Allowable stress of SA-193 (B7 or B7M) studs in B31.3 systems (or ASME Section VIII - Div 1)

[auzie5]

B31.3 Paragraph 302.3.2 (a) states for bolting materials that: Design stress values at temperature for bolting materials shall not exceed the lowest of the following:

(1) One-fourth of specified minimum tensile strength at temperature.
(2) Two-thirds of yield strength at temperature.

(See page 13 of B31.3 2010 edition. Note that there are five other criteria to verify are not lower than the two criteria summarized above but (1) and (2) will help simplify the problem at hand.)

Furthermore, Table A-2 lists Design Stress Values for Bolting Materials.

If we take a SA-193 Gr. B7M stud at 800 deg F we have the following:

Max Tensile Strenth = 86.7 ksi
• ASME Section II, Part D Table U – page 471 in 2011 edition.
• ¼ * 86.7 ksi = 21.7 ksi

Max Yield Strength = 56.3 ksi
• ASME Section II, Part D Table Y-1 – page 563 in 2011 edition.
• 2/3 * 56.3 ksi = 37.5 ksi

Design Stress = 18.5 ksi
• ASME B31.3 Table A-2 – page 206/207 in 2010 edition.


Question 1:

If a flange needs to be bolted together using a hydraulic tensioner (so we can avoid discussing the conversion of bolt torque to bolt preload) what is the limiting bolt stress value?​

The obvious answer is to use the lowest of the three values listed above but B31.3 Paragraph 302.3.2 (a) does not make reference to Table A-2 (Design Stress Values for Bolting Materials). Furthermore, most gasket manufacturers require higher bolt preloads than can be attained using any of the above values to properly seat the gasket (for most B16.5 flanges).

Question 2:

I’ve seen many hydraulic tensioning procedures that recommend tensioning until a stud stress of 45 ksi is achieved. Would it violate B31.3 code to tension a B7M stud to this level which is above all of the three stress values listed above?​

Although 45 ksi is below the room temp yield strength of an SA-193 Gr. B7M stud (105 ksi) it is very close to the yield strength at 800F.

Thanks in advance for any comments.

 

[bernoullies123]

302.3.1 (a) directs you to the Table A-1 and A-2 for bolting material stresses. Again they note that those are determined from the next clause as you have defined in your original post. Interesting that the numbers aren't aligned as you pointed out.

 

[bernoullies123]

Flange design in 31.3 refers you to ASME BPVC VIII appendix 2. The bolt up is addressed here as well as a pretty good explanation about bolt up stress is covered in non mandatory appendix S which should satisfy your query.

 

[auzie5]

Thanks for the helpful comment bernoullies123!

From ASME BPVC Appendix 2 I am directed to UG-23 (Maximum Allowable Stress Values) which stipulates the use of ASME Section II, Part D, Table 3 for bolting (page 336 in 2011 edition). These are the same values listed in ASME B31.3 Table A-2 (page 206/207 in 2010 edition).

It seems like an obivous requirement that we need to adhere to allowable stress values when designing bolted connections; However, I encourage you to review any common bolt tensioning procedures. Most recommend tensioning to higher stresses than those limited by B31.3/BPVC.

Take for example:

Type: ASTM A193 Grade B7M Stud
Size: 1" dia.
Design Temp: 800 deg F
Allowable Stress: 18.5 ksi

Therefore, max allowable tension in the stud is limited to:

F = S*A

Where F = max stud tension, S = allowable stress at temp., and A = tensile stress area of 1" B7M stud (from ASME B1.1 Table 6 - page 72 in 2003 edition).

F = (18,500 psi) * (0.60572 in^2)
= 11,205.82 lbs

Compare the above value to the tension value recommended by API 6-A, Appendix D, Table D1 & D2 for flange bolting:

http://www.sigmafasteners.com/category/stud-bolts

F = 31,802 lbs which leads to a stud stress of 52,502.8 psi...well above the allowable 18.5 ksi but still below yield at 800 deg F (i.e. 56.3 ksi).

Follow the link provided (above) and review the API "TORQUE FOR FLANGE BOLTING (B7/L7 STUDS)*" table half way down the page. All recommended tensions lead to a stud stress of approxamtely 40-50 ksi which is what I've seen used in industry. I will dig up the official API 6-A code to verify these values.

But the question remains: Is everyone out there breaking code?

 

[auzie5]

Correction:

Note that the table I referenced in the link was for B7 studs. Yield of B7 at 800 deg F is 73.9 ksi. So with the API recommended 31,802 lbs we still have a stud stress 52,502.8 psi which is still above the allowable 21 ksi and below the 73.9 ksi yield at 800 deg F.

 

[TGS4]

The Code design bolt stress values are for the design of flanges ONLY.

Please refer to Appendix S in ASME Section VIII, Division 1. This has been discussed ad nauseum on this and other fora in eng-tips. Also, please refer to ASME PCC-1. You will definitely NOT seal with the design-level assembly bolt-stresses.

 

[auzie5]

Thanks very much TGS4.

I had read through Appendix S but couldn't get past the second line in paragraph two:

"The considerations presented in the following discussion will be important only when some unusual feature exists, such as a very large diameter, a high design pressure, a high temperature, severe temperature gradients, an unusual gasket arrangement, and so on."​

However, ASME PCC-1 is much more definitive about the practical reasons for exceeding the design stress values of bolts in flanged connections.

I guess I'm still a little confused about why Section VII Div 1 doesn't require us to use the "actual" bolt tensions that will be realized when designing our flange connections. Everyone seems to be in agreement that higher bolt preload is required for tightness so wouldn't it be more prudent to use "actual" bolt tension values to ensure the flanges/gaskets are designed adequately? Or at the very least to have a record of the actual loading they will undergo during construction/operation?

In any case, I appreciate all the feedback; this has been a brain teaser that has been bothering me for the last couple days. For my own records, would you mind posting a link to the previous discussions on this (or other) forums that you referenced?

Again, thanks for taking the time to consider my request for clarification.

KDW

 

[TGS4]

thread292-95664
thread292-112403
thread378-318004
thread292-321532
among many others. The search feature here works rather well...

 

[Yarik]

Hi gents, very helpful info is here, thanks.

auzie5, tensile stress area in your calcs, shouldn't it be doubled, since you use a stud (2 times more of threads per inch, since there are two nuts)?

other question I would like to ask.
I would like to calculate a force that would be applied on a gasket when the bolts are tightened to the maximum allowable load (stress). Please, suggest if the way I do it is right.
1. I determine the contact area of the flange (A[sub]c[/sub])
A[sub]r[/sub]=PI/4*(OD[sup]2[/sup]-ID[sup]2[/sup])
A[sub]r[/sub]=area of the ring
OD and ID are inside and outside diameters

A[sub]h[/sub]=PI*R[sup]2[/sup]*number of holes
A[sub]h[/sub]=area of all stud holes

A[sub]c[/sub]=A[sub]r[/sub]-A[sub]h[/sub]

2. Determine max tension force (F) of every stud (shown in the post of auzie5). I am just not sure about the tensile stress area per stud, as I've indicated at the beginning.

3. Determine equivalent pressure applied on the flange surface:

P[sub]e[/sub]=F*n/A[sub]c[/sub]
n=number of the tightened studs

The reason for all the calculations is to compare "tightening" ability of two designs:
One with 96 off 1" B7 studs and bigger contact area
and another one with 48 off 3/4" B7 studs and smaller contact area.

Would the approach be correct?
Thank you.

 

[TGS4]

Yarik - you really have things messed up. May I suggest that you start by drawing for yourself a free-body. Deal only with forces - leave the areas aside for a moment. A 2D-axisymmetric FBD will suffice to start.

 

[Yarik]

TGS4
Thanks for the directions. The reason for me doing these calculations is to compare two designs, as I have mentioned.
I am sitting with a vessel, flanges of which were redesigned (thickness increased, size and amount of bolts decreased,contact surface of the flanges decreased). As a result, the seal did not work and the vessel was leaking on the flanges after a while (the gasket was just pushed out of the flange). I try to determine why the designer went this way and try to see whether or not he was wrong in regards to the changes that were made to the flanges.

I understand that my approach to the forces transfer is not correct. But for a comparison of two designs, types of forces applied would be the same. The only two variables would be contact area and force applied to the studs (smaller stud means a smaller tightening force that may be applied to it, which is limited by the maximum allowed stud tension).

I would like to see if the values of the applied force per area unit are the same in both designs or not. If not, which one is smaller. That's it, with no deep specifics. Am I still wrong with this approach?

 

[TGS4]

Troubleshooting bad-actor flanges is much more involved than your binary approach. When I solve bad-actor flange problems, there are at least a dozen aspects involved, and sometimes as many as 20 factors. I would recommend that you find an expert in the field to have them assist you.

 

[Yarik]

Thank you for the advise, again. There is no doubt that should this case follow a legal route, than an expert will be contacted.
And in regards to other aspects, you are right. There are at least two more contributing factors I am aware of: 1. Right bolting sequence was not followed (also was not specified by the designer, though). 2. Torque on the studs was not measured (nor specified).
Looks like the guy paid attention to the applied pressure on the flanges only, when redesigning flanges. Three different types of gaskets were pushed out. We had to manufacture an inner and outer rings and put an O-ring in between to solve the problem. Will see for how many years it will work.