Pressure test a cast pipe tee 1

[Ralph2]

Hello
Can some one advise or point to any information on this problem.
A cast tee nominally 10" ID needs to be pressure tested to 4200 psi. Our plan is to temporarily weld blind flanges on the 3 open ends.
The question .......what thickness of flanges (plate we will cut) do we need to use.

Thanks for any information
Ralph

 

[aberta]

You can calculate the thickness from the PressureVessel Codes. If you do not have one, I would suggest you get someone to do that. I presume your cast tee is cast steel, and the metal thickness will be well in excess of 1". Welding something to it, pressure testing it and them removing it to re-weld will damage the weld area even if you normalize it afterwards. Also when welding you will require preheat, postheattreatment and post heat controlled cooling.
I would strongly recommend you get the tee x-ray tested, and after field welding and normalizing, x-ray the welds. If you have to repair weld ot tee, it must be normalized after repair.
If you are still inclined to pressure test the tee, weld pipe stubs long enough so you can weld stubs to pipe in field. Your flanges and fittings should be class 1500.
Hope this helps you.

 

[Ralph2]

Thanks..
The casting is currently 1" (or so) too long at each end, we will be machining a weld prep to the required length when it is all complete. Most (?) of the "damage" we incur by welding will be machined back, as well the "tee" will be eventually welded in place so is this still a concern?
From a catalog I find that a class 1500 10" blind flange is 4 1/4 inches thick... Go with this??? It seems excesive considering that the wall of the casting is only 1 1/2..

 

[unclesyd]

[red]DON'T[/red] weld Blind Flanges to the Tee for those pressures. I is better to weld a pup piece and a pipe cap of the correct ratings.

I've got to go out, hopefully somone will jump in and help you.

 

[nbucska]

How are you going to connect the T to the rest of the
piping? certainly not by welding !

Test it after you finish the macshining, etc and
with quality control assure that not too many
fail.


<nbucska@pcperipherals DOT com> subj: eng-tips

 

[aberta]

Ralph2:
Unclecyd is correct. Go with a pup piece and end cap rated for class 1500. Alternative is to weld a WN flange to the tee and then bolt blind flanges to the welded flanges. You must use class 1500 gasket and cover the flange separation faces with a shield so that if the gasket leaks, no one would get hurt. Use the smallest size pump (hand pump) to minimize water content accumulated in the system. Take measure you donot stand near the system being tested. This is dangerous work and is not to be taken lightly.
Your fill valve should be welded bonnet (no bonnet gaskets), small size ( say 1/2" class 1500). The only problem you have now is to weld the valve to the cap/pup/flange. The best is to weld a 6000# weldolet to the pup. Use a bevel end valve to weld to the weldolet.Unclecyd recommendations are correct and maybe you should confirm this with him.
When pressurizing, increase and hold pressure in small increments to ensure drastic failure does not take place.
One final comment, I would spend the time to persuade your party to accept X-ray. If your dimensions and grade of material is to ANSI standards, this would be the best course of action.
Pressurizing systems to these high pressures can damage the casting which will go undetected. Most parties would accept design pressure test (2800 psi) with radiograph.
I hope you fully understand what I am trying to emphasize. Take alternative route.
All the best and have a nice week-end.

 

[unclesyd]

aberta essentially passed on my concerns of personal safety, yours and others.

You have not mentioned the operating parameters that the tee will see nor the specification and material of construction of the tee.

If you can please post the above information especially the particulars of the tee.

Let's hold up until trying to give you a little more information to work with until we have at least the material of construction of the tee.

 

[Ralph2]

Thanks to all.
I appreciate all the feed back and safety concerns.
Ralph

 

[aberta]

Unclesyd
What is your opinion about hydrostatic testing in general. I have found that a material or fitting may be damaged visually to an unacceptable level and still pass the hydrostatic test at 1.5 times, if necessary. A weld of relatively poor quality will also pass similarly.
NDT testing (ultrasonic or radiograph)on the other hand provides the most measure of quality.
Quality inspection during fabrication provides more quality assurance and than hydrostatic testing 1.5 times. ??

 

[unclesyd]

I think the pressure test is still needed even if not seeming to serve any useful function other than verify that a component will hold a certain pressure at that moment in time and keep the AI happy. Two of our biggest contentions was with high pressure jacketed pipe, we could test all day and not replicate inservice thermal expansion and the procurement for Cl free water for testing large SS vessels in intricate internals. The vessels always had to be tested at the fabricators with no good water and we had 3 sources of water on site. But being a requirement by the code and standards we had to do it and do a good job while doing same. It is just a little insurance.

I agree with the your statement other than to say the pressure test is just another step in the quality assurance program. I think that the emphasis should be on the QC up front as you state instead of depending on the hydro to tell what quality component we have. This seems to be the trend as I hear all the time: “Why worry we are going to hydro”, and then can’t do the pressure test correctly. I’ve seen more time and effort getting a system or pipe spool ready to test than the actual test including hold time.

We have been fortunate over the years in having very little equipment mistreated during fabrication or operation where a pressure test would have shown anything.

By doing the work up front as you state it would seem that pressure testing is redundant. Like wise I've seen some partially completed or corroded welds, very poor castings, and so on pass a pressure test. In 50 years of being around pressure testing I personally have never seen a failure at the pressure boundary on newly fabricated equipment from a pressure test. I have seen pictures and done failure analysis on other people’s failed components, not many, but enough.
I have seen a couple of rips on in-service equipment that was known to have corrosion problems and a few on new heat exchanger tubes.

We had the case of the unwelded long seams on SS pipe by a particular manufacturer that caused us a lot of grief. None of the defective pipe has ever showed up on a pressure test at fabrication or during a leak test at 1.1x during outages prior to starting up. We actually tested several failed sections up to the point where we would should start deforming the pipe and still no failure.
We either discovered it on visual inspection, during welding, or by inservice failures. More of the latter.

The only time I”ve see a pressure test help evaluate a vessel was in the evaluation of two Horton Spheres in anhydrous NH3 service. It was only effective due to being used in conjunction with Acoustic Emission testing equipment.

With all of the above said, I still think we should carry on as prescribed by the Jurisdictional Authorities. I don’t think with the liability situation as it is today I wouldn’t deviate from the proscribed procedure unless it is chiseled in stone, 3 duplicates.
.
Pressure testing still has to be done properly with direct control of someone so that the testing procedure in itself is safe.

My main concern has been and is for personnel safety as the equipment can be repaired or replaced.

 

[aberta]

Unclesyd, thanks for your comments; you have a wealth of knowledge and experience which is invaluable, especially to newer engineers. It is also nice of you to share it.
Thanks again very much.

 

[Ralph2]

Update on our latest thoughts.
This project has jumped through various options based on the suggestions above. The manufacturer, was and then was not going to do the welding (and supply the pups and end caps)
Now it is back in our court........ My thoughts on an alternative way to do this are now being studied by an engineer at one of our other branches. I am basing my plan on the fact that a hydraulic system would routinely see these pressures and sizes.


We are going to make 3 caps (essentially plugs with a shoulder and sized for the pressure) then use "O"(or similar) rings to seal to the inside of the Tee. This surface is machined to a finish that a "O" ring would seal on. We could hone it if necessary.
Then make essentially a metal enclosure. Our current plan is to cut a rectangular hole just big enough to allow inserting the tee with the end caps in 4 inch plate. Will make 2 of these and weld spacers to keep them separate by 2~3 inches Place the Tee with end caps in place inside this rectangular hole. Shim with plate to reduce the gap from the caps to the frame Add some locating bars to ensure the Tee stays centered

Place the whole assembly in a concrete pit with a steel (1”) plate over and pressure up.

This way we avoid any welding to the tee (which is SS of some sort)and the cost of temorary piping. As well because we have we have two to test (different sizes) we can incorporate the design for the next one.

????????
Ralph

 

[btrueblood]

Umm. I have some reservations about pressure testing pipe tees in the first place. Piping generally does not conform or comply to ASME pressure vessel standards. The main reason for this is that axial stresses in pipe and fittings are assumed to be reacted by end blocks and supports along the length of pipe runs. Thus, a tee fitting, if tested by capping the open runs with flat flanges, bolted to the fitting and not reacted by other means, will apply a tensile stress across the tee in two directions. I've done some FEA of this situation on cast iron pipe tees and came to the conclusion that this would be a fun test to have someone I don't like do.

In regards to the last post from Ralph2, it looks like he is planning to react the caps on the tee fitting to an external support (4" plate). This eliminates my concern above, but: is this a reasonably good simulation of how the fitting and its connections will be supported in use/operation of your system? If not, you may want to re-consider.

Ben T.

 

[boo1]

Normally piping systems testing procedure shall be in accordance with requirements or ASME/ANSI B31.3 Ch VI 345
The test pressure shall be no more than 110% of design pressure. 345.1

Recommend the pressure shall be gradually increased until a gage pressure which is the lesser of one-half the test pressure is attained, at which time a preliminary check shall be made, including examination of joints in accordance with ASME/ANSI B31.3.
There after, the pressure shall be gradually increased in steps until the test pressure is reached, holding the pressure at each step long enough to equalise piping strains and not less than 10 minutes.