I am in the process of reconfiguring some piping. The piping in question is eight small bore (less than 2" diameter) pipes that serve as drains on a 400 MW power boiler. These drains are from the primary and secondary superheater inlet and outlet headers, the reheater inlet and outlet headers, economizer drains etc .... Under normal operating conditions, the drain valves on these systems are closed at the top of the boiler. All of these drains are routed to a vented blowdown tank. The measured temperature of these pipes is approximately 200 F.

When the boiler shuts down for its yearly one week outage, all drain valves are opened and the above noted drain pipes are now at 500 F.

The sustained stress for this piping system is well within the Code Allowable. The expansion stress at 200 F is within Code Allowable. The expansion stress at 500 F is 65,000 psi and hence exceeds the Code Allowable.

My observation is that since the 65,000 psi stress only occurs once per year, I, as a designer, am inclined to accept this system because by my interpretation, this piping would be "good" for several hundered full thermal cycles.

This forum's comments please.

Regards,

Pat LaPointe P. E.
Neill and Gunter Limited
Fredericton, New Brunswick, Canada

Pat,

There are two aspects of your problem. One : How do you ensure that 500F will be there only once a year? Is 500F your design condition /upset condition as per LDT? If so, you have to qualify your system at that stress range.You can talk to your process and instrumentation group regarding the frequency of this happening.

Now coming to your observation : If it has 50OF only once a year and hence your system will be subjected to lesser no. of cycles during it's design life span, my opinion is like this:

31.1 gives an expression for allowable stress range upto 7000 cycles which reduces with more no. of cycles.So the code does not give any favorable guideline for very less no. of cycles.I have seen that for materials like A106B, The 1.25(Sc+Sh)meets the fatigue curve for 7000 cyles.Based on this argument you can check that for the computed stres what is the allowable no. of cycles and I am sure your system will pass (even after subtracting the mean stress effect).

But this will be a fatigue way of qualyfying your system . This method is tacitly underlined in the fact that I just mentioned about A106B, but if your system has a third party inspection, I don't think you can use the code for justifying your analysis, unless the inspection party accept this method.

I have pondered on this problem , no. of times as I encountered this problem in many projects.But I had to qualify the system even though my engineering judgement told me to accept those systems.

I wish there was some kind of Larson -Miller parameter to prove our case like the way we have for creep analysis.

A.Bhattacharya

Stress Analyst

Bechtel Corporation

If high SIF of fittings is the cause of overstress,is there any room for relaxing SIF by a careful review and justifying the use of lower SIF? Just a thought!

Regards,

Mohammad Aziz
Black&Veatch SPC

Hello Patrick,

I may be useful to review the discussion of the allowable stress range concept in a previous thread:

http://www.coade.com/ubb/Forum1/HTML/000057.html

As Anindya points out, if you have a calculated stress range magnitude of 65ksi, you can't rationalize it away - it does not comply to the requirements of the current B31.1 Code regardless of the number of cycles in its lifetime. Also, the pipe material will actually yield before this "calculated" (ergo, fictitious) stress is realized.

Some things to think about:

What is the failure mechanism that you are trying to protect the systems against? The B31 Codes are fatigue-based codes (since 1955) and the (allowable stress range) rules therein will provide protection against fatigue failures as long as the system "shakes down" to elastic response in a FEW operating cycles. If you are really seeing calculated stresses that exceed yield upon the application of the design temperature, then it is possible that you will have plastic deformation of the pipe during the blow-down event. Then there is the possibility that you will have a stress reversal when the system temperature returns to ambient and this might result in a reversed plastic deformation ("cold"). If this happens once a year, you might wonder how many years it would take for the alternating plastic response to cause enough racheting to fail the pipe.

Is the temperature RANGE (delta T) really from the (theoretically "stressless") installed temperature (e.g., 70 degrees F) to the 500 degree design temperature (delta T = 430 degrees)? Or does it include the contraction temperature excursion from "installed temperature" to (system shut down on a cold winter night) say -20 degrees F. (delta T = 520 degrees)? According to the Code, you do have to evaluate the 520 degree F. temperature stress range but this "Code analysis" assumes that the system will "shake down" to elastic response in a few cycles. If the installed to 500 degree temperature excursion does not result in a stress greater than (hot material) yield and the installed to -20 degree temperature excursion does not result in a stress greater than (cold material) yield then it is likely that racheting will not occur. Then, the fatigue rules would apply but you still would have to show a calculated stress range that is less than the Code allowable stress range. It is interesting to consider that the total temperature excursion from -20 degrees F. to 500 degrees F. is the REAL full temperature cycle. How often does this really happen? All lesser (than 520 degrees F.) temperature excursions ( say 70 degrees F. to 500 degrees F.) are "partial" cycles and the paragraph 102.3.2 (C) rules for "equivalent full temperature cycles" apply.

The B31 Mechanical Design Committee passed rules several years ago that will allow (with small numbers of thermal cycles) stress range "reduction" factors (i.e., "F") greater than 1.0. It remains to be seen when if ever this will get into the various "book sections".

I think that you are going to have to simply accept the limit on allowable stress range as the Code presents it and "build the church for the Easter crowd" by redesigning to get
your "once per year" stress range below the Code allowable stress range.

Just my opinion.

As usual, John has covered the code/technical aspects of the issue very well, so I can add nothing there.

I will add that, as a practical matter, since you are talking about lines that are 2" and less, it shouldn't take a lot of work to add enough flexibility to that system to knock your stresses down and not have your sleep troubled by it.

One other thing to consider, regarding SIF: Since you are using small bore pipe, I'm guessing it's socket weld construction. It may be worth pursuing with your design team about making it all butt weld.