As context for what I write below, I’d like to say that my bias is that I have an ever greater appreciation for the breath and depth of the B31 code and what it has and continues to accomplish.

However, my experience has been that most owners and engineering managers do not have the same perspective.

They ask:

1. Why does it take such extensive and time-consuming analysis to complete a stress analysis now, when 20-30 years ago it was done quickly with simple methods?

2. I can show you dozens of steam piping systems where the tees have other fittings, flanges, or valves welded directly to them. Why were those systems successfully accomplished then, and I can assure you no shell/brick finite element analysis or prototype was built to prove the tees were okay, but you need these methods to accomplish the same now?

3. Such systems are still working today, decades later. Do you hear of tees failing in carbon steel steam piping? The code is overly conservative and highly specified with no benefit over the designs of decades ago.

4. “The code is not precluding designs that do not reproduce these end conditions…” (John B.) No, it is not precluding such designs, but doesn’t provide an answer within the code as to whether a very common piping fit-up is okay or not. Why was it okay by code to weld other fittings/flanges/valves to tees decades ago (agreed the code does not, per se, state it can not be done so now), but now says “may require special consideration”, read “the SIFs may be higher, may be lower”, so the code can not tell me whether that tee that I’m looking at right here in this 20 year old station is okay or not? Tees are often fitted to other fittings/flanges/valves, so would not share the perspective that such configurations are a small percentage of systems that are evaluated.

5. “Owners need only to insist upon systems that have been shown by EXPERT ANALYSIS to comply with ALL THE DESIGN REQUIREMENTS of the Codes and Standards (and to be fabricated, erected, examined and tested in compliance with the Codes.” (John B.) The same could be said for a code that says: Build it so it doesn’t fail. What good is a code that doesn’t address common piping arrangements other to say you need to go to another code or conduct special analysis to know if the system is will not fail.

The Luf in me would like to respond: “Well, piping designs can be produced quickly and appear to work on paper. Just give a recent graduate the software and tell them you need the system designed in 2 weeks and you’ll get your quick design. Do you feel lucky? Just because you can weld hunks of metal together, doesn’t mean you have a structurally sound system. Go ahead and make my day.”

But I’d like to bridge the gap of understanding a chunk better than that.

Here’s what I am finding on these steam stations with tees that don’t meet Table D300, Note 13: The cold startup expansion stresses are at, or slightly above the code allowable, 75 to 150%. But the switchover between one leg of the station to another have stresses of 500 to 700% of the allowable stress. The switchovers will occur two to three dozen times in the life time of the system. I speculate that the systems were designed for the cold startup stresses, but without regard to the switchover stresses.

I understand that I will have to gut the system I am to revise and provide the spacing around the tees, or gut the system and perform a shell/brick finite element analysis on each tee that does not meet the Table D300, Note 13 requirement, or perform NDE to check the current status of the piping and perform a API 579 fitness for service analysis to see what can be left in place and added to.

I’d like to yet ask:

1. “…since the flexibility would be diminished the Code SIF may still be conservative.” (John B.) But may not be conservative. So, how can I learn more about how the Flexibility Factor, k, factors into all of this? I don’t see any code equations using k, so is it used in the linear elastic beam theory equations? I didn’t see Rodabaugh use it in WRC 329, though he discusses it. Why does the code give the values but not explicitly use it … did I miss where it is used?

The responses were very helpful. Thank you John L. and John B. for taking the time to respond and for all that you wrote. Richard – I usually need my spouse to translate many of Luf’s references, but I did get his reference to burning trains on my own.