Modelling Cables using LUSAS software 2

[Asakura]

I'm using the FEA software package LUSAS to compare cable stayed bridges for my dissertation.
I'm new to FEA and am struggling getting my hand calculations to match the results for a basic element.
I am trying to model an simply cable hanging from two pins and the answers of way off.

Does anyone know of any relevant literature or helpful tips for modelling cables in this software?

Many thanks

 

[ukbridge]

Hey Asakura,

Interesting dissertatin topic!

Its hard to help you specifically without a sketch of what you meab by a cable hanging from two pins. You should also give moment and axial force diagrams and a deflected shape. You could uplod then as screenshots to imgur.com and paste the links.

Other helpful information would iclude....

Geometry
Mesh/Element used
Material properties
Support Properties
Loading used.

I expect your problem lies with not using "moment releases" within LUSAS. This ensures that there is no moment continuity at the ends of a line with an adjacent line. The Applications and Examples manual will be helpful and includes some worked examples/problems similar to what you want to do and eaborate on "moment releases".

If youre still stuck, post here and I'll try and help, I was also a student once! Good luck, ukbridge

 

[ukbridge]

Two more things:

1. If youre analysing a 2D problem make sure you use 2D elements as a general rule.
2. I would mode the cables as "Thick beam Elements BMI21". Make sure they have a single divisoon/element otherwise they will behave like a beam. Bar elements may work as well but use the former.

Also check your hand calculations as well just in case. Look at the deformed shape, why/why isnt it what you expect? Are you getting moments? etc Do your reactions sum to the applied loading etc.

gl

 

[ThomasH]

Hi
When you say that the results are "way off", what does that mean?

I am not familiar with Lusas but as a general rule, cable elements require a nonlinear analysis. If you analyze a cable as a "simply supported beam" (linear analysis) the loads will probably be transferred as moments and shear forces. If you instead use a nonlinear analysis (large deformation) the loads should be transferred as axial forces in the cable.

Could that be the problem when you say "way off"?

Thomas

 

[Asakura]

Thanks for helping me out! I've uploaded the details of what im getting so far on imgur here are the links

http://imgur.com/ggQF7ea

- for the results (vertical supporting forces)

http://imgur.com/Id1FEoi

- for the mesh details

http://imgur.com/usVkY7z

- for the geometric properties

http://imgur.com/FWsZfnn

http://imgur.com/LqUkWs0

- for the axial and moment diagrams

My main worry with these results is the fact that the vertical forces are higher than they should be by 10kN and i dont understand why im getting that difference!

Thanks again for any help you can give me

 

[ukbridge]

Asakura,

Yes, I see what you're trying to accomplish now. Unfortunately this isnt something I have experience modelling.

As ThomasH said, I expect you'll need to use some form of geometrically nonlinear analysis .

Its worth giving copying what youve posted here and sending it to the LUSAS support team along with your .mdl file to take a look at. They should be able to help you make sure your model os set up correctly, and are generally helpful people. Really your supervisor should be doing this but I digress...

Best of luck.

 

[ThomasH]

Hi
Most of the dialog boxes are unfamiliar since I don't use Lusas.

But if the vertical reactions don't match the applied forces and there is no dynamics involved (inertia forces), there there is something wrong. Also, for a cable analysis to work properly you need the deformations to be updated during the analysis (non-linear geometric). It can work with a linear analysis if you initial geometry is close the the non-linear solution.

I suggest that you contact Lusas support. My impression is that this is not a complex analysis in terms of non-linear. I think it is some software specific issue and that can be difficult for anyone not using to Lusas to solve. There can be a lot of parameters and settings that are software specific. I work with another software and there are a lot of tricks that can be used to make the solver behave as I want it to.

Good Luck

Thomas

 

[Denial]

I have zero experience with LUSAS.[ ] However from the content of this thread I get the impression that it does not offer a specific cable element.[ ] If this is the case you will have a lot of difficulty modelling true catenary behaviour using any form of "beam" element.

 

[ThomasH]

Hi again
From what I have seen previously on Lusas site I think the software should be able to handle cables. On the other hand, cable analysis can be very non-linear and include large deformations, or not. Part of the cable analysis is to find a stable geometry that gives equilibrium. So how large the deformations become depends on how good the initial assumed geometry is.

Based on the dialog boxes you have enclosed in this thread your cable elements have bending properties, that is not a cable but instead a beam. A real cable element only requires properties for axial deformations.

Since Lusas comes in different configurations (or packages) I think you should contact Lusas support. They know what capabilities you specific package has. And if it can handle cables they should be able to help you set it up. The example you have is fairly simple.

Good Luck

Thomas