Error making element tables: Can not get data in database!

[sladisch]

Hi,

I have a problem with an existing model. When I try to run my analysis it says that it can not access the database. It seems to be only one part wich is effected. After the solver is running I can see that there are a lot of elemts missing. I'll try to attach some Images.

Here is the error message:

Autodesk (R) Simulation Nonlinear Analysis Decoder for AGSDB
 2013.02.00.0010-W64/X64 02-Oct-2012
 Copyright (c) 2012, Autodesk, Inc. All rights reserved.

 Model [C:\Users\ ... \RS-Typ5.ds_data\1\ds] opened 

 Zapping NODES, ELEMENTS, NODECOND, and ELEMCOND tables ...
 Start to make nonlinear analysis AGSDB model ...
 Made 593343 node tables
 Start to make nodal condition tables ...
 Start to make element tables ...
   Element Part(29): Brick element
 Error: Can not get data in database!
        Skipped the rest.
   Made 2542 element tables
   Element Part(28): Brick element
 Error: Can not get data in database!
        Skipped the rest.
   Made 5234 element tables
   Element Part

[sladisch]

Pushing this Problem ...

I have another model with the same problem. I have no idea when or how the problem occurs.

Any Idea how I can solve this problem?

Regards

[AstroJohnPE]

This is just a guess, but I think you are encountering a timing issue. Some step in the process is trying to open a database, but another process still has the database open. I encounter similar things during an analysis. I'll get an error message that one of the temporary files could not be deleted. (This is during a transient stress analysis. In my case, it is not a problem that the file cannot be deleted.)

Is the problem repeatable? If you exit the software, restart it, and open the model, does the exact same problem occur? (The same error in the same part with the number of elements created, and so on.)

Maybe it is related to your cores, such as if you have a dual or quad core computer, or if you have hyperthreading activated. Unfortunately, I do not know enough about this aspect to provide any help.

[sladisch]

Hi!

First of all thanks for your reply.

I tried to close the software, reboot the system (maybe a memory allocation problem) but no change. Even creating a archive and expand it into a new model didn't solve the problem. The curious thing is, that another scenario in the same model runs very well.

I'm trying to run an MES-Analysis using the default parameter set. I'm not sure if I can reproduce it with a fresh model.

Regards

Stefan

[sladisch]

... even remeshing the model with a different mesh size doesn't solve the Problem.

[AstroJohnPE]

Hi Stefan,

Looking at the image attached to the original post triggered a question. Is the element type for the (blue) part correct for the type of mesh? For example, the default mesh type is brick: solid elements with 8, 6, 5, or 4 nodes. if the mesh type was not changed to tetrahedron, and if the element type was set or changed to tetrahedron, then the software is expecting to find only 4 node elements. It would "choke" when it found all of the other configurations (8, 6, and 5 node elements).

Otherwise, you would need to attach an archive of the model and hope that someone else can see something. The model looks large, so follow the instructions in "Create, Post, or Provide an Archive of your model" to make it smaller if needed.

[sladisch]

Hi John,

I checked the element types and the mesh type with no result. Everything seems to be well defined.

I uploaded my model to my dropbox. I whould be glad if anyone could check the model.

Dropbox link to my model

Regards

Stefan

[AstroJohnPE]

Hi Stefan,

When I suppressed all of the parts except for part 1, I was still getting the same errors! I could not see any problems with the input, so I simplified the model by deleting the surface to surface contact. The model was able to do the "Check Model".

My suggestion is to delete all of the contact, do a "Analysis > Analysis > Check Model" to confirm that that will work. Then redo the surface to surface contact.

[sladisch]

Hi John,

I tested to build a new model from scratch but with the same DWG-File from Fusion but I ran into the same Error.

After reading your Post this morning I deleted the contact pairs an tried to run an analysis with no success. There where errors saying that a loadfactor was not defined for several elements. The Elements which are effected are located on one side of the cross member where the second surface load is applied. After deleting this force I was not able to apply the force again. Simulation says that it is not possible to apply a force to a shell surface.

After deleting this force the preprocessor throws no error anymore. But when the solver (iterative) starts it terminates after two minutes with an error. See below:

 ** Nonzero entries initially in upper triangle matrix = 38092455
 ** Parallel AMG Iterative Solver Solution

       Time        Percent     DT       L Iter. Residual
  0.00000000E+00   0.0000     0.100000  1   0   0.0000E+00

 ** ERROR READING DISPLACEMENTS
 ** Error (1)
 ** Job started           : 2013-11-04  08:53:11
 ** Job finished          : 2013-11-04  08:55:37
 ** Total elapsed time    : 2:25 (minutes:seconds)
  !! Program aborted !

Any ideas?

Regards

Stefan

[AstroJohnPE]

Wow -- what a problem! By any chance, was the surface force applied to the model in a different design scenario with a different analysis type? I can imagine this scenario might be true:

• in linear static stress, apply a surface force to the plate. (Plate and shell elements are very close to each other.)
• copy the design scenario to a new one.
• change the analysis type to MES. MES may in fact not support surface force on shell elements, but the FEA Editor may not detect the conflict until you try to apply a surface force while using MES.

Here are my only quesses about the errors you are encountering:

1. A problem with the motherboard?
2. A problem with the hard disk?
3. Are there any nonstandard characters in any folder that the model is stored? I can envision one analysis type working but another one having problems with the path. See what happens if you save the model to a short directory such as C:\Temp or C:\Junk.
4. I think it is unlikely to be an administrator rights problem.