Is it possible to reduce the stress by machining off material? 2

[Christine74]

For a given loading condition, is it ever possible to reduce the amount of stress seen in the part by machining off some of the material (ignoring of course the loading from the weight of the material that gets machined off)?

I ask because I noticed that the ASME Appendix 2 flange calculation procedure will often calculate a lower stress value if you machine down the flange hub (say from a hub angle of 15 degrees to 10 degrees), which to me seem counterintuitive.

Is there any reason to believe this is possible?

Thanks,

-Christine

 

[GregLocock]

Yes sort of. I know of cases where a welded joint was failing because somebody didn't copy somebody elses design properly, and omitted a hole in the structure near the weld.

Cheers

Greg Locock

 

[unclesyd]

If I understand you correctly, the 10? gives a longer hub which results in lower stress values seen by the flange.

I don't have any current flange catalogues but at one time there was a difference in hub length between SS and CS flanges due to the different allowables for each material.
The SS being the longer.

Anecdotal:
In the 60's this caused us a lot head scratching when a engineer ordered 16 48" SS flanges for pressure vessels with 300 psi @ 600°F. He used a CS flange dimension. By doing calculations by logs instead of a slide rule we were able to get the flange in by less than 1% using the code allowables. The actual values were about 15% higher than code and this made everyone feel better. They are still in service.

 

[MintJulep]

Yes. Common applications include cutting a groove at a step diameter change in a shaft to reduce stress concentration factor or reducing the diamter of the shaft of a bolt promote more favorable stress distribution.

A less common application is to taper the thickness of two plates for a few inches on either side of a butt weld to reduce the stress gradient.

 

[MikeHalloran]

Imagine that you have an assembly comprising two rectangular bars, joined by two bolts, one near each end, and that by means that are unimportant for this example, you apply a load that tends to separate the bars, and is concentrated near their centers. I.e., like the cross section of a flanged joint in tension.

Now, as the two bars bend away from each other, two fulcrums are formed at their ends. The bolts then see a magnified load, the magnification factor being roughly the ratio of the flange outer radius to the bolt edge distance.

If the bars are relieved so that they cannot touch each other except at points inboard of the bolts, i.e. like the step in a typical flange face, then the axial load on the joint cannot be magnified.

[ I've often wondered why engine connecting rods are not similarly relieved outboard of the bolts' inner tangents. I'm guessing that maybe the strap of the 'big end' is not stiff enough to carry a serious moment load to the rod bosses, and just deforms itself. ]




Mike Halloran
NOT speaking for
DeAngelo Marine Exhaust Inc.
Ft. Lauderdale, FL, USA

 

[TheTick]

The best example that comes to mind is drilling a hole at the end of a crack to stop a crack from propagating (learned this in high school from a drummer who fixed his cymbals this way).

Anyhoo...

There are many cases where a smaller shape can be stronger if it reduces stress-concentrating features.

 

[fredt]

Most of the above examples are stress concentration related. Cases not related to stress concentration are when the additional mass imposes a greater force on the structure than is off-set by the extra cross-section. This is common in rotating machines where reducing material by profiling can significantly reduce stresses. Taper beams are another common example.

It is also possible to have a redundant member on a structure so that the member will itself fail although the structure as a whole would be sound if the member was omitted.

 

[SnTMan]

Christine74,

For your Apx 2 calculation are you SURE the hub slope was the only change? Did mean gasket dia, bolt circle, or anything else change?

If so you are not comparing the same thing, as the applied moment may have changed.

Also, one of the moment arms will change (ht maybe?) based on the different hub large end thickness.

The shape factors involved in Apx 2 are pretty complex and it is not always obvious what effect a change in hub shape will produce.

Compare your calculations carefully to see what's going on.

Hope this is helpful.

 

[Christine74]

Thanks, guys.

"For your Apx 2 calculation are you SURE the hub slope was the only change? Did mean gasket dia, bolt circle, or anything else change?"

Yes, I held the flange O.D., I.D., the bolt circle, the gasket width, G , y , m, # bolts, bolt diameter, flange thickness, allow. stresses, hub length , go, etc.

I only changed g1, and increasing it seems to increase the calculated stress if the angle is too steep. Above 20 degrees the stresses become quite high.

I believe it has something to do with one of the shape factors, but I'm not positive about that. I'll look into it more.

-Christine

 

[SnTMan]

Christine74,

Sounds like you are pretty thorough. The reason I mentioned those things is some flange software can change some of them without your intervention, for example will give a smaller bolt circle based on a smaller g1.

Check your moment arm hd (not ht, I'm sitting by a Code book now!) and see if total moment is higher on the flange with the larger g1.

If not, then it probably is the effect of the shape factors. As I said, its not always obvious what effect they will have. I have seen the effect you mention any number of times, and it can be a puzzler.

Good luck.

 

[SnTMan]

Well, there I go again. ACTUALLY, the moment arm hd DECREASES with increasing g1, assuming a constant bolt circle. Also ht would appear to decrease given a constant bolt circle (R decreases). The loads stay the same, decreasing the total moment.

In an example I just ran, only the radial flange stress increased, the others decreased. Is SR or perhaps ST+SR governing your design?

Have to sign off now, will look into this more tomorrow.

Interesting!

 

[XELR8]

Christine74,

The answer to your question is yes, stresses can be reduced at a discontinuity because the they get redistributed into other areas.

Be more specifc about your problem and I will elaborate on the results.

Can you tell me the: Internal pressure, any external loads on the flange and their direction, flange class, pipe size, wall thickness and face?

 

[valintine]

You are not really reducing the stress in the material, rather reducing the stress at a given point, and as you say redistributing it to the rest of the material! This works well as it is always better to distrubt the stresses rather than have a build up in one area leading to failure!

This only works if the stress you were origionally talking about was built up in a single area, i.e. a corner or angle.
If the stresses are in a flat surface or uniform part, I can not see how removing any material would help reduce the stress, apart from the decrease in wieght from the removal of the material!

 

[SnTMan]

Hello All. Got a chance to get back to this problem.

Here is some data for a flange I ran (Codecalc) with varying g1 thickness.

P 250 t 4.125 h 1.3125
T 400 G 40.75 g0 0.5
A 48 N 0.5 (flat face)
B 40 m 3.75
C 46.125 y 7600
C.A. 0.125 Ab 20.112 (48 - 7/8 Dia)

Here are results which I charted in Excel (couldn't figure out how to paste the graph here). The quantities Mo and SRo I have scaled just to improve the appearance of the graph by reducing the data range. "o" denotes operating stresses, I didn't chart gasket seating stresses but it shows the same pattern except the total applied moment is constant since it depends on a single moment arm, hg, which is not a function of g1.


g1 0.5625 0.6875 0.8125 0.9375
.25*Mo 21843 21429 21015 20601
SHo 23130 17555 19833 21359
10*SRo 5890 7000 7750 8260
STo 13646 12685 11931 11318
SAo 18388 15120 15882 16338

The interesting thing is the hub bending stress SHo and average stress SAo are minimized at g1 = 0.6875".

For Appendix 2 flange design three separate stresses are calculated, bending stress in the hub SH, radial stress in the ring SR and tangential stress in the ring ST. The average stress SA is a function of these three.

I am not familiar with the THEORY BEHIND Appendix 2 flange design, so I can't offer an explanation at this point other than to say it is an effect of the shape factors.

Enjoy! (sorry about the data formatting)

 

[Christine74]

Thanks for the info, SnTMan.

I checked your calculations against our in-house software and got the same results.

I also found that the optimal hub angle for your flange is 7.68 degrees. In other Appendix 2 flange calculations that I've run, I recall that the optimal hub angle has always been somewhere around 8 degrees.

Interestingly, standard ASME WN flanges typically have hub angles much higher than this, which is a major reason why they usually won't pass the Appendix 2 calc's.

-Christine

 

[SnTMan]

Well, it was kinda fun!

I agree with you about std ASME flanges, plus they generally have far more bolting than a Apx 2 flange for given T & P.

I have noticed that regardless of widely varying design conditions, for a given g0, your get very few distinct hubs, that is you tend to get the same thing all the time. Program logic must seek to produce a hub that minimizes stresses, although I can't say just how.

For the class of work I am used to it is not usual to have "R" greater than what is required for a given bolt dia. Therefore when you increase g1, you increase the bolt circle. If G is held constant total moment usually increases, and for me has masked the effect of varying g1. This is what lead to some of my confusion earlier in this thread.

"It ain't what you don't know that gets you in trouble, it's what you know that ain't so."