Table of Contents ECNDT '98 Session: Railway Copenhagen 26 - 29 May 1998 Hear Mr. Schneider live in Copenhagen (This is a 20 minutes speech as a streaming Real Audio file. Download the Player) Study the viewgraphs while listening Mr. Schneider's speech

Real Audio Presentation: Ultrasonic Evaluation of Stresses in the Rims of Railroad Wheels Paper of the proceedings E. Schneider *, R. Herzer Fraunhofer Institut Zerstörungsfreie Prüfverfahren (IZFP) Saarbrücken, Germany * Corresponding Author: Eckhardt Schneider Head of Department Process integrated Nondestructive Testing at Fraunhofer Institut Zerstörungsfreie Prüfverfahren IZFP, University Bldg 37, D-66123 Saarbrücken Phone: +49 681 302 3840 Fax: +49 681 39580 Email: schneider@izfp.fhg.de Homepage: http://www.izfp.fhg.de/english/ NDTnet: Q Net Stand

I am afraid that I will not be able to give you a total different topic - we are talking about the same thing (previous speaker).

The fact is that since the passenger trains are running faster and faster, there is also a strong need that the cargo trains which needs to thread into those schedules of the passenger trains have to go with higher speeds as well. As well as there have to carry higher loads, hence the breaking energy is getting higher. That again is introducing some changes of the residual stress in the wheels (as already explained by the previous speaker).

Fig.1: Circumferential stress [MPa] and stress profiles in a wheel, heavily braked in a braking test stand (left part) and in a wheel used in traffic (right part). European Railroad Research Institute: Document ORE B169/RP2, Utrecht (1989).

Session:

This for example is a cross section of the upper part of a wheel, of the rim of one of those wheels. The axis would be here about, this is the area where the contact with the rail is happening. The numbers you can see in here are the numbers in Mega Pascals - their describing the stress in the hub direction. Those are tensile stresses which like to make the diameter larger - they would like to open the wheel. But in this particular case those numbers are extremely high. That is not the usual situation, that has been achieved by a very drastically breaking sequence, in the breaking test stand. Point is that in this area, as well as in this outer edge, the higher stresses are always measured, that means from the fracture mechanical point of view everybody would like to get a very precise information about the stress state in the running surface as well as in this corner area. So the fasten demand was to find a nondestructive tool to evaluate the surface stresses. Surface stresses in so far important as you may know as there are always small little tiny cracks, they are always there. They are result of the very heavy interaction between this surface of the wheel and the corresponding surface of the rail so that you always have those cracks.

The point is now to make sure that those cracks are not in a field between high tensile stress. So that those cracks which are not harmful, otherwise may not growth, or may not course a break or a break down of the wheel. It is not possible by no nondestructive tool to evaluate the stress states in those surface near area under industrial conditions, hence we need it to proof that the stress which are changed by the breaking conditions in this surface area find somehow corresponding changes in the part of this volume.

This has been done in cooperation with the German railroad (the DB AG) and what you can see here is more or less the final result already. What we are doing here is pretty close to the presentation of the previous speaker, we going to use ultrasonic waves which are started at a certain distance from this level and the distance is organized in such a way that the outer rays of the ultrasonic beam are just covering this very highly interesting area. And than the automated measurement is moving this transducer step by step, millimeters of bits, to this lowest position which is find out by the system by itself because that is the area where the ray is just catching this corner here.

Now there are two major area of applications for a system like this. One is to optimize the production, that is now showing here. This results are found in new wheel. In a new wheel you see that the residual stress profile in a slight compressive area. I do not agree with the results of my previous speaker who said that the stress in the radial direction is zero. Our results shown that the stress in the radial direction is not zero, but the radial is not affected by breaking sequences. So you have already a radial stress in the very beginning of the living of the wheel, but this radial is not influence by breaking. Only the hub stress of the tangential is influence by breaking. Hence we like to show that what is the outcome of the birefringence defects of the birefringence theory which has been has been explained by the previous speaker as well. We prefer to show the difference between the stress in the hub minus the stress in the radial direction.

Since this stress is not very influence by the breaking this gives you mainly the change of the stress in the hub direction. But on the other side as I will explain it later again this difference of the principles stresses they are driving forces of the cracks. Hence the fracture mechanical people they are much more happy to get this information because than they can calculate the ..... whether or not a crack which is already there may growth under this load or not. What you can see in here, the results in a new wheel, and you can see how this stress level changes after this different breaking sequences. Again, those measurements are made in a test stand, this is not a usual breaking load but you can see that there is a very consequent increase of the stresses. Some other suppliers thought if this is the case I might do some additional heat treating in order to start with a stress profile which is higher in compressive. So that this profile of this compressive stress needed to be broken up at first before than the dangerous tensile stress may growth up.

So we found in another supplier's wheel results like this, there is certainly a much higher compressive stress in this wheel. But as you can see a little later the same thing or after the second of breaking we are ending up in the same region so that the more expensive wheel so that of this sort is really not paying of. As you can see in a little while, you have it after the 200 stop breaking, you are in the same area of tensile stresses as in other cases.

Hence the German railroad was very happy to have a system of this sort, in order to make those test on those wheelsets their had just bought from the suppliers, in order to see so that the stress profiles of those new wheelset are in a very acceptable range.

Couple of those measurements had been made in the measurement test stand and this now a series of measurement which was done on real live. You see that the maximum stress is by far not as high as it used to be in previous transparences. Three brand new wheelssets have been mounted on a train, this train was moved along the usual track. But a breaking program was applied as it usually applied for a cargo train going downhill at the southern side of the St. Gotthard tunnel.

You can see that just this one breaking sequence, it takes about roughly two hours, this breaking sequence is changing the stress profile from the very new wheel into this profile. In case that all wheels are contributing to the same amount to the total breaking energy of the train.

There is no danger for this wheel because the stresses do usual not being higher than 300 MPa. Only in case that for some reason this breaking load is only carried by a few wheels, by a few axis, may be the compress of the breaking system is not working equally on all those cargo wagons, even in case that one of those wheels have to carry a higher load, then that shift to very high residual stresses.

In cooperation with the fracture mechanic institute, in cooperation with the German Railroad DB there is a limit value evaluated, this was evaluated being 300 MPa for certain type of wheel sets. As long as the residual stress is lower than that, there is no danger, even larger cracks are tough enough so that this stress is not enough in order to make them growing. All those measurements all those transparency has been made in 1992. Since January 1993 at first 5 then another 5 up to now 25 systems of this sort or the previous type has been delivered to the German, to the Swiss, to the Austrian, to the Spain railroad, as well as to different manufacturers.

The system consist of manipulator (its hardly to bee seen - please excuse that) but just belive me there is a little transducer inside which is moved automatically from the upper to the lover position along this radial trace. We are making measurements according to the same approach as has been described at each millimeter of depth in order to have a profile like that. This is from another result, one of the result which is shown to the operator which is stored to the computer base and which is also described in the certificate which is automatically printed by the software. I am showing you that, because I would like to point out that, it is very important to do this continues measurement of the stress in the depth direction.
There are other devices where are only one or two points in this depth position are measured, this looks to me very dangerous. For example, see this particular result, it is a very nice stress profile for a new wheel. After some month in use this stress profile changed a bit, you see in a the very surface near area, you can imagine that energy is coming from the side, the change of the stresses is of course coming from the running surface going to the depth. Here you see already a change into the tensile direction which is compensated lower here. Another period of time later the wheelset has a profile like this, you see that the whole thing is now already shifted into the tensile direction with very high tensile stresses in the surface nearest area.

A little later if this would go up then this wheel would not be allowed to be used again, since the critical value of 300 MPa is reached, and that is indeed a true decision because this is the very surface near area where we have problems with cracks. On the other side see this wheel, this wheels already has after some months of use has already overcome this threshold, so that this wheel is not allowed to be used again, but look the stress is here in the area of about 15 mm since deeper in the surface near areas, because of the plastic deformations which has been enhanced, which has been facilitated by the heat, there was plastic deformation reducing the stress in the surface near area. So by my very earnest opinion I rather would like to go with a train having this kind of stress profile since the cracks in here are in the stress environment which is very slow compared with the same cargos .... , where we have now a wheelset, where we have significantly higher stress profile in the surface near area.

Right now the decision are only made by this profile, but we are cooperating with the German Bahn and we are suggesting an additional measurement, the measurement of the crack depth. Meaning if there is the possibility to evaluate the actual crack depth of the deepest crack than you have the two real important information in order to use the fracture mechanical approach saying, with this crack length, with that high stresses, than we need to move this wheelset out of service. Much more precise, much more reliable, because that takes the real material situation in account and not just only the stress profile.

All this are pretty much the same as already described by the previous speaker. We have no problem texture, as long as we are using the new wheelsets. The new materials are 7... and better, they are forged their annealed so that the texture influence is indeed neglectable. Within the accuracy of about +- 20 MPa the texture influence is such a small amount in this inaccuracy that it is really neglectable.

That is not the case in older wheels. Unfortunately some older wheels are still in use and those older wheels have two problems. In one case we very often cannot get any echo, because the course grains as well as the porosities in the material are causing such a high ultrasonic attenuation that the received amplitude is very very small. Hence with the same measurement, I just described, we measure together with the time of flight also the attenuation, we just measure the amplitude height of the received echo.

Fig. 3: Change of ultrasonic signals amplitude with the distance from the running surface measured using a wheel of a new type (solid line) and of an older (dashed line) type.	Fig. 4: Change of the ultrasonic signals amplitude with the distance from the running surface measured using a wheel of an older type.

Now this again the distance from the surface, and this is now not stress, and not time of flight, this is the ultrasonic amplitude of the received echo. What you can see is here - a very nice strong echo, a decay in that area where we have this little notch on the outside, a little mark, which is going around the circumference and just under knees of that again a very high amplitude. That is a typical amplitude profile for new wheelset, prime for wheelset with a new material, with a new technology. That is an older type of wheel which is not yet manufactured but this kind of wheels are still in use. And this in particular is again a realistic wheel, lots of wheels of this type are still running since they have been used for a long while and since a lots of specially equipped cargo wagons do have this kind of wheels.

What the systems now knows is - it measures the amplitude, and as soon as the amplitude is above a threshold the systems measure then in those areas were the amplitude is over this threshold. Meaning in this particular case the system would measure from about 8 mm until to about 55 mm.

In this wheel our system is only able to measure from about lets say 9 mm to this position, it cannot measure in between here, and here is another result for that. In this wheel there is really no measurement possible; in this particular case the software says there is a wheel with a microstructure anomaly meaning there is something hidden in the material: there is flaw, there is ... , there is some reason in the microstructure which is eating those ultrasonic energy which is somehow not good.

In other wheels we have texture and this texture as detected by measuring the ultrasonic amplitude and the time of flight with the change of the polarization of the shear wave. This is a typical result of a very heavily textured material. The physically explanation goes along that arguments that texture is very often also go along with the fact that grains are extended, are longer, in one direction than their are in other. Hence the ultrasonic scattering is frequency dependant but also dependant on this relationship of wavelength and grain size. This is a result in a material without any texture.

Fig. 5: Ultrasonic amplitude versus the change of shear wave polarization from the radial (0°) into the circumferential (90°) direction as measured using a wheel with a strongly developed texture (left part) and using a wheel with no or a slightly developed texture (right part).

A study based on about 100 wheels that confirmed this program and we have now made some threshold approximations in order to tell you wether the texture is very strongly dependant, then we cannot do any stress measurement, or wether the texture is present, but is below a certain threshold, than we can evaluate the stress, but the inaccuracy is not +- 20 MPa, but about +- 50 Mpa.

I would like to summarize with a few advantages of the system I just described. This system evaluates the stress state and stress variations. It is like the system of the previous speaker a very easy handling system. And the point is that this system is, at least to our knowledge, the only system which is absolutely totally automated. The only thing the operator needs to do is, he needs to fix the manipulator on the wheelset, then the measurements are doing everything by themselves, also the last step is done, writing the document with the stress profile, wheelset number, and things like that.

That concludes my talks and I thank you very much. |Top|