FEA Modeling Pipe-to-Vessel Interface 1

[SMontgomery119]

Hello,

I am looking for FEA software to help model a pipe-to-vessel interface.

As displayed in the attached drawing, the pipe has two 90-degree bends shortly before interfacing the pressure vessel. The pipe/vessel interface is supported with a 24"x24" and 3/16" thick plate. The piping fluid temp. variation is 95 to 400-degree F.

The problem is a recurring weld failure at the pipe/plate interface.

Any fea software suggestions to redesign the pipe-to-vessel interface?

Below are some packages I've found:
Chempute - nozzle pro
Algor - pipepak
Codeware - nozzle pro

Are there other suitable fea programs?

Thanks,

Steve M.

 

[TGS4]

Please, please, please do not "buy" a piece of software, especially FEA software to solve your one-of problem. The level of skill required to PROPERLY perform an FEA is significantly more than you might expect.

Please, just find a consultant who does this all the time and pay them to solve your problem. In the end, it will cost less, AND you will know that it is done right.

 

[FILGUEIRAS]

SMontgomery119,

I cannot see the attachment. So, I missed some details, as the size of things, pressures, number of cycles, and the like. It will be good, if tou provide them.

About tha problem, have you considered to split it in two, calculate the flexibility of the piping, then apply the results to the nozzle ( it´s a nozzle, isn´t it? )?

 

[SMontgomery119]

Yes, it is a nozzle, i.e. a 20-foot pipe, including 2-90-degree bends, attached to a vessel.

It appears that nozzle pro will model the pipe w/ beam elements.

The failure is in the weld at the pipe/vessel support plate interface. The load is a 95-to-400 degree F temperature differential.

Perhaps, spliting the model in two as you say may be the way to go. Using a beam element to model the pipe a the pipe/vessel interface doesn't seem right.

--Steve M.

 

[FILGUEIRAS]

There are still informations missing. Allow me to do some assumptions:

- the pipe has not a large diameter - comparing with lenght;

- the temperature differencial occurs from pipe ( higher ) to vessel ( lower );

- the number of temperature cycles is low ( like maintenance cycles );

If so, I would analyse the line with a flexibility software and apply the results do the nozzle, per WRC-107. As your line has only 2 bends, may be you find interesting some vintage graphs by MW Kellog, and calculate the line by hand. Very exotic, these digital days - but, believe, I´m not kidding.

If not, you may be involved in a huge problem, mainly if your product can explode, like in Flixborough.

Then it will be better to follow TGS4's advice, and hire a specialist.

 

[MJCronin]

S&M...

I suspect that this is not a real "pressure vessel" designed to any particular code or standard.

I also sam sure that the 24"x24" is far too wide and thin to offer much help in a distribution of the load across the face of the vessel. You don't need an FEA program to tell you this...

As a geezer who is rapidly approaching the end of the ride....let me suggest to you:

"When in doubt make it stout"

I suggest that you install a thickened nozzle at the interface, with full "ASME Style" penetration welds and a flanged end. Make it extend about 6-9 inches from the face of the vessel. Flange your piping and connect it to this new nozzle. Try for about double the wall thickness of your attached piping system.

If you feel that your nozzle is still too flexible, install triangular gusset plates at four places around the nozzle. Full fillet welds all around.

Your goal here, of course, is to distribute the stiffness (and thus the deflection)of the nozzle into your thin reinforcement plate.

Remember, full penetration welds ( not those crappy, easy to make fillet welds)are required.

My opinion only

There are other solutions.....

-MJC

 

[LSThill]

SMontgomery119 (Mechanical)

Also, for your education: FE/Pipe & Nozzle Pro by Paluin Research Gorup

http://www.paluin.com

application software for your solution.

Also, calculate the stress in circular reinforcing support rings: MJCronin (Mechanical) "install triangular gusset plates at four places around the nozzle" utilizing the analysis developed by Blogett found in DESIGN OF WELDMANTS in Sectin 4.7. (Look or ASK for the Solution Book of Problems that go with Blogett found in DESIGN OF WELDMANTS)

You need to review support section moment of inertia, neutral axis and affected lingth of the pipewall.

LSThill
Leonard Stephen Thill