe the frequency relationship of the loading (2)
● Use a DLOAD entry to combine the loadings (RLOAD1s)

TABLED4 ENTRY FOR THIS SAMPLE
N  X–X  i
 1 
Y =  A i
---------------------- 
 X2 
i= 0
● Set X1 = 0. X2 = 1. X3 = 0. X4 = 1000. (More than highest freq of interest)
● Since X is in cycles per unit time, it must be multiplied by 2to get radians per
unit time - enter 2
● The input force is mr2 - we will enter mr using DAREA entries, but need the
2 term
● We want to use the second (or the 2) term from the TABLED4, This can be
accomplished by setting A1=0.0 and A2=(2)2

DYNAMIC RESPONSE SENSITIVITY
● Allows calculation of sensitivities of selected dynamic
responses with respect to changes in design variables
dri/dxj.
● These quantities provide insight on which areas are most
effective in driving the design.
● Sensitivity coefficients can be requested with the DSAPRT
case control command
● Analysis disciplines:
● Direct frequency response
● Modal frequency response
● Modal transient response

DESCRIPTION OF THE DESIGN MODEL
● Identify structural properties to be related to design
variables (DVPREL1, DVPREL2).
● Identify response quantities of interest (DRESP1,
DRESP2).
● Specify bounds on responses and (optionally) screening
criteria (DCONSTR, DSCREEN).
● Select output frequencies or time steps in Case Control
(OFREQ or OTIME sets).

INPUT FOR OPTIMIZATION MODEL
 Executive and Case Control
 SOL 200
CEND
TITLE = Sample dynamic analysis model
set 999 = 358,471
DISP(phase) = 999
SUBCASE 1
ANALYSIS = MFREQ
DESSUB = 100 $ constraints
DESOBJ(min) = 300 $ design objective - minimize driver’s response
DLOAD = 1
METHOD = 1
FREQ = 1
Bulk Data Section
 BEGIN BULK
.
.
include ’car.blk’
include ’springs.blk’
include ’optim1.blk’
.
.
ENDDATA

PARTIAL LISTING OF CAR.BLK
Input for Dynamic Loading
EIGRL,1,-1.0,100.
DLOAD,1,1.,1.,11,1.,12
RLOAD1,11,20,,,,111
RLOAD1,12,30,,,40,111
DPHASE 40 358 2 90.
DAREA 20 358 1 3.
DAREA 30 358 2 3.
TABLED4,111,0.,1.,0.,1000.
,0.,0.,39.478,ENDT
$
$ PLUS THE REST OF THE MODEL DESCRIPTION
$

PARTIAL LISTING OF SPRINGS.BLK
Input for Springs
 CROD 1002 1002 402 1402
CROD 1012 1001 825 1825
CROD 1022 1001 358 1358
CROD 1032 1002 869 1869
$
PROD 1001 1000 1000.
PROD 1002 1000 800.
$ select material so that value of PROD
$ is spring stiffness,
$ therefore, E = l = 10.
MAT1 1000 10.
Input for Shock Absorbers

 $
$ add dampers for shock absorbers
$
$ front
cvisc 2011 2001 825 1825
cvisc 2021 2001 358 1358
$ back
cvisc 2001 2002 402 1402
cvisc 2031 2002 869 1869
$ damper properties
pvisc 2001 10. 0.
pvisc 2002 5. 0.
$

PARTIAL LISTING OF OPTIM.BLK
Define Design Variables
desvar,1,frntdamp,1.,.1,10.
desvar,2,reardamp,1.,.2,20.
desvar,3,frntstif,1.,.4,2.
desvar,4,rearstif,1.,.5,2.5
Link Design Variables to Properties

$
$ relation between properties and variables
$
dvprel1,101,pvisc,2001,3,1.,,,,+dv101
+dv101,1,10.
dvprel1,102,pvisc,2002,3,1.,,,,+dv102
+dv102,2,5.
dvprel1,103,prod,1001,4,4.,,,,+dv103
+dv103,3,10.
dvprel1,104,prod,1002,4,4.,,,,+dv104
+dv104,4,8.
$
Define Design Constraints

$ require that maximum tire displacement be .5 inches
$
dconstr,101,200,-.5,.5
$
$ require that maximum driver displacement be .25 inches
$
dconstr,102,201,-.25,.25,0.5,25.
$ combine constraints into set 100
$
dconadd,100,101,102

PARTIAL LISTING OF OPTIM.BLK (Cont.)
Select Response Quantities
$ select displacement Y at driver seat and mount point as
$ response quantities
$
$ mount point
$
dresp1,200,disp,frdisp,,,2,,358
$
$ define driver’s seat disp as a response
$
dresp1,201,driver,frdisp,,,2,,471
Define Objective Function - SRSS of Driver Response

dresp1,300,srss,frdisp,,,2,RSS,471

WORKSHOP #16
OPTIMIZATION USING FREQUENCY

RESPONSE

SOLUTION TO WORKSHOP #16
$
$ input file to optimize (minimize) the response of a car to a $
$ rotating imbalance - V68 - June, 1994 $ DATA RELATED TO FREQUENCY RESPONSE
$ use modal approach - up to 100 hz $
$ DLOAD 1 1. 1. 11 1. 12
SOL 200 RLOAD1 11 20 111
diag 8 RLOAD1 12 30 40 111
CEND DPHASE 40 358 2 90.
TITLE = Sample dynamic analysis model DAREA 20 358 1 3.
SUBTITLE = Rotating force due to tire out of balance DAREA 30 358 2 3.
LABEL = perform optimization to minimize driver response TABLED4 111 0. 1. 0. 100.
set 999 = 358,471 0. 39.478 ENDT
DISP(phase) = 999 FREQ1 1 .5 .5 100
SUBCASE 1 $
ANALYSIS = MFREQ ENDDATA
DESSUB = 100 $ constraints
DESOBJ(min) = 300 $ design objective - minimize driver response
DLOAD = 1
METHOD = 1
FREQ = 1
BEGIN BULK
eigrl,1,0.,100.
doptprm,desmax,25
include ’car.blk’
include ’springs.blk’
include ’optim1.blk’
param,post,0
$

SOLUTION TO WORKSHOP #16 (Cont.)
$ $
$ springs.blk $ add dampers for shock absorbers
$ $ front
CONM2 3001 1402 1.E8 cvisc 2011 2001 825 1825
CONM2 3002 1825 1.E8 cvisc 2021 2001 358 1358
CONM2 3003 1358 1.E8 $ back
CONM2 3004 1869 1.E8 cvisc 2001 2002 402 1402
$ cvisc 2031 2002 869 1869
GRID 1402 159.870 11.0000 -14.3750 13456 $ damper properties
GRID 1358 20.6116 11.0000 -15.1000 13456 pvisc 2001 10. 0.
GRID 1825 20.6116 11.0000 -54.9000 13456 pvisc 2002 5. 0.
GRID 1869 159.870 11.0000 -55.6250 13456 $
$ $ end of springs.blk
CELAS2 1001 10000. 402 1 1402 1 $
CROD 1002 1002 402 1402
CELAS2 1003 10000. 402 3 1402 3
CELAS2 1011 10000. 825 1 1825 1
CROD 1012 1001 825 1825
CELAS2 1013 10000. 825 3 1825 3
CELAS2 1021 10000. 358 1 1358 1
CROD 1022 1001 358 1358
CELAS2 1023 10000. 358 3 1358 3
CELAS2 1031 10000. 869 1 1869 1
CROD 1032 1002 869 1869
CELAS2 1033 10000. 869 3 1869 3
$
$ properties for springs
$
PROD 1001 1000 1000.
PROD 1002 1000 800.
$ select material so that value of PROD is spring stiffness,
$ therefore, E = l = 10.
MAT1 1000 10.

$
$ file - car.blk
$
$ MODEL COURTESY LAPCAD ENGINEERING
$ CHULA VISTA, CALIFORNIA
$
GRID 1 79.0000 56.0000-2.00000
GRID 2 157.000 40.1000-1.30000
.
.
CQUAD4 29 2 22 19 12 21 0.00E+0 0.00E+0
CQUAD4 30 2 21 20 11 22 0.00E+0 0.00E+0
.
.
PSHELL 1 1 0.025 1
PSHELL 2 1 0.200 1
.
.
MAT1 11.000E+73.759E+63.300E-12.600E-41.370E-57.000E+1.3
$
$ end of car.blk
$

$ $
$ beginning of optim1.blk $ select displacement Y at tire and driver seat as
$ $ response quantities
$ data for design sensitivity $
$ $ maximum tire displacement be < +/-0.5 inches
$
$ define design variables dconstr,101,200,-.5,.5
$ dresp1,200,disp,frdisp,,,2,,358
desvar,1,frntdamp,1.,.1,10. $
desvar,2,reardamp,1.,.2,20. $ define driver's seat disp as a response
desvar,3,frntstif,1.,.4,2. $
$
desvar,4,rearstif,1.,.5,2.5 $ require that maximum driver displacement be < +/- 0.25 inches
$ $ between .5 and 25 hz
$ relation between properties and variables $
$ dconstr,102,201,-.25,.25,0.5,25.0
dvprel1,101,pvisc,2001,3,1.,,,,+dv101 dresp1,201,driver,frdisp,,,2,,471
$
+dv101,1,10. $ combine constraints into set 100
dvprel1,102,pvisc,2002,3,1.,,,,+dv102 $
+dv102,2,5. dconadd,100,101,102
dvprel1,103,prod,1001,4,4.,,,,+dv103 $
+dv103,3,10. $ define objective = minimize srss of response
$
dvprel1,104,prod,1002,4,4.,,,,+dv104
dresp1,300,rss,frdisp,,,2,rss,471
+dv104,4,8. $
$ $ end of optimization input
$
$ end of optim1.blk
$



RESULTS OF OPTIMIZATION
● Design criteria satisfied in 10 cycles.
● Results follow:

OBJECTIVE FUNCTION HISTORY

DESIGN VARIABLES HISTORY

WHEEL RESPONSE
Before Optimization
After Optimization

DRIVER’S SEAT RESPONSE
Before Optimization
After Optimization

CONCLUSION
● With minimum effort, the design is modified to satisfy constraints and minimize
selected responses.
● SOL 200 is a valuable design tool for dynamic analysis.

BLANK PAGE

SECTION 20
TEST-ANALYSIS CORRELATION

INTRODUCTION TO TEST-ANALYSIS
CORRELATION
● MD Nastran results and test data may not match due to
modeling and testing uncertainties.
● Common saying: “No one believes the analytical results
(except for the modeler), and everyone believes the test
data (except for the experimentalist).”
● Sources of modeling uncertainties:
● Physics of the problem being simulated
● Boundary conditions
● Material properties
● Joint flexibility
● “As-built” versus “as-designed”
● Damping

INTRODUCTION TO TEST-ANALYSIS
CORRELATION (CONT.)
● Goals of test-analysis correlation:

● Assess the degree of correlation between MD Nastran results and
test data.
● Refine the MD Nastran model to match test data.
● Person doing the correlation must understand the test

data, the MD Nastran results, and the uncertainties in
both.

INTEGRATED TEST-ANALYSIS PROCEDURE
● Four phases of an integrated test-analysis procedure:
● 1. Pre-test planning (test simulation)
● 2. Data acquisition. (Capture raw data, such as,
accelerations.)
● 3. Data reduction and analysis. (Process raw data into
quantities of interest, such as, mode shapes.)
● 4. Post-test evaluation. (Assess goodness of fit between
test data and MD Nastran results; refine the
MD Nastran model to match the test data.)
● The analyst is involved in phases 1 and 4, and the

experimentalist is involved in phases 2 and 3.

PRE-TEST PLANNING
● Create a baseline MD Nastran model to determine optimal excitation
and measurement locations. There are two methods for doing this:
● Simulation and inspection
● Cross-orthogonality check
● Simulation and inspection. Use MD Nastran to simulate the test, and
choose input and output locations that give maximum response.
● Cross-orthogonality check:
● From a proposed set of measurement locations create an A-set, use
Guyan reduction, and compute the mode shapes normalizing to a unit
modal mass. Call this set of vectors (as in “test” modes). Output the
modes that span the frequency range of interest and the A-set mass
matrix.
● Then, remove the A-set, repeat the modal calculation computing
responses for the A-set DOFs for the full model, and output these
modes (a, as in “analysis” modes).
PRE-TEST PLANNING (Cont.)
● In a third MD Nastran run, read both sets of results and
compute:
T
t M aa
● If the proposed measurement locations (A-set) are

adequate, then the resulting matrix has 1 on the diagonal
and 0 as off-diagonal terms. If the off-diagonal terms are
not 0, the proposed measurement locations are not
adequate and a new set must be formulated. (In actuality,
off-diagonal terms less than 0.05 are acceptable.)
● DMAP alter for pre-test planning, premaca.vxx is on the
MD Nastran delivery.
PRE-TEST PLANNING (Cont.)
● Modal effective mass1 and modal kinetic energy2 calculations can also be
made in MD Nastran to ensure that the test specimen is well understood
before testing.
● Test and analytical locations need to align fully (location and coordinate
direction) to facilitate pre-test planning and post-test evaluation. RBARs,
MPCs, and alternate output coordinate systems can be used to “line up” the
locations.
● Once the excitation and measurement locations are verified, then the test
should be simulated to ensure that the test specimen is not overstressed
during testing.
T 2
1. For the i-th mode, effective mass = 
i M D m
where M = mass matrix and Dm = rigid-body vector.
2
2. For the i-th mode, kinetic energy = Mi . This shows the energy
distribution within a mode.

POST-TEST EVALUATION
● Compare MD Nastran results to test data by inspection or
cross-orthogonality checks.
● Inspection—graphics
● XY plots:
● Plot test data and MD Nastran results together and assess the
degree of correlation in magnitude and frequency content.
● Make sure that the curves plotted together represent the same
spatial location and direction.
● Structure plots:
● Plot test and analytical mode shapes together and assess the degree
of correlation. (Plots can also be animated.)
● Inspection can also be applied to non-graphical data such
as, comparison of measured and computed resonant
frequencies.
POST-TEST EVALUATION (Cont.)
● Cross-orthogonality check
Quantitative check on the degree of test-analysis correlation is the cross-
orthogonality check defined by
T
t Maa a = COR
where t = test modes

Maa = A-set mass matrix
a = analysis modes computed for the A-set
Off-diagonal terms should be less than 10% of the diagonal terms in order to
have a reliable match between test and analysis.
Modes may be “switched” which is reflected by large off-diagonal terms in
COR, the correlation matrix.
● If there are discrepancies between test data and analytical results, it may be
desirable to refine the MD Nastran model to get a better match.
● COR example:
Consider the 2-D beam model shown below. Assume that
accelerometers are placed on the beam at every other grid
point. The translational accelerations are measured in the
x and y directions.
y
x
1 2 3 4 5 6 7 8 9 10 11
= grid point
= A-set point

 Test Frequencies and Mode Shapes
Grid Point Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6

2x 0 0.15 0 0.45 1 0
2y 0 0 0.15 0 0 0.3
4x 0 0.45 0 1 0.95 0
4y 0.15 0 0.75 0 0 1
6x 0 0.7 0 0.7 -1 0
6y 0.4 0 1 0 0 0.05
8x 0 0.9 0 -0.15 -0.95 0
8y 0.7 0 0.45 0 0 -0.85
10x 0 1 0 -0.85 1 0
10y 1 0 -0.75 0 0 0.25
Frequency
(Hz) 14.2 50.5 85.2 144.1 245.3 257.3

 MD Nastran Frequencies and Mode Shapes
Grid Point Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6
2x 0 0.254 0 0.773 0 1.316
2y 0.038 0 0.207 0 -0.526 0
4x 0 0.736 0 1.681 0 1.316
4y 0.309 0 1.183 0 -1.759 0
6x 0 1.147 0 1.203 0 -1.316
6y 0.77 0 1.62 0 -0.078 0
8x 0 1.445 0 -0.266 0 -1.316
8y 1.34 0 0.747 0 1.524 0
10x 0 1.602 0 -1.516 0 1.316
10y 1.957 0 -1.145 0 -0.474 0
Frequency
13.91 49.45 86.38 150.59 244.03 256.5
(Hz)
 Test and MD Nastran frequencies are close

Post-Test Evaluation (Cont.)
ID PRETEST, DYNOTES
SOL, 103
compile xread $
ALTER ' read.*mxx.*phix'$
The locations MATPRN MXX,,,,//$

MATPCH MXX,PHIX,,,// $
CEND
of the aset TITLE= Cantilever Beam for Normal Modes

SUBTITLE=MODES CASE CONTROL
DISP=ALL
SPC=1
points are in METHOD=100

BEGIN BULK
PARAM, AUTOSPC, YES
the same GRDSET, , , , , , ,345

BAROR, , , , ,0., 1., 0.
GRID, 1, , 0.0, 0.0, 0.0
geometric GRID, 2, , 1.0, 0.0, 0.0

GRID, 3, , 2.0, 0.0, 0.0
GRID, 4, , 3.0, 0.0, 0.0
location as in
GRID, 5, , 4.0, 0.0, 0.0
GRID, 6, , 5.0, 0.0, 0.0
GRID, 7, , 6.0, 0.0, 0.0
GRID, 8, , 7.0, 0.0, 0.0
the test article GRID, 9, , 8.0, 0.0, 0.0

GRID,10, , 9.0, 0.0, 0.0
GRID,11, ,10.0, 0.0, 0.0
CBAR, 1, 1, 1, 2
CBAR, 2, 1, 2, 3
CBAR, 3, 1, 3, 4
CBAR, 4, 1, 4, 5
CBAR, 5, 1, 5, 6
CBAR, 6, 1, 6, 7
CBAR, 7, 1, 7, 8
CBAR, 8, 1, 8, 9
CBAR, 9, 1, 9,10
CBAR,10, 1,10,11
$Use Master Set DOF, to reduce to Test DOF
ASET1, 12, 2, 4, 6, 8, 10
SPC1, 1, 123456, 1
PBAR, 1, 1, 0.01, 0.016, 0.016
MAT1, 1, 3.E7, , 0.3, 7.7
EIGRL, 100, , 10.E3, 6
ENDDATA

ID COR, DYN.NOTES
TIME 30
SOL 100
COMPILE USERDMAP, SOUIN=MSCSOU, NOLIST, NOREF
ALTER 2 $
$----------------------------------------------------------
$ DMAP TO COMPUTE CROSS-ORTHOGONALITY
$ INPUTS FROM MSC.Nastran RUN: MXX (A-SET MASS)
$ PHIX (A-SET MODE SHAPES)
$ (PREVIOUS M/N RUN USED MATPCH TO PUNCH DMI ENTRIES)
$ INPUT FROM TEST: PHITEST (A-SET MODE SHAPES)
$ OUTPUTS: UNITCHK (UNIT MASS CHECK)
$ COR (CROSS-ORTHOGONALITY MATRIX)
$----------------------------------------------------------
$ READ DMI INPUT
DMIIN DMI,DMINDX/PHIX,PHITEST,MXX,,,,,,,/ $
$ VERIFY INPUT MATRICES
MATPRN PHIX,PHITEST,MXX,,// $
$ MULTIPLY PHIX(TRANS)*MXX = PHITMASS
MPYAD PHIX,MXX,/ PHITMASS /1///$
$ MULTIPLY PHITMASS*PHIX = UNITCHK
MPYAD PHITMASS,PHIX,/ UNITCHK // $
$ PRINT TITLE AND UNITCHK
MATPRN UNITCHK,,,,// $
MESSAGE // ’CHECK ON UNIT MASS’/ $
$ MULTIPLY PHITMASS*PHITEST = COR
MPYAD PHITMASS,PHITEST,/ COR // $
$ PRINT TITLE AND COR
MATPRN COR,,,,// $
MESSAGE // ’CROSS-ORTHOGONALITY MATRIX’/ $
ENDALTER
CEND
TITLE = CROSS-ORTHOGONALITY CHECK
BEGIN BULK
DMI,PHITEST,0,2,1,0,,10,6
DMI,PHITEST,1,1,0.,.0,0.,.15,0.,
,.40,0.,.70,0.,1.0
DMI,PHITEST,2,1,.15,0.,.45,0.,.70,
,0.,.90,0.,1.0,0.
.
. (rest of PHITEST)
.
DMI MXX 0 6 1 0 10 10
DMI* MXX 1 1 9.6249 9976E-02
* 3 1.92499999E-02
DMI* MXX 2 2 1.3886 7334E-01
* 4 2.67470982E-02 6 -1.37760025E-02
* 8 4.32037748E-03 10 3.94638191E-05
.
. (rest of MXX)
.
DMI PHIX 0 2 1 0 10 6
DMI* PHIX 1 1 -1.02694275E-17
* 3.79799381E-02 -3.21330419E-17 3.09179097E-01 1.62494801E-17
* 7.69559503E-01 -1.60461922E-17 1.34014440E+00 -4.83748987E-17
* 1.95740056E+00
DMI* PHIX 2 1 2.5367 3702E-01
* -3.71932167E-18 7.36189783E-01 -2.11487186E-17 1.14664245E+00
* -1.32509834E-17 1.44485378E+00 -5.04831637E-19 1.60163271E+00
* 1.19194476E-17
.
. (rest of PHIX)
.
ENDDATA

MATRIX COR (GINO NAME 101 ) IS A DB PREC 6 COLUMN X 6 ROW SQUARE MATRIX.
COLUMN 1 ROWS 1 THRU 6 --------------------------------------------------
ROW
1) 5.1380D-01 6.7188D-19 1.8822D-03 -1.6038D-17 8.3091D-03 -1.5467D-16
COLUMN 2 ROWS 1 THRU 6 --------------------------------------------------
ROW
1) -1.0247D-17 6.1972D-01 3.1542D-18 -5.7035D-03 -1.0390D-17 -8.8287D-09
COLUMN 3 ROWS 1 THRU 6 --------------------------------------------------
ROW
1) -1.5467D-02 -6.3725D-18 6.3014D-01 -2.7673D-17 -4.6145D-03 9.5503D-17
COLUMN 4 ROWS 1 THRU 6 --------------------------------------------------
ROW
1) 2.5192D-18 1.1390D-02 8.1214D-17 5.8007D-01 -1.0636D-16 7.5993D-03
COLUMN 5 ROWS 1 THRU 6 --------------------------------------------------
ROW
1) -1.0452D-17 5.3899D-03 -2.1283D-17 -1.3447D-02 1.2668D-14 7.4474D-01
COLUMN 6 ROWS 1 THRU 6 --------------------------------------------------
ROW
1) -3.2927D-03 -2.5409D-18 8.7221D-03 -2.3220D-16 -5.6143D-01 9.6348D-15
THE NUMBER OF NON-ZERO TERMS IN THE DENSEST COLUMN = 6
THE DENSITY OF THIS MATRIX IS 100.00 PERCENT.
^^^CROSS-ORTHOGONALITY MATRIX

SUMMARY
1. COR matrix shows good agreement in mode shapes. (Off diagonal
are small compared to the diagonal terms.)
2. Mode 5 (test) is Mode 6 (analysis) and vice versa. This is shown by
large off-diagonal terms in 5,6 and 6,5.
3. The magnitude of diagonal terms are not 1.0 because the test modes
are not normalized to unit modal mass. (They are normalized to a
maximum component of 1.0.)
4. Some effort is required to reformat test data for DMI input.
5. For a large number of modes, additional DMAP may be written to
simplify the cross-correlation output.
6. DMAP alter for test-analysis cross-orthogonality, postmaca.vxx, is on
the MD Nastran delivery.

MODEL REFINEMENT
● There are three ways to update an MD Nastran model to match test
data:
● 1. Using brute force
● 2. Using the sensitivity matrix
● 3. Using design optimization
● All methods update MD Nastran model parameters such as I and A for
BARs and t for QUADs. Base flexibility can also be a parameter if it
was explicitly modeled (such as with ELASs).
● Brute force:
● Make model changes based on inspection of the results and make an
educated “guess” as to the type of changes necessary.
● After the model is updated, another analysis is made and the MD
Nastran results are again compared to test data to see if the match is
better.
● If the match is not better, then make more changes and repeat.
MODEL REFINEMENT (Cont.)
● Sensitivity matrix:
● Sensitivity is the gradient of response with respect to model parameter. “Responses”
are MD Nastran-computed results, and “parameters” are property values.
 R
S
P = P 0 = -------
P
Response
Po
Parameter
● Each term in the sensitivity matrix S is given by

S ij = Ri P
j
● where Ri is the i-th response and Pj is the j-th parameter.

● The greater the magnitude Sij, the greater the response sensitivity.
● SOL 200 in MD Nastran computes the sensitivity matrix.
● S can be used by inspection to indicate which parameters need to be changed to
change the response.

● S can also be output and used in a least-squares sense to
minimize the difference between test and analysis as
follows:
Pn = Po + (STS)-1ST(Rt - Ra)
where Pn = updated parameters for test-analysis match
Po = parameters from original model
S = sensitivity matrix
Rt = test response
Ra = analytical results

● Design optimization:
● Implemented in SOL 200
● Want to minimize the difference between test and analysis but with a model that changes least
from the baseline model
● Use DEQATN to write an objective function that is the weighted difference between test data
and analytical results.
test d ata model params
2
 
2
E = wt WR i 
RT i – RA i  + wp WP j
PF j – POj 
i= 1 j=1
● where RTi = test data
● PF = final parameters
● RAi = analysis results
● POj = original parameters
● WRi = test weighting
● WPj = parameter weighting
● wt = weighting for test as a whole
● wp = weighting for model parameters as a whole
● SOL 200 minimizes the objective (E) subject to design constraints.

● Example:
● Disk drive enclosure with 1406 grid points, 1354 plate elements, four
design variables (plate thicknesses), and four measured flexible modes
● Minimize the difference between computed and measured resonant
frequencies.

● Example (Cont.)
● Test data weighted 100 times more than the baseline model parameters.
● Put upper and lower bounds on allowable plate thicknesses.
● Five iterations to converge
● Results:
Frequencies (Hz)
Test Baseline Final
346 234 346
1307 892 1295
1567 1165 1460
1678 1267 1695
Plate Thicknesses (in.)

Baseline Lower Bound Upper Bound Final
.08 .05 .125 .125
.12 .05 .20 .075
.10 .05 .15 .084
.20 .10 .30 .227

Error versus Design Cycle
Plate Thickness versus Design Cycle

● Can also use SOL 200 to match frequency response test
data.
● Can match static results as well.
● For best results, use as much test data as feasible
(including the weight of the structure).
● Also, limit design variables to only the parameters that are
truly uncertain.

REFERENCES
 Blakely, K. and Dobbs, M., “Integrated System Identification: The Union of Testing and Analysis,”
Proceedings First International Modal Analysis Conference, November 1982.
 Kientzy, D., Richardson, M. and Blakely, K., “Using Finite Element Data to Set Up Modal Tests,”
Sound and Vibration, June 1989.
 Blakely, K., “Updating MSC.Nastran Models to Match Test Data,” Proceedings MSC World Users’
Conference, March 1991.
 Blakely, K. and Rose, T., “Cross-Orthogonality Calculations for Pre-Test Planning and Model
Verification,” Proceedings MSC World Users’ Conference, May 1993.
 Blakely, K., “Matching Frequency Response Test Data with MSC.Nastran,” Proceedings MSC World
Users’ Conference, June 1994.


MD Nastran - Dynamic Analysis

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MD Nastran - Dynamic Analysis

Uploaded by

Copyright:

Available Formats

MSC.

MD Nastran - Dynamic Analysis

Part Number: MDNA*R2*Z*Z*Z*SM-NAS102-NT1 Copyright 2007 MSC.Software Corporation

Copyright 2007 MSC.Software Corporation

Copyright 2007 MSC.Software Corporation

NAS102, Section 1, March 2007

OVERVIEW OF DYNAMIC ANALYSIS

Finite Element Interfacing

NAS102, Section 1, March 2007

SINGLE DOF SYSTEM

●Fundamental and derived units

NAS102, Section 1, March 2007

● Error in units is the number one cause of modeling

● Most common errors are in mass and damping units.

● MD Nastran contains no built-in set of units. The

● Consistent Units: e.g. N, t, mm, s or N, kg, m, s

NAS102, Section 1, March 2007

SINGLE DOF SYSTEM - UNDAMPED FREE

SINGLE DOF SYSTEM - UNDAMPED FREE

Disp lace m ent

NAS102, Section 1, March 2007

SINGLE DOF SYSTEM - DAMPED FREE

●Fraction of critical damping

●Damped natural frequency

NAS102, Section 1, March 2007

SINGLE DOF SYSTEM - DAMPED FREE

● The usual analysis case is underdamped.

● Structures typically have damping in the 0-10% range.

NAS102, Section 1, March 2007

SINGLE DOF SYSTEM - UNDAMPED

where = forcing frequency

NAS102, Section 1, March 2007

SINGLE DOF SYSTEM - UNDAMPED

• P/k is the static response.

NAS102, Section 1, March 2007

SD OF Syste m Freq ue n cy (Hz)

NAS102, Section 1, March 2007

SINGLE DOF SYSTEM - DAMPED

● The transient solution decays rapidly and is of no interest.

SINGLE DOF SYSTEM - DAMPED

• Magnification factor 0 (no response)

NAS102, Section 1, March 2007

NAS102, Section 1, March 2007

MULTI-DEGREE OF FREEDOM SYSTEM

● Now the equation of motion becomes

NAS102, Section 1, March 2007

NAS102, Section 1, March 2007

NAS102, Section 1, March 2007

FINITE ELEMENT DYNAMIC MODELING

NAS102, Section 1, March 2007

• Quick Reference Guide

MD NASTRAN DOCUMENTATION (Cont.)

● Documentation also available in online form

NAS102, Section 1, March 2007

TEXT REFERENCES ON DYNAMIC ANALYSIS

• W. C. Hurty and M. F. Rubinstein, Dynamics of Structures, Prentice-

TEXT REFERENCES ON DYNAMIC ANALYSIS

NAS102, Section 1, March 2007

NAS102, Section 1, March 2007

NAS102,Section 2, March 2007

NAS102,Section 2, March 2007

● Examples in this course will use free format and replicators.

NAS102,Section 2, March 2007

● Bookkeeping is done via the matrices that define the relationships

Part Number: MDNAR2Z*ZZSM-NAS102-NT1 Copyright 2007 MSC.Software Corporation