Sheave and rope

Fatmac

Hi guys I've been trying to model a rope running around a sheave. I've struggled with trying to model rope before, I'm trying to produce an element that has tensile stiffness but no bending stiffness, I've looked at a series of short beams, trusses and solids and not really come up with any nice solution. My best approach so far is to use a thin 'belt' so the stiffness in one plane is small. I'd like to progress the series of beams or trusses but run into contact issues with the trusses and don't have simple approach to connect the series of beams with no stiffness at each junction. I've added 2 samples of my rather poor efforts so far. Any ideas would be welcome

disla

First is to make the contact work.

Contact problems between curved surfaces require a good initial node position.
I would recommend you to start with finer mesh on the belt and use the project nodes into cylinder tool to adjust the positions to a common circle.
Remove any non-useful element to speed up the analysis.
Use reduced integration as it is less stiff.
Define a maximum time step to help the convergence. Start with 1mm not 50mm to check contact works.
Swap Master and Slave. Slave shouldn’t be the largest and coarser surface.
Look at the deformed shape if it fails. Will tell you what is going wrong.

Fatmac

thank you disla for looking at this - i had a little play with this and managed to get more sense out of it by swapping the contact surfaces as u suggested. I'm afraid some of the other suggestions u made were beyond my level of understanding. "Project nodes into cylinder tool is way above my league"- i will take a look at the latest manual and see if i learn more about it there. I assume the process u outline is to produce a better matching set of nodes to minimise gaps and overclosure?? I was also not too sure about using "reduced integration" - I assumed u meant changing the elements to an "XXR" type -I tried to use C3D8R elements, but in this model did not converge with my efforts where as the standard elements muddled thru. I have attached where i got to and u will se where I tired to change elements- the saved file has that option suppressed. The deformed shape now certainly looks ok to me. I did wonder if u had any other approaches to modelling rope?

disla

"Project nodes into cylinder tool is way above my league"

It's a Mecway meshing tool (Project onto surface). The idea is to prepare both surface's nodes accurately following a desired shape. You have available a cylinder similar with your scenario.

Your contact stiffness seems too small. There is some penetration of the belt inside the Sheave, but the contact seems more clear now.



I can see one issue with this set up. Unless you are interested in the first engaging of the rope on the sheave, your spring is increasing the tension linearly. I guess a loom is supposed to be working with constant tension.

I did wonder if u had any other approaches to modelling rope?

Not really sorry. The material definition and parameters for rope is completely unknown to me.

disla

By other hand. If you still do not have clear the material model for the rope I would do the tests on a much simple and smaller model. Without rotations. Tensile test.