bailout -1

[Walterke]

(Using nx nastran on siemens NX 7.5)

Sometimes when I'm performing a simulation on a file, when the calculation is complete it gives me the error 'no results are found'. In 99% of the cases, changing the parameter bailout to '-1' will fix the problem and the result will show me what I did wrong (usually, there are too many degrees of freedom because a part is not connected properly or things like that)

My question is: What exactly is the difference between calculating with bailout 0 and bailout -1? What exactly is being calculated when bailout is on 0 and it won't give me any results?

This question probably comes from lack of knowledge of the subject but I'm hoping someone here can explain this to me.

Thanks in advance.

NX 7.5.5.4 with Teamcenter 8 on win7 64
Intel Xeon @3.2GHz
8GB RAM
Nvidia Quadro 2000

 

[dwallace1971]

Normally, when the solver encounters a pivot ratio error, indicating an underconstrained model, it will exit the program and give you those fatal pivot ratio errors we all love. When you set bailout = -1, you instruct the solver to continue with the solution when it encounters such an error. This allows you to check your displacements to determine what is unconstrained.

"On the human scale, the laws of Newtonian Physics are non-negotiable"

 

[Walterke]

Do you know of a way to let the program tell you whether or not there is an error like that and gives you the option to either continue the calculation or abort it?

NX 7.5.5.4 with Teamcenter 8 on win7 64
Intel Xeon @3.2GHz
8GB RAM
Nvidia Quadro 2000

 

[ESPcomposites]

Walterke,

I don't think so. I think the intent is that this should not be something to ignore and you must consciously override it. If they made it too easy to continue, then users might take that path rather than fixing (or thinking about) their models.

Companies probably wouldn't want their users to have this option either, as it would encourage bad practice. The deck should also be something that anyone can run and get the same answer (on the same version), without intervention after the run starts. This way, if you share the deck, the person on the other end would get the same result and know what adjustments you may have made.

Brian

http://www.espcomposites.com

 

[Walterke]

I just think it's a waste of time that you have to redo the first part of the calculation.
I don't see a difference between:
1)making calculation
2)results not found
3)manually setting bailout to -1
4)again make calculation
5)get result to see what's wrong

and:
1)make calculation
2)pivot error: do you want to continue(+whatever warning message you find necessary)?
3)yes: parameter bailout gets set to -1
4)calculation continues
5)get result to see what's wrong

The only difference is that the first way costs more time.

NX 7.5.5.4 with Teamcenter 8 on win7 64
Intel Xeon @3.2GHz
8GB RAM
Nvidia Quadro 2000

 

[dwallace1971]

Does it really take that much time? The pivot ratios are checked immediately after the matrices are built and before the solution begins, so any errors are generated early in the process. It seems to me that if there were a popup menu that said something like "you have pivot ratio errors, do you want to continue", many analysts, particularly the inexperienced, might be tempted to simply push the button and drive on. As it is, at least one is encouraged to learn what the error is and what caused it.

"On the human scale, the laws of Newtonian Physics are non-negotiable"

 

[dwallace1971]

Additionally, one should always sketch out free-bodies for all the components in an assembly before beginning to build his model. Doing this will greatly reduce the likelihood of pivot errors in the first place.

"On the human scale, the laws of Newtonian Physics are non-negotiable"

 

[Walterke]

The pivot ratios are checked immediately after the matrices are built and before the solution begins, so any errors are generated early in the process.

Siemens NX 7.5 's NX nastran seems to do the entire solution regardless of there being pivot errors or not. As I said in my opening post, both calculations (bailout 0 or -1) take the same amount of time to complete.

The calculation starts with a "solution information" tab giving some random numbers of which I have no idea what they mean.

Then after a few seconds/minutes (depending on the size of the calculation) a new tab opens (sparse Matrix Solver) showing a graph describing the amount of equaltions completed with "supernode" (again, no idea what those are) on the X-axis.

NX 7.5.5.4 with Teamcenter 8 on win7 64
Intel Xeon @3.2GHz
8GB RAM
Nvidia Quadro 2000

 

[BlasMolero]

Dear Walterke,
Not exactly, the nx nastran solver stop when realized that the matrix is singular, the moment depends of the model size, if contact are present or not, etc..

I have reported to SIEMENS guys that I don't like the way that NX Advanced Simulation manages the nx nastran solver error, I have suggested many times to pop-up in the screen that an error exist, this is the way that runs FEMAP pre/post-processor when any solver error exist, the user receives in the screen a window where he/she can revise the F06 file for error explanation or close the window.

Well, let's hope in the future this changes, because novice users are really frustrated when no results exist, and not any error appear in the screen ... well, patience!!.

Best regatds,
Blas.

~~~~~~~~~~~~~~~~~~~~~~
Blas Molero Hidalgo
Ingeniero Industrial
Director

IBERISA
48011 BILBAO (SPAIN)
WEB:

http://www.iberisa.com

Blog de FEMAP & NX Nastran:

http://iberisa.wordpress.com/

 

[Walterke]

Thanks Blas,

I'm glad to know there are people thinking the same way I do.

NX 7.5.5.4 with Teamcenter 8 on win7 64
Intel Xeon @3.2GHz
8GB RAM
Nvidia Quadro 2000