My question is pertaining to B31.3 requirement for OCCASIONAL stress.

According to B31.3 requirement the "SUM OF THE longitudinal STRESSES due to
sustained loads such as pressure AND weight and OF THE STRESSES produced
by occasional loads" be added.

We can achive this by :

1) Stress in case W+P+U which for linear systems will be equal to :

a) Stress in a case { (W+P+U)-(W+P)}computed by vector difference of loads, i.e. ALGEBRAIC in CAESAR II terminology + stresses in case (W+P) combined by CAESAR II method of SCALAR.

b)Stress in case { (W+P+U)-(W+P)}computed by vector difference of loads, i.e. ALGEBRAIC in CAESAR II terminology + ALBEBRAIC combination with stresses in W+P if the line is in one direction only.

However linear systems and line in one direction only are hypothetical situations so CAESAR II meets the CODE requirement by option (a) above.

My questions are:

1)Are we deviating from the code requirement if option (b) is opted for ? If so is it because of the use of the word "AND" in the code requirement?

2) If (1) is correct, why the CODE requirement is like this? Why not by the option (b)? That is I would like to get the rationale behind the code requirement.

On this forum we have code committee members like John Breen , John C luf.I would request them to throw some light on this.

Best regards

Working heavy hours right now but I will state that B31.3 provides little specific guidance other than what you have quoted....

What do I do??? I assume non-linearity and write my load cases accordingly... see the OnLine help (Technical reference manual) for this....

Basically you add you occasional load to an operating case and then subtract that form the same operating case without your occasional load... this gives you the occasional stresses then you add them via scalar to the SUS case...

Remember the code book is not a cookbook and it does not provide "consulting" information.

I have put in a series of RFIs on OCC stresses that may get tossed out as consulting.

This is another way of the committee saying hey you should know the answer to this yourself...

Thanks John.However my question is not about how to do it in CAESAR II but why the requirement of a SCALAR i.e. sum of the stresses in the individual cases and not ALGEBRAIC i.e.stresses computed from vector sum of forces and moments.
I agree and personally I do it by the way you have suggested, but that does not answer my question.

I want to know the rationale behind the code requirement on this issue.

Regards

Ahh…. I see the point of your question…..

Well First when CAESAR II calculates code stresses for B31.1 or B31.3 (Perhaps others as well) the calculated stresses are unsigned (neither + or -) See for instance B31.1 Para 104.8.4.


When I looked up scalar on Wilkepedia I found more information than I wanted or needed so I snipped the following out…

“In physics, a scalar is a physical quantity which assumes a single value which is independent of the coordinate system being used to describe the physical system. In this sense it is a "real" quantity and not an artifact of the coordinate system. For example, the distance between two points in space is a scalar. It does not depend on one's choice of coordinate system.”

So when you give the program instruction to add these two calculated unsigned code stresses together as scalars you are saying to add them together as unsigned values…

John,

First of all my thanks to you for taking out time from your busy schedule to answer my question.

In my opinion stress is a tensor of rank 2, it cannot be a scalar quantity ( i.e a tensor of rank zero ). The sign of stress is important only if tensile and compressive membrane stresses exist as bending stresses are sign independent.

However categorizing stress as a vector, scalar or tensor quantity was not my intent.

The intent of my question is:

For occasional stresses , B31.3 requires that the sum of the longitudinal stresses due to pressure , weight etc. and the stresses due to occasional loads like wind, earthquake etc. should be less than 1.33 Sh.

It appears from the above wordings that the code wants the combination done in a way which CAESAR II terminology describes as SCALAR i.e. stresses from individual load cases be summed.

What I want to know is that :why the stresses be not arrived at by the ALGEBRAIC way i.e performing a vector sum of the forces /moments for cases like W+P and U and then computing the stresses from the vector sum .

ALGEBRAIC and SCALAR method of summing stresses arrive at the same values if the line is in one direction only ( a hypothetical case )and will differ in all other cases.

I want to know why the code wants for sustained +occasional case, the combination be done by the SCALAR and not the ALGEBRAIC way i.e I would like to go to the physics of the problem to understand this rationale of code requirement.

Best regards

The B31.3 code does not use the word algebraic when talking about sustained and occasional stresses. Whereas it is specifically mentioned when discussing thermal displacements, conversely the code says to sum or add together the sustained and occasional stresses. These stresses when calculated per the code are unsigned therefore when added together scalar is what is used.

Therefore based on the murky and gray B31.3 code, Coade recommends and uses scalar quantities for the summation.

B31.3 does not in the base code specifically state what or how the sustained stresses and occasional stresses should be combined. I suggest if you would like to, write an RFI (request for interpretation) to the B31.3 code committee… (my guess is you will get an answer which may not shed anymore light on it then what’s been written here)

I encourage all persons to write well thought out and well written RFI’s. This is one of the driving forces that can cause code revisions. It has occurred that we have had to answer an RFI in a manner that we did not really agree with (Remember RFI’s are based upon solely what is written in the code as well as past interpret ions)… anyway after answering the RFI we have then taken out an AI (agenda item) that finally results in a change in the code.

As for why…. Well the current method is “simplified” which is the basis for the code as a whole. Its’ rules come from a time before wide spread computer usage. Certainly the current method is simplistic and yields an answer that should be “conservative” in nature, and I guess I would also say that the method works based on all the systems designed by it and how they have performed to date.

Too bad your so far away you would do well to make one of our meetings....

John,

Kindly accept my heartfelt thanks for your response.

Yes, it is sad even though we ( people living in this part of the world) want to participate in these meetings, we cannot ,as we are separated by a long distance.

These meetings would have given us opportunities to interact with experts like you ,Mr Breen, Mr, Tom Vaan Laan, Mr.Richard Ay , Mr. Dave Dielhl and all other stalwarts in this field , something we miss in this part of the world.

So we look into these forums for advices from people like you.

Thanks once again.

Best Regards

Anindya

Anindya,
I have a related question to the discussion on Ocasional stress combination.

While doing Dynamic response spectra analysis, what is a preferred stress combination method for Static (P+W) and Dynamic Stress (Dynamic results from Responce spectra). Caesar gives SRSS and ABS methods. I am doing B31.3 analysis and Could not locate any reference for this in it.

Thanks
Naveen Verma
Black & Veatch, India

Naveen,

Before answering your question, I would like to summarize the discussion in this thread.

It seems for sustained and occasional stresses the CODE intent was to go by the SCALAR method ( the way it is defined in CAESAR II )rather than ALGEBRAIC ( remember these two methods give the same value of stress for unidirectional systems ).I feel that there is no"strong word" in the code statement for occasional stress that make the SCALAR" combination mandatory ( a fact John C luf acknowledges ), although the use of the word"and" makes it suggestive that perhaps that is what was "wanted". So for the time being we will consider that the CODE requires the sustained and ocacsional stress be done by the SCALAR method.

So using the dynamic analysis, in the static dynamic combination,we have two options, ABS and SRSS. In both method the stresses from the load cases are added . ABS adds the stresses ( taking positive sign only)which is identical with SCALAR combination stresses if and only if there are no compressive stresses in the sytem. ( membrane pressure stresses are tensile and bending stresses are sign independent), but SRSS performs the square root of square summation which is less conservative and certainly not the CODE intent ( we can argue on the CODE intent of SCALAR or ALBEGRAIC in this matter, but not on SRSS ).

So go ahead with ABS.

Regards

Anindya,
Thanks for explaination for combination method.
However let me be more specific in asking static and dynamic combination method.

I have D1 D2 D3 dynamic shock load cases (X, Y, Z direction seismic resoinses) of which I need to combined all of them with Static Sustain stresses.
As you suggested static and dynamic shall be combinde by ABS method, but I am not sure of doing "D1+D2+D3+SUS" by ABS method. Is it okay to perform SRSS on D1, D2 and D3 and then combine that result with SUS by ABS method.

Thanks in advance for your great help
Naveen

Are you doing a static equivalent of dynamic method or a static and dynamic method separately ?

If it is static equivalent analysis of dynamic ,I would recommend SRSS combination for D1, D2 and D3.

If it is a separate dynamic analysis , I would recommend :
The dynamic modal combinations should best be done using CQC ( Complete Quadratic combination ) method, which is not available in CII. The closest to that is DSRSS, so this is what I would recommend.Than with the static analysis SUS case make the combination using ABS.

Also make sure that you follow that you follow the directional/modal/spatial combination in that sequence.By the way I hope you are aware of the terms "modal-spatial-directional combinations " and their respective meanings.

Regards

Thanks Anindya,

I am doing dynamic seismic response spectra analysis and trying to combined seismic stresses with SUS to qualify for code stresses.

I am somwhat aware of "modal-spatial-directional combinations" method which I read thru NRC 1.92 and NRC 1.62 and Caesar technical manual. I will appreciate any help in getting more inforamtion on this.
Coming back to Static/Dynamic combination, would you suggest that Dynamic load (D1,D2,D3) shall be combined as SRSS and then added to SUS by ABS method in dynamic analysis.

Thanks for your great help
Naveen
verman@bv.com

Modal combination refers to addition of modal values .Total soln. is a sum of the modal solutions ( maximas in the response spectrum concept).

Directional combination refers to addition of values from same directional excitations like say you have two excitations in X direction.

Spatial combination refers to addition of values from different directions like say X and Z.

I would recommend a SRSS combination method for Spatial as they show "statistical independence"( I don't think the concept of "directional" is applicable in your system as your system is excited in X, Y and Z directions with no multiple excitation in the same direction)and a DSRSS method for modal combination with modal preceeding spatial.
Use ABS combination for SUS and OCC cases.

Regards