Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in
  • Congratulations GregLocock on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Need some hardware configuration advice. 1

Status
Not open for further replies.
MdBanks

MdBanks

Military
Feb 23, 2011
8
My computer here at work is on a secure network that blocks many of the useful features on this site. So, please excuse me if I skipped ahead in a few FAQ protocols.

I am attempting to perform a stress analysis for a part of a ship(A loading ramp). I have to simulate an asymmetrical load and I would also like to look into some fatigue analysis.

I'm having problems getting the mesh to work. The fact that each attempt takes my computer 3+ hours to run before I see what parts fail to mesh, is really setting me behind schedule.

I've been simplifying parts as I can and as I study each failure, but it seems the list gets longer with each attempt.

When I set the global mesh to a coarse and applied mesh controls the the smaller parts, I had the longest run time and the greatest number of failed parts(finally got a memory error).

I'm using Solidworks 2010 Premium. Any suggestions on typical harware or typical model simplification techniques could help me.

What is the industry standard computer config in use for this?

Should I apply shell elements? And if so, will this affect my ability to apply split lines and simulate varying asymmetrical loading scenarios?

 
Replies continue below

Recommended for you

Sort by date Sort by votes
Currently I am still meshing it as a solid.

I tried using the curvature-based mesh and got a part interferance error. That prompted me to look closely at the construction of the model. I found dozens of geometry intrusions.

After spending a few hours consulting with the engineers (making sure I'm not comprimising the design) and a few hours cutting and trimming parts, I've finally gotten the thing to mesh.

I've run one stress study, but the structure appears to just fall apart. It's as if the global bonding condition isn't applying.

Now all I have to do is get the contact sets to work properly.
 
Upvote 0 Downvote
A ramp is best shelled and meshed. It is composed of components that are very thin compared to their length and width. Long mesh time is a sure sign that you are using the wrong approach to the problem.

To fix your current problem you need to check that all components are meshed together with no cracks. To do this you need to display the mesh outline only. Unfortunately in Simulation/CosmosWorks you cannot do this and you cannot easily fix the problem even if you know it is happening.

I am guessing that you are trying to mesh an assembly. You probably shouldn't be doing this for this type of structure. If it takes 3 hours to mesh it may take a week to run. Just how many DOF does the mesh have?

TOP
CSWP, BSSE

"Node news is good news."
 
Upvote 0 Downvote
DOF 17395,047

Current study has been running 1hr 53 min and shows 61% completion of surface to surface bonding task.

Overall completion bar doesn't even show progress yet.
 
Upvote 0 Downvote
I'm running a pretty powerful rig, but I think you are right about the time.

What average DOF do you typically look for?
 
Upvote 0 Downvote
I can't say without seeing your model.

Anything more than 100,000 dof for the full ramp would be a lot. You should be able to achieve that with a shell model.

To get a shell model you will have to find the mid planes of all the plates, etc and mesh those.

TOP
CSWP, BSSE

"Node news is good news."
 
Upvote 0 Downvote
The upper surface of the ramp has raised ridges.

When I define the midplane of the ramp surface, do I lose the added strength of those features?

What if I shelled using just the upper surface, would that work?

 
Upvote 0 Downvote
"Have you ever used shell elements before?"

My experience with shell elements is limited. But your inquiry begs the question - Why do you ask?

My book tells me that I can, in fact, create a surface from midplane OR either face, but my book is for SW 2009 and I'm using SW 2010. Who knows what's changed in a year?

My question is, How do I convert this model to shell elements and retain the added structural rigidity of these ridges?

Also, I don't see a function allowing me to define contact sets for the shell elements with the remaining solid components. Does SW imply the contact conditions previously established for their associated solids?

Thanks for the help and advice!
 
Upvote 0 Downvote
If you are doing finite element you have an obligation to know the capabilities of the method which includes the use and behavior of the different element types and what they can and cannot do. Part of this obligation is creating test models to determine if the functionality of the software gives the correct answer in the situation to which you are applying it. I can't answer that question for you. Solid elements are not appropriate for what you are doing.

What book are you referring to? SolidWorks help will explain mid-plane extraction for whatever release you are on. You will have to be familiar with surfaces and the mesher in SolidWorks in order to obtain a correct mid-plane representation of your ramp.

You can define contact on shell elements in Cosmos/M, but probably not in CosmosWorks or Simulation. I am not sure why you need contact for a ramp. Is it warped?

TOP
CSWP, BSSE

"Node news is good news."
 
Upvote 0 Downvote
Obligations.....duly noted. Thank you.

I am referring to the Solid Works manuals for SW 2009. SW 2010, which is what I am actually using, is a bit different. I was actually asking if you were familiar with SolidWorks 2010 and how it may differ in the way it handles shell elements.

The structure of the ramp looks a lot like an airplane wing. It has upper and lower skins with a skeletal rib structure inside. Some of the structural members are complex extrusions. There are many contacting surfaces to bond.











 
Upvote 0 Downvote
This kind of structure is where your skill as an analyst comes in. You would not, in general, use contact elements where one skin contacts another. You may or may not use bonding. I wouldn't.

The discussion in Bruhn* might be helpful in this regard in deciding what detail to include and what to leave out. When modeling the entire structure you are after global results. You can apply the localized forces to a smaller more detailed portion of the structure using the cut boundary displacement method which is not available in SW Simulation AFAIK.

*Design of Flight Vehicle Structures

TOP
CSWP, BSSE

"Node news is good news."
 
Upvote 0 Downvote
Looks like I've got some reading to do.

Thanks.
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

roldy
  • Locked
  • Question
Need guidance for a structural simulation
Replies
0
Views
40
roldy
roldy
Ateyx
  • Locked
  • Question
Need some help performing an impact test
Replies
1
Views
113
academix
academix
AFG03082015
  • Locked
  • Question
Could someone explain the stress intensity plot in Solidworks? 1
Replies
1
Views
54
cbPVme
cbPVme
seba7777
  • Locked
  • Question
Can somebody check my simulation?
Replies
1
Views
37
RathodAsit
RathodAsit
RathodAsit
  • Locked
  • Question
Help Needed Doing Vibration Analysis
Replies
0
Views
54
RathodAsit
RathodAsit

Part and Inventory Search

Sponsor

Back
Top Bottom