DN1500 pipe with DN900 branch

Hi everyone,

I am not sure how to model this so I am asking the question. The geometry is as follows.

I have a DN1500 pipe running horizontal. From two places a 45 deg DN900 branch is connected to the main pipe. I am not sure what is the correct way of modelling this. there are two options, both discussed below.

1. Model DN1500. Model DN900 as a reinforced fabricated tee connection right at the junction of DN1500 and DN900, OR
2. Model DN1500. Model a rigid for the branch till it comes out of the DN1500 pipe. Then model DN900 branch.

My problem is, if I model as option 2, where should I place tee connection? At the intersection of DN1500 and rigid or between rigid and DN900?

Thanks

Dharmit

In our standard practice , for big size line we follow option 2 as you mention.
Let us consider 10, 20 30 are node point of header , where 20 is the tee point.
For Branch 20 to 200 is rigid & 200 to 210 branch pipe.
We first check default Caesar SIF for header & branch from Environment-Review SIF at intersection node.
We then manually put SIF value as follow
Node 10 to 20 , put SIF at 20 ( Caesar default header stiffness)
Node 20 to 30 , put SIF at 20 ( Caesar default header stiffness)
Node 200 to 210 , put SIF at 200 ( Caesar default branch stiffness, since here branch is 45 degree we multiply default value with 1.5).

Thanks shr,

My next question is what will be the length of the rigid from 20 to 200?

Simple trigonometry will tell me that the length will be 762 / cos 45 = 1077mm.

762 = 1524 / 2 (1524 is the diameter of the header).

But when we input this length for the rigid, the 3D model doesnt look correct as half branch is outside the header.

Is this correct or I should calculate the length so that the model will show full branch which goes inside the header.

Thanks

Dharmit,

Read class I branch flexibility to get more insight into this issue.Also one of C2 NEWSLETTERS( Don't remember which one, you have to search a little bit)addresses this issue.

Regards

Sorry Anindya, consider myself dumb but I didnt follow. May be I was dosing when I read your reply .

Could you please elaborate on the Class I branch flexibility. What should I read?

I found the Newsletter that answers the big diameter intersection issue. The issue was December 1998. This gives excellent insight on how to model a straight tee. Quiet impressed by it. But my problem is for a 45 degree Tee. Please go through the question in my original post as I think that the question is still not answered.

Thanks Anindya for directing me to proper source.

Dharmit

Dharmit

what is the D/T ratio of the piping you are working on.

31.3 says:

"Stress intensification and flexibility factor data in Table D300 are for use in the absence of more directly applicable
data (see para. 319.3.6). Their validity has been demonstrated for D/ T ≤ 100"

You may have to look at evaluating the SIF using a Finite Element Analysis.

Neelam,

The OD of pipe is 1524mm and the thickness of the plate that will be used to fabricate is 16mm. Thus I can say that D/t ratio will be < 100 (it will be near 100 but still less than that)

Regards

Dharmit

Details of Class 1 Branch Flexibility can be found in Subsection NB Of ASME B&PV code Sec III.

A 45 Tee can be modelled using the same philosophy as a 90 degree Tee.

Regards

Anindya,

I do understand that a 45 tee will be modelled as a 90 tee but please read the following



My question is what is the correct way of modelling a 45 tee.

Jeepers Creepers Dharmit, what's not to get?

The 3D model is just graphics, making it look right doesn't make your Caesar model correct.

You should model as a rigid element from the centerline intersection of the header and branch to the branch centerline intersection of the header wall. From that point model the branch as pipe. Each pipe element shown as a cylinder in the 3D graphics with a cylinder length equal to your element length and a cylinder OD equal to the pipe OD and that cylinder is placed on the centerline of your piping. So, when you skew the cylinder as you have and have a node at the branch centerline and header OD, in your graphics part of theat cylinder will be outside the header and part will be inside the header, but the node on the centerline of the element should be at the header wall.

Hopefully you are using more than just Caesar to evaluate this system.

Beam element models and graphics are dissimiliar in nature if you are just looking at the pictures and fail to understand the implication of what a beam element model is all about you are NOT qualified nor competent in doing this work!

Read B31.3 on the qualifications of a designer.

Thanks for all who commented on the post and had queries regarding my competiencies to do the Stress Analysis.

I know what a 3D Beam element is capable of doing in a Stress Analysis (Atleast thats what I am feeling, but some think that I dont know what I am doing) and I also know the limitation of the 3D Beam element. I also know that 3D graphics is just a representation and not what is in reality.

If someone can direct me to some article or literature that caters for SIF values of 45 Tees. I am looking for an article which says that SIF values of 45 Tees should be taken equal to Straight Tees (Except C2 website, Newsletters etc unless it refers to another academic article or book).

If someone doesn't have that reference material and still wishes to help me read through the entire reply why I posted this query on this forum.

For a straight Tee, the SIF values taken (SIFo and SIFi both) should be different than a 45 Tee. Well that is what I am feeling (if someone would like to comment on that, they are welcome to do so). The only reason why I am feeling that is because the hole cut in the header for a straight Tee will be different than 45 Tee and because the Area welded at the junction of a Tee is different I thought (if someone wants to comment on this, they are welcome) that the SIF values should be different. As far as I know (does someone wants to comment on this) ASME B31.3 doesn't cater for 45 Tees and if we put the SIF values for a Straight Tee how much different is the value.

I didnt find a single response stating that a Finite element Analysis should be carried out for a DN1500 header having DN900 branch in a 45 degree Tee. Just because C2 doesn't allow or B31.3 doesn't cater for a 45 Tee doesn't mean world stops there. It also doesn't mean that there is nothing beyond the stated Codes and that we should take what is similar to the given situation (taking SIF of Straight Tee on a 45 Tee) and move on.

I think that most of the users on this forum are having more experience than what I am having and I would like to thank all those who have given me right direction. I am sure that there will be atleast one person who can guide me in the right direction rather than question my competiencies. Everyone should also remember that it is always difficult unless you know the answer. And for a new commer in the field, it is like a black box.

Thanks

Dharmit

Dharmit,

By doing a search in this forum you will find that this topic has been covered a few times. The most relevant, is "SIF for lateral tee" Borzki 29 Sep 2005. There are a number of references to research papers on this topic and that FEA should be used. A big surprise to me was the reply from Chuck Becht (ASME B31.3 committee member) in this topic thread that stated:

"Laterals have been found in testing and analysis to have a lower SIF than 90 degree branch connections. The opening size is larger so the moment divided by the moment of inertia of the section where the branch connects to the run pipe yields a lower line load than for a 90 degree branch. The B31.3 committee has discussed putting in a simple statement that the SIF for 90 degree branches can be conservatively used for laterals."

I was always told previously by my peers that SIF's for lateral tees are worse than for straight, and a factor such as 2 should be used in the absence of code direction, this before the use of FEA analysis. However, the above statement as far as I know, does not appear in any code or published industry paper, which is what you/we are after. Laterals have been used for many decades, so I wonder why the codes would not want to address them if in all cases the SIF's for laterals are less than for straight tees and that by simply assumming the same would be conservative.

So the best bet in my opinion is to get the tee analysed by FEA to get the actual SIF's especially for such a tee as yours which is very close to the limit of the code's applicability, and way beyond the size range and D/t ratio of the tested tees that the B31.3 code SIF's are based upon.

Regards
Mike

Thanks Mike,

I will use the approach of FEA in this case.

Thanks again

Dharmit

To Mike

Did you find out SIF from FEA?
If you find out it, Please advise me that how to get SIF from FEA result.
Is there any Formula to get SIF?

Regards
DW KIM

45 deg laterals are frequently used in relief systems made from thin wall stainless. I agree that SIF's should be equal or lower to a 90 deg where the branch to header ratio is close to 1. However when the ratio is significant, big header, small branch, then local shell flexibility must be a major player. Looking at the acute angle crotch, a bending load will concentrate stresses at back of an un-controlled fillet weld. Not a good situation.

The only place I have seen laterals addressed in a national Code is the French Codeti, where tee sifs are similar to B31.3 but with a factor of
sin(a)^3/2 in the denominator, where (a) is the acute angle.

To establish sifs using finite elements, look at FE/Pipe software by the same author as Caesar II, the venerable Tony Paulin, at www.paulin.com.

DW KIM,
I have not come across a lateral tee for quite some time now, so have not used FEA to investgate SIF's for these tees. Even the new relief systems I look at these days have no laterals, with the only place I see laterals being on open drain systems, which is non critical not analysed piping. As MoverZ has mentioned above, FEPipe would be a good package to use for finding SIF's on piping components. Also as mentioned above the french code gives a factor close to 2 for a 45 degree tee which ties up nicely with the factor I have used in the past, and this would appear to be conservative. So in general I would use a factor of 2 for 45 degree laterals knowing that this will be conservative. However, in cases where the D/t ratio is getting close to code limits or above, large header to branch ratio (per MoverZ) and for more acute angled tees, I would look into using FEA/FEPipe to calculate SIF's.
Regards

You do not mention your application. If your application however happens to be potable water, I know AWWA Manual M11 provides at least some basic design procedure in that realm with formulae for making 90 degree as well as other angled outlet connections like 45 degree or other angle wyes by fabrication to large steel water pipelines. Many services per this reference (particularly when employing lighter pipes and higher pressures) require reinforcement that is available in many forms (perhaps this could be used in any case to compare your results?)
I believe you are correct in your intuition that at least the most simplistic wyes (for example running a certain thickness of outlet pipe into a certain thickness of parent pipe at say a 45 angle) are different in at least a couple localized stress respects to similar fabrication 90 degree outlets. One respect as has been stated is the amount and shape of metal removed from the parent pipe/intersection, and I believe another arguable point is a different stress concentration geometry at least in the local area of the more acute groin or crotch area intersection angle of the wye. While perhaps there have rarely been problems with bursts of such specials or fittings in general, I think I have read many years ago that early in the last century (when there was little available information relating to design, nor test results, and certainly not FEA/sophisticated software etc.), some of the very early problems with at least some early (more brittle?) material fittings were encountered preferentially at wyes. This does not necessarily mean that there is anything wrong with modern codes designed around modern materials and welding processes etc., or for that matter anything said by those who write them, but it would appear to be at least an indication that in general large wyes (and/or what comes to bear on them in operating pipelines?) could be in general relatively more vulnerable from a stress standpoint than many other fitting configurations and thus should probably be carefully constructed.