Vibration Induced Fatigue, Small Bore Instrument Tees, Incompressible Flow

[Doove117]

Hi
I am a draftsman who is studying to become an engineer so please excuse my ignorance.

We have a job and have subcontracted out pipe stress analysis, including an assessment to EI guide on vibration induced fatigue.

It is 4" titanium pipe with a 4" to 2" reducing tee then a little downstream a 4" to 1.5" Weldolet branch. Both these short (<300mm) branches are dead legs with pressure instruments in.

As the fluid is water at ambient temp. and the pumps are centrifugal and velocity is approx 3m/s, I can't really see why these branches would be subject to vibration. I don't have a copy of the EI guide though I'm working on it for my birthday present from my wife along with the B31.3 Complete Guide!

I ask as the consultant can't seem to explain it either which is a bit worrying, but it might be my ignorance and him assuming I know more than I do!

Thanks for any explanation of this vibration failure risk in advance.

 

[hacksaw]

based on my own experience it is not likely a problem for the flows you describe, the customer or their engineer is usually tasked to identify the risk factors before your team sees the design

 

[BigInch]

Centrifugal pump bearing wear can produce at least some vibration.
With the cost of titanium pipe, I can imagine that somebody doesn't want any problems at all... with anything.

I hate Windowz 8!!!!

 

[Doove117]

Thanks for the response guys.

The consultant has done the LOF risk calcs according to the Energy Institute guide and they have indeed come back saying these lines should be braced.

As I don't really get the physical failure mechanism I'm a bit unsure about the bracing design. I had thought a weldolet type branch fitting didn't usually need bracing but not sure if the EI guide takes this into account.

Nor am I sure if 'bracing' can be achieved by a clamp support around the small bore pipe close to the branch, like a rubber lined clamp.

The pipe is ambient temperature seawater and about 3 barg; customer specified Sched 80S Grade 2 titanium pipe. Bit OTT I suppose but it's built to last so I don't want to have a branch weld fail, however if I can avoid buying in Ti bar stock and welding ribs on the branch then I will!

Staying with my run of confusion; as these welds must be subject to the same vibrational energy as the branch /reducing tee welds why won't they fail?

Any advice?

I'm thinking I need the EI Guide sooner rather than later.

 

[davefitz]

Some thoughts, based on below listed facts.
1. typical thermocouple intallations in small pipes use forged or cast tees.
2. titanium has a low value of E ( youngs modulus) and that may adversely affect hi cycle fatigue issues, I am not sure of teh fracture mechanics parameters for Ti.
3. typical failures of thermocouples ( also injection quills or spray nozzles) start with a sharp thermal stress initiating a crack, then fluid-elastic vibrations generate a hi-cycle fatigue failure- once a crack is initiated, it has a stress concentration factor of 5, and fracture mechanics-type calculations are needed to predict failure.
For example, the failure at the Monju (japan) liquid sodium cooled breeder reactor occurred at a thermocouple, whose detailed design included a "pre-existing crack" as part of the weld detail.

Bottom line is , provide a detailed design without a pre-existing crack and , if severe thermal transients are expected, then other design detials must address that phenomena==> this implies a finite element model must then be made of the detailed design to confirm thermal stresses will not develop an initiating crack in its normal commercial lifetime.

"Whom the gods would destroy, they first make mad "

 

[Duwe6]

And hire a GOOD welding inspector. Weldolets need to have full-fusion, i.e. a visible root. Lazy welders/fitters set the olet down on the pipe and then weld around it - thus no root, and now that "pre-existing crack" effect. And it only takes one.