Need HELP on Generating mnf. Files for CBeam Elements!!

[Liangyiteng888]

Hello,

I currently need to use the Cbeam element to model a flexible beam and import the mnf. file generated into Adams in order to do a multibody analysis; however, the mnf. file can only be generated when 3 elements are used. If 4 or more elements are used, mnf. file will not be generated. I have no idea at all how to solve it. I have checked the .bdf file for both 3-element and 4-element (shown below) and cannot find anything wrong.

3-element:
$ NASTRAN input file created by the Patran 2014.1 64-Bit (Student
$ Edition) input file translator on May 24, 2016 at 10:45:53.
$ Direct Text Input for Nastran System Cell Section
$ Direct Text Input for File Management Section
$ Direct Text Input for Executive Control
$ Normal Modes Analysis, Database
SOL 103
CEND
$ Direct Text Input for Global Case Control Data
TITLE = MSC.Nastran job created on 23-May-16 at 23:16:36
ECHO = NONE
$ Using Nastran default values for RESVEC
GPSTRAIN=ALL
GPSTRESS=ALL
ADAMSMNF FLEXBODY=YES,FLEXONLY=YES,OUTGSTRN=YES,OUTGSTRS=YES,
MINVAR=FULL
SUBCASE 1
SUBTITLE=Default
METHOD = 1
VECTOR(SORT1,REAL)=ALL
SPCFORCES(SORT1,REAL)=ALL
$ Direct Text Input for this Subcase
BEGIN BULK
$ Direct Text Input for Bulk Data
PARAM POST 0
PARAM PRTMAXIM YES
DTI UNITS 1 KG N M S
EIGRL 1 26 0 MASS
$ Elements and Element Properties for region : 1mbeam
PBEAML 1 1 ROD
.0398942
$ Pset: "1mbeam" will be imported as: "pbeaml.1"
CBEAM 1 1 1 2 0. 1. 0.
CBEAM 2 1 2 3 0. 1. 0.
CBEAM 3 1 3 4 0. 1. 0.
$ Referenced Material Records
$ Material Record : Al
$ Description of Material : Date: 23-May-16 Time: 23:12:34
MAT1 1 7.+10 2.7+10 3000.
$ Nodes of the Entire Model
GRID 1 0. 0. 0.
GRID* 2 .400000005960465 0.
* 0.
GRID* 3 .800000071525574 0.
* 0.
GRID* 4 1.20000004768372 0.
* 0.
$ Loads for Load Case : Default
SPOINT 5 THRU 30
QSET1 5 THRU 30
ASET1 123456 1
$ Referenced Coordinate Frames
ENDDATA 2abed2a9

4-element;
$ NASTRAN input file created by the Patran 2014.1 64-Bit (Student
$ Edition) input file translator on May 24, 2016 at 12:38:31.
$ Direct Text Input for Nastran System Cell Section
$ Direct Text Input for File Management Section
$ Direct Text Input for Executive Control
$ Normal Modes Analysis, Database
SOL 103
CEND
$ Direct Text Input for Global Case Control Data
TITLE = MSC.Nastran job created on 23-May-16 at 23:16:36
ECHO = NONE
$ Using Nastran default values for RESVEC
GPSTRAIN=ALL
GPSTRESS=ALL
ADAMSMNF FLEXBODY=YES,FLEXONLY=YES,OUTGSTRN=YES,OUTGSTRS=YES,
MINVAR=FULL
SUBCASE 1
SUBTITLE=Default
METHOD = 1
VECTOR(SORT1,REAL)=ALL
SPCFORCES(SORT1,REAL)=ALL
$ Direct Text Input for this Subcase
BEGIN BULK
$ Direct Text Input for Bulk Data
PARAM POST 0
PARAM PRTMAXIM YES
DTI UNITS 1 KG N M S
EIGRL 1 26 0 MASS
$ Elements and Element Properties for region : 1mbeam
PBEAML 1 1 ROD
.0398942
$ Pset: "1mbeam" will be imported as: "pbeaml.1"
CBEAM 1 1 1 2 0. 1. 0.
CBEAM 2 1 2 3 0. 1. 0.
CBEAM 3 1 3 4 0. 1. 0.
CBEAM 4 1 4 5 0. 1. 0.
$ Referenced Material Records
$ Material Record : Al
$ Description of Material : Date: 23-May-16 Time: 23:12:34
MAT1 1 7.+10 2.7+10 3000.
$ Nodes of the Entire Model
GRID 1 0. 0. 0.
GRID* 2 .300000011920929 0.
* 0.
GRID* 3 .600000023841858 0.
* 0.
GRID* 4 .900000035762787 0.
* 0.
GRID* 5 1.20000004768372 0.
* 0.
$ Loads for Load Case : Default
SPOINT 6 THRU 31
QSET1 6 THRU 31
ASET1 123456 1
$ Referenced Coordinate Frames
ENDDATA 4f161a8c

The only difference is that 4-element has one more grid point...

Please help and I appreciate anyone who can give me suggestion.

 

[dmapguru]

Simple straight line beam models can present problems for eigensolution because of ill-defined mass. By default, MSC Nastran computes diagonal mass matrices for all elements, and very few ever have rotational masses, even when the coupled mass formulation is requested with the user parameter PARAM,COUPMASS,1. You can’t use COUPMASS with your model as a limitation in the MNF generation capability requires diagonal mass; if you try to use COUPMASS, you’ll get a FATAL error.

For elaborate 3D models, this does not pose a problem because even global rotational modes of an elaborate model consist in translational motion of many dof, but simple beams, particularly collinear ones, can have rotations without translation. In addition, for MNF generation, the condensation process couples the mass matrix, but little for simple small problems.
Try adding small rotational inertia terms to the GRID points using CONM2 entries like this:
CONM2,92,2
,1-6,,1-6,,,1-6

You can also force the solver to use a different eigensolution method. By default, the Lanczos method is selected, but this method is adapted to solving huge problems. At the handful of dof level, it can fail. Try using either the advanced Householder or modified Givens methods of eigensolution for small problems, i.e. instead of the EIGRL entry, use

EIGR,1,AHOU,,,26

or

EIGR,1,MGIV,,,,26

If you are just trying out this technique, consider a slightly more complicated model which is not collinear; my guess is most of these issues will disappear.