Random Vibration Analysis Setup Using Large Mass Method Versus Direct Method

Anonymous — Tue, 05 Dec 2017 21:26:33 GMT

Greeting All,

I am running into an issue where I am seeing hugely different stress values between using the two different methodologies for setting Random Response in Nastran. The Large Mass Method gives me a peak stress of 40 ksi in my part and the "Direct Method" gives me a peak stress of 6.70 ksi. The contour plots as well are immensely different.

So for the Large mass method, I am doing the following:

1) Creating a large mass element at the base excitation node (1e8) and tying back to fastener holes using RBE.

2) Using PSD input values as g^2/hz x (386.4)^2.

3) Constraining All DOFs of the base excitation node except for the direction to vibe in.

4) Also applying a force (in lbs) to the base node of 1e8 in the direction to vibe in.

5) All my material cards have mass density units and all my concentrated masses are mass units.

And the Direct Method, I am doing the following:

1) Taking out the large mass element and force and just applying an enforced acceleration at the base excitation node of 386.4.

2) Constraining all 6 DOFs at the base node.

3) Using PSD input values as g^2/hz (i.e. not scaling).

4) All my material cards have mass density units and all my concentrated masses are mass units.

Reading these two methods, what am I doing wrong?? Why are the stresses so different and contours between the two methods?? Anybody please help.

Thanks,

George

Re: Random Vibration Analysis Setup Using Large Mass Method Versus Direct Method

KubliJ — Thu, 07 Dec 2017 00:27:32 GMT

Hi George,

Sorry, not sure what the issue might be. Where is the max stress on each model? If they are near the applied load then that might explain the difference. Otherwise, could you upload the model for me to look at? If you are worried about sharing the model, you can host it on a third party sight and PM me a link.

Thanks,

James

Re: Random Vibration Analysis Setup Using Large Mass Method Versus Direct Method

Anonymous — Thu, 07 Dec 2017 16:07:46 GMT

James,

I can send you the Case Control for the two input decks (Large mass Method and Direct Method). And maybe you can examine the files and see if I am missing any command lines or something of that nature.

I copied in the case control for the two files below. Keep in mind that the node IDs in acceleration request sets and element IDs for the force request sets differ and that the FEA entities are all deleted due to proprietary issues. These are files for exactly the same model...no additions have been made structure-wise. The only part that needs attention is the sections that are shown. I also just left the CONM2 large mass element in the large mass method file...all other nodes, properties and elements have been removed from each of the two files.

DIRECT METHOD:

XYDATAGENERATE,1,123,1,GRID
NASTRAN SYSTEM(319)=1
ID Dual,Femap
SOL SEMFREQ
CEND
TITLE = Random AY
ECHO = NONE
DISPLACEMENT(SORT2,PLOT,PHASE,PSDF) = ALL
SET 1 = 176019,417868,560678
ACCELERATION(SORT2,PLOT,PHASE,PSDF) = 1
SET 2 = 85 THRU 100,102 THRU 103,
    105 THRU 135,142 THRU 147
FORCE(SORT2,PLOT,PHASE,CORNER,PSDF) = 2
STRESS(SORT2,PLOT,PHASE,CORNER,PSDF) = ALL
SPC = 2
FREQUENCY = 1
METHOD = 1
SDAMPING = 2
RANDOM = 200
SUBCASE 1
DLOAD = 3
BEGIN BULK
$ ***************************************************************************
$   Written by : Femap with NX Nastran
$   Version    : 11.3.2
$   Translator : NEi Nastran
$   Date       : Wed Dec 06 19:54:58 2017
$ ***************************************************************************
$
PARAM,FREQRESPRSLTOUT,OFF
PARAM,RANDRESPRSLTOUT,OFF
PARAM,RANDRESPINVLEVEL,2
PARAM,RIGIDELEM2ELAS,OFF
PARAM,ELEMGEOMCHECKS,OFF
PARAM,PRGPST,YES
PARAM,OGEOM,NO
PARAM,AUTOSPC,YES
PARAM,GRDPNT,0
RANDPS       200       1       1      1.      0.       5
$ Femap with NX Nastran Function 5 : Qualification 18.84 Grms g2/hz
TABRND1        5     LOG     LOG                                        +
+            20.     .08     54.     .32    679.     .32   2000.     .04+
+       ENDT
CORD2C         1       0      0.      0.      0.      0.      0.      1.+FEMAPC1
+FEMAPC1      1.      0.      1.
CORD2S         2       0      0.      0.      0.      0.      0.      1.+FEMAPC2
+FEMAPC2      1.      0.      1.
$ Femap with NX Nastran Function 2 : Q-damping = 50
TABDMP1        2       Q                                                +
+            20.     50.   2000.     50.ENDT
PARAM,HFREQ,2000.
EIGRL          1           2000.                                    MASS
$ Femap with NX Nastran Load Set 3 : Random AY
RLOAD2       101     102                       6            ACCE
SPCD         102 560678       2   386.4
DLOAD          3      1.      1.     101
FREQ1          1     20.     20.      99
FREQ4          1     20.   2000.    .035       7
TABLED2        6      0.                                                +
+             0.      1.      1.      1.    ENDT
$ Femap with NX Nastran Constraint Set 2 : All DOFS
SPC1           2 123456 560678

LARGE MASS METHOD

XYDATAGENERATE,1,123,1,GRID
NASTRAN SYSTEM(319)=1
ID Dual,Femap
SOL SEMFREQ
CEND
TITLE = Random Y
ECHO = NONE
DISPLACEMENT(SORT2,PLOT,PHASE,PSDF) = ALL
SET 1 = 32105,162722,627436
ACCELERATION(SORT2,PLOT,PHASE,PSDF) = 1
SET 2 = 74 THRU 95,106 THRU 115,
    117 THRU 123,126 THRU 131,
    134 THRU 139
FORCE(SORT2,PLOT,PHASE,CORNER,PSDF) = 2
STRESS(SORT2,PLOT,PHASE,CORNER,PSDF) = ALL
SPC = 3
FREQUENCY = 2
METHOD = 1
SDAMPING = 2
RANDOM = 200
SUBCASE 1
DLOAD = 2
BEGIN BULK
$ ***************************************************************************
$   Written by : Femap with NX Nastran
$   Version    : 11.3.2
$   Translator : NEi Nastran
$   Date       : Wed Nov 29 14:45:19 2017
$ ***************************************************************************
$
PARAM,FREQRESPRSLTOUT,OFF
PARAM,RANDRESPRSLTOUT,OFF
PARAM,RANDRESPINVLEVEL,2
PARAM,RIGIDELEM2ELAS,OFF
PARAM,ELEMGEOMCHECKS,OFF
PARAM,PRGPST,YES
PARAM,OGEOM,NO
PARAM,AUTOSPC,YES
PARAM,GRDPNT,0
RANDPS       200       1       1      1.      0.       3
$ Femap with NX Nastran Function 3 : Qualification 18.84 Grms
TABRND1        3     LOG     LOG                                        +
+            20. 11944.4     54.47777.59    679.47777.59   2000.5972.198+
+       ENDT
CORD2C         1       0      0.      0.      0.      0.      0.      1.+FEMAPC1
+FEMAPC1      1.      0.      1.
CORD2S         2       0      0.      0.      0.      0.      0.      1.+FEMAPC2
+FEMAPC2      1.      0.      1.
$ Femap with NX Nastran Function 2 : Q damping = 50
TABDMP1        2       Q                                                +
+            20.     50.   2000.     50.ENDT
PARAM,HFREQ,2000.
EIGRL          1           2000.                                    MASS
$ Femap with NX Nastran Load Set 2 : Random Y
$ Femap with NX Nastran Function 1 : Load vs Frequency
TABLED2        1      0.                                                +
+            20.      1.   2000.      1.ENDT
RLOAD2       101     102                       1            LOAD
FORCE        102 162722       0      1.      0.    1.+8      0.
DLOAD          2      1.      1.     101
FREQ1          2     20.     20.      99
FREQ4          2     20.   2000.    .035       7
$ Femap with NX Nastran Constraint Set 3 : Random Y
SPC1           3   13456 162722
+         30800. 30800. 27000.
CONM2        125 162722       0    1.+8      0.      0.      0.

topic Re: Random Vibration Analysis Setup Using Large Mass Method Versus Direct Method in Inventor Nastran Forum

Random Vibration Analysis Setup Using Large Mass Method Versus Direct Method

Re: Random Vibration Analysis Setup Using Large Mass Method Versus Direct Method

Re: Random Vibration Analysis Setup Using Large Mass Method Versus Direct Method