I am facing a contractual problem about welding inspection by ultrasonic test. The "owner or client" specification say that 5% of all complete penetration groove weld shall be inspectioned. The problem is that we have detected around 40% of the defect joints (20 joint inspected - 8 joints with some slag inclusions). The lenght of the defect is minimum (around 420 mm in 26890 mm of weld length inspectioned). I said to my client that the real percent of weld defect is 420/26890 x 100, because the discontinuity detected is the same (slag inclusion), but he said that the percent of "weld joint" with defect is 40%, so he wants that the ultrasonic inspection shall be increase to 100%. The code doesn´t talk anything about this increase. I think that this percent the error do not require that increase of inspection. I would like to know other view points about this. The term spot weld do not apply for this case.
Regards.
I agree with your position Miguel. How could a weld with short length be compared to be the same as a weld of say several feet (or meters) long? The words you used were "5% of all complete penetration groove weld" and not "...groove welds". If the specification says "weld" I would believe it is total length of weld, if it says "welds" then it sounds like each weld is treated equal regardless of length. To treat welds of different length or volume equal is not a true measure of quality. The proper practice would be to determine defect rate per unit length or volume of weld.
Marty
Was your 40 Joints 5% of total joints or was your 40 joints 5 % of the total length?
Without having the specification and seeing "exactly" what is stated I couldn't begin to say anything specific about your situation.
In my OPINION, and based on no more than the information you provided>>
IF the requirement is to inspect 5% of the joints and 40% of the joints are unacceptable. THEN
That would warrant closer investigation of ALL similar joints welded by that process and or welder. OR
IF the requirement was to UT 5% of the weld length and 1.5% of that was unacceptable then chances are that 8067 mm of more rejectable weld exist. THEN
That would warrant closer investigation of ALL similar joints welded by that process and or welder if that amount of deefective weld could cause a problem. (Not a QC Decision)
The contract documents I have dealt with would clearly define the scope of NDT required either by length or by joint both for initial and follow up. The engineer for the project should be consulted if the specifications and other documents are unclear.
In some situations, the specifcations may miss some items that qualified engineers should decide. I would like decisions regarding situations not covered in specifications taht apply to the bridges I drive over, the planes I fly in and the playground equipment my children play on based upon proven engineering practice vs. interpretation by non-engineers.
If I feel the code or specifications are unclear. I get it resolved by the design authority.
The thought that the defective welds are a "contractual" problem may indicate the need for some "engineering" consultation.
Have a good Day
Gerald Austin
By -
Date 04-09-2002 03:40
Sir:
If you are on a D1.1 project, the extent of examination needs to be part of the information furnished to bidders (will usually be part of a specification). Refer to paragraph 6.1.1 of D1.1.
I have seen many times just this sort of disagreement when extent of examination (and particularly the protocol for additional examination when defects are detected when the origninal extent of examination is less than 100%) is not clearly defined in the contract documents. There is no right or wrong way to determine additional examinations when it is not specified by the code. You will need to resolve this disagreement with your client if the contract documents do not already specify how to determine additional examinations.
Mankenberg
As stated by some of the other respondents, (Big word hey!) it is difficult to really "judge" this without seeing the contract documents and all the codes and specifications related to it. I do however believe that there is a basic principle that you could apply.
When less than 100% inspection is asked for, the inspection is really a "sampling plan". The idea being that if you find no rejectable defects in the inspected welds, you have a certain probability that there will be no rejectable defects in the rest of the welds. If however you find rejectable defects, you need to "extend" your sampling plan to find with a greater certainty what the rest of the welds look like. Typically it would be required that for each rejectable length of weld, twice as much should be inspected. If you again find rejectable welds in this sample, you would go for 100% inspection.
The question you have to answer is therefore not what percentage of welds failed, but rather whether there were ANY rejectable welds found. You obviously need some kind of a code or specification to tell you what is acceptable and what is rejectable.
You are talking about slag inclusions. What size are they? Do they lie in a pattern etc. Your acceptance criteria must tell you what is acceptable.
Regards
Niekie Jooste
What code of construction are you using Miguel? If it's the ASME B31.3, your spec would be referencing random examination (5% min) of "individual welds" as stated in para. 341.4.1(b)(1). In addition, if this is the seneriao as you've described and you do have welds that fail to meet the minimum quality requirements of the code, the inspector is undoubtedly refering to para. 341.3.4 which can take you into complete examination (100%)as per para. 341.3.4(e)(2)
Seldom has pinpointed an important detail: which code governs the design and construction of the "thing" you've welded? (I've said "thing" because you don't mention what it is). Seldom mentioned the ASME/ANSI B.31.3 Code for Oil Refinery and Chemical Plant Piping.
I'll add three more: ASME VIII for Pressure Vessels, API 650 for Oil and Oil Products Storage Tanks and AWWA D100 for Water Storage Tanks.
In any of those cases the Codes are quite precise on what to do if the radiographs aren't satisfactory, not leaving margin to confusion or wrong interpretations.
Giovanni S. Crisi
Sao Paulo - Brazil
Giovanni, Seldom
The mandatory code is AWS D1.1. The welded connections have been made in columns and beams. We have used double and single V joint type with backgouging before weld the other side. However at the project begin some welders do not clean the joitn root.. so some slag inclusion (from smaw proccess. "tack welds") was detected in the ultrasonic examination. The code indicate that if any NDT other than visual examination is required the type and extension of the examination shall be clear in the contractual specifications. the specifications in this case say "5% of all full penetration welds shall be ultrasonic examinated". it do not say anything about retesting.
I determined the % defect by the relationship between the defect lenght and the weld inspectioned lenght. but the client say that a splice in a W 18*103 beam (with a defect 10 mm length) is complete bad and take yours average by joint and not for weld length.
I want to re-iterate the principle:
What is the acceptance criteria.
All codes have acceptance criteria for NDE. If the NDE performed has no acceptance criteria in the code, then you must establish some with the purchaser. If any part of your weld fails to meet the acceptance criteria, then your component has failed the inspection. It is irrelevant what percentage of the length of weld failed.
You mentioned that you had a defect with a 10mm length. If this is a crack-like defect, then I must agree that it would be considdered rejectable. You would have to excavate it and repair. In addition, you should perform additional NDE to give you assurance of the quality of the rest of your welds.
The bottom line is: WHAT ARE YOUR ACCEPTANCE CRITERIA? They will rarely be based on the amount of welds with defects as a percentage of total weld inspected. They are generally based on the absolute size of the defect or defects if they are clustered close together. In addition, crack-like defects are generally rejectable irrespective of their size.
Regards
Niekie Jooste
Niekie. I understand you. I have tested these joints by AWS table 6.2 acceptance criteria and I have found some slag inclusions rejectables according the table 6.2. I think you are right that if I find a minimum length of slag inclusion rejectable according the acceptance criteria all the joint is rejectable regardless its length. so I should perform additional NDE, but how much have to be this increase? my client is requesting 100% of ultrasonic test.
Regards
Miguel
As stated in another post, typically the approach is to select twice as much length of weld for each rejected weld. If these also contain rejectable defects then 100% inspection becomes a requirement. All rejectable defects have to in any case be excavated and repaired.
We have to remember that the end-user wants a product that will be able to do the job. The inspection itself is not a means of "improving" the welds. It is merely a means of obtaining a reasonable probability that you have good welds. If the welds are not good, then you need to find them all and repair them.
Regards
Niekie
Hi Miguel,
Coming from the ASME and API world and not being an aficionado by a long shot of the D1.1, your question has tweaked my interest. My first thoughts were that your question was a pretty straightforward one and it would be if it were ASME or API but not AWS! Nope, it appears they like to see it done the hard way.
If all that is stated in your spec is that "5% of all full penetration welds shall be ultrasonic examined" as you’ve written, then you have a case as per the requirements of the D1.1. It would appear to me that the Owner had the foresight to require random tests in order to give him a reliable indication of overall weld quality. This would be great but the problem that arises as I see it is he didn’t have quite enough foresight! He apparently neglected to follow through with a specific procedure defining requirements for additional testing (similar to ASME progressive) due to any defects disclosed during the 5% examination that exceeded the acceptance criteria.
So the only question there is to answer is- were those 8 welds rejected or not? If the 8 welds were found to meet the minimum quality requirements of the specification, then the Owner’s request is a deviation from the specification and under the auspices of para 6.4.5 and good luck to both.
In my opinion IF they were rejected and subsequently repaired as per para 6.6.4, the specification has been have satisfied by virtue of “specified welds” as stated therein. Having said that, I’d direct you to the “Commentary” section of the Code and specifically para C6.4.4. which continues to address both specified and non-specified NDT. In your case the Owner has found that approximately 40% of the work (40% of a sampling depicts 40% of the whole) can be assumed to contain defects that likewise exceed the maximum quality allowances of the Code. The Owner, due to his aforementioned lack of foresight, has found himself caught between the proverbial rock and the hard spot. Meaning he’s caught between para 6.4.5 and his budget! The Contractor though, is also caught, he’s caught between his budget and his reputation for being forthright!
We all know what random sampling is and what it’s meant to provide both the Owner and the Contractor. It would appear that an agreement could be worked out that would lessen the impact of continued NDT and repair for both parties. One effective method of obtaining a middle ground could be to identify who welded the joints that were rejected. Possibly, any additional NDT could focus on those welders work rather then those welds (welders) found that consistently met the minimum quality requirements during the original 5% sampling.
as per AWS D1.1-96:
6.15.3 Spot Testing. When spot testing is specified, the number of spots in each designated category of welded joint to be tested in a stated length of weld or a designated segment of weld shall be included in the information furnished to the bidders. Each spot test shall cover at least 4 in. (100 mm) of the weld length. When spot testing reveals indications of rejectable discontinuities that require repair, the extent of those discontinuities shall be explored. Two additional spots in the same segment of weld joint shall be taken at locations away from the original spot. The location of the additional spots shall be agreed upon between the contractor and the verification Inspector.
When either of the two additional spots show defects that require repair, the entire segment of weld represented by the original spot shall be completely tested. If the weld involves more than one segment, two additional spots in each segment shall be tested at locations agreed upon by the contractor and the verification Inspector, subject to the foregoing interpretation.
Rob. The term spot does not aplly here, because the specification says that the joint selected shall be inspectioned around its full length. Regards. Miguel.
Miguel,
It is very common in CA to scan a percentage of structural welds. There may be additional criteria in the form of building codes, depending on your project's location. Jurisdictional requirements (e.g.: 1997 UBC, Section 17) are included with the contract documents along with a requirement for compliance with the applicable AWS Structural Code(s) and any other project specification. The AISC-ASD manual is typically included in these specifications. Maybe the IBC is a requirement. Many countries have adopted the International Building Code.
The 97 UBC, subsection 1703 specifies NDT for siesmic zones 3 & 4 (most of the west coast). NDT may be reduced to 25% if the welder has a 5% or less reject rate. The fact is most Engineers require 100% NDT. When percentages are specified, then they also require the full length of any one weld unless the weld is greater than 3' in length. The weld size (E) determines how much weld >3' are considered to be one weld.
I think the point of what does your AC state kind of drives home the responses. I would inquire about the project specifications and or any applicable local governing jurisdiction requirements. If your at a fab shop that is shipping this to Anywhere, USA or Someplace, England then you'll need to know what the requirements are for the future home of that steel.
