% Allowable Stress 5

[Eduard47]

Sometimes back I encountered a pipe stress model that generated one of the highest at 95% of the allowable on operating load case. I was then a bit worry of the closeness to the maximum allowable limit untill somebody told me that it is okey. Technically it is stil within the allowable and is supposed to be consider safe. My mentor has told me 2 years ago the lower the better and of course it makes sense to me. My question from the experts around is: Is the % allowable be limited to figures below 100% for any longterm effect (say not > 90%? Any effect on fatigue stress, creep etc. that may cause failure on the future? Thanks again for a professional help.

Eduard

 

[oswandi]

I think your friend right. If your result in calculation max. not more than allowable stress, you can use it. Same or lower is better because there is enough safety factor in allowable stress. Oke? thank you.

 

[Eduard47]

Thanks for the support Oswandi. Let me clarify the question one more time for the rest cause I think I did not technically raise the question right: Is the code stress % be kept a distance far way below 100% of allowable stress? Thanks again.

Eduard

 

[prex]

Of course, no: the code allowable limit is there as a limit and that's it.
What sometimes happens is that people doing calculations and limit checks with rules of thumb prefer to take a margin with respect to code limits to somewhat cover their errors!
Of course these behaviours are not covered by code: only an as far as necessary detailed analysis done by a professional can be used to satisfy code limits. prex
motori@xcalcsREMOVE.com

http://www.xcalcs.com

Online tools for structural design

 

[ralphsare]

i agree with the position taken by prex. why perfrom a s tress calc - to esure that we keep the stresses in the pipe and fitting within the code allowable limit (the "shall not exceed" clause). my zen master says - the calculated stresses are not necessaritly the "real" stress but are ratther "code" stresses based on decades of compromises and simplification.

 

Hello all,

Well, look at S-N curves for various materials. The lower the stress, generally the more stress CYCLES can be tolerated in the piping system's design life.

If you look at the B31.3 Code, paragraph 300, Severe Cyclic Service, you will see that you should limit your calculated stress range (Se)to 0.8 Allowable syress range (Sa) if you will have more than 7000 cycles in the design life.

Again, I advize seminars on B31 Piping Design Concepts and the book by Glynn Woods and Roy Baguley (CASTI Publishing) and the book by Dr. Chuck Becht (ASME Publications) to get a good understanding of the background and basic principles.

Best regards, John.

 

Hello,

One other thing after reading the original question again.

The ASME/ANSI B31 Pressure Piping Codes provide an allowable stress range for the calculated expansion/contraction (displacement) load cases and they also provide an allowable stress for the calculated sustained longitudinal stresses due to weight and internal longitudinal pressure (i.e., "additive stresses&quot. The B31 Codes DO NOT provide an allowable stress for the "Operating case" which would include thermal, weight and pressure loadings all together. The Mechanical Design Committee is currently looking at providing an allowable stress (range) for operating case loadings but it will be a while before it gets into the Code "books".

regards, John.

 

[trilinga]

How close you can go to the allowable stress limit during design is often the designer's dilemma. A true designer does not get the satisfaction of an effective (and economical) design if the calculated stresses are very low. At the same time, due to the fear for any possible incidental or unaccounted loading during the service life of the system one does not want to be too close to the allowable limit. Normally the value is chosen via-media.

In structural designs we consider the calculated stress between 80 to 90% as safe and satisfactory. In some cases where the loads are defined more accurately with less probability to change during the service, values upto 95% can be acceptable.

Of course, it is better to have the stress levels lower for specific systems subject to load cycles considering the fatigue strength.

However, the value is left to the discretion of the designer.

 

[Eduard47]

I appreciate your response John and I am very much interested on this issue. It took me a while to formulate the question. And again here you are just right because there is no code stress checked for operating case and I don't understand actually why (when temperature is a factor in stress calculation). Anyway if then the code stress checked are for sustained, expansion and, you happen to get 99% code stress, are you not going to go back to your layout and play with the support configuraton to lower the code stress%? Somebody has adviced me that the closer a component is to the allowable limit, the more likely it is a candidate for fatigue failure if cyclic thermal changes goes into a higher counts. So now here, as I heard this "cycles" it just keep me a bit distance to making a conclusion. And so part of my add-in time in the past and in fact even now), I would go back to my model, play with the layout and reduce the stress code to no more than 93%. I do appreciate the criticism and ideas, thanks a lot for the expert advice of all of the above. Regards.

Eduard

 

[PAN]

Eduard47,
The actual installation may be different from the original design (due to field adjust or some modification at site). The desiner should re-check the pipe stress analysis model after installation in the case of close to allowable stress in the first stage.

This is not quite often to modify again after recheck the model. However, in my experience, the contractor run pipe stress close to allowable stress. After the installation and recheck the model was changed and exceeded the allowable stress.

 

[ralphsare]

how does eduard47 thread jive with para 302.2.4(1) that state "... its permissible to exceed the pressure rating or the allowable stress for pressure design at the design temp of ...". thats over 100% .

 

[StressGuy]

That's going to end up simply being a reduction in the safety factor to allow the system to be able to handle some kind of short term excursion without it having to become the "design" condition.

Allowable stress is just that - allowable as defined by the code. That doesn't mean that exceeding that stress automatically results in catastrophic failure.


Q for John Breen:

While the code doesn't explicitly define an operating allowable, isn't it essentially given (refering to B31.3) by equation 1b in 302.3.5d

Sa = f(1.25(Sc+Sh)) - Sl

This equation incorporates both the displacement and sustained stresses and can through some simple rearrangment be setup to say that:

Sl + Se < f(1.25(Sc+Sh))

Obviously, the displacement and sustained stresses need to be checked against their individual equations since this equation doesn't tell you whether or not you've satisfied the sustained and displacement stress criteria.
However, it does seem like f(1.25(Sc+Sh)) is going to be the maximum combined stress allowable. Is the committee considering a lower value? Edward L. Klein
Pipe Stress Engineer
Houston, Texas

All opinions expressed here are my own and not my company's.

 

[trilinga]

Stressguy

Just a word about the second para of your post.

It is true that allowable stress is not the ultimate and exceeding the value will not cause failure. The effect is only a reduction in the Factor of Safety. However, the Factors of Safety have been provided to take care of other uncertainities such as material imperfection and unnoticed defects in fabrication and erection. These factors have been prescribed in codes as a result of considerable studies and statistical data analysis.

Hence, we, design engineers generally do not encroach into that. The margin we keep below the allowable stress is to take care of the uncertainities in loading only.

 

[ralphsare]

if your operating under severe cyclic condition, then your worry is more than the calculated stress. in this conditon need to focus more on the nuts and bolts of the piping layout details such as the branch and nozzle connection, welding approach, etc. rarely one will encounter an exposure to more than 7000 cycles (one cycle per day)for the 20 year life of the faiclity. Code are set of rules (not design guideline) and mostly its all about compliance Its either youre in compliance or not (allowable has been exceeded or not and the thought of the AI ).

again, revisiting 302.2.4 (f).1
Subject to owners approval, its is permissible to exceed allowable by 33% for 10 hours incursion not once but 10x over. Thats 100 hrs events/year. It may not be your operating case but thats still count a lot and exceeding the allowable has been permitted.

 

[joshikv]

Well, let me mention here that Code gives only guidelines for stress analysis and allowables. It does not specify the rule or binding to stick to these limits.

As a good engineer / designer we have to consider some unseen aspects like fabrication changes, any changes that happen on site, material inconsistency. So I consider, it better not to cross 90% for the expansion stresses and may be 60 to 70% in sustained stresses. Especially when you have a high temperature, high pressure system, why to take risk?

Regards,
Kaustubh

 

[ralphsare]

the code is a guideline and not rules? not accordng to glynn woods (casti books), the AI and the code committee.
Introduction of B31.3-1996 Edition

Second paragraph opens .. &quot;Rules for each section reflect ...&quot;

Then follows by the statement &quot;The designer is cautioned that the Code is not (repeat not) a design handbook; ..&quot;

Let me cite the thought of glyn woods (type in bold by him) page 13 of his PIping lecture (asme education and training).

The Code does not provide any instruction as to how to design anything.

 

[ayfarm]

I would not use the allowable stress method if there is great risk factor involved in the application, rather I would analyse the stresses using finite element method.

 

[trilinga]

ayfarm

I think your post needs a correction. Basically, the allowable stress we discuss here is for the design, whereas finite element method is a method of analysis which gives the actual stress in a loaded system.The discussion is about how much close these stress (as calculated through FEM or otherwise) can be to the value allowed by the codes.

 

[ayfarm]

As I stated, the issue here is purely judgemental. PE's have differing level of confidence in their work. I myself would assess the risk factor involve when close to the fine line, as is the case. The statement I made was just my own way to approach the problem and may not be relevant to what the basic question is. Furthermore, the finite element method does not produce the &quot;actual&quot; stress as mentioned above, it only &quot;calculates&quot; the predicted stress level at any given loading scenario. Only a physical load test can give you the actual stress if the means to measure it is absolutely accurate!