Dear all,

I was once told that +Y restraints are forbidden by B31.3 code and that I should use "Y" supports instead. Please explain to me the reason behind this, and the advantages and/or disadvantages of "Y" over "+Y" restraint.
I would also like to know if there are special cases where +Y restraints are allowed (e.g. can an existing +Y support be modelled as a "+Y" restraint in Caesar II).

Thanks

You have been misinformed.... The code does not care what sort of supports or restraints you chose to use. However the code does require an evaluation of the state of sustained stresses (and thereby occasional stresses) in all states of support.

So if you have a +Y that goes to zero load in the operating state this support is no longer active and therefore the sustained stress evaluation should be extended to include the state of supports in the "hot" condition. A so-called hot sustained case should be run removing the support which has lifted off.

This subject I covered in detail in a Coade newsletter which can be found here... http://www.coade.com/newsletters/jan01.pdf

So there you go....

Dear John,

Thanks a lot. Your posting was indeed very helpful.

I was working for an offshore project where we were instructed to use 'Y' restrain instead of '+Y ' just to check whether in load summery there is any +load or not .If so we were told to use hold-down support in that specific location.

Habib

yes often times a fixed Y is advantageous...

I am asking this question to try to learn something not to question the use of a fixed Y. What is the advantage to holding down +Y support that wants to lift up off of its support. It would seem that holding the pipe down and restraining its thermal growth would put additional stress on the line.

the right support in the right place is my only answer to this.
Hold downs should only be used if required. Why they are required is an expansive question.

i would resist the thought to shove supports be they hold downs, rest etc just 'because'
Each support should suit its own purpose.

I am always carefull that if i want a Y support to actually resist vertical displacements or forces, then i model in a realistic vertical gap to mimic its real life situation.

I am a a beliver in that the model should reflect the design as accuratly as possible.

I have lost count of the number of times i explain to piping designers and stress engineers
to be carefull in how the rest/hold down support design is comunicated between parties.
The wrong inclusion in the design of a y support instead of a +y support is infuriatingly frequent.

Malcolm,
I use +Y supports in almost all cases. I had a situation recently, however, where a Y support was required. We had a 16" HP steam line running several hundred meters along a sleeperway. At two locations the line rose about 7m to get over roads. On the bridge the line lifted as much as 75mm off the supports making it difficult to guide. The solution I used was to leave the shoes off the sleepers adjacent to the bridges and use Y supports with guides on the top of the bridge. This forced the vertical rises each end of the bridge to grow down instead of pushing up. I have also used Y supports with guides where a line has expansion joints to stop the lines springing up and destroying the joint. In addition, some clients use a guided shoe that consists of a small angle or plates that hang over the flange of the shoe. In that case I model the support as a +Y, Guide with gap, and -Y with gap.

Piping systems back in the good old days were floated... very little restraint was provided lines either horizontally or vertically.

This provided as low a thermal displacement stress as could be acheived by the layout however these sytems were subject to large displacements imposed by even small dynamically imposed loads (Fat Guys leaning on Insulation etc.)

Some control of the piping both vertically and horizontally is desirable presuming it does not create an excess displacement stress... For instance a minor lift off support can be restrained verticaly with little change in the displacement stress thereby negating the need for a spring can (assuming the support cannot simply be relocated)

Its all based on experience and opinion I guess... everybody has an opinion and sooner or later everybody gains experience!

To use Y or +Y for ALL SUPPORTS in a model rather than only using a +Y when it is ACTUALLY REQUIRED is a question of preference. Caesar like similar programmes first does a run to check which supports have lifted off. For supports not lifted off, it changes all the supports to +/-Y ie. "Y". Like many older folk, I would put in a Y where it was obvious that the support would not lift off. e.g. Along a piperack. It saves typing a "+" and saves a few milliseconds of calculation. 20 years ago this processing cost real money. I would then check the output for negative vertical loads to confirm my initial assumption. The more common alternative now is to use +Y for every support that is not physically held down either because "it is more correct and more accurate" or because it produces a prettier picture. In a crowded model the latter could be quite a valid reason.
Conclusion : Go with the flow. At the end of the day, Caesar sorts it out and it makes not a jot of difference. It also saves on red pencil.

I prefer a +y or full y based on where I need them.. that way there is no confusion as to the intent of the design...

The important thing is lift off must be considered and accounted for.

Gaps that are absolutely critical must be accounted for in the design details...

I think an important point here is that however you model your system, you are accepting a batch of assumptions - either your own or the program's. As a user, your first look at the output (I won't call them results yet) should be to confirm that the assumptions are correct before plowing though all the other numbers.

Let's take this Y versus +Y model (assuming Y is vertical). I know several users who will not use a +Y; citing that if the pipe lifts off at that location, it is poor design. They review the results looking for a large +Y load on those restraints - numbers that may also indicate poor support location. Now look at those who are comfortable with the +Y and liftoff. The program will allow the pipe to liftoff the support with the support load going to zero. But these people should now look at deflections to learn how far the pipe moves up off the support. And maybe they should also re-evaluate sustained stresses with those inactive supports removed.

All this review already and I have yet to look at stresses.

My point is, establish your own pattern for model construction and output review. You might think that modeling "everything" (as in gaps, friction, even rod swing) might eliminate time in confirming the output but, as others have said here, there's lots of tolerance in the field that is not incorporated in the analysis. When it comes to modeling what actually gets built, sometimes you might be fooling yourself with such "sophistication".

Hi, suppose that reaction force due to the sudden opening of a PSV causes lift-off of the first support of discharge pipe.

If I avoid support lift-off by hold-down, have I to remove +Y, to leave -Y and to evaluate sus+occ stress?

I'm considering static analysis, of course.

Thanks for Your kindly reply.
Best regards.

A +Y restraint in your model will allow liftoff.
A Y restraint will allow no vertical movement.
A -Y restraint will prevent the node from moving up but allow it to move down.
Do you really want to put a structure in that reflects a -Y restraint? I don't think so.
Why not use a Y restraint?

Dear Mr. Diehl,
thanks for Your reply.
When I use hold-down, my structure reflects a Y restraint, of course, but I haven't explained my doubt very well.

If pipe doesn't lift, +Y supports pipe weight and there's a force acting downward on this restraint.

If pipe lifts, +Y doesn't supports pipe weight, there isn't a force acting on this restraint and I have to check sus stress by removing +Y restraint.

If I use hold-down because pipe tries to lift, pipe itself can't lift and there's a force acting upwards on restraint; in this condition hold-down works as a -Y restraint and there's nothing that supports pipe weight (I think and perhaps I'm wrong).

My question is:

When I use hold-down because pipe tries to lift, have I to check sus stress by -Y?

If I use hold-down to avoid that pipe lifts when a PSV opens, have I to check sus+occ stress by -Y in my static analysis?

If not, why?

Thanks.
Best regards.

If, physically, in the field, there is a device that prevents the pipe from moving up and no device to prevent the pipe from moving down, then you model a -Y support in CAESAR II and use that for the sustained stress evaluation.
If, however, you have a device that prevents pipe from moving up or down, then a Y restraint would be used throughout - even for the sustained case.
I believe it is your intent to evaluate sustained stresses based on the supports that are active in the operating condition. To do so, restraints should be modeled as they would appear in the operating system considering both the orientation (e.g., +Y or -Y) of these active supports and perhaps even any gaps that occur in the operating state.
Simply because your operating state may show a +Y load on a hold down does not mean you do not model the active support as a -Y restraint if there is also a restraint under the pipe, ready to support the deadweight distribution.