where expertise comes together - since 1996

Web's Largest Portal of Nondestructive Testing (NDT)
Open Access Database (Conference Proceedings, Articles, News), Exhibition, Forum, Network

All Forum Boards
Technical Discussions >
Material or Machine Problems
Career Discussions
Job Offers
Job Seeks
Classified Ads
About NDT.net
Articles & News

Phoenix Inspection Systems Limited
Design and manufacture ultrasonic Transducers, Scanners and Custom Solutions for NDT inspections. Innovators in NDT technology

961 views
07:08 Oct-03-2002
K.C.Fosha
Material or Machine Problems

Recently the company I work for did a floor scan on a tank. Several areas where identified, and repaired. Another floor scan was preformed, and several more areas where noted. (This time I watched.) Again, areas where marked and repaired. Upon compleation of the repairs and inspections, a hydro test was performed. Hydro was acceptable. (No leaks within 24hrs.)

After water was emptied from tank, tank was put back into service, and product was put into the tank. Tank was about half way full and started to leak. A few days later during the inspection, through holes where observed in one area of the tank floor.

The question I have is: Could the MFE Unit have missed these indications during scaning? (I know that the area was scaned, because I saw it scaned.) If so how?

Could something have happened in the material that would have coused this? (No top side corrosion noted on floor. Failure was from bottom side.) The only thing that was different between the hydro-test and when product was put in the tank was two heating coils where turned on. (To about 250 degrees F.)

If anyone has any ideas, let me know.


 
01:15 Oct-03-2002

Peter Large

NDT Inspector
Anglo Worldwide Technology Ltd,
United Kingdom,
Joined Apr 2000
3
Re: Material or Machine Problems : Recently the company I work for did a floor scan on a tank. Several areas where identified, and repaired. Another floor scan was preformed, and several more areas where noted. (This time I watched.) Again, areas where marked and repaired. Upon compleation of the repairs and inspections, a hydro test was performed. Hydro was acceptable. (No leaks within 24hrs.)
.
: After water was emptied from tank, tank was put back into service, and product was put into the tank. Tank was about half way full and started to leak. A few days later during the inspection, through holes where observed in one area of the tank floor.
.
: The question I have is: Could the MFE Unit have missed these indications during scaning? (I know that the area was scaned, because I saw it scaned.) If so how?
.
: Could something have happened in the material that would have coused this? (No top side corrosion noted on floor. Failure was from bottom side.) The only thing that was different between the hydro-test and when product was put in the tank was two heating coils where turned on. (To about 250 degrees F.)
.
: If anyone has any ideas, let me know.
.
Does happen occaisionally. Have used technique very successfully, but have missed holes. Was the through hole surrounded by underside corrosion? Sensitivity correct? Scanning technique good enough to cover 100%? What diameter of hole through floor? Anywhere near floor plate lap joint that may have caused sensors to be raised up or tilted? Near annulus plates? A big worry but it does happen.



 
01:18 Oct-03-2002

Peter Large

NDT Inspector
Anglo Worldwide Technology Ltd,
United Kingdom,
Joined Apr 2000
3
Re: Material or Machine Problems : : Recently the company I work for did a floor scan on a tank. Several areas where identified, and repaired. Another floor scan was preformed, and several more areas where noted. (This time I watched.) Again, areas where marked and repaired. Upon compleation of the repairs and inspections, a hydro test was performed. Hydro was acceptable. (No leaks within 24hrs.)
: .
: : After water was emptied from tank, tank was put back into service, and product was put into the tank. Tank was about half way full and started to leak. A few days later during the inspection, through holes where observed in one area of the tank floor.
: .
: : The question I have is: Could the MFE Unit have missed these indications during scaning? (I know that the area was scaned, because I saw it scaned.) If so how?
: .
: : Could something have happened in the material that would have coused this? (No top side corrosion noted on floor. Failure was from bottom side.) The only thing that was different between the hydro-test and when product was put in the tank was two heating coils where turned on. (To about 250 degrees F.)
: .
: : If anyone has any ideas, let me know.
: .
: Does happen occaisionally. Have used technique very successfully, but have missed holes. Was the through hole surrounded by underside corrosion? Sensitivity correct? Scanning technique good enough to cover 100%? What diameter of hole through floor? Anywhere near floor plate lap joint that may have caused sensors to be raised up or tilted? Near annulus plates? A big worry but it does happen.

Forgot to ask, were you using MF leakage or straight forward ultrasonic testing? I presumed MFL.
.



 
06:38 Oct-04-2002

John O'Brien

Consultant, -
Chevron ETC ,
USA,
Joined Jan 2000
278
Re: Material or Machine Problems There are two significant components here
A) it is possible that operator/instrument can be at fault. A number of industry round robin trials have identified that operator and procedure is more important with MFL than first thought. It is possible to get widely varying inspection results with the same instrument but changing operators. This is one reason why API 653 is producing a new Appendix G on qualification of MFL scanners and operators.

B) Just beacuse you scan an area does not mean that a change will be detected. As with most NDE techniques MFL has detection thresholds. It is widely reported that small isolated defects such as MIC have been missed by MFL Floor Scanning. Localised pits less than 3 mm in Diameter are commonly missed by most floor scanners.

A comprehensive floor inspection must include a number of components A) Coverage of all areas, B) assessment of expected corrosion mechanisms C) NDE to reduce the potential to miss any of the mechansisms D) Control of inspection methodologies.

Regards

John O'Brien


 
03:38 Oct-07-2002
Richard Kazares
Re: Material or Machine Problems Peter, excellent question: you have otherwise good material with a small anomaly - located in a large area. With MFE (as with UT) finding a very small problem in a VERY big area is problematic, at best.

If however, (as is normally the case - FORTUNATELY) the area is surrounded by corroding material (thinning from ether surface) then the problem is much simpler - PARTICULARLY FOR Automated UT - since the corroding areas are easily detectable from large area scans (C-Scans) where there is assurance that there is full coverage. This may (or may not) be the case with MFE.

In our experience with Automated UT system (such as our own LSI System) - a scanning STRATEGY is VERY important to create the HIGHEST probability of detecting corroding (problematic) areas - with a reasonable scan time and cost.

We operate our own systems with a dual-mode inspection strategy - Moderate resolution is applied for a complete 1st pass of the entire area - and then selecting any suspect areas and performing VERY high resolution UT C-Scans - in those areas only.

In addition, for an even higher probability of detection of small through holes, several random high-resolution scans - in otherwise "good" material areas can also be performed to assure that the original moderate resolution data is not misleading - or has not somehow canceled real damage.

Combining high-resolution Automated UT scanning with Acoustic Emission testing (which is a GLOBAL inspection methodology) is also another means of reducing the probability of missing heavily corroded areas - which require high-resolution UT scanning.

Again, NO METHOD is perfect (100%!!) in detecting small through holes in otherwise "good" material - but the dual-strategies cited here - at least have demonstrated a significantly higher probability of detection that any single method, alone.


: : Recently the company I work for did a floor scan on a tank. Several areas where identified, and repaired. Another floor scan was preformed, and several more areas where noted.(This time I watched.) Again, areas where marked and repaired. Upon compleation of the repairs and inspections, a hydro test was performed. Hydro was acceptable. (No leaks within 24hrs.)
: .
: : After water was emptied from tank, tank was put back into service, and product was put into the tank. Tank was about half way full and started to leak. A few days later during the inspection, through holes where observed in one area of the tank floor.
: .
: : The question I have is: Could the MFE Unit have missed these indications during scaning? (I know that the area was scaned, because I saw it scaned.) If so how?
: .
: : Could something have happened in the material that would have coused this? (No top side corrosion noted on floor. Failure was from bottom side.) The only thing that was different between the hydro-test and when product was put in the tank was two heating coils where turned on. (To about 250 degrees F.)
: .
: : If anyone has any ideas, let me know.
: .
: Does happen occaisionally. Have used technique very successfully, but have missed holes. Was the through hole surrounded by underside corrosion? Sensitivity correct? Scanning technique good enough to cover 100%? What diameter of hole through floor? Anywhere near floor plate lap joint that may have caused sensors to be raised up or tilted? Near annulus plates? A big worry but it does happen.
.



 
04:19 Oct-07-2002

Joe Buckley, Sonatest Plc

Consultant, ASNT L-III, Honorary Secretary of BINDT
Level X NDT, BINDT,
United Kingdom,
Joined Oct 1999
515
Re: Material or Machine Problems (We are the European distributors for MFE. In North America please contact them directly)

The main question here is how was the scanner sensitivity set?

It is important to remember that Magnetic Flux leakage (as used by MFE) is essentially a volume-sensitive technique.) it gives good sensitivity to typical areas of thinning and corrosion but is inherently not very sensitive to 'pinhole' type defects. It can find holes (down to 2-3mm or so in 10mm plate) but this requires higher gain settings, which will give unacceptable backgound noise if significant corrosion is present.

In practice isolated pinholes are seldom seen in oil storage tank floors, But like all NDT methods, you need to have an understanding of the expected defect mechanisms, and of the limitation of the techniques used .

Did you remove the defective floor plates - What was the defect geometry like? The fact that the failure was from underneath suggests that its not a product related corrosion issue, so normal assumptions about defectshape may not be appropraite. However it should be possible to simulate the defects in a steel plate and verify that they are detected.

The extra 200 degrees or so will cause quite a lot of movement. Its quite feasible that defects will open up (or close)

Joe


: Recently the company I work for did a floor scan on a tank. Several areas where identified, and repaired. Another floor scan was preformed, and several more areas where noted. (This time I watched.) Again, areas where marked and repaired. Upon compleation of the repairs and inspections, a hydro test was performed. Hydro was acceptable. (No leaks within 24hrs.)
.
: After water was emptied from tank, tank was put back into service, and product was put into the tank. Tank was about half way full and started to leak. A few days later during the inspection, through holes where observed in one area of the tank floor.
.
: The question I have is: Could the MFE Unit have missed these indications during scaning? (I know that the area was scaned, because I saw it scaned.) If so how?
.
: Could something have happened in the material that would have coused this? (No top side corrosion noted on floor. Failure was from bottom side.) The only thing that was different between the hydro-test and when product was put in the tank was two heating coils where turned on. (To about 250 degrees F.)
.
: If anyone has any ideas, let me know.
.



 


© NDT.net - The Web's Largest Portal of Nondestructive Testing (NDT) ISSN 1435-4934

Open Access Database, |Conference Proceedings| |Articles| |News| |Exhibition| |Forum| |Professional Network|