Building an arch in SystemModeler

Question

I'm trying to learn SystemModeler and struggling a little...

Building on the chain pendulum example from Wolfram, I've happily made a chain with revolute joints with spring/dampeners connecting the elements. With a single end fixed in space.

What I would like to do is set the geometry into a simple arch, with rotational springs and model this under a variety of loadings and accelerations. The model below is simplified, eventually I'd like to do this with numerous links and springs to model a smoother curve.

My attempts are below: I'm fairly confident I'm missing something quite elementary, but I've no idea what it is... any helpful suggestions welcomed.

and

I know SystemModeler isn't used a great deal here, but I saw there were other questions relating to it so thought I'd chance my arm.

I'm using SystemModeler 3 on MacOs. I know my support conditions don't quite match my diagram as I've included additional springs and dampeners, but I'm not convinced this is the issue.

Answer 1

Having posted a very similar question at the official SystemModeler forum, I received an answer that has shown where I was going wrong. Essentially when a chain of revolute joints are placed in a single plane, it will form a planar loop. To overcome this, there is an additional form of revolute joint that should be placed into the system at least once to allow the system to become solvable.

This type of revolute joint doesn't appear to allow for the flanges to be enabled (in version 3 anyway) so you cannot add springs and dampeners to this joint directly.

By adding this different type of revolute joint at some position in the chain, the chain can then be solved. Diagram below, showing that the fixed2 support is placed using a directional vector as the World is set, this is important to prevent duplication of points in space.

After running the model I get the arch as below:

One thing I wonder is if the revolute joint needed to make the planar loop solve could have dampening added through inclusion of additional joints at the end as shown below.

It seems to run OK when compiling the model but I need more time to check the results are sensible.

Now I've got a working principle established, I hope this post will help anyone else learning planar structures on SystemModeler too. I'm loathed to answer my own question, particularly given that it's come in part from the official forum so I'll leave it open for a week or so to give others a chance to add something else.