where expertise comes together - since 1996 -

The Largest Open Access Portal of Nondestructive Testing (NDT)

Conference Proceedings, Articles, News, Exhibition, Forum, Network and more

where expertise comes together
- since 1996 -
5651 views
Technical Discussions
Heydar Alakbarov
Heydar Alakbarov
21:03 Jun-02-2016
Detectability of pulse-echo UT method

We use classic geometric approach of ultrasonic weave propagation for explanation of results obtained using pulse-echo UT method. It works well when size of reflector (size of area perpendicular to the direction of wave propagation) exceeds size of wavelength.
In the training materials of NDT Resource Center (https://www.nde-ed.org/EducationResources/educationresource.php ; direction Home-Education Resources-NDT Course Material-Ultrasound-Physics of Ultrasound-Wavelength/Flaw Detection) I found statement “A general rule of thumb is that a discontinuity must be larger than one-half the wavelength to stand a reasonable chance of being detected”
For practical use it means that detectability of pulse-echo method ( the minimum size of reflector could be detected by the method) is one-half of the wavelength (assuming that we use transduser with very narrow frequency band).
Do anybody have some arguments pro- or contra- about this statement.
Thanks in advance.

    
 
 Reply 
 
Anmol Birring
Consultant,
Birring NDE Center, Inc., USA, Joined Aug 2011, 738

Anmol Birring

Consultant,
Birring NDE Center, Inc.,
USA,
Joined Aug 2011
738
21:37 Jun-02-2016
Re: Detectability of pulse-echo UT method
In Reply to Heydar Alakbarov at 21:03 Jun-02-2016 (Opening).

This is a common misconception. Detectibilty is not based only on wavelength but on signal to noise ratio. Reflectivity from flaw depends on size, orientation,shape etc.. Noise is produced from material microstructure. In addition, one has to consider loss from diffraction that depends on ratio of wavelength to flaw size

    
 
 Reply 
 
Diego
Consultant,
Freelance, Spain, Joined May 2013, 188

Diego

Consultant,
Freelance,
Spain,
Joined May 2013
188
22:20 Jun-02-2016
Re: Detectability of pulse-echo UT method
In Reply to Heydar Alakbarov at 21:03 Jun-02-2016 (Opening).

Hi Heydar.
It is the wavelength which determines the smallest detectable reflector in a material, usually taken to be approximately l/2, but also the ability of the ultrasound to penetrate into material structure.
If you want to employ the beam geometry method (6 dB, 20 dB or maximum amplitude) to determine the dimensions of a reflector, then the dimension to be determined shall be larger than the width of the beam at the reflector location.

    
 
 Reply 
 
Andrew Cunningham
NDT Inspector
Canada, Joined Jun 2008, 238

Andrew Cunningham

NDT Inspector
Canada,
Joined Jun 2008
238
22:23 Jun-02-2016
Re: Detectability of pulse-echo UT method
In Reply to Diego at 22:20 Jun-02-2016 .

The broad sweeping statement of wavelength is WRONG, the amplitude of a reflector is governed by orientation. A large defect not orientated to the angle of the probe may be detectable, whereas a tiny reflector perfectly orientated to the probe will send back a screamer. The blip on the screen that you can never find again is the prime example.

    
 
 Reply 
 
Heydar Alakbarov
Heydar Alakbarov
22:50 Jun-02-2016
Re: Detectability of pulse-echo UT method
In Reply to Anmol Birring at 21:37 Jun-02-2016 .

Anmol,
If I use 1 MHz (with very low band wide) compression wave transducer to inspect ceramic I will have very low noise level. What size of flaw which reflecting surface is strictly perpendicular to beam axis can I detect?

    
 
 Reply 
 
Anmol Birring
Consultant,
Birring NDE Center, Inc., USA, Joined Aug 2011, 738

Anmol Birring

Consultant,
Birring NDE Center, Inc.,
USA,
Joined Aug 2011
738
23:25 Jun-02-2016
Re: Detectability of pulse-echo UT method
In Reply to Heydar Alakbarov at 22:50 Jun-02-2016 .

You will have to run such tests as I have no idea on the background noise produced from the ceramic you are using.

    
 
 Reply 
 
Heydar Alakbarov
Heydar Alakbarov
23:34 Jun-02-2016
Re: Detectability of pulse-echo UT method
In Reply to Anmol Birring at 23:25 Jun-02-2016 .

Anmol,
Detectability depend on used wavelength. That is why to inspect concrete we use transducers operating under hundreds kiloherz and for examine ceramic we use 5 MHz and more.

    
 
 Reply 
 
Ed Ginzel
R & D, -
Materials Research Institute, Canada, Joined Nov 1998, 1266

Ed Ginzel

R & D, -
Materials Research Institute,
Canada,
Joined Nov 1998
1266
00:16 Jun-03-2016
Re: Detectability of pulse-echo UT method
In Reply to Heydar Alakbarov at 23:34 Jun-02-2016 .

Heydar, you are now making general statements and introducing small bits of information each time! Ceramics can be of many types...some are so porous as to be impractical to test (even at very low frequencies...never mind 1MHz). Then you made a reference to testing concrete using kilohertz frequencies. What is the size of the aggregate? will the aggregate be identified as aggregate or will it produce a response you think is a flaw (void)? Chalk, limestone and dolomites can have velocities as low as 2300m/s while other aggregates (e.g. granites) are reported as up to 6000m/s. How big does the aggregate have to be in cementaceous martix order to differentiate it from the acoustic impedance of a "relatively" small void.
Off line I provided you with a reference (Krautkramer's Ultrasonic Testing Of Materials) that is considered a good reference in the discsussion on scatter/reflection. It had some caveats stated in the section I quoted you on scatter. Was there something you thought was incorrect in Krautkramer?

We still have no idea what you are testing and what you are trying to detect!
This concept of "What is the smallest defect I can detect" is a silly question that people are still asking. We have already discussed this before and the question is wrong....the question should always be "What is the largest flaw that I might miss?"

    
 
 Reply 
 
massimo carminati
Consultant, AUT specialist
IMG Ultrasuoni Srl, Italy, Joined Apr 2007, 691

massimo carminati

Consultant, AUT specialist
IMG Ultrasuoni Srl,
Italy,
Joined Apr 2007
691
10:46 Jun-03-2016
Re: Detectability of pulse-echo UT method
In Reply to Heydar Alakbarov at 21:03 Jun-02-2016 (Opening).

The halfwave related statement is definitely untrue. Just an example: 2 MHz compression wave in steel means wavelength of about 3 mm. Halfwave is 1,5 mm. With 4 MHz i half of that.
In forging inspection for critical components, the MDDR required is often less than a mm. Which frequency would you use? 4 MHz, for sure. Well try to detect 0.8 mm FBH with 4 MHz at 1 meter distance....attenuation and SNR are almost always prohibitive, therefore you should use 2 MHz. But then comes beam spread.... so we must increase probe size...
This is to explain how not just frequency has to be taken into account when calculating minimum detectable defect.

    
 
 Reply 
 
Ed Ginzel
R & D, -
Materials Research Institute, Canada, Joined Nov 1998, 1266

Ed Ginzel

R & D, -
Materials Research Institute,
Canada,
Joined Nov 1998
1266
13:49 Jun-03-2016
Re: Detectability of pulse-echo UT method
In Reply to massimo carminati at 10:46 Jun-03-2016 .

Massimo, you pointed out good practical considerations. Back in 2012 Prof. Svilainis and Wieslaw Bicz provided good replies to the query I made asking about where this half-wavelength rule of thumb arose.
http://www.ndt.net/forum/thread.php?admin=&forenID=0&msgID=45745&rootID=45501#45745
As Wieslaw pointed out, Rayleigh scattering will always occur. But at some point, the accumulated noise (including system electrical noise) and attenuation will be, as you pointed out, the practical limit and we will simply not be able to differentiate the tiny signal that arises from a small point diffractor and the system noise.

    
 
 Reply 
 
Anmol Birring
Consultant,
Birring NDE Center, Inc., USA, Joined Aug 2011, 738

Anmol Birring

Consultant,
Birring NDE Center, Inc.,
USA,
Joined Aug 2011
738
16:46 Jun-03-2016
Re: Detectability of pulse-echo UT method
In Reply to Ed Ginzel at 13:49 Jun-03-2016 .

Ed, There is a lot of misunderstanding on this topic of flaw detectibility. There are sereval variables involved. I will write a paper on this subject

    
 
 Reply 
 
Diego
Consultant,
Freelance, Spain, Joined May 2013, 188

Diego

Consultant,
Freelance,
Spain,
Joined May 2013
188
22:48 Jun-04-2016
Re: Detectability of pulse-echo UT method
In Reply to Andrew Cunningham at 22:23 Jun-02-2016 .

Mr. Cunningham.

The first part of my previous post was in relation to the first post of Mr. Alakbaro, and it is from a general view of the ultrasonic principles; the second part of the post was to indicated that, if the beam geometry method is employed to size a reflector, it should be keep in mind that it is not possible to determine any dimension of the reflector that is smaller than the width of beam at the location of the reflector.

Having written the above, I do not understand why you say "The broad sweeping statement of wavelength is WRONG..."Please read you what I show below; this is a verbatim copy of the page 13 of Lesson 2 in The Classroom Training Handbook for Ultrasonic Method, it edited by ASNTC.
"In practical situations, the smallest discontinuity you can find with ultrasonic testing is about 1/2 lambda (wavelength).
Therefore, to detect smaller defects, you will need transducer that produce higher frequencies"

And finally I suggest you, please do not write capital letters (WRONG). This indicates that you're screaming, and it says nothing in favor of your manners.

    
 
 Reply 
 
Gokul Hrishikesh
NDT Inspector, PCN UT Level 3
India, Joined Jul 2013, 73

Gokul Hrishikesh

NDT Inspector, PCN UT Level 3
India,
Joined Jul 2013
73
11:39 Jun-05-2016
Re: Detectability of pulse-echo UT method
In Reply to Diego at 22:48 Jun-04-2016 .

Gentlemen,

I have experienced time and again that at higher frequency there is good sensitivity but I doubt it is 1/2 wavelength exactly or not... all the articles say that sensitivity is 1/2 wavelength I have no clue where that value is coming from... also back in 2004 in my level 2 cases I learned the same from my mentor.

For example :https://www.nde-ed.org/EducationResources/CommunityCollege/Ultrasonics/Physics/defectdetect.htm

My understanding is that, lower wavelength is so sensitive that it produce lot of reflections and scattering at large grain boundaries hence causing larger amount of noise(grass/hash) and attenuation.
But when we move on to lower frequency reflections and scattering at grain boundaries reduces(because now it is less sensitive to detect it with larger wavelength) and the impact is seen on attenuation and signal to noise ratio since no energy is wasted in scattering and internal reflections from grain boundaries.

I will stick to my understanding that higher frequency can detect smaller reflectors which lower frequency cannot (and that I am very sure if we are talking about lower thickness and small grain material)

What challenges the detection in higher thickness is other parameters like attenuation at larger beam path if we use higher frequency I don't expect any energy reaching at further locations (we have to consider that the attenuation will exist equally on the reflected energy from a small reflector)..... Practically we will not come across this issue since we would have chosen correct frequency for calibrations.
Attenuation, system generated noise or noise by EMF..... I agree with ED

I will be very attentive to learn better on this topic.

Kind Regards,
Gokul

    
 
 Reply 
 
Anmol Birring
Consultant,
Birring NDE Center, Inc., USA, Joined Aug 2011, 738

Anmol Birring

Consultant,
Birring NDE Center, Inc.,
USA,
Joined Aug 2011
738
15:58 Jun-05-2016
Re: Detectability of pulse-echo UT method
In Reply to Heydar Alakbarov at 21:03 Jun-02-2016 (Opening).

Let us take the case of detecting a 1 mm flaw in a steel at a depth of 20 mm. At 5 MHz L-waves, Lambda = 1.2 mm, so lambda/2 = 0.6 mm

1. The 1 mm will be easily detected with a 5 MHz L-wave probe in fine grain steel
2. When the same 1 mm flaw is in coarse grain material, signal to noise ratio with the same probe will degrade. It may not be detectable in coarse grain forging and definitely not in a casting. And at longer metal path detectibility degrades
3. Now let us go back to the fine grain steel and use 20 MHz probe (0.3 mm wave length) and lambda/2 = 0.15. Most likely no detection. So this wavelength /2 is not a correct approach.
4. Another issue is diffraction from flaw. When wavelength (lanbda) is equal to or greater than flaw size (d) then lanbda/d, diffraction start playing role and energy is lost . What it means that there is higher chance of detecting disoriented flaws at lower frequencies is more than at than higher frequencies. So the orientation of disk type reflector is more critical when using a 10 MHz probe rather than with a 2 MHz probe that will be more forgiving.
5. Probe size plays a role when the sound path increases eg more than 100 mm. Large probe size means less beam spread and increased SNR.
6. And finally focusing done in immersion testing (and with PA) improves detection provided the material is clean, fine grain.

In conclusion, there are many variables that affect flaw detection, but the bottom line is to maximize signal to noise ratio. lamda/d is only a factor for diffraction effects

    
 
 Reply 
 
Gokul Hrishikesh
NDT Inspector, PCN UT Level 3
India, Joined Jul 2013, 73

Gokul Hrishikesh

NDT Inspector, PCN UT Level 3
India,
Joined Jul 2013
73
21:47 Jun-05-2016
Re: Detectability of pulse-echo UT method
In Reply to Anmol Birring at 15:58 Jun-05-2016 .

At 5MHz we get noise from grain boundaries (coarse grains)
At 20MHz we get noise from grain boundaries (fine grains)

It is indeed a proof that when the frequency is higher it starts reflecting on grain boundaries because of the higher sensitivity to grain boundaries.
Lower the frequency and it is gone in other words less sensitive to grain boundaries.

http://www.olympus-ims.com/en/resources/white-papers/applications-high-frequency-transducers/

In the above reference you can see an imaging of 50MHz with a focal spot of 0.05 mm.... which a lower frequency cannot do.

Regarding orientation..... flaws can also be invisible if they are parallel to energy axis even if they are very very large...

    
 
 Reply 
 
Gokul Hrishikesh
NDT Inspector, PCN UT Level 3
India, Joined Jul 2013, 73

Gokul Hrishikesh

NDT Inspector, PCN UT Level 3
India,
Joined Jul 2013
73
21:55 Jun-05-2016
Re: Detectability of pulse-echo UT method
In Reply to Anmol Birring at 15:58 Jun-05-2016 .

I also have to say that when it comes to diffraction I am in total agreement with what you.

    
 
 Reply 
 
John Norman
Consultant, owner of business
NTS Ultrasonics Pty Ltd, Australia, Joined Oct 2012, 111

John Norman

Consultant, owner of business
NTS Ultrasonics Pty Ltd,
Australia,
Joined Oct 2012
111
03:04 Jun-07-2016
Re: Detectability of pulse-echo UT method
In Reply to Heydar Alakbarov at 21:03 Jun-02-2016 (Opening).

Hi Heydar.

Great question. You have certainly got people thinking. My take on the answer is as follows. The theory of Rayleigh scattering was developed many years ago and most probably for small reflectors in a liquid, or some other "ideal" situation. Technical literature from sonar often considers scattering.

The reflectivity of a small reflector is expressed as a scattering cross-section and is proportional to the "equivalent volume" of the reflector times the fourth power of frequency (D^3*f^4 where D is the equivalent diameter and f is the ultrasonic frequency). So, it is very sensitive to wavelength. In the Rayleigh scattering regime, the shape of the reflector is "reduced" to an equivalent sphere by some mathematical process. The scattering "pattern" has two lobes: one lobe scattering backwards towards the ultrasonic source and one scattering forwards. Very little scattering to the side.

In the ideal Rayleigh scattering situation, e.g. a hard reflector immersed in a liquid, the minimum detectability is still an open question and I have references suggesting wavelength/2, wavelength/8, and wavelength*Pi/2. For a fixed size reflector there is a dependence on wavelength, on the angle between the transmitting and the receiving transducer, and the signal to noise ratio of the ultrasonic system.

Scatting theory for an ideal situation is useful to know when you work in sonar or ultrasonic particle detection and fields like that. In the real world of NDT, the scatterer of interest (i.e. the small defect of flaw) is usually embedded in a matrix of other scatterers. So, the theory of scattering does not just apply to the small scatterer you are trying to locate, but also the grain boundaries around it, the aggregate in concrete, the fibres in a composite, etc. The scattered waves from your reflector may be scattered again by other nearby boundaries, i.e. multiple scattering.

It is very difficult to adapt the theory of Rayleigh scattering to real-world NDT, and there is probably very little point in attempting to do so.

Those people who have answered your question with NDT oriented examples considering attenuation, wavelength, etc, have given you good advice. The practical minimum detectability is going to depend on the particular inspection you are attempting and you are going to have to do some investigation yourself with probably some extrapolation of your results. You will come up with some sort of estimate of minimum detectability with some statistical measure of how reliable your estimate is.

John Norman

1    
 
 Reply 
 
Anmol Birring
Consultant,
Birring NDE Center, Inc., USA, Joined Aug 2011, 738

Anmol Birring

Consultant,
Birring NDE Center, Inc.,
USA,
Joined Aug 2011
738
03:45 Jun-07-2016
Re: Detectability of pulse-echo UT method
In Reply to John Norman at 03:04 Jun-07-2016 .

Clarification on Rayleigh Scattering

Rayleigh scattering is applicable for (lλ/d) > 10. For a 1 mm defect, this would mean a λ=10 mm or a frequency of 590 KHz. This not the frequency range used in ultrasonic testing and therefore not applicable to flaw detectibilty.

    
 
 Reply 
 
John Norman
Consultant, owner of business
NTS Ultrasonics Pty Ltd, Australia, Joined Oct 2012, 111

John Norman

Consultant, owner of business
NTS Ultrasonics Pty Ltd,
Australia,
Joined Oct 2012
111
05:20 Jun-07-2016
Re: Detectability of pulse-echo UT method
In Reply to Anmol Birring at 03:45 Jun-07-2016 .

Hi Anmol.

For most discussions of practical NDT, you are correct, but there is nothing special about ultrasonic NDT that makes the basic idea of Rayleigh scattering invalid. For a smaller defect, use a shorter wavelength.

This discussion is very relevant to NDT. In ultrasonic rail flaw detection, for example, there are contractors with automated inspection equipment who charge a premium to conduct a "cleanliness survey" of a rail system. The idea being that they take "high resolution" ultrasonic measurements to find microscopic inclusions that may lead to defect initiation. Many of these inclusions will be less than 1 mm equivalent diameter. Such surveys use ultrasonic frequencies in the 2 MHz to 5 MHz range.

Rayleigh scattering is discussed in plenty of ultrasonic NDT text books including Krautkramer's "Ultrasonic Testing of Materials", "Ultrasonic Testing" edited by J. Szilard, and although it is not called Rayleigh scattering, it is discussed in "Ultrasonic Waves in Solid Media" by Joseph Rose. These discussions are generally about attenuation and Rayleigh scattering is discussed in the context of multi-scattering from gain boundaries, etc. However, nobody suggests that scattering theory is not relevant to ultrasonic NDT.

John Norman

    
 
 Reply 
 
Wieslaw Bicz
Engineering,
PBP Optel sp. z o.o., Poland, Joined Feb 2009, 256

Wieslaw Bicz

Engineering,
PBP Optel sp. z o.o.,
Poland,
Joined Feb 2009
256
14:54 Jun-07-2016
Re: Detectability of pulse-echo UT method
In Reply to Heydar Alakbarov at 21:03 Jun-02-2016 (Opening).

The actual discussion (and the previous one) is caused by the wrong, but popular among NDT people assumption, that detectability has something to do with wave length. But in reality it has almost nothing to do with it - it is a myth. Resolution (ability to differentiate between two adjacent points) has to do with wavelength, and may be this is the reason for such wrong assumption.

Detectability has to do with the amount of sound, that is going back from the object that should be detected to the receiver and the ability of receiver to differentiate between the signal and noise. Additionally the acoustic signal generated by the surroundings of the object, that should be detected is also important - if it is too large, for example because of the grain in the material, it will make the differentiation between the useful signal and grain scattering impossible.

Exactly speaking, the detectability has to do with the intensity of sound at the position of the object, that should be detected (i.e. with the intensity generated by the sender), with numerical aperture and sensitivity of the receiving transducer, sensitivity of electronics used (it has also to do with the signal detection method - the use of lock-in amplifier and bursts allows for example to detect signal, that is more than 130dB lower than noise) and naturally with the size, shape and impedance of the reflector (object, that should be detected). Wavelength could play some role, but since it is easily possible to demonstrate, that objects, that are much smaller than wave length can be detected - also objects that are scattering in all directions, the inclusion of wave length do not have any influence.

My comment about Rayleigh scattering, made in previous discussion was made only because somebody wrote, that small objects will not reflect sound (wave should "bounce around them"??!!), and this statement is naturally wrong.

Separate question is naturally, how important is this for UT "in the field"? From my point of view it is important, that people using ultrasonic methods are aware of the facts, described above and not believing, that wavelength has something to do with detectability, because such "knowledge" is misleading. Everybody, who has focusing transducer and standard equipment can easily make some tests, showing, that it is possible to detect objects, that are significantly smaller than wave length. It is more difficult with plane wave transducers, since they have very small numerical aperture, but not impossible.

I hope, that this clarification will answer all important questions concerning the detectability.

1    
 
 Reply 
 
Anmol Birring
Consultant,
Birring NDE Center, Inc., USA, Joined Aug 2011, 738

Anmol Birring

Consultant,
Birring NDE Center, Inc.,
USA,
Joined Aug 2011
738
16:28 Jun-07-2016
Re: Detectability of pulse-echo UT method
In Reply to Wieslaw Bicz at 14:54 Jun-07-2016 .

Diffraction from the flaw is controlled by wavelength when the flaw size is in the wavelength size range , so wavelength is one of the important factors.

    
 
 Reply 
 
Wieslaw Bicz
Engineering,
PBP Optel sp. z o.o., Poland, Joined Feb 2009, 256

Wieslaw Bicz

Engineering,
PBP Optel sp. z o.o.,
Poland,
Joined Feb 2009
256
17:13 Jun-07-2016
Re: Detectability of pulse-echo UT method
In Reply to Anmol Birring at 16:28 Jun-07-2016 .

This statement is true, but only for the case, if the size of the reflector is comparable with the wavelength. In this case you can talk about diffraction. In the case of smaller reflectors, that are scattering in all directions (in half space or more - when no classical diffraction can be considered) only following factors are playing the role:
1. Intensity of sound at the position of reflector;
2. Numerical aperture of the receiver (seen from the point of view of the reflector);
3. Sensitivity of receiver (transducer and electronics);
4. Size of the reflector.

I am repeating all this, because it is clearly visible, that the myth about the connection between the wavelength and detectability is alive and misleading.

It is really very simple to make some tests, showing, that this myth is only a myth and nothing else.

    
 
 Reply 
 

Product Spotlight

FD800 Bench Top Flaw Detectors

The bench-top FD800 flaw detector range combines state-of-the-art flaw detection with advanced mater
...
ial thickness capabilities. Designed for use in the laboratory these gauges are the tool you need for all your flaw detecting needs.
>

Echomac® Small

Available with up to eight channels of electronics to detect and evaluate thickness, flaws and eccen
...
tricity, this UT tester is housed in a convenient, smaller cabinet. This instrument can be used in conjunction with bubbler or immersion tank systems, or with a test bench or in laboratory applications. The Echomac® Small is available in the FD4, FD6 or FD6A versions.
>

Robotic laser shearography enables 100% inspection of complex, flight-critical composite structures

An article in “Composites World Magazine” showcases Non Destructive Testing of aero-structures
...
with Laser Shearography. Over the years Dantec Dynamics has supplied many solutions for the aerospace industry. Referring to specific customer projects several of these cases are examined to outline the advantages of using Laser Shearography for automated defect detection.
>

High-end Ultrasonic Flaw Detector with 32:128PR PAUT and 2-ch TOFD: SyncScan 2

SIUI’s newly launched SyncScan 2, is a high-end ultrasonic flaw detector with 32:128PR PAUT and
...
2-ch TOFD, which can maximize your efficiency for PA and TOFD. ● Support PA/TOFD/UT ● 32-ch PA is more suitable for inspection on extra-thick wall and high-attenuation material. ● 32-ch PA and 2-ch TOFD work simultaneously. ● Support PR mode for corrosion inspection.
>

Share...
We use technical and analytics cookies to ensure that we will give you the best experience of our website - More Info
Accept
top
this is debug window