# Strain on NPT thread in high pressure chamber

1. Jun 3, 2013

### gpsimms

Thanks in advance to any help on this:

I am working in a combustion lab and we are designing a high pressure tube reactor. My background is only in math and I have never taken a materials or design course before, so I apologize that this is probably a very basic application and a very classical problem--but I need help.

We want to pass tubing through the chamber wall, and to do that we are going to tap female 3/8 NPT threads into the wall. Inside the chamber, we plan to go up to pressure of at least 50 atm.

The wall in which I want to tap the hole is 3/8" thick, and I want to be sure that this thickness allows enough threads to engage and secure the male NPT piece so that the threads do not strip.

It is easy to use the pressure in the chamber and the area of the tapped hole to determine the force (F) against the male NPT.

Then I divide that total force by the tensile stress area, which I got from a formula A=pi/4(d - .9743/n)^2, where d is the same diameter I used to calculate F, and n for 3/8 NPT is 18.

Finally, I divide the total force I calculated F by the tensile stress area to get tensile stress.

I compared this result to the yield strength of stainless steel and find that the yield strength is approximately 34 times the tensile stress, and the design is 'safe.'

OK here are my questions: In the above calculations, there is nothing about how many of the threads need to be engaged. Is the tensile stress formula assuming all threads are engaged? Is there a "rule of thumb" about the minimum number of threads needed for this formula to be applied? Also, I think the formula is probably meant to be used to for straight, not NPT threads. Is there a significant difference? The threaded section of the NPT piece I am planning to use is longer than 3/8", which is the wall thickness I am hoping to use. If only 3/8" worth of threads (about 5-6 threads) are in the wall, should that be sufficient?

Of course, my feeling is that I am safe, since the number I calculated has a safety factor of 34, but in the interest of not blowing up our lab and everyone in it, I'd like to be a little more sure on how to do the proper analysis. : )

Thanks!

simms

2. Jun 3, 2013

### etudiant

At 50 atm, you are entirely correct to be concerned.
It is very difficult to get a fitting securely mounted to a 3/8th thickness of metal, especially if you are doing combustion tests that may distort elements.
Afaik, it is customary to provide backups for such fittings, so you can be confident that things don't go pear shaped if something bumps the sensors. Just a couple of doubling plates should do the trick.

3. Jun 3, 2013

### gpsimms

So I think I need to more completely explain what we are doing. The chamber has a 1" thick wall, and we want to pass 1/4" quartz tubing through the 1" endplate where the combustion products will go into a vacuum chamber and molecular beam mass spec. It is a pretty complicated problem, we need to have a good seal at the endplate because the mass spec needs to be in a vacuum chamber, plus the products being very hot causes problems as well. Our plan to seal around the quartz tube through the endplate is to use a 3/8" NPT to 1/4" swagelok piece with graphite ferrules. Unfortunately, this whole piece is too long to fit completely in the 1" endplate, and our quartz nozzle cannot protrude past the 1" endplate. The swagelok end is about 1" exactly, so we want to add a plate to the NPT side. We want it to be as short as possible, so there is not too much space for cooling as the combustion products pass through the endplate (with add-on plate). We hoped 3/8" would be sufficient to safely fit the NPT end, but do you think I should recommend to my professor that we go to 1/2"? I am going to see the machinist tomorrow, so I will get his input as well.

Thanks so much for the quick reply!

simms

4. Jun 3, 2013

### etudiant

Seems like a fascinating problem, how to get the combustion samples stat to the mass spec, when the combustion phase is 50 atm and the mass spec chokes at 0.01 T.
I think absolutely you should go to 1/2", but your design challenge is really one that you should get expert help on from your lab. The quartz tube seems another possible trouble spot, although presumably you are doing pulsed or deflagration type combustion, not explosions.
In any case, there has to be some group or person in your institution who is responsible for safety/liability and you should get their formal blessing. Their overall construction requirements are above my pay grade. It is much better to call in available support beforehand if you have doubts than to dismiss them and possibly experience an accident.

5. Jun 4, 2013

### gpsimms

Thanks for the feedback. I am just preparing a presentation to my professor about what I have found. Don't worry, safety is the obvious priority.

I am still curious, though, I found a formula for length of engagement Le = 2*At/(.5*pi(D-.64952/n))

Where At is tensile area as calculated above, D is diameter, and n is threads per inch. My understanding is this formula determines the length of engagement needs to ensure the stainless fitting fails before the threads sheer off. This is not *exactly* what I want to figure (especially because the answer is almost 5/8" which is much longer than I would like for the inner plate thickness). What I would rather know is if there is some sort of formula for calculating the # of threads required to handle a particular load.

Anyone pointing me in the right direction on this would be appreciated!

Thanks etudiant for all your help so far,

simms

6. Jun 13, 2013

### 256bits

I am not exactly understanding how you want to design your fitting. Is it that you want to drill a hole in the wall of the vessel, tap it with a thread to accept a threaded pipe.

I am just curious also why you selected NPT threads.

You must know that is a tapered thread with the depth of thread engagment fixed by the nature of the thread, so calculating the required thread depth for a certain diameter of pipe is fixed and cannot be altered.

One problem you may have with NPT is sealing. Normally one would use something such as teflon tape wrapped around the threads. If you have a high temperature, you may want to investigate whether the connection can be adequately sealed and does not deteriorate with use.

The formulas you are using do apply for UN threads, and might not be totally applicable for NPT.

7. Jun 13, 2013

### gpsimms

Hey 256, Thanks for the reply.

You are correct as to the design. Basically, to get through the chamber wall we have an NPT to swagelok fitting. The NPT end will screw in to the tapped hole of a 4/10" thick plate (orginially, I planned 3/8", but after discussion we decided to go with 4/10").

What I want to be sure of is that the chamber can go to high pressure without shearing the NPT threads off and shooting the fitting into the lab. I did fear that the formulas I cited did not apply to npt, but have not found others, and am not sure exactly how to analyze.

To answer your question, we selected NPT so we could get a better seal, and we plan on using it with teflon tape as you say. We do understand that it will get hot and there may be oxidation issues. The plan is to switch the piece out often between experiments. That is also part of the reason we are using 3/8" NPT, because we are only using 1/4" tubing, and are going to try to drill out the NPT end with enough room to wrap the tube with a layer of thin ceramic paper, which will hopefully help the connection last under pressure/heat.

Thanks for the info and help!

8. Jun 13, 2013

### 256bits

If you drill out the NPT side, make sure the wall thickness is able to withstand 50 atm.
Plus, you also should determine the maximum temperature of the fitting and apply a degridation factor from room temperature for the stress, ( and for your pressure chamber if you haven't already done that)