Insufficient Data in File 1

[mechengineer18]

Does anyone out there know how to trouble shoot the error: Insufficient Data in File? I am trying to analyze a pull test fixture made up of a 12' x 4.5' base of I-beams and a thick plate with a shackle mount attached and braced with I-beams cut at 45 degree angles on either end. These add up to 13 total parts. The 80,000 lb load is applied to the face of the hole in the shackle mount and the bottom of the base I-beams are restrained.

I successfully analyzed a smaller assembly which did not include the I-beam base, only the plate, the shackle mount, and the I-beam braces. I have tried making these as one sub-assembly and the base as another as well as all of the parts in one assembly but I keep getting the same error.

Everything goes along fine until I click on my results. Has anyone been able to overcome this issue?

Thanks,
Mike

 

[mechengineer18]

Excuse me, that was Insufficient Data in Results File.

Mike

 

[jte]

What kind of elements are you using? If solid tet's then either the "draft" or high order tet's will result in a lot of elements. Wondering how many you have? I haven't run across this type of problem, but suspect you're running out of disk space. Is the solver staying in RAM or going out of core? If it is going out of core then you're using the hard disk to store the stiffness matrix which then gets deleted at the end of the run. Maybe see if you can periodically check the hard disk properties to see how much space is available as the problem runs. If you're running out of space, then it is likely that the output files are getting truncated. If you're using high order tet's, try solving using the low order ("draft" in the mesh options) tet's and see if it works. Not as good an answer, but it might provide a clue as to what's going on.

jt

 

[mechengineer18]

jte,

Thanks for the reply. I had been using a solid, high quality mesh with the default mesh size. When I changed to Draft quality I did get results although on occasion Solidworks would shut down.

I was not keeping track of the number of elements but since using draft settings I have been getting around 150,000.

I will try to run those tests to check on memory usage.

Our SolidWorks vendor claims that my system is adequate to run Cosmos but the lengthy (2-3 hour) computation times and these errors lead me to believe that it is not. Do you think that an Intel Core2Duo, 1.80 GHz proccessor with 2 GB of RAM is enough to run these types of calculations?

Thanks,
Mike

 

[mechengineer18]

jte,

I have 20 GB of hard drive allocated to this process.

Mike

 

[jte]

Mike-

150k elements is a bunch! I try to keep my models below 20k - but then, mostly for nonlinear. For linear runs, I would expect 150k elements to take all of 10 minutes. Then again, when I started working (Cosmos/M version 1.65 about 15 years ago) the things we do now would have seemed like science fiction. Wasn't unusual for runs to take over 24 hours for a 10k node linear run. I won't even mention the system crashes which destroyed the entire model database.

I run a similar machine as you but with 3 GB RAM. It can make a big difference if you exceed available RAM and have to hit the hard drive. Try to convince your boss that the ROI on more RAM is quick given your billing rate. Like 1 hour savings.

Let me know how the testing works out. The fact that you were able to run using low order elements is encouraging. That indicates that there is not some modeling issue going on.

jt

 

[mechengineer18]

Thanks JT.

I am getting the same error on another similarly sized test. I think we will need to upgrade quick.

Mike

 

[DLTguy]

(1) The SolidWorks vendors are not very knowledgeable about COSMOSWorks. They always try to put the blame on the hardware (graphics card, disk space, memory, etc.). Please contact the COSMOSWorks technical support folks in Santa Monica directly. They will be able to help you better. I am saying this from my experience on a similar problem four months ago. The engineer in SolidWorks in Santa Monica helped me in 3 hours and my vendor took three days to tell me that it was problem with my computer. It turned out the patch had not been installed correctly.

(2) Here's what the technical engineer in Santa Monica told me -- In the results folder you will find a text file with OUT extension. Open it and check the last line. If the last line says something about time to complete, then the analysis has completed well and there is no problem with memory or disk space.

(3) In your case you can use the "Weldments" in COSMOSWorks to model the beams. Analysis time will be faster.

Regards,
DLT

 

[mechengineer18]

DLT,

Thank you.

Do you have contact info for the tech support people in Santa Monica?

Are you sure it is the OUT extension file? My system does not recognize that as a text file although it does for the LOG file and it lists the Nodes, Elements, DoF, and Results time.

Weldments are much quicker but I sometimes have trouble getting different materials to match up by trimming one of them.

Thanks again,
Mike

 

[DLTguy]

The phone number on their website is 310 309 2800

http://www.cosmosm.com/pages/company/contacts.html

Regards,
DLT