Double Shoulder Connection Thread Form and Design Question

[Walther1522]

Hello all,

I have some questions concerning threads that I need some help with.
I am designing an improvement for a tool for O&G industry to handle max 40,000 ft-lbf of torque and 30,000 lbf of axial force.
The component is maxed out at its current OD of about 3 inches and minimum ID of about 1.5.

I have the machinery hand book and API spec hand book but I still have several questions some fundamental concerning threads.
I chose to start off with a 4 TPI, 3/4 TPF stub Acme double shoulder thread. This particular connection calls for double shouldering.

My first question is does anything in particular need to be done to the thread form of a stub acme for it to be able to double shoulder with no issues?
The reason for the taper is that a straight stub acme has been used before and it was cracking at the thread relief so I have decided to taper it and run the thread out similar to an NPT thread to also create a sealing surface as this is a high flow section of the tool. Which leads to my next questions, I have read that stub ACME threads are centralizing if this is so would this negate the sealing properties that a taper intends to contribute to the design?

Secondly, what formula is best to use for the shear area of a custom thread form and what is the best practice for calculating shear area and torque to shear for a tapered thread? I have seen several answers to this question from calculating it as a straight thread using the smallest diameter, to calculating it as a straight thread using the AVG between the smallest and largest diameter of the last full thread. The API handbook provides a formula that includes the taper per inch within it but I am getting vastly different answers from the API formula for tapered threads vs the straight thread machinery handbook formula when using the practices above that conceptualize a tapered thread as a straight one IE AVG diameters method, Minimum Diameter method.

Lastly, I have seen tools similar to this one using a custom modified 10 deg thread that work really well and that raises a question for me, why does that thread work? What is the benefit to having the flank angle less or more in this type of application?


Thank you for any help,

 

[SwinnyGG]

My first question is does anything in particular need to be done to the thread form of a stub acme for it to be able to double shoulder with no issues?

If the thread form is manufactured to the same tolerances and a normal API thread which is double-shoulder compatible (or any one of the wildcats) than the actual threadform itself doesn't really matter. Your manufactured parts have to hit the same shoulder-to-shoulder tolerances either way, and when made up the pin will rotate in the box until the thread faces are in contact. Threadform doesn't matter much as long at the shoulder arrangement is designed and manufactured properly.

Secondly, what formula is best to use for the shear area of a custom thread form?

The standard formulas still apply. As long as the thread taper is the same angle from root to tip, when you average the pitch and major diameters, the thread angle doesn't matter - you always get the right number.

I have read that stub ACME threads are centralizing, would this negate the sealing properties that a taper intends to contribute to the design?

Tapered threads (ie, NPT) provide sealing via contact between the tip of the thread and the thread root... you don't want that in your application. You can seal via tip/root contact, or you can seal via the conventional tapered sealing face that premium double shouldered connections use; you cannot use both methods at the same time. Tip/root contact in NPT connections relies on plastic deformation of the threadform to 'pack' the root with material. This works fine for connections which aren't subject to any mechanical load after install; you're operating in a different universe, and you definitely don't want plastic deformation of your threadform, as it will make your tools useable for a single makeup before they are junk.

What is the best practice for calculating shear area and torque to shear for a tapered thread? I have seen several answers to this question from calculating it as a straight thread using the smallest diameter, to calculating it as a straight thread using the AVG between the smallest and largest diameter of the last full thread. The API handbook provides a formula that includes the taper per inch within it but I am getting vastly different answers from the API formula for tapered threads vs the straight thread machinery handbook formula when using the practices above that conceptualize a tapered thread as a straight one IE AVG diameters method, Minimum Diameter method.

All of these methods you've listed 'work', with varying levels of conservatism. Using the smallest OD is going to be extremely conservative; using the average is still conservative but less so; using the largest diameter is likely closest to reality (as the base of the thread is where these joints will typically fail). Based on the level of study of these applications, I would shade much more toward using API methods over something out of MH.

What is the benefit to having the flank angle less or more in this type of application?

As thread taper approaches 0°, ultimate tensile strength of the assembly goes up, and ease/speed of joint makeup goes down, and vice versa. I would assume that 10° was chosen because it would provide the maximum strength without compromising makeup speed, but that's just an assumption.

 

[3DDave]

This is the sort of joint that will practically require FEA to properly analyze. In typical thread analysis it is assumed the body the thread is attached to is infinitely rigid radially, but in this case the relatively thin wall deforms, allowing the thread to move with it. This shifts the contact away from the thread root, increasing the bending stresses.

I tried to read this paper:

https://www.researchgate.net/public...thread/link/5cb6e0ad299bf120976acd3e/download

. I don't see exactly what the influences are doing to the threads, but I can see that the effects they are examining aren't hand-calc compatible.

It's interesting - the first stop must be at the bottom of the box. Further tightening then stretches the box and compresses the pin. the first stop apparently is the top of the box. Further tightening then compresses the box and stretches the pin. This changes the thread pitch on both components which continuously changes the thread elastic deformation along the joint. When the second shoulder makes contact it starts to shift that pitch change back the other way.

Edit after reading

https://www.iadd-intl.org/media/files/files/0d5ecb5d/What_Keeps_Your_String_Together_2_.pdf

more closely.

 

[3DDave]

Also - it might be worth asking Mark Garrett at

https://drillpipe.com/contact/

I found a presentation he did on the general topic. No guarantees, of course.

 

[Walther1522]

@SwinnyGG

Thank you so much for your answer. It is incredibly helpful.
So it appears I was wrong for one of the purposes of tapering a connection here.

I was under the impression that tapering a connection serves the following:
-Sealing properties
-Greater length of engagement for a given total length of pin so greater total thread shear area.

the make up face at the shoulder for my particular connection will be flat so 0 deg. Does this mean that tapering this connection will provide no extra sealing advantages that a straight thread would provide?
Is there any other advantages that tapering a pin and box has over a straight thread?

Again thank you in advance and for your previous answer.

 

[Walther1522]

@3DDave

I have seen that paper and similar others to it as well. In the API handbook there is a bit about deformation of the pin and box and to back up this concept when you look at connections that have been made up several times you do see that they are compressed so that is definitely in line with the paper.

Sadly, I can only perform axisymmetric FEA's as I am limited to using solid works. It is useful but from what I can see impossible to simulate make-up torque and even regular torque. I do my best to conceptualize how torque axially pulls the connections and set the pre conditions that way but I must admit its definitely an a approximation and not as accurate to reality as I'd like.

 

[SwinnyGG]

Is there any other advantages that tapering a pin and box has over a straight thread?

The standard API pin/box system is tapered because it drastically speeds up makeup time. Watch a video of drill pipe being strung out, ie:

https://www.drillingformulas.com/making-drill-pipe-connection-by-throwing-a-spinning-chain/

And pay attention to when the pin drops into the box initially - without any rotation the pin is already 90% of the way home, and it's also already self centered by the two tapered threadforms contacting each other. The joint is then seated with only a few turns on the pipe at the top of the stack. This is possible because of the taper. If this was a straight thread, to makeup the joint you'd first need to get the pin and box aligned, which would be difficult in normal working conditions, then you'd have to turn the top section of pipe 50+ turns (or whatever, based on pin length and thread pitch) to seat the joint. This would slow down the process by an order of magnitude..

The taper has nothing to do with sealing. Sealing is provided by the shoulders, or by separate tapered sealing faces - always independent of the threadform.