Selection of branch fitting 1

[NovaStark]

According to ASME B31.3, a fitting can be chosen from other standards such as MSS SP-97.

However these don't go into detail about the actual selection process.

For example, when would one just opt for weldolet instead of a sweepolet connection ? If it is simply up to the designer, what factors would influence this ?

 

[XL83NL]

Service conditions is one reason, price and availability some others. Then there’s ease of installation.
The latter three reasons would quite often, if not always, make the choice for a weldolet more logical. However, the (sometimes huge) stress risers in the vast weld volumes of weldolets can make them susceptible for fatigue, which is where a sweepolet may be more favourable.
Codes are silent on these details and leave this up to designers. If you’d for example review a Shell piping class you’ll find that sweepolets (or contour insert fittings) are found in their branch tables, but only at few spots. Although such specs are ‘based on safety, economics, installation details’ (and more bla bla), they first and foremost are based on the best possible (safe) solution. I doubt if economics really play a big role in their decision matrix.
There’s probably a lot more to tell on this subject, which I leave to the more experienced members here.

 

[Rodddxl]

I have been doing piping for over 10 years and have never used a sweepolet. Weldolets are pretty much the standard. Another thread has a good explanation of the differences.

http://www.eng-tips.com/viewthread.cfm?qid=369098

 

[LittleInch]

If you read the above past I think you'll find its the other way around, I. e. why would you opt for a sweepolet over a weldolet.

Cost, technical requirements will guide you as well project documents and specifications.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[pennpiper]

NovaStark,
1, I am now retired but spent over 45 years in the Piping Design profession. So it could be said that I have enough experience to be able to comment on this issue.
2, There can be places/situations where Sweep-O-Lets might be used however these places are few and far between. One case is where the pressure and the velocity are both very high. Case in point are Compressor Suction & Discharge Pulsation Dampers connections.
3, For most normal piping the other Branch methods are easier and cheaper to make. If you look at Sweep-O-Let (pictured in the link shown below) and think about the weld prep and the final weld you will be impressed with the care and precision required for the connection. This Branch Weld would also tend to be subject to a more extensive NDE examination process.
4, If you decide to use the Sweep-O-Let Branch make sure your system really needs it.

https://www.google.com/search?q=Swe...sAQILQ&biw=1024&bih=633#imgrc=NblhnD5wUac68M:

Sometimes its possible to do all the right things and still get bad results

 

[NovaStark]

Thanks for the information guys.

I asked specifically about the sweepolet since in the particular piping I am referring to transports steam and methane to a reformer.

In that piping run, sweepolets are used as the isometric states. However the plant designer/specification just states that it should be an integrally reinforced branch weld fitting which is generic.

This piping is the only place where I have seen a sweepolet used. Everywhere else would be a fabricated tee with repad, or weldolet and so on. And since we did have a failure on a sweepolet here within a year in service, I am concerned as to why this was chosen.

(I can't really get into too much detail about the failure in a public forum however if anyone is interested, maybe you can actually assist me in a not as public forum)

 

[GD2]

NovaStark,
A 'Piping Class' which I understand you call 'Plant Specification' essentially give the pipe and fitting material specification with manufaturing standard. For example when O'lets are specified as MSS-SP-97, it simply means that you can use any of the fittings manufactured under the Standard. It will not specifically select the type of O'let in a piping run.
In the case of the noted pipe run, sweepolet was selected over other fittings. A good reason is that a sweepolet has lower Stress Intesification Factor (SIF) than weldolet and longer fatigue life. Stress engineer would not generally look more into welding and NDE difficulty. A Sweepolet must have passed the stress whereas a weldolet may have failed.In my opinion, it wont be prudent for you to substitue the sweepolet with a weldolet whout doing a stress analysis.
On another note, you should be looking at the final IFC drawings for the final fitting specification. stress isometrics are intermediate drawings made for the stress engineer.

Ganga D. Deka, P. Eng
Canada

 

[NovaStark]

A 'Piping Class' which I understand you call 'Plant Specification' essentially give the pipe and fitting material specification with manufaturing standard. For example when O'lets are specified as MSS-SP-97, it simply means that you can use any of the fittings manufactured under the Standard. It will not specifically select the type of O'let in a piping run.

Yeah that's why I wanted to know what factors were considered.

In the case of the noted pipe run, sweepolet was selected over other fittings. A good reason is that a sweepolet has lower Stress Intesification Factor (SIF) than weldolet and longer fatigue life. Stress engineer would not generally look more into welding and NDE difficulty. A Sweepolet must have passed the stress whereas a weldolet may have failed.In my opinion, it wont be prudent for you to substitue the sweepolet with a weldolet whout doing a stress analysis.
On another note, you should be looking at the final IFC drawings for the final fitting specification. stress isometrics are intermediate drawings made for the stress engineer.

Unfortunately the intermediate stress isometrics aren't readily available to me. But in one of the threads on this site, I did see someone mention about the SOL SIF being less than the WOL. However with this being a branched connection for an instrument indicator (so just a valve, gauge and tubing),I am not seeing a lot of stress being present at that junction.

 

[MJCronin]

My two cents ....

I too have 43+ years dealing with piping. pressure vessels, tanks etc.

I only seen sweepolets installed on a header in a high pressure steam system.... at the entrance to a code safety valve.

It was my understanding that, in certain circumstances, the SOL would prevent noise and vibration hat might be generated by the valve.

MJCronin
Sr. Process Engineer

 

[GD2]

NovaStark,

Small reducing branch connections (specially instrument connections) are the most vulnerable points in a piping system prone for failure due to high vibration issue (which is linked to the fatigue life). And you said that the WOL had failed out of all the other Olet fittings. You need to think seriously about the vibration issue and provide a good support to the branch line. Usually Sweep O-let have better fatigue life than WOL but if you can use a reduced welded tee, it is still better. The SIF value goes as: Welded Tee < Sweep O-Let < WOL. In terms of fatigue life it goes as: Welded Tee > Sweep O-Let > WOL

On another note, I missed to advise you that you must look at the 'As-Built' drawings rather than IFC Dwg. to know the as-built condition of the piping.


Ganga D. Deka, P. Eng
Canada

 

[LittleInch]

In short, we can't really get into the head of the piping designer, but can only conclude that the reasons may well be associated with excessive stress or vibration on a weldolet and maybe the schedule was such that a reducing tee couldn't be obtained in time or that the header pipe was so much bigger than the branch (you haven't mentioned sizes anywhere).

Without the full story - we have about 15% here - it's not possible to second guess the rest other than offer a few pointers.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[NovaStark]

@LittleInch

That's basically what I wanted to know as I can't expect anyone to know what the designer was thinking especially considering that I have some questionable drawings here!

Essentially on this same line, all the branched connections call for a sweepolet so 16x8 sweepolet, 16x 3/4 etc yet the piping specification that my isometric refers to does not even have sweepolet listed yet it is installed. So my isometric from my EPC has listed sweepolet, the iso from the designer lists integrally reinforced branch welded fitting (which is any fitting such as weldolet)and the pipe spec lists weldolet.

 

[Gator]

"So my isometric from my EPC has listed sweepolet, the iso from the designer"

What's the difference between the EPC isometric and the designer isometric? I don't understand why there are two (maybe I missed something).

Is this a manually drawn iso?

the plant designer/specification

Can you elaborate on this?

 

[NovaStark]

What's the difference between the EPC isometric and the designer isometric? I don't understand why there are two (maybe I missed something).

Is this a manually drawn iso?

Can you elaborate on this?

The plant designer's iso shows the layout with material specification however, no piping specification is given in that generic fittings are highlighted such as "integrally reinforced branch weld fitting".

The EPC would have supplied As Built drawings which for this line specify all the connections, in this case sweepolets and states the piping material specification used.


So for all the piping on my plant, the designer would have specified a code which points to a specific piping specification such as AAA for hydrogen piping which would give what fittings can be used, piping sizes and schedule etc.

So my EPC would have supplied an As Built drawing, and specified that they used sweepolets and piping specification AAA however in AAA, there is no mention of a sweepolet.

 

[Gator]

I still don't know why there are two isometrics ("plant designers" and "EPC"*) since you say there are no as-builts ("The EPC would have supplied As Built drawings"). Or am I misunderstanding something?

Are the iso(s) hand-drawn? I want to rule out that this isn't some kind of silly issue related to actual components not being entered into the spec before piping was modelled.

* Which could be the same company.

 

[NovaStark]

I still don't know why there are two isometrics ("plant designers" and "EPC"*) since you say there are no as-builts ("The EPC would have supplied As Built drawings"). Or am I misunderstanding something?

The EPC provided all the as-built isometrics for our plant. However the plant designer (different company) provided the isometric for this line as it is part of a packaged unit. Thus two isometrics.

Are the iso(s) hand-drawn? I want to rule out that this isn't some kind of silly issue related to actual components not being entered into the spec before piping was modelled.

* Which could be the same company.

I did find out that it was the EPC who decided to use the sweepolet. Which I am going to try to find out exactly why.

 

[hacksaw]

Nova,

Sounds like you need to have a chat with your contractors, as they are already covered by proprietary agreements,

For independent assessment, you will have to furnish quite abit of detail, well beyond the trifles described thus far.

There are more than a few firms that deal with nozzle failures especially involving various sorts of flow-induced vibration,

good luck,

 

[moltenmetal]

I can tell you that I've been involved in the design of piping for my entire career of over a couple decades, and I had to look up what a "sweep-o-let" was because I had never seen one before!

I can tell you that they are rare indeed for linesizes below 8", which is where I've always lived.

Despite never having seen one before, I understood their advantages- and disadvantages- instantly when I saw the part on Bonney Forge's website. They are an entrance loss, stress intensificaton/fatigue failure mitigation tool, and a fine alternative to the other solution to such problems, which below 8" is typically the installation of a reducing tee, requiring new butt welds in the run of the line. The use of a reducing tee becomes very rapidly unfavourable as line-size increases. This device reduces not just the cost of the tee, but also eliminates the two full circ butt welds to a much smaller single butt weld, albeit one which could be difficult to examine by radiography I would imagine.

Whenever stress intensification, fatigue/thermal cycling or entrance losses are not a primary concern, people are going to use either another type of o-let (butt weld-o-let, sock-o-let etc.) or a stub-in connection because the cost of the components and welding will be lower.

 

[GD2]

NovaStark,
Good luck if you can get any meaningful response from the EPC explaining why they used the Sweepolet. The project had been completed, all drawings and documents handed over to the Owner and engineering team had moved to other projects.
Remember that it's next to impossible to predict flow-induced vibration by Design and Stress Engineers in the design phase. It's the Plant Operation/Maintenance team that has to fight the dilemmas after the project is handed over.
If you don't have your internal resource, try to contract another EPC to investigate and give you a solution. There is surely unacceptable load applied to the weldolet, may be by flow-induced vibration or flexibility issue between the branch and the header pipe.

Ganga D. Deka, P. Eng
Canada

 

[NovaStark]

@GD2

Yeah it has been a challenge, I recently learned that the pipe fitting was actually changed from WOL to SOL and no one at the EPC could have given me an answer.