Hi Pete,

Sorry to get into the discussion so late. My esteemed colleagues, writing as they have above, have certainly provided many excellent observations. Not much more needs to be said. But.......

Well, usually when A COMPANY disallows the use of some type of construction (via their company Standards) it is because they have had many bad experiences. In the case of seal welad, it is understandable that there have been bad experiences because most of then have been done badly - not good workmanship. Bad welds CAN do more harm than good sometimes. It is important to remember that a seal weld is, first of all a weld, and that fact brings with it certain fabrication and inspection responsibilities. The B31 Piping Codes, however do not disallow seal welds - but have tried to point out some of thier limitations and pitfalls.

Usually we hear this question asked as: "what is the difference betwen a seal weld and a full penetration weld - and why are seal welds "bad" "?

Seal welds are defined in ANSI/ASME B31.3, Process Piping, paragraph 300.2 as follows: "Seal Weld - a weld intended primarily to provide joint tightness against leakage in metallic piping." The definition given by ANSI/ASME B31.1, Power Piping, paragraph 100.2 is as follows: "Seal weld: a weld used on a pipe joint primarily to obtain fluid tightness as opposed to mechanical strength".


Further, B31.3, paragraph 311.2.6, states: "Seal welds (para. 328.5.3) may be used only to prevent leakage of threaded joints and shall not be considered as contributing any strength to the joint". B31.1, paragraph 111.5, states: "Seal welding of connections, including threaded joints, may be used to avoid joint leakage but the welding shall not be considered as contributing any strength to the joint". It can be seen that seal welds, per se, are only used in the B31 piping Codes with threaded joints. If a torsional moment is acting in a way which might "loosen" or "unscrew" a threaded joint, maybe the designer should consider another type of joint (i.e., a girth butt weld".


Paragraph 328.5.3 addresses fabrication and it adds:

"Seal welds shall be done by a qualified welder. Seal welds shall cover all exposed threads"...

Welder qualification is addressed by paragraph 328.2. If seal welding is done, paragraph 331.1.3 should be reviewed to determine if heat treating is required (what base metal alloy is being used?). Paragraph 335.3.3 mandates that seal welds shall be made with no thread compound. Of course, in B31.3, there are limitations on where (what fluid services) threaded joints (with or without seal welds) can be used. B31.1 includes similar service requirements and limitations. So, no matter that the weld is not a structural weld, its use carries with it all the associated fabrication requirements of any other type of weld.


Obviously, the implication in these paragraphs is that the seal weld should not be considered a structural weld, that is it does not fulfill the requirements for structural integrity. The term "full penetration weld" is not defined by B31.3. All welds that are required by the structural design of a B31.3 piping system shall conform with the requirements of paragraphs 328.1 through 328.6. The purpose here is to develop the full strength of the base material.


In the case where ....."these pipes are not intended to be used as a pressure piping system" (e.g., they are part of a structure that is not a system covered by the B31 Code for Pressure Piping, and frankly, if there is no pressure I don't know what would be "sealed" by the weld), then perhaps the AWS Structural Welding Code, ANSI/AWS D1.1 should be consulted. However, even if you are welding up some handrail using threaded pipe, only the engagement of the threaded joint itself can be considered in determining the strength. A weld added to prevent rotation will not add strength to the joint.

Having said all of that, one should consider other issues involving threaded connections. Each thread is an incipient crack in the pipe wall. If there is any cyclic bending (and maybe internal pressure) loading involved you would not expect a long service life. When a seal weld is made over the threads (especially if it does not cover all the threads), the inspector MUST be very careful to look for weld "undercut". With small bore piping, the (field) welder usually changed his electrode angle several times as he makes a full circumferential girth weld. Each time he stops the weld and restarts it he has another opportunity to undercut the weld. I would suggest that you peruse B31.1, Paras 127.4.4 (Fillet Welds) and 127.4.5 (Seal Welds) and especially Figure 127.4.4(A)(d). Remember ther is a reason why B31.1 specifies a larger stress intensification factor for certain fillet welds (and threaded joints)in Appendix "D".


Also. if one would have to "grind-out" a fillet weld every time an instrument is to be routinely changed, it would seem to be a labor intensive maintenance procedure - one which begged for an alternative. After several change-outs (and rewelds)the local base metal would likely be embrittled and the area would have a high likelihood of failure. As the physicians "Code" states: "first, do no harm".

In the case of using a fillet weld, not as a seal, but as a method of preventing rotation in a threaded joint, there is no reason to disallow it BUT only if ALL the requirements of the Code are complied with. However, for all the reasons implied above, designing a mechanical locking method would be a better option.


At risk of adding complication to the issue, I will say we have seen the term "cover welds" used when refering to the fillet placed over a groove weld in the fabrication of some pressure vessels and piping. A properly made fillet, covering any other properly made weld will contribute to the strength developed by the completed weld. Note that B31.3 et. al., allow the designer to "take credit for" this weld area of metal in the area replacement rules for fabricated branch connections (para 304.3.).

Well, for somebody who did not have much to say, I seem to have said all of it. I hope that some of this will be of some assistance to you.


Of course, the information offered above are only my opinions and they do not represent the opinion of the ASME, or the B31 Code Committees.

Best regards, John Breen.