Point A-------------------------------------------------------------------------Point B
When we inspect the shaft using 0 Degree probe from point A to point B we received a back signal from 940mm and when inspecting the same from point B to point A we received back signal from 573mm.
The same back signal appears for all axles and the height of the signal is different each time although the calibration is same.
In actual or theoretical fact, we were supposed to receive the back signal at 1745mm and not as per case above. By this it safe to assume that the signal never reached the other end of the shaft when we send the signal from one side.
The inspection was performed at scan 360 degree circumference around the shaft using 0 Degree probe and a 45 degree probes at the back signal location (940mm and 573mm) respectively where we did not find any indication. We have few questions:
1. Why we received back signal from 940mm, 573mm and not from 1745mm? Is it because of the material composition or material properties problem? Is it because of the forging problems?
2. Would it because of the characteristic of UT itself that doesnt reach the other point of axle?
3. If there is any indication, its should be appear when we scan at the 940mm and 573mm locations at 360 Degree using 0 Degree and 45 degree probes. Any idea on why these indications did not appear?
Thanks. Please give your opinion and advices. Your co-operation much appreciated.
I can assure you that this is not the first time such observations have been made in the testing of railway axles , particularly while testing in the axial direction. Normally some beam bending etc are involved in this problem. But certain information is essential before we make a proper assessment of the observed phenomena.
Usually these type of observations are made while testing axles that are in service.
Please confirm whether these are in-service or new axles (If these are new axles then I shall be asking a whole lot of other information later)
If they are in service please the give the approximate period of running.
Please give the bearing positions, brake positions (widths also), wheel positions. Please also indicate if there any other steps.
Please give the position of the transducer on the side face of the axle when you are picking up these echoes. Also what is the result when you keep the transducer at the centre of the side face?
Please give the gain being used and the position of the reject control when you are getting these echoes. Whether there are any other small echoes simultaneously appearing.
As you are aware that the testing in axial direction of the axle is done to detect transverse cracks.
You can definitely get the back echo from the other end of the axle (1745) if you place the transducer in the middle of the side face and increase the gain sufficiently. (assuming that the flaw detector can pump sufficient pulse energy.) But this may not serve the intended purpose of the axial testing.
Usually it is enough if it can be proved that you are picking up transverse notch of a given depth as per your specification, say about 5 mm depth, at the centre of the axle. The testing is done from both ends.
On the other hand if one is setting the gain to receive the back echo from the other end of the axle with 2 MHz transducer, please be assured that you will be getting an array of sundry echoes from every where, but much before the back echo and also some after the back echo. The picture depends on the type of flaw detector the probe damping characteristics, the reject setting etc.
Please provide the information that I requested above so that we can understand the mysterious echoes.
this is a tricky problem but my first instinct is that you have a planar crack in the centre of your forging. This assumes that you have eliminated PRF and that this is plain forged bar.
The fact that you cannot detect it using a single shear wave probe indicates that it does not extend to the surface and and the 360 test using the 0 degree probe would indicate that it has no reflective component in that direction (it is at 90 degree to the longitudinal axis of the bar). You may want to see if there is a slight drop in back wall echo in the region when scanning along the bar. The only way to confirm a central planar crack (other than the axial test) is to use 2 45 degree shear wave probes either by through testing facing each other on opposite sides of the bar and looking for a loss of signal, or a tandem test with both probes in line facing in the same direction approximately a bar diameter apart and looking for a return signal.
Please let us know what your final conclusions are.
I think the echo raising at 940 mm could be explained by sound beam edge wave mode change at axle surface.
Longitudinal alpha of 78 degrees (divergence 12 deg) on axle surface would lead to transversal beta of 32,5 deg. Theoretical angle of divergence on your probe is 10.4 degrees. Further on these values lead to sound paths 497(L) + 243(T).
On screen you'll see the sound path per longitudinal velocity, so transversal soundpath must be corrected to meet longitudinal velocity. This leads close to the sound path you have mentioned.
The 573 mm echo can not be explained same way.
1. is the forging an cilinder constant diameter 205 mm, also roughly shaped?
2. are the extremity surface planar and perpendicular to the forging axis?
3. are any wheel or flange mounted with interference on O.D.?
More detailed dimensional drawings would be needed to confirm, But It seems fairly certain that these signals originate from reflections/ Possible mode conversions at the inner edge of the two raised sections.
A larger diameter probe would give a narrower beam and thus less signal from these areas.
This might make it possible to confirm what is happening
All the dimensional drawings could be downloaded from ndt forum website.
I tend to believe yor advice however I have some doubts. Could you explain why these reflections/ Possible mode conversions at the inner edge of the two raised sections appear before back wall. These refections or mode converted sound waves have to travel more distance than the stright beam from point A to point B so they ,ust appear after back wall isn't it?
The sound will not have to go as far as the far end of the axle, It can be reflected back towards the sending end - either directly (the step near end B looks sharp enough to act as a corner reflector) or by multiple reflections from different surfaces.
In general with Rail axles you will always get these 'artefacts' which need to be be recognized when you are testing. Modern flaw detectors usually have a 'reference image' facility, so that the echoes from a 'standard axle' can be shown in a different colour from the live data. In the old days they used to draw on the screen with a chinagraph pencil.
I am fully agree with Joe. This is a typical situation in the axial inspection of axis. Discover the travel of the sound can be a funny play, but it is more important the exactly registration of the screen, to compare the ecos in the future inservice inspection.
Attention: the ecos situation can change after the montage of the wheels!
I agree with Joe and Ezio. I believe you problem to be mode converted signals from the journals making it back to the transducer.
I do have a question as to why different results are obtained from Point A and Point B. Are these in service or new axles? Are the ends drilled for end caps or only center drilled?
If they are drilled for end caps, this may explain why the difference in signal, ie different locations could pick up different mode converted signals.
Lastly, what code are you working to? AAR? GOST? other? We have quite a bit of experience with working for an axle manufacturer here in the states who works with most of the codes in the world. Your problem is not an uncommon one. As stated previously, you must determine the location of the nonrelevant signals and recognize them as such then inspect the parts with that knowledge.
Please could you advise me which flaw detector could show the echoes in different colours? epoch - III, IV?
We have been asked to develop an operating procedure and acceptance criteria by our client as we (both client and us) are not aware of any existing standard. Is it possible to get a copy of these AAR & Gost standards?
Finally, I would like to ask you gentlemen; whether there is any new technology available to carry out axle inspection? I have come across "shaftest" tool (developed by SGS Australia) and the inspection results looks like a b/c scan to me. Appreciate your advice on any new technology like shaftest so that we can purchase the equipment
In addition to above, I have now checked the echo locations with 45 & 60 degree probes in tandem and through positions (as suggested by Peter earlier) and I could not find any discontinuities! Based on the above I presume that these echos are nonrelevant indications
The standards I referred to are developed by the individual countries railroad oversight agencies (public or Private). AAR is the American Association of Railroads and they are an industry body who polices the rail industry in the USA. Department of Transportation (DOT) is the gov't oversight here in the US.
There should be a government oversight or private organization in your country who looks over the rail industry. You should look to them for your specification. If not, let me know and we will help, if possible.
Training and Consulting Services
Quality Testing Services, Inc.
Dear Jaga, we have in Italy many people who use multichannel UT for inspection of rail axles. You can contact my friend Sarti (email@example.com).
You can find in the ndt.net database many articles on rail axel inspection, for exemple from Italy, of prof Beretta and Mr Cantini.
See also Workshop "Damage Tolerance of Railway Axles" organised in Milan October 13-14, 2008 from ESIS TC24 (Structural Integrity of railway Components", chaiman Prof Beretta (firstname.lastname@example.org)
I have tried to view the dimensions on the drawings but they are illegible. This makes it very difficuly to try and track where these reflected signals have come from. Try using triangulation in your thinking......... subtract the diameter from your reading and see where that puts you on the axle.
The ShafTest was developed by CCI Pope now Bureau Veritas Australia.
You ask many questions which lead to many more.
Now regarding colour echo displays. Do you want to have the capability to change the colour of the echoes of the whole display? Or would you like the feature that the colour of the echoes will change due to beam path length?
You may find some railway axle inspection techniques under the following references:
NDT.net - February 2001, Vol. 6 No. 2
Recent Advances in Ultrasonic Inspection of Railway -Axles and - Wheels
H. Wüstenberg, A. Erhard, N. Bertus, T. Hauser (BAM - Berlin)
H. Hintze, M. Schüßler (DB AG Kirchmöser )
Corresponding Author Contact:
http://www.aaende.org.ar/sitio/biblioteca/material/T-076.pdf June 2003
Ultrasonic testing of railway axles with phased array technique
Wolfgang Hansen ( Agfa NDT GmbH, Robert-Bosch-Str. 3, D-50354 Hürth, Germany, now a GE company)
Hartmut Hintze (Deutsche Bahn AG, Am Südtor, D-14774 Brandenburg-Kirchmöser, Germany)
Hermann Schubert(Agfa Gevaert do Brasil Ltda, R. Alexandre Dumas 1711, 3. andar - São Paulo SP, Brazi, now probably a GE companyl)
An improvement for the classical problem of structure and geometry based echo indications at the axle inspection had been one of the reasons to apply a phased array approach in combination with a C-scan like presentation of the results as described by the references.
"Please could you advise me which flaw detector could show the echoes in different colours? epoch - III, IV? "
With respect to Mr. Jaga's question about Panametrics-NDT Epoch series flaw detectors, the older models that you mention (Epoch III, Epoch 4) do not have the feature in question, but newer models Epoch 4 PLUS, Epoch XT, and Epoch 1000 do. The "Peak Hold" function allows you to save a reference waveform on the screen, with the live waveform superimposed in a different color for comparison.
The Epoch LTC is a current model with a color display, but it is our simplest and least expensive flaw detector and I'm afraid that it does not offer certain features found on some other instruments, including the Peak Hold function.
I would also tend to agree with Joe's conclusions on this one. We ave done many axles on large mining vehicles and have found that mode conversion is the most common false signal found and we have had to really develop procedures with that awareness.
It is primarily caused at the angled transition points or areas of case hardening or dissimilar grain structure. An easy way to verify if your sound is going where you want it is to have someone assist and move their hand along the shaft with oil or couplant. Their hand should cause an amplitude drop in the region of your soundpath bounces. This helps in establishing if your skip is where you expect it to be, and you can even check how many skips are occurring in your piece. I've found that it often shows a mode conversion causing skip location that we didn't take into account initially. Hope this might help.