Sail rigging analysis

natefred

Hello,
I am performing analysis for a square-rigged sailing rig, which I have modeled as follows:

1. Sail canvas modeled as isotropic shell/membrane elements, 0.039" thick
   
2. The edges of the sail are reinforced with rope, which is modeled as beam elements, which have nodes common with the perimeter edges of the sail's shell elements (these were co-located and have been merged)
   
3. Wooden mast and yardarm modeled as beam elements
   
4. Stays & shrouds (ropes) supporting the mast are modeled as beam elements
   
5. Sail handling lines (also ropes) modeled as beam elements
   
6. Displacement xyz constraints applied at the mast step (base of mast), and where the standing and running rigging are secured
   
7. Uniform pressure applied to the sail
   
8. Gravity applied
   

This is my first attempt to use Mecway to analyze ropes or thin shell elements subject to membrane effect or relatively large displacement, and I am encountering difficulties.
Initially I assigned all elements representing ropes as Tension only truss elements. However, the internal static 3D solver would not converge, I expect due to large displacements.
I then tried nonlinear analysis using CCX solver. To do so I had to turn off the tension only selection for the ropes. I have been unable to achieve convergence with this approach either.
Any advice how to go about this? Perhaps I need to pretension the rigging elements, or try load ramping? Thanks for any recommendations you may have.
Here are images of the model with & without loads/BC's:

Fatmac

nice project. I've spent (wasted?) many hours modelling masts and trial and error has lead me to the following conclusions. My comments are not some much based on science - rather making many errors and looking for work arounds!
-start simple, make that work first (in this case perhaps set up the mast alone with a simple concentrated load and get that to work b4 adding the complexity of the sail
-get the static linear model to solve first, if u can't get that to solve then there is little chance of getting the non-linear analysis to work.
-I avoid using beams as rigging elements (unless I'm applying a UDL type load to them), I use truss elements where possible. Particularly in the non-linear analysis they have tendency to spin around there own axis and require more constraints to converge on a solution.
-be careful when using shells in non-linear analysis, if they have any node-surface connections I normally have problems
-I have had little to no luck with tension only elements. The simplest approach i have is to run the analysis, they simply remove any rigging element that is in compression or apply a very elastic material property to it. I have developed my own methods of dealing with rigging elements in tension -I split the truss element in 2 and apply a small side load - if it goes into compression the small side load is enough to cause it to 'buckle' and often allows the solution to converge. Other approaches I have had some success is by prebending the rigging element again allowing it to buckle but this often requires something to stabilise it before convergence will be permitted.
-I apply pretension by either moving the mast foot up before the loads are applied (as how it is done in the real world) or by shrinking the appropriate rigging elements by cooling them down.
if u post u file I'm happy to take a quick look -and good luck!

natefred

Thank you for the input, Fatmac! I will try your suggestions.
File attached, let me know if you have any feedback.

Fatmac

I had a little play with ur model -I changed a few things and added far more complexity than needed, but it may give u some ideas how to deal with some of the issues u are likely to come up against. I did make an attempt with tension only elements, but as usual I could not make them work properly with the NL analysis, so I took the easy option of making the cables that went into compression a very elastic material. I normally apply rig loads as UDLs to spars or cables so messing around with pressure loads on sails was a good learner. I left the sail as a shell but imagine using membrane elements is the way to go. I have no experience with membranes and really struggled to successfully get them into the model, however i did manage eventually but I only got them to work in the linear static analysis, so I leave that for another day!. Some of my additions were not that polished and I had to use some dirty tricks to make the NL converge but it should give u some pointers. I think the pressure I have applied here is representative of 40 knts AWA. It can be varied by adjusting the wind speed term in the pressure input-see pict Srig2