I'm trying to model an existing expansion loop on a steam line. Known geometry (radii are given) - Large radius on 1st bend, smaller radius on 2nd, a straight run, then same smaller radius on 3rd bend and same larger radius on the 4th and final bend. In the 'x-z' plane. I want to add +Y-restraints at the 1st bend (start), 2nd bend (middle), 3rd bend (middle), and 4th bend (end). When I add a '+Y' to various nodes in the bends - they are hanging out in midair according to the model. Methods so the '+Y's are back on the pipe? (Consider adding to your Expansion Loop tutorial)

You have probably modeled things correctly. In the input processor, all items attached to a bend plot at the bend's tangent intersection point.

For a bend at node 20, with start node at 18 and mid node at 19:

- to place a restraint at the bend start, put it at node 18.

- to place a restraint at the mid point, put it at node 19.

- to place a restraint at the end, put it at 20.

Look at the plots from the OUTPUT MODULE. The restraints are shouwn in the actual position modeled. I direct all my personnel to only plot from the OUTPUT for reports, reviews, etc...

Thanks for the feedback....I'll check the Output plot. When I get that far. :rolleyes: Next stumbling block....Has anyone modeled back-to-back bends in an expansion loop where the various radii are different? Specifically, the radii on bends 1 and 4 are the same (17'-10", if you want to know), and bends 2 and 3 are the same (10'-10"). Meaning back-to-back bends (1 and 2) and (3 and 4) have different radii. My model makes it thru the 'single step' Start Run if I make the radii on back-to-back bends the same (1 and 2 must be the same but they are not). If I enter them as they really are, I get fatal error 45E, "Trying to specify a point twice" (the endpoint of the 1st bend). Recall, there is a real straight section conveniently separating bends (1 n 2) from (3 n 4) - no fatal errors there.

Suggestions to resolve the back-to-back, non-equal radii bends issue? A co-worker suggested adding a very small straight section (say, ~1") between the back-to-back bend....but didn't I read somewhere here that such small sections play havoc with flexibility? (and it didn't appear to work, anyway).

The reason that the input processor only knows about the bend tangent point is "speed". Actually trying to determine where things are at "the editing stage" of the model would really slow down the processor.

Modeling back-to-back bends is not a problem at all. Left alone, if you do this, the processor will not put an auxiliary bend node at zero degrees if it won't fit. However, if there is room, you will get a node at zero degrees - the start of the bend. Now, if you then shorten the straigth length (sufficently), then the start of the 2nd bend is on top of the end of the 1st bend, which is what the error message is trying to tell you. All you need to do is delete the node (and angle) at zero degrees on the 2nd bend.

It worked! Thanks. Now, to see how well the model's displacements match what we see (which started the whole thing....) (No doubt more tweaking to follow, but progressing now.)

A subtle point here. I believe you generated the error because you specified the second bend before defining the first bend in this back-to-back bend pair. That's why you got the zero angle node placed on the second bend that conflicted with the far end (the 90 degree angle) node of the first bend. As Rich points out, if you lay in the first bend first, CAESAR II will not build the zero angle node on the second bend as it is redundant.