NPT Hoop Stress Calculation 1

[spacecowboypaur]

Hello All:
The application is screwing a NPT fitting into a boss on a cast pressure vessel. The fitting is sometimes cracking the boss during installation, and I would like to 1: Calculate the hoop stress from the NPT and 2: Be able to control the strength of the material at the boss to prevent this from happening again. Thanks.

 

[Cockroach]

This is the result of excessive applied torque. You need to look through the various threads in this forum for the equations to interference fit, then compute the torsion required to makeup the thread for an allowable interference.

As a quicker alternative, just ask the manufacturer for the recommended makeup torque.

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[spacecowboypaur]

The part is a very complicated die cast part, and the boss that the NPT's mate to is the last area to see material. I tested these and found that there is a high variability in the strength of the material in boss. I wanted to get some sort of correlation between turns of an NPT to hoop stress, and also some way to test that small circular boss for it's material strength.

 

[scastillo]

When you say the fitting is sometimes cracking the boss am I correct in assuming you are repairing or replacing the boss? If so, have you thought about converting both Boss and fitting to a SAE connection? That should help eliminate the excessive stress.

 

[spacecowboypaur]

The boss is just a cast boss, I'm not sure if I am using that term correctly, it is an outlet of a diaphragm pump that has a machined NPT thread in it. When a fitting is inserted into the outlet, it will occasionally crack the housing.

 

[JohnBreen]

Hi spacecowboypaur,

What Cockroach is alluding to is that we believe it is NOT a hoop stress problem. The most likely cause of the failure is torsional shear at the local area of the boss. This is especially likely with the brittle material.

If you really HAD to try to convert the act of driving (turning) the NPT fitting (like a wedge) into the boss into its component stresses, it would be very complicated as the geometry (with threads) would be difficult to model accurately. Also, each thread is an incipient "crack" and with the brittle material involved a break (failure) along the line of a thread would be most likely.

You might try to establish a failure torque experimentally and from that place a torque limit on the fitting but I would not be surprised if experiments did not produce characterist repeatability. If you used brittle coatings with the experiments you might be able to come to some conclusion about the distribution of surface stresses.

Good luck.

Regards, John.

 

[scastillo]

OK, so the entire housing is trashed?

If getting a housing with a SAE port is not feasable you might try getting an NPT/SAE adapter. I know Swagelok makes these with both male and female threads or combos. Have the adapter threaded into the boss at the right torque, then use the SAE for the periodic hookups.

I admit this doesn't answer your first question, but it may help with your second (prevent from happening again). We experienced similar problems at our plant with SST NPT fittings being over-torqued in plastic housings. As soon as we switched to SAE ports the problem went away.

 

[spacecowboypaur]

First off, thank you to everyone for replying to my question.
Yeah, I ran into the problems in calculating the stresses because of the thread geometry and it sounds like there is no easy solution to get an accurate stress calculation in this situation. I did try an adapter setup to relieve the direct stress on the casting by using a straight thread to NPT tread adapter with an internal seal. That did work, but was (as I'm sure you're all familiar with) shot down by marketing. The direction from that point was to make the casting stronger, as it was in the past before we changed vendors/casting tools. I had the vendor change the gating on the part, and that made the outlet stronger, but the concern was that if we switch vendors again, there is no way to control the strength of the metal at that particular point.

 

[Cockroach]

You will then need to weld repair the boss area and recut the box threads. Using a sealant such as threadlok may help you with the lubrication aspects of the threads during makeup. I would also be torquing to industry specifications on the pin.

If marketing has a problem with that, then perhaps they can carry out the engineering analysis! Obviously you have a sense of duty to demonstrate due diligence and possible catastrophic failure if the piece fails. Protection of public property and upholding the safety code are paramount to whatever marketing can bring to the table!

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[teranko]

NPT threads are tapered with published maximum insertion depths to provide adequate joining strength and yet prevent splitting the fitting that the pipe is being threaded into. Monitoring the torque is not a reliable method for determining the induced hoop stress as it varies depending on how clean the threads are, whether or not pipe dope is used, etc.