case i: pipe element node 70( in horizontal) , rigid element node 500.

node cnode
500 70
+y

for the above condition, in sustained case the +Y restraint didnt work.

case ii:

node cnode
70 500
+y

for this case, the restraint works in sustained case but not for the operating cases.

why??

This is a little hard to comprehend without seeing your model, and knowing the relationship between nodes 70 and 500. However -

• A CNODE is nothing more than an association of degrees of freedom.
• When you set things up such that "Node 70, CNODE 500, +Y", you are defining that "Node 70" is restrained to "Node 500" in the "+Y" direction. This means that 70 can lift off of 500.
• When you set things up such that "Node 500, CNODE 70, +Y", you are defining that "Node 500" is restrained to "Node 70" in the "+Y" direction. This means that 500 can lift off of 70.

So, depending on your loading, you would get two completely different behaviors from your system with these two different restraint specifications.

It's all relative...

You first say 500 is connected to 70 in the +Y, then you say 70 is connected to 500 in the +Y.

500 CNoded to 70 is +Y is identical to 70 Cnoded to 500 in the -Y. What you did was flip over the restraint.

Think of it this way...

When you don't supply a CNode, the assumed "connecting node" is EARTH. I capitalize it because it is very big, rigid and unmovable. (Actually we default to 1E12 for stiffness and that's pretty stiff too.) So your restraint reports show pipe load ON EARTH (as opposed to the restraint load on pipe). Think of EARTH as the other end of the restraint.

Now you add a CNode. All response is now related to the CNode rather than EARTH. A +Y restraint can move in the +Y direction AS RELATED TO THE CNODE. If the CNode "wants" to move more in the +Y direction than the Node, the restraint will remain active. If the Node "wants" move more in the +Y direction than the CNode, the pipe will disengage from the support and lift off.

So you can see why your two methods of input are opposite in effect.

Whether or not you have a firm grasp of this concept, it is important to do what you are doing - review the results to make sure the response is what you expect and question the analysis if necessary.