Low temperature per B31.3

By pataterchip (**)

Date 03-22-2012 21:32

What chapter are you working to? High pressure, fluid service?

By fschweighardt (***)

Date 03-23-2012 01:33

Normal fluid service, Right now I'm in chapter III trying to nail down base metal and weld metal required tests for -320 and -425 deg F. on 304L SS

By fschweighardt (***)

Date 04-04-2012 14:50

Bump to the top

By 803056

Date 04-05-2012 15:52

I don't have my B31.3 handy, but based on my experience; most ASME procedures are qualified on plate. Things get a little more complicated when notch toughness is invoked. In some cases the testing must be done for each product form and material specification, and each combination. As you already know, you have to refer to the construction code to determine if and when CVN testing is required. Each service category has an opportunity to tweak the requirements of how the procedure is qualified and what can and cannot be done in production.

Who is calling out B31.3 normal service for a cryogenic application? The individual classifying the fluid service should determine if notch testing is required.

OK, you forced me to find my copy of B31.3 (2006 is handy) and dig through all the muck to see what I can see. Assuming it is "normal" fluid service, look at Appendix A. I located 304L and notice there are several notes. Notes 1, 5, 6, 14, and 36 apply depending on whether the 304L pipe meets ASME SA269, SA312 or SA358 requirements. Let's assume it meets SA269 where the most notes apply. Of those notes listed, note 36 seem to be applicable for both SA269 and SA358 materials. Depending on the condition of heat treatment, it may have to be impact tested if the service temperature is lower than -20 degrees F. If the pipe meets SA312 or if it is solution heat treated the lowest temperature listed in Appendix A-1 applies (-425 degrees F) and the pipe can be used down to that temperature before impact testing is required.

Remember what the letters ASME stands for, Always, Sometimes, Maybe, or Except. Go back to paragraph 323.2.2(c) where it references Table 323.2.2. In the table, column A list conditions for the base metal where impact testing is required and conditions where the weld and HAZ must be impact tested. It chases you back to paragraph 323.2.2 and to notes (3) and (6) found at the end of the table. Note (3) requires you to know the stress ratio as well as the temperature which must be between -20 and -155 degrees F, and note (6) says you don't have to test is the width of the impact specimen obtainable isn't wide enough. Your service temperature is lower, so you are still on the hook.

Keep in mind that paragraph 323.3 states that each product form and material specification must be tested. So, the raw materials must be ordered with the required testing or the user can do it (right!). Then the welds must also be qualified by impact testing. This gets very expensive, very quickly because you have to check the design to determine what combinations of material specification and alloys are welded together. Each combination has to be tested from what I read. These requirements override the requirements of paragraph 328.2 (f) which allows for testing based on P number. That being the case, back to testing each combination of material specification and alloy type. Another example where the construction code trumps the requirements of Section IX.

The only out I see is note (6), but consider the fittings and valves you will be using. If the fitting or valves are thick enough, you still may have to do impact testing depending on the material specification involved.

The need for impact testing is the design engineer’s responsibility. The CWI, in general, is not trained to determine when impact testing is required. The use of ASME codes is more involved than D1.1 or API 1104. If you are actually working with ASME, I suggest you attend several of the short courses offered by ASME. The ASME short courses are not inexpensive and you need to determine whether you can justify the cost of buying the applicable code sections and time needed to attend the ASME training courses. If you are not normally involved with ASME and you cannot justify the cost of proper training, you may be wise to steer clear of it.

Welder qualification is a different story, it is fairly straight forward and not that complicated. That is something most CWIs can handle, but qualifying a welding procedure is much more involved especially when working to the multitude of piping codes and the various fluid service classes. The CWI typically isn't usually trained, experienced, and rarely provided with sufficient information to properly maneuver through the maze called ASME.

Best regards – Al

By jon20013 (*****)

Date 04-05-2012 19:29

Al said it perfectly in one sentence: The need for impact testing is the design engineer’s responsibility. While B31.3 calls out impacts for low temperatures (original poster has already noted exceptions based on certain attributes), it remains the designers responsibility to specifically require.

That said, I would expect very little problems with austenitic stainless, even at cryogenic temps. Why not just do them, they're not hugely expensive,that way your covered if they are needed?

By fschweighardt (***)

Date 04-08-2012 17:38

let me get into the office and looks ome of this up guys

By fschweighardt (***)

Date 04-09-2012 11:31 Edited 04-09-2012 20:08

On this one I am part of the design team, and we have a project where most of the fabrication will be done by an outside vendor, an an internal employee will be doing the tie-in. I have made my way along the same lines as Al outlined to note 6 of Table 323.2.2 and it tells me that if I cannot extract a Charpy sample of any of the approved sizes from my coupon, I don't have to do the impact tests (for stainless steel, it is actually a lateral expansion per table 323.3.5) The thing I run into, is most of the PQR testing I get from vendors is the same as the Performance qualification test, and is run on ~2.5" pipe as they are trying to kill 2 birds with one test. Not much chance of getting a Chapry specimen of any size out of that little pipe.

As it stands, I will have to specify materials that are supplied in the soln. heat treated condition (which is usually the case, but you gotta spec it)

Table 323.3.1 par A-5 give some guidance on the fact that Charpy samples do not have to be taken from each lot, nor from material from each job, procided that appropriate testign has been done. they go on to say that all you need is same type and grade of material and of the same thickness range. This may give an opportunity to just make up some small samples just wide enough for 3 charpys. probably will need a couple of thickness ranges, but that should be easy enough.

I was contemplating having the small supplementary coupons welded up for the impact/lateral expansion tests and having them appended to the existing WPS. Problem now becomes the fact that all of the supp. essential variables are needed and probably not recorded.

So, for the $64K question, do you make the vendor run a plate WPS/PQR and get the Charpys from that coupon, and then have the standard Performance qual run on the appropriate size pipe

Or do you follow the letter of the code and make yourself exempt from impact testing based on note 6 as above.

Personally I am leaning toward plate charpys, and separate performance qualification. Last thing I need is a ruptured liquid hydrogen line and less than perfect documentation.

By js55

Date 04-09-2012 17:19

Fred,
Those are the engineering decisions that can be tough to make. But I would suggest that if you have concerns go conservative, especially if they are asking for -425. You won't get a representative -425. You cannot get the specimen from the bath to the machine fast enough.

By 803056

Date 04-09-2012 18:01

I know of an environmental test chamber at the Johnson Space Center that is being refurbished to go down to -400 degrees F. You might solicite your congressman to get some time in the chamber.

Just kidding.

Best regards - Al

By fschweighardt (***)

Date 04-09-2012 20:22

I can actually get pretty good low temp data with liquid helium, I should be able to hit right around the -425 mark with samples starting at ~-452. If there is any justice (which there never is) I should be able to use P8 (304/304L/316/316L anyway) tested at -425 for my services at -320 with no problems. It kinda sucks having the temps in Table A-1 go so nice and low, and then have to come back and do a bunch of impacts for way higher MDMT

By js55

Date 04-10-2012 14:49

This whole -425 thing has been in debate for some time due to the inherent inacuracy. The specimen will heat heat up really fast. Liquid helium is certainly the optmimum way to go but I would still guess you are hotter than -425 by the time the specimen is removed from the bath, placed in the anvil, and smacked with the hammer.
If you are operating as a design spec and not code compliance you might consider notched tensiles as they do for nickel alloys and such under B31.3.

By fschweighardt (***)

Date 04-11-2012 01:44

yeah, it will be tuff, but we are going to have more and more full compliance piping, and we do so much low temp. I gotta get after the copper guys as well