Re: Weld Flaw Sizing Methods : With regard to weld testing and sizing flaws.
: I am aware of the following ultrasonic sizing methods eg.
: * 6db Drop
: * 20db Drop
: * Maximum Amplitude Method
: * Tip Diffraction Method
: Would it be possible to get a general comment from the UT experts that covers the following points:
: - What are the main sizing methods used for manual UT scanning of welds.
: - From a practical point of view, are a combination of sizing methods generally used to obtain flaw dimensions.
: - Do sizing methods differ from country to country.
: - What are the accepted practical tolerance for flaw sizing (eg +-1mm ?).
: - What are the main advantages & disadvantages of each sizing method.
: Thank you,
: Jeff Phillips
Mr. Phillips: You have broached a much debated topic! The 4 options for sizing that you mentioned are now probably the most common options and I suspect we could add things like ALOK and AVG the latter being more area than linear based. Let me address the 5 questions you asked and qualify this as opinion, since the feeling and practices established on this matter are quite entrenched in some quarters.
1- What are the main sizing methods used for manual UT scanning of welds.
Roughly speaking we group sizing methods into 2 types, either amplitude based or time based. Since the earliest inspections it has been easiest to relate large amplitude signals to flaw sizes. A direct relationship relating amplitude and defect "size" was developed by Krautkramer (AVG or DGS in English). Other amplitude based sizing techniques relate size to amplitude drop from a maximum as related to a change in probe position. These include your Maximum Amplitude, 6 and 20 dB Drop and ALOK methods.
We have more recently recogonised Time based methods as useful sizing options (forward scattered technique TOFD and back scattered TOF methods using both single and bi-modal probe configurations). TOFD might be better considered a mechanised rather than manual UT scanning technique. Similarly SAFT, used for lateral sizing, is more likely to be used in mechanised systems where B-scans are being evaluated.
2.- From a practical point of view, are a combination of sizing methods generally used to obtain flaw dimensions.
This depends on your point of perspective. If you are a diligent technician or researcher you might want to evaluate every indication in detail. If you are a manufacturer you want to just meet the code requirements so too much detail is not desirable. In nuclear in-service inspections flaws are often located by manual scanning. Occasionally flaws are located that the manufacturer missed and the unit has been in service for some time. Had the flaw been detected in the shop it would have been repaired. In-service repair is VERY inconvenient so now the owner looks for rationale to avoid it. Elaborate sizing options are then used to explain to regulatory bodies why the flaw is not critical. Then a combination of methods is used to allow more precise engineering of the failure analysis calculations.
Similarly, defect length is usually a significant criteria for accept/reject of a weld. Since most codes' acceptance criteria for UT are conversions of Radiographic acceptance criteria, the defect lengths are based on "workmanship" therefore they are conservative. Typically 10-12mm would be the maximum allowable length for some flaw types. However, for short flaws, the normal dB drop method measures only beam width not defect length. If this became a problem then someone would look to a combination of methods to look at size flaws when it appeared it would be costly to repair the weld. This is not always just an excuse by the manufacturer to cut cost (although it might be one reason). If a repair is made on many small flaws because the Code is very conservative, the weld metal structure could be degraded and a loss of strength result (even with all unacceptable Code defects removed).
3.- Do sizing methods differ from country to country.
It would be more accurate to say that "preferred" or "more frequently used" methods have national or regional differences. In North America we are having a difficult time making changes to the rules that regulate how we do UT. Codes dictate what methods will be used. Pressure vessel code, structural steel construction codes, etc. made and used in North America still base sizing on dB drop methods. ASME now permits (Section XI) other sizing methods to be used in certain cases. In North America the 6 dB drop is preferred. It was only when I met technicians from the UK did I learn of the 20dB Drop and MaxAmp methods. I suspect AVG is more commonly required in continental Europe. Asia has many codes using a 6dB drop method too. In Australia I think the 20dB drop is preferred.
Tip diffraction methods of sizing are not usually a "first line" option. For convenience of recording, the amplitude drop methods are probably the first choice. The Japanese have developed tip diffraction methods to areliable level since the modelling algorithms by Harumi explained many of the signals on the UT scope but it normally involves a defect be detected and assessed by "traditional" methods before tip diffraction is used.
4.- What are the accepted practical tolerance for flaw sizing (eg +-1mm ?).
I doubt we can reasonably expect to assign such a quantity to amplitude based methods considering the variables involved. For Tip diffraction methods a "practical" tolerance is often a percentage of wall thickness and a lower limit (under good conditions) might be +/- 0.5-1mm. Frequency and probe characteristics and actual flaw size will dictate a lower limit of the flaw size that can be sized. I.e. when looking to size a surface braking flaw with a 2mm high profile, the frequency and probe damping may not be adequate to see a time difference between the tip and "corner" reflection. Type of flaw will also dictate accuracy since planar defects usually provide a single set of signals to assess and multi-faceted flaws will provide several signals which might detract from the overall accuracy.
5.- What are the main advantages & disadvantages of each sizing method.
Here I will again be general in my reply.
Any amplitude drop based method is subject to the normal variables such as reflectivity, angle of incidence, coupling effects, surface conditions, geometry of part, frequency, etc. On the ideal reflectors such as flat bottom holes and side drilled holes the theory is fine, but for real defects the theory quickly breaks down as we try relating the change of amplitude with change of position in the beam,.
For time based methods the variables are not quite so bad but assumptions must be made and these should be reasonable based on good information from echo dynamics to characterise the flaw. Orientation of the flaw with respect to the beam will still be a major factor.
Amplitude based options are fast and relatively convenient methods to position and rank a flaw but generally have poor sizing accuracy. TOF options are usually better for accuracy but time consuming and difficult to incorporate as part of a general pre-service inspection.
Two gentlemen with the Australian DND, H.R.Chin Quan and I.G.Scott, did a study on the NDT operator and factors affecting their performance. They came up with 8 conclusions in their study. The last 4 of their conclusions relate in one way or another to your general concerns for flaw sizing.
1. The statement of a minimum detectable defect size is of use only in fixing the lower bound of detection process; it is not an efficient assessment of either operator or technique. Valid statements can only be made in terms of probabilities and confidence levels, for defect sizes relevant to the task envisaged. These can thereafter be sensibly applied to real problems concerned with the reliability of finding defects.
2. An entirely different set of problems arise when the operator is asked to make measurements. Present NDT equipment is not designed for this purpose and the operators' training is unsuitable. The requirement arises when improvements in structural reliability are sought using NDT and fracture mechanics.
3. Caution is needed in assessing various comparisons. Tests on operators in the laboratory can be extrapolated to the field but not necessarily without qualification. Comparison techniques should be made with care because changes in the conditions may produce entirely false indications and some comparisons are invalid, e.g. when attempts are made to compare unsuitable techniques. No universal solution to these problems has been found and, of necessity, workers need to examine their own situation with care.
4. The need to improve the overall results obtained by NDT is apparent-if improvement is not achieved then some replacement process is likely to be developed.
(From: Operator Performance and Reliability, by H.R.Chin Quan and I.G.Scott of the Department of National Defence, Australia, in series Research Techniques in NondestructiveTesting, ed. R.S.Sharpe)
If you would like more information the literature on this subject is very extensive.