Re: Small Void/Inclusion detection

[ Ultrasonic Application Workshop ]
[ Follow Ups ] [ Post Followup ] [ Help ]

Re: Small Void/Inclusion detection
Posted by: Linas Svilainis , E-mail: svilnis@rf.ktu.lt, on May 23, 1997 at 21:29:38:
In Reply to: Small Void/Inclusion detection posted by : Gerald Feldewerth on May 23, 1997 at 20:21:21:
: Are standards and techniques available to detect very small voids or inclusions in pure metals using standard C-scan equipment?? The size I am trying to detect is 0.05 to 0.1mm. If this can be accomplished can anyone tell me where I can get flat bottom holes made to use for calibration standards.
Dear Gerald,
I've seemed paper dealing with problem like yours.
They use the phenomenon that reflectors exhibit particular
frequency template, depending on their size (and,
unfortunately, the shape). So, after inspecting just
time domain (A-scan) frequency response it is possible
to tell whether it was the defect of desired(?) size
or not. They went even further. The exciting pulse
was the complex signal, which contained inverse of
the spectra of reflection from defect of targeted size.
In such case just the reflection from defects of targeted
size will give the largest reflection. C-scan (I mean a
MaxPlot C-scan) will exhibit such defects reflections.
One can go even further. The spectral processing can
be applied to C-scan data in spatial domain, or, if
you can collect the whole volume of 3-D space(time,x,y),
you could apply the correlation algorithms or neural
networks. Might be SAFT'ing in connection with frequency
signature analysis could eliminate the orientation
problem... But...
One has to consider few factors, limiting such application.
One problem is the facilities required to do such task.
Good reference is right here - Tom Nelligan's paper is
presenting the time and storage requirements for ultrasonic
automated systems:
http://www.ultrasonic.de/article/wsho0597/nelligan/nelligan.htm
Another problem is related to the small size of the
defects you are looking at. It would be good to have
the inspection wave of order of the defect you are
looking for. But in such case you'll have to switch to
frequencies of 20-50MHz. In such case the attenuation
will raise rapidly. Also it should be denoted what you
mean by pure metals. If the level of "grain noise" is
high, then your defects could be masked by the structure
of the metal.
Calibration blocks...do not know. We usually use the
holes produced by electrical spark. But the shape
of boottom of such hole could not be predicted. Might
be laser-drilled hole will suite?
Good luck,
Linas

Follow Ups:

Post a Followup:

Name:

E-Mail:

Subject:

Comments:
: : Are standards and techniques available to detect very small voids or inclusions in pure metals using standard C-scan equipment?? The size I am trying to detect is 0.05 to 0.1mm. If this can be accomplished can anyone tell me where I can get flat bottom holes made to use for calibration standards. : Dear Gerald, : I've seemed paper dealing with problem like yours. : They use the phenomenon that reflectors exhibit particular : frequency template, depending on their size (and, : unfortunately, the shape). So, after inspecting just : time domain (A-scan) frequency response it is possible : to tell whether it was the defect of desired(?) size : or not. They went even further. The exciting pulse : was the complex signal, which contained inverse of : the spectra of reflection from defect of targeted size. : In such case just the reflection from defects of targeted : size will give the largest reflection. C-scan (I mean a : MaxPlot C-scan) will exhibit such defects reflections. : One can go even further. The spectral processing can : be applied to C-scan data in spatial domain, or, if : you can collect the whole volume of 3-D space(time,x,y), : you could apply the correlation algorithms or neural : networks. Might be SAFT'ing in connection with frequency : signature analysis could eliminate the orientation : problem... But... : One has to consider few factors, limiting such application. : One problem is the facilities required to do such task. : Good reference is right here - Tom Nelligan's paper is : presenting the time and storage requirements for ultrasonic : automated systems: : http://www.ultrasonic.de/article/wsho0597/nelligan/nelligan.htm : Another problem is related to the small size of the : defects you are looking at. It would be good to have : the inspection wave of order of the defect you are : looking for. But in such case you'll have to switch to : frequencies of 20-50MHz. In such case the attenuation : will raise rapidly. Also it should be denoted what you : mean by pure metals. If the level of "grain noise" is : high, then your defects could be masked by the structure : of the metal. : Calibration blocks...do not know. We usually use the : holes produced by electrical spark. But the shape : of boottom of such hole could not be predicted. Might : be laser-drilled hole will suite? : Good luck, : Linas

Optional Link URL: Example: "http://www....html". Alternative write "http://www.ultrasonic.de/wshop/attach/filename.htm" if you transfered a filename.htm via FTP

Link Title: is important! Information on my Home Page

Optional Image URL: http://www....gif

[ Ultrasonic Application Workshop ] [ Help ]

Optional Link URL:	Example: "http://www....html". Alternative write "http://www.ultrasonic.de/wshop/attach/filename.htm" if you transfered a filename.htm via FTP
Link Title: is important!	Information on my Home Page
Optional Image URL:	http://www....gif