cancel
Showing results for 
Search instead for 
Did you mean: 

Simulation on NX with the solver NX NASTRAN

Experimenter
Experimenter
I am looking for someone who could help me for modal analysis with the software NX and the sovler NX NASTRAN. Indeed, I would like to represent the hydrostatic pressure exerted by a fluid on a tank's shell. As a result, would someone know how could I do ?

For informations, it's better to consider the liquid as a virtual mass. Therefore, the virtual mass carrys on the tank's shell a hydrostatic pressure represented by MFLUID. But i don't know how to write this in the INPUT FILE (.DAT). Particularly in the case control together with in the main bulk and the DECK NASTRAN. I use the solution type SEMODES 103 RESPONSE SIMULATION, and the hydrostatic pressure takes part in the static subcase. At this moment, I just did a real pressure hydrostatic, id est it's represented by PLOADD4 like a force, but for modal analysis, it's not exact if I launch a calcul with this force. Indeed, we are not interested in the effects of hydrostatic pressure, but we are interested in frequential responses of the system.I need your help, if someone has some ideas or contacts, tell me.


NB : when i will have the MFLUID in the case control, Where I have to put MFLUID DEFINITION in the main bulk ? And HOW ?

Furthermore, I would to increase the disk memory, so I did the following statement :

$*$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$*

$* NASTRAN

$*

$*$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$*

INIT MASTER(S)

NASTRAN BUFFSIZE=32769

NASTRAN ITER=NO

$*

$*$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$*

$* FILE MANAGEMENT

$*

$*$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$*

ASSIGN SCR1='SCRATCH.SCR1'DELETE

ASSIGN SCR2='SCRATCH.SCR2'DELETE

ASSIGN DB1 ='SCRATCH.DB1'DELETE

ASSIGN DB2 ='SCRATCJ.DB2'DELETE

INIT SCRATCH LOGI=(SCR1(250000)),SCR300=(SCR2(250000)

INIT DBALL LOGI=(DB1(250000),DB2(250000)

$*

$*$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$*

$* EXECUTIVE CONTROL

$*

$*$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$*

ID,NASTRAN,nouveau1sim1-solution_2

SOL 103

TIME 999

DIAG 44

CEND

$*


Do you think it's enough to launch calcul with big models ? If not, could you give me another solution please.
I am waiting for your responses, and I am looking for your help.
Could you give me some help please ? Or may be we could call us.
Thank you for your attention and for your time.
Best regards.
Tibo
8 REPLIES

Re: Simulation on NX with the solver NX NASTRAN

Creator
Creator
Virtual mass has nothing to do with hydrostatic pressure. It models the amount of added mass that you get from submerging a structure in a fluid, but it does not apply any load to the fluid. NX Nastran doesn't have a hydrostatic pressure input that automatically calculates the pressure as a function of the distance below the surface. However, FEMAP does allow you to define a pressure as a function of location, which makes it fairly easy to define a pressure that varies lineary as distance from surface.
Paul Blelloch, Ph.D.
Director, Aerospace Analysis
ATA Engineering, Inc.
11995 El Camino Real
San Diego, CA 92130
(858) 480-2065

Re: Simulation on NX with the solver NX NASTRAN

Creator
Creator
In reading your post more carefully I see that you're actually doing a modal analysis. What effect do you want to include in your modes? Is it the added mass of the fluid or the effect of the pressure loading on the modes. If it's the former then MFLUID is what you want. If it's the latter you want to set up a static load case with the hydrostatic pressure and use the STATSUB Case Control to use that as a static load case for the modal calculation. If you want to include the mass effects, the MFLUID is what you want. I don't believe that NX supports MFLUID, though FEMAP does. The short version is that you need an MFLUID case control that points to an MFLUID bulk data card, which in turn points to one or more ELIST cards to define the shell elements that are loaded with fluid. You'll need to read the Nastran Quick Reference Guide, but MFLUID allows you to enter the density of the fluid and the location of a free surface if you have a free surface. This does not create a hydrostatic pressure though, it just adds mass.
Paul Blelloch, Ph.D.
Director, Aerospace Analysis
ATA Engineering, Inc.
11995 El Camino Real
San Diego, CA 92130
(858) 480-2065

Re: Simulation on NX with the solver NX NASTRAN

Experimenter
Experimenter
Thank you very much Pbelloc. Indeed, I would like to include the mass effect of the fluid. Therefore, MFLUID seems correct for this type of study. According to your mail, I should create in the case control a new mass named MFLUID, in the first subcase STATIC (instead of hydrostatic pressure)? Furthermore, MFLUID has to point to MFLUID bulk data card. So, I have to precise the ID of MFLUID in the subcase, but how can i write the MFLUID in the subcase STATIC knowing that STATIC SUBCASE can only accept LOAD whereas MFLUID is not a load ? In addition to these preview points, In the BULK DATA ENTRY, how can I precise the free surface ? Should I create a label before loading the input file (.dat) ? What is the form of the writing of MFLUID in the bulk data ? What is the ID for MFLUID in the case control knowing it's a load for the static subcase.
I'm going to show what I have done for the hydrostatic pressure :
ECHO = NONE
BGSET = 100
SUBCASE 1
LOAD = 21
SUBCASE 2
LOAD = 100
SUBCASE 3
LOADSET = 103
METHOD = 103
STATSUB = 2

Now, for the MFLUID, I thought about the following sentence :
ECHO = NONE
BGSET = 100
MFLUID = ? (I don't know)
SUBCASE 1
LOAD = ? (for the MFLUID)
SUBCASE 2
LOAD = 100
SUBCASE 3
LOADSET = 103 (corresponding to the gravity)
METHOD = 103
STATSUB = 2


Now for the bulk data entries, what i have to pu for MFLUID ?
Should I put first of all MFLUID and characterics associated to the fluid and the surface, how can i do ? (I read the PDFs of SIEMENS, and it's not efficient for me) And for the second time, should I put LOAD correspondig to MFLUID so that the case control can understand what we refer to ?
Thank you in advance, you are really helpfull for me.
I am looking forward to read your response.
Regards

Re: Simulation on NX with the solver NX NASTRAN

Creator
Creator
MFLUID has nothing to do with the static load, so it belongs in the modal subcase. In fact I wonder if you really want the static load. Do you want to account for differential stiffness effects of the load on the modes? If all you want is the mass leave out the static subcase. If you do want to include the effect of the static load, just put MFLUID in the modal subcase. What you want is something like:

SUBCASE 3
METHOD = 103
STATSUB = 2
MFLUID = 100
BEGIN BULK
MFLUID, 100, ....
ELIST,....

The mass has nothing to do with the static load, so the set ID for MFLUID has nothing to do with the set ID for the load.
Paul Blelloch, Ph.D.
Director, Aerospace Analysis
ATA Engineering, Inc.
11995 El Camino Real
San Diego, CA 92130
(858) 480-2065

Re: Simulation on NX with the solver NX NASTRAN

Experimenter
Experimenter
Thank you,
So If I put MFLUID in the case control, and that there was only Hydrostatic Pressure (H.P) in the static subcase, I have to dismiss the static subcase if i don't want to include the effect of static load, is that correct if i have no static subcase ? But if i keep the H.P in the static subcase, and in addition to that I put the MFLUID in the third subcase (modal subcase) is that correct ? Because I would like to include the effect of the static load. So, I'm going to show you what I think about what I have to do :

ECHO = NONE
BGSET = 100
SUBCASE 1
LOAD = 21
SUBCASE 2
LOAD = 100
SUBCASE 3
LOADSET = 103
METHOD = 103
STATSUB = 2
MFLUID = 100

After, there is the bulk data entries.
So, how can I point to the ELIST Card ?
And, how can i describe MFLUID for the free surface, liquid density ? Should I create a label number in the CAE INTERFACE ?
What is exactly ELIST ?
I know these questions are very stupid, but, in the Quick Reference Guide, and in the Installation Guide of NX NASTRAN, all these things are not really understandable.
I am really waiting for your response.
Thank you very much.
Friendly and Respectfully,
TIBO

Re: Simulation on NX with the solver NX NASTRAN

Creator
Creator
Your case control setup looks correct, though there's no purpose to your 1st subcase since you're not using it for anything. The QRG for the MFLUID tells you how to set it up. I forget the exact syntax, but you have a set ID which matches the set ID on the MFLUID Case Control, fluid density, free surface height, a coordinate system ID (must have Z axis pointing up out of the surface if I remember correctly) and one of two ELIST entries. One is for elements that are loaded on both sides by fluid and one for elements that are only loaded on one side. There are also a couple of symmetry options, which should be both set to N if you aren't using symmetry options. The ELIST card is simply a list of elements. It has a set ID pointed to by the corresponding MFLUID card.
Paul Blelloch, Ph.D.
Director, Aerospace Analysis
ATA Engineering, Inc.
11995 El Camino Real
San Diego, CA 92130
(858) 480-2065

Re: Simulation on NX with the solver NX NASTRAN

Experimenter
Experimenter
Thank you very much Mr PBELLOC,
Do you an example for the bulk data card for MFLUID and ELIST ?
Regards

Re: Simulation on NX with the solver NX NASTRAN

Innovator
Innovator
The user interface for a virtual fluid mass analysis is simple and straightforward. The sketch below illustrates some of the features.

The fluid/structure interface is defined with ELIST Bulk Data, which specify a set of wetted TRIA3 and QUAD4 elements that define the structural portion of the fluid boundary.
Each fluid volume is defined on an MFLUID Bulk Data input, which defines the fluid density, the ELISTs, and other boundaries.
A set of MFLUID volumes are requested in the case control request MFLUID = SID.
If pressure outputs are desired they will be printed if the case control request MPRES = is used to define a set of elements. The elements must also be active in an ELIST/MFLUID volume.
An example set of input data for a small problem is shown below:

Case Control
MFLUID = 25MPRES= ALLetc.Bulk Data
1 ,2 ,3 ,4 ,5 ,6 ,7 ,8 ,9 ,10$MFLUID, SID , CID , ZFS , RHO , ELIST1, ELIST2, PLANE1, PLANE2MFLUID , 25 , 0 , 115.4, .0246, 255 , , S$ELIST , ELID ,E1 , E2 , E3 , etc.ELIST , 255 , 1 , THRU , 25ETC...In the input above we have defined a fluid volume with a free surface normal to the basic z direction. The xz plane (PLANE1) is a plane of symmetry and the density is 0.0246. ELIST set 255 defines the positive faces of elements in the range 1 through 25 (missing numbers are allowed) as the structural/fluid boundary.
The connected elements may be wetted on one or both sides by the same volume. For instance a baffle plate extends partially into a tank and have two sides in the same fluid volume. If a plate completely separates the two parts of the tank, it is recommended that a separate volume be used for each side. Each side of an element should be listed on a separate ELIST.
Special terms are calculated if enclosed fluid volumes do not have a free surface or a plane of anti-symmetry. Otherwise, the incompressible fluid would produce a spurious low frequency mode with a mode shape corresponding to the volume change. This mode will not affect the structural response, but it is eliminated. Because the free surface allows the fluid volume to expand, it does not have these problems.
A free surface is defined as an x-y plane in any local rectangular coordinate system. The user simply specifies a value of z as the upper limit of the fluid volume. Different fluid volumes may have different levels and orientation. It is not required that the surface coincide with the element properties since partially wetted elements are allowed.

The virtual mass fluid option may be used in all NX Nastran dynamics solutions, including the following special approaches:

    Superelements – (Residual Only )
    Nonlinear Analysis – (SOL 129)
    Optimization – (does not create sensitivity matrices.)
Best regards,
Blas.

Figure 4-3. Virtual Mass Terminology