I'm getting confused trying to model the axial movement that a Victaulic flexible coupling can absorb. Say I have a vertical piece of pipe. The location where the coupling is intended will be node 20 CNODED to 21 (node 10 is below 20, node 30 is above 21, and the model is input from 10 to 30 in the +Y direction). It is also intended that the coupling be installed so that the gap between the ends of the pipe which are to be connected is a maximum. During operation the gap will close by 0.135" which is 75% of the range listed by Victaulic. How to CNODE the two nodes in the Y direction? I think both the +Y and -Y diections have to be defined but I'm confused as to how to do it. Maybe I shouldn't define the Y direction at all but check the Y movements after CAESAR runs against the 0.135"?

Yes, you'll need both a +Y restraint and a -Y restraint to completely define the Y relationship between 20 & 21.
What makes things confusing is which is the Node and which is the CNode.
First - let's say that this is a "absolute" +Y restraint at 21 - that is, there is no CNode specified - the node is connected to "earth" in the +Y direction. This would indicate that Node can freely move in +Y but cannot move -Y. Now, let's say that Node 21 is hovering 1 inch above earth. It is still a +Y restraint but the gap of 1 inch must be closed before the +Y restraint becomes active. That's our gap.
Now let's relate 21 to 20 rather than to earth. This is now our "relative" restraint - where 21 moves with respect to 20.
Your nodes 20& 21 will be defined in your input as coincident. As I see it, Node 21 can grow down 0.18 inches before "clashing" with 20. This 0.18 is the gap that must be closed before the restraint becomes active. But is it the +Y restraint of the -Y restraint? Referring to the "earth" example, it's a +Y restraint.

So, two restraints:
Node 21 CNode 20 +Y GAP=0.18
Node 21 CNode 20 -Y

And this will either confuse or confirm your understanding but it says the same thing:
Node 20 CNode 21 +Y
Node 20 CNode 21 -Y GAP=0.18

No matter how you model it, be sure to confirm the intended response in the output.

Great - thanks for your help Dave

R. James
LU '74