B31.1 and 10,000+ psig Tube and Fittings - Manufacturers sell items with stresses above yield 1

[Killstadogg]

Hi all,

I'm designing an ASME B31.1-2012 system that operates at 10,000 psig. After considering possible fluid transients I selected a design pressure of 12,000 psig. I've never done anything at this high of a pressure before so it's been sort of a learning curve figuring out what fittings, hoses and tubing are available. The system is essentially a custom high pressure spray apparatus. I realize that B31.1 isn't really the code to use for this application but the client requires it. End of story there.

However, to get to the point here, I'm finding out that even though manufacturers such as Parker and Spir Star make products that are rated to 15,000 psig and beyond, I'm finding that per ASME B31.1 straight pipe minimum wall thickness calculations that these products would experience yielding.

The minimum wall thickness per ASME B31.1 para 104.1.2 essentially determines the minimum allowable wall thickness based upon the stress at the inner pipe wall. By using thick-walled pressure calcs, I'm finding that the stresses present in the inner pipe wall of off-the-shelf components are above the yield strength of the specified material. However the stresses present at the outer pipe wall are below yield strength.

For example let's look at this Parker MS15-211 stainless steel tubing. They claim the max working pressure is 43,000 psig, so we'll use that as the design pressure. The material is 316 SS, which has a specified minimum yield of 30 ksi. We'll just assume this is the maximum allowable stress before yielding - which normally we'd use the allowable stress value given in ASME B31.1, Appendix A. The OD of the tubing is 1 inch, so using B31.1 eq 7, and we get tm = (43000*1)/(2*(30000 + 43000*0.305)) = 0.499 in
where:
y = 0.438/(0.438+1.0) = 0.305

This result is essentially telling us that the tube is not acceptable at this pressure with this yield strength (OD - 2*tm) = very small.

However, the ultimate tensile strength is not reached because if we plug in 70 ksi into B31.1 eq 7 we get tm=0.259, therefore the ID could be as large as 0.48 inches.

The major caveat here is that this high of a yield stess is only present at the innermost surface of the pressure boundary. Per thick-walled pressure vessel calcs, and the distortion energy failure theory, yielding occurs at the inside surface of the pressure boundary, but not the outside surface.

So, to finally get to my question here: is it reasonable to justify the components are acceptable for use in a B31.1 system because yielding doesn't occur at the outside surface and that the ultimate tensile stress is not reached anywhere within the component wall? The calculations show that 'yielding' occurs at the inside surface, however it doesn't seem that the 'yielding' metal would have anywhere to move to, considering the other material is staying put. Perhaps this material essentially strain-hardens. Considering the components that Parker and other manufactures make seem to do just fine in regular industry, I personally don't see a problem. I really just need a solid argument to justify components per B31.1 para 104.7.2, which requires "... calculations consistent with the design criteria of this code".

Any and all advice is more than welcome!

Thanks

 

[BigInch]

If you use the equation given in the design codes and it checks out, your design is usually code acceptable. It must also be an acceptable solution to the engineer in charge, so unless you have reason to believe the equation is not applicable to your specific situation and you feel that it produces reasonable results, stop there.

Usually we see the other extreme, the equation given in the code does not show acceptable results and people try to justify the use other formulas that will give lower stresses and acceptable results to get it approved.

 

[Killstadogg]

Thanks BigInch,
The equations provided in B31.1 show that the components DO NOT pass. However these B31.1 equations are based on the stress at the inside surface.

Thick-walled pressure vessel equations (non-B31.1) show that yielding occurs on the inside surface but not the outside surface. They also show that the ultimate tensile stress is not exceeded anywhere.

 

[TGS4]

There are minimum-specified margins required for both yield and ultimate - so designing to yield would be a non-starter.

I know you said B31.1, but B31.3, Appendix K is intended for high pressure. Perhaps you could use the formulae there, concurrent with the allowable stress basis from B31.1?

When it comes to thick-wall stuff (pipe or vessels), I tend to prefer the approach in Section VIII, Division 3. A proper assessment needs to done using an elastic-plastic non-linear approach - because the behavior is fundamentally elastic-plastic non-linear.

You also ought to be asking hard questions about cyclic service, too.

 

[moltenmetal]

What on earth you're doing with those pressures under B31.1 is a mystery to me entirely...

 

[btrueblood]

"The material is 316 SS, which has a specified minimum yield of 30 ksi. "

I added the italics. Austenitic stainless steels can be work-hardened to yield strengths over 80 ksi. I would assume the manufacturer has done this (via autofrettage) to produce tubing with the stated working pressure limits, but you could probably confirm that with a phone call to their technical support people.

 

[Killstadogg]

btrueblood - Good point, these parts are probably autofrettaged. There's where I could make my argument. I will ask these manufacturers if autofrettage is part of their process.

 

[TGS4]

One correction to me earlier post - for these pressure you should be using B31.3 Chapter IX, using the allowable stresses in Appendix K. The equations that you are using in the Mathcad sheet that you posted, are incorrect - equations 34a through 35d are more appropriate.

On that basis (specifically equation 35a), those pressures/thicknesses require an allowable stress of 52,100psi, which even for autofrettage SA-316 TP316 seems aggressive (UTS is only 70ksi). The allowable stress basis in B31.3 Chapter IX is 2/3 Sy.

If you are considering autofrettage, then VIII-3 is the only way to go.

 

[moltenmetal]

If the components are "cone and thread" fittings on heavy wall tubing, indeed the tubing is usually what is referred to as "1/4 hard", not fully annealed. Note that the strengthening is LOST again if the tubing is used above a certain temperature.

 

[BigInch]

Cherry picking criteria from the most favorable code and combining designs amongst them is not allowed. Once you chose a code, you have to live with the provisions of that code, unless you can prove that the provision does not apply to your specific case.

 

[TGS4]

BigInch - I wasn't promoting cherry-picking. However, B31.1 does not have rules appropriate for such high pressures; but B31.3 does. In my opinion, the selection of B31.1 in this instance is a mistake - compounded by an application of non-Code equations coupled with non-Code design margins.

If this were me, I would apply B31.1 Chapter IX or VIII-3.

Note that even within the various Codes are opt-out clauses that permit evaluation of out-of-coverage situations. Sometimes they are prescriptive: B31.1 304.7.2, while others are not: VIII-1 U-2(g).

 

[saplanti]

This might give you better explanation about the tube selection criteria of Parker:

http://www.parker.com/literature/In...ogs/Bul 4200-TS - Feb2010 - Final Low Res.pdf

 

[saplanti]

These are others:

http://mdmetric.com/4300/Metric Tube.pdf

http://www.ihservice.com/PDF's/Tube Selection Chart.pdf

So you need to be very selective of the code to be used for your system. This will change the wall thickness accordingly.
In case you mention the usage of the tube with brief explanation might give better idea to understand, and experts can provide adequate advice.