ASME BPVC VIII Div 1 Max Allowable Stress 4

[JustinME]

I have a copy of the BPVC VIII Div 1, and a Professional Engineering study book. I can use these two to calculate minimum allowable pressure vessel thickness. However, I have not found what the ASME max allowable stress is for a broad range of materials. Searching the internet, I have found a few [unverified] sources stating that, for Div I, the max allowable stress is the materials UTS/4 (eg if UTS is 100,000 kPa, then the max allowable stress for ASME calculations is 25,000 kPa).

Is the UTS/4 rule of thumb valid? If so, why would the max allowable stress be based on ultimate tensile stress and not yield stress? It seems to me that using yield stress as a benchmark for max allowable stress would allow for greater accuracy in regards to predicting the minimum allowable material thickness.

 

[david339933]

UTS/3.5 = Allowable stress.

 

[david339933]

Should add that Stress Values are found in ASME BPVC Sec II Part D.

 

[JustinME]

Thanks David.

I'm not so concerned about WHAT the max allowable stress is. I am curious as to WHY the max allowable stress is what it is. Why base the max allowable stress on UTS and not YS?

 

[cbPVme]

It isn't simply UTS/3.5. It's the minimum of 2/3 yield and UTS/3.5. This is only applicable for materials not within the creep range. The rules also change a bit for austenitic stainless steels.

This is discussed in more detail in ASME II-D, Appendix 1.

 

[metengr]

I am curious as to WHY the max allowable stress is what it is. Why base the max allowable stress on UTS and not YS?

ASME B&PV code committee adopted design by rule years ago to provide suitable margin for safe, long term design of boilers and pressure vessels in service. With that said, ASME B&PV code rules are strength based design using UTS/3.5 or 67% of YS, whichever is lower for time independent service.

 

[JustinME]

Thank you cbPVme & metengr.

I was not aware of the 67%YS requirement. That fully satisfies my curiosity

 

[JStephen]

In older versions, it was UTS/4, that was revised to UTS/3.5 several years back, so that's why you see that in books or internet.
UTS would correspond more or less to the bursting pressure.

 

[JustinME]

I say this for the sake of completeness and for any future readers... not to be confrontational. It may come across that way, but that's not the purpose.

The UTS requirements safeguards against a burst failure, and the YS requirement safeguards against a low or mid cycle fatigue failure.

The UTS/3.5 requirement may be complete in a practical sense, in that perhaps no available pressure vessel steel will have a low cycle fatigue failure when the applied stress is under UTS/3.5.

However, the UTS/3.5 requirement is incomplete on a logical & theoretical basis. Given my limited experience in pressure vessel design, I prefer to have the logically complete set of requirements.

 

[TGS4]

JustinME, those are some very interesting claims you've made there; especially the one about the design margin of 1.5 on yield being there for low to mid cycle fatigue failure. I've been doing pressure vessels for over 20 years and have been a member of the Code Committees for 13+ years, and that is the very first time that I have heard such a claim - care to back that up with a reference or a technical basis?

You are aware that there exists rules in the pressure vessel codes that deals with fatigue, from the lowest of cycles to the highest of cycles? And that there are multiple other failure modes besides burst?

 

[DSB123]

TGS4,
I concur with your comments. I don't think you will get a technical basis for the claims.

 

[JustinME]

If you start a fatigue test with a sample stressed at YS, you will find that the part fails at a fairly low cycle. If you test another sample at 90% of YS, you will find a greater cycle count to failure. If you repeat that sequence using incrementally lower values of stress, you will find that the cycle count to failure continues to increase. If you are testing steel, you will find that the parts stop failing, and achieve infinite life somewhere around 50% of YS (varies, depending on material).

"Low cycle" and "mid cycle" are relative terms. What I consider to be a low cycle for my products is likely different than what others may consider low for other products.

I have in my mind other failure modes besides burst, because I consider burst pressure to be the failure that occurs around UTS on virgin material. There is also a fatigue failure, which I do not consider burst even though the event may be a high energy event like a burst. There is also a corrosion failure, which you may argue is burst because the stress is now at UTS given the thinner material... but I am still going to use different wording because the process leading to the failure is different.

If the ASME BVPC did not have a requirement that was linked to YS, then I would be disappointed. However, my curiosity and question has lead to a learning event for me.

That is my thought process. My thought process has proven valuable, because I discovered the 67% UTS requirement... which I suspect should exist based on my thought process. I am not going to argue any more.

 

[TGS4]

Wow, I'm not sure where to start. My friend, you need to start doing some serious research into this topic before you start making claims such as this.

First, regarding a supposed "endurance limit". Such a beast just doesn't exist, either in the literature nor in the pressure vessel code. Both the smooth bar and welded joint S-n fatigue curves have a negative slope at the maximum cycles (10^11 for smooth bar and 10^7 for welded joint). And for the smooth bar curve, the stress amplitude at 10^11 is around 7ksi for carbon steels.

Second, but again using the smooth bar curves, if you take a typical carbon steel such as SA516-70, with a minimum specified yield of 38ksi, at 38ksi stress amplitude (not range, but amplitude), the allowable cycles is roughly 1000 cycles. 90% is 34.2ksi, which results in roughly 1200 cycles. At 2/3 yield (25.3ksi), the allowable number of cycles is roughly 15000 cycles. However, the general membrane stress in a pressure vessel is often much less compared to the peak stress, which could be 3 or more times higher. And the cycle life at 50% yield (19ksi) is around 2e6 - not infinite life by a long shot.

Third, fatigue is absolutely nothing like burst. It's a crack, and considering that most vessel designs are leak before break, a through wall crack will result in a leak, which is absolutely nothing like a burst.

The 2/3 margin on yield, if you actually research other industries and other codes, ebbs end being a fairly typical margin: structural codes use it, for example. It provides a margin against gross section yielding, which is somewhat relevant for materials that exhibit a yield plateau, and less so for materials that don't.

Finally, before anyone feels the bed to remind me that the ASME fatigue curves have a supposed margin of 2 on stress and 20 on life, let me remind you that those margins are on the mean of the data (that contains substantial scatter), and actually best approximates a minimum curve.

Hopefully this has been instructive for you and you learn something.