·Home ·Table of Contents ·General

Ultrasonic Testing of Axle sets of Diesel-Engine trains Janez Grum Faculty of Mechanical Engineering University of Ljubljana, Aškerceva 6, 1000 Ljubljana, Slovenia Viktor Jemec Secondary School Domžale, Cesta talcev 12, 1230 Domzale, Slovenia Contact

ABSTRACT

The railway traffic requires safety and reliability of service of all railway vehicles. Suitable technical systems and working methods adapted to it, which meet the requirements on safety and good order of traffic, should be maintained. For detection of defects, non-destructive testing methods - which should be quick, reliable and cost-effective - are most often used. Inspection of characteristic parts is carried out periodically in accordance with internal standards or regulations; inspections may be both regular and extraordinary; the latter should be carried out after collisions, derailment or grazing of railway vehicles.
Maintenance of railway vehicles is scheduled in accordance with periodic inspections and regular repairs. Inspections and repairs are prescribed according to the criteria of operational life, limited by the time of operation of a locomotive in traffic or according to the criteria of operational life including the path travelled.
The present paper treats testing of axles in wheel and axle sets of diesel-engine trains; attention is devoted to theoretical as well as practical work. A suitable perfected experimental system for the ultrasonic testing of axles of diesel - engine trains was elaborated. To this end we had at our disposal unserviceable axles of wheel and disassembled axle sets, where at individual critical points it was possible to simulate defects at various depths. On the basis of the defects, the applicability of individual ultrasonic probes was assessed. Testing was thus successfully carried out with each of the selected ultrasonic probes at the shorter and longer side of the gear wheel of the driving axle from different fields, and with the apparatus setting at 1.5 and 2.5 m. Each of the variants selected was recorded for subsequent analysis of the results. Finally, based on the procedures selected, geometries of the known types of ultrasonic paths were calculated for a comparison with the results of the experimental work.
Keywords: Ultrasonic testing, Experimental system, Axle - testing, Diesel engine trains, Screen display

1. INTRODUCTION

The most heavily loaded parts of railway vehicles are the axles in the wheel and axle sets. Their damage or breakdown would, in fact, entail considerable material losses and endanger human lives. Consequently, the former Yugoslav Railways had already issued an internal standard [1] which specifies a compulsory application of ultrasonic testing to the axle sets. This standard was amended by the former Railway Administration Ljubljana which issued the relevant instructions [2]. Maintenance of railway vehicles is scheduled in accordance with periodic inspections and regular repairs. Inspections and repairs are prescribed according to the criteria of operational life, limited by the time of operation of a locomotive in traffic or according to the criteria of operational life including the path travelled.
Slovenian Railways have a set of railcars of series 813/814, which was first manufactured by the Italian firm Fiat, and later in co-operation with DO TVT "Boris Kidric" from Maribor. The vehicles required the introduction of a systematic inspection of individual assemblies and/or parts. Already in 1976, by means of non-destructive testing the first cracks were found on the axle sets. The finding that the axles in the wheel and axle sets belonged to the most heavily loaded parts of railway vehicles led to their systematic inspection. The data available on the state of the axle sets indicated that there were numerous defects located at the transition points from wheel hub edge to axle axis.
The following kinds of damages have been found on 152 axles after 10 years of service of the axle sets:
• defects at the inner edge of the transition from wheel hub to axle: 75.6 %,
• defects at the transition from gear wheel to axle: 24.4 %.
In addition to the relevant proportion of damages, the operational life of the axle sets also needs to be established. Consequently, data on axle quality and display of the frequencies of the operational life provide very important information. Generally, it may be stated that the tendency towards increase in the number of damaged, i.e. discarded, axles is within expectations, and that numerous axles of high quality since they can last 2 to 3 times the guaranteed operational life. Unfortunately, it has to be stated that much more attention should be paid to causes of damages and discarding of the axles before their operational life of 4 years has elapsed. It is supposed that the material quality is low, which would require elaboration of the relevant regulations on material quality before being released to production; therefore, only some manufacturers perform testing of the axles in the wheel and axle sets with the least favorable cyclic fatigue loading.
In order to improve safety in railway traffic, the Service for Defectoscopy of Slovenian Railways in Ljubljana has set itself a task of improving the reliability of testing and ensuring earlier detection of defects in the axles. To this end, several years ago, the first systematic analyses were made, which permitted the elaboration of instructions for ultrasonic inspection. It should also be mentioned that the inspection method applied to the front, using normal probes, is rather unreliable since at the transitions - points there are various reflections and transformations in the form of wave - motion. Three questions seem to need answers, i.e.:
1. Is it possible to detect defects at all critical points by transmission from the same side of the axle?
2. At which size - depth of the defect can appropriate identification be achieved?
3. Which of the normal ultrasonic probes is suitable for efficient identification of defects?
For ultrasonic testing, three driving axles and one running axle were selected. With them three different sizes of artificial defects, which were transverse notches on the cylindrical part of the axles, were simulated. The selected areas with the notches were those characteristic areas where cracks normally appear. These areas were, therefore, called critical areas.

2. EXPERIMENTAL SYSTEM

A suitably perfected experimental system for the ultrasonic testing of axles of railway vehicles was elaborated. The conventional non-destructive testing has encompassed the examination of railway engines and vehicles with only standard ultrasonic devices having standard ultrasonic probes [3]. Ultrasonic testing operators must be qualified to correctly evaluate the echograms at the cathode tube of the ultrasonic device. It is also expected of the operator to have a good theoretical knowledge, so that he can successfully compare the signals calculated from the echograms [4]. When using a standard ultrasonic device, the operator is facing a demanding task of evaluating the state and usability of the machine element, i.e., test specimen, even though at the final stage he does not have any document or echogram available which would confirm his decision taken when testing a material. A photograph of the screen display or an impression of the cracked surface can serve as evidence for a crack detected at the axle surface or after an additional penetrant testing.

The process of ultrasonic testing and also evaluating is a distinctively subjective process of testing itself as well as of decision-making or evaluating of the test specimen. The quality of the state evaluation of the test specimen depends on the operator, particularly on his knowledge and qualification as well as caution, accuracy and responsibility in performing the ultrasonic examination. With the intent to exclude subjective influences in testing the railway-vehicle axles, a corresponding experimental system was adapted, which is capable of saving echograms to floppy discs and enables further computer processing [5, 6]. The system proposed permits documenting of testing results, the making of different reviews and analyses over a short or long period of time. Such a systematic approach to ultrasonic testing also enables the development and perfection of the procedure in such a way that the axles may be tested faster and with a greater quality.
The experimental system for non-destructive testing of railway-vehicle axles consists of the following components:

• a defectoscope of the home firm ISKRA, which has been later substituted for defectoscope USL-32 of the firm Krautkrämer;
• an A/D converter of ultrasonic signals;
• a personal computer and a printer.

Fig 1: Experimental system for non-destructive testing of railway-vehicle axles.

Signal capturing is performed manually with normal, 2 MHz or 4 MHz ultrasonic probes, as well as with special ultrasonic probes with different angles for testing from the axle front. Axle front testing is used in the cases when various machine elements attached to the axle limit direct contact of the ultrasonic probe with the axle.
Consequently, for axle testing the standard ultrasonic probes which were modified in such a way that the angle of incidence of the ultrasonic wave could be varied were chosen. With a computer program elaborated for this purpose [7, 8], theoretical calculations of ultrasonic signals on the basis of input data can be performed and a necessary printout of signals for individual, i.e., chosen, axles of railway engines can be made. Theoretical signals are then compared to the signals captured on the new, flawless axles. The theoretical signals can be determined on the basis of a corresponding representation of the shape and size of individual axles and the physical model for the calculation of the path of ultrasonic waves between the reflections in the test specimen. The program is made for general use, which means that the size and individual characteristic shapes of axles, i.e., machine elements, to be tested must be input.
In view of modern computer and processing technology, it is possible to complement the experimental system by connecting the standard ultrasonic device to a computer, which, with an adapted software program, permits a quick, real-time calculation of ultrasonic paths. In order to do that, a computer program should include the following:

• a selection of corresponding data regarding the size of the axle, i.e., machine element;
• a selection of corresponding data regarding the shape of the axle, i.e., machine element;
• location of the probe at the test specimen;
• ultrasonic device measurement range settings;
• recorded echogram for the given axle, i.e., machine element;
• computer-aided verification of the theoretical and actual ultrasonic signals;
• decision about the type and size of defects;
• capability to save echograms and other testing data;
• statistical processing of data and making of recommendations;
• perfection of testing method.
The computer aids in the calculation of theoretical signals in the given machine element to be tested with the selected ultrasonic probes. The second step is the comparison of the actual and theoretical signals, which allows the prediction of the state of the tested machine element. All theoretical signals as well as the captured ultrasonic signals can also be compared.
The minimal computer configuration consists of a PC AT, a graphic card, the MS DOS operating system, and a printer of A4 size. With the proposed experimental system it is possible to execute the entire procedure with a computer program named ULTRAZ, which is written in Turbo Pascal.
To be able to make a successful comparison of the actual, i.e., captured, ultrasonic signals and the theoretical ones, first the ultrasonic device for the communication with an oscilloscope had to be adapted. Later an A/D converter, which permits a conversion of signals into a digital form for a future display, was integrated into the experimental system. The captured ultrasonic signal was also modified for a future computer processing [9, 10].

3. COMPUTER PROGRAM FOR THE CALCULATION OF THEORETICAL SIGNALS AND THEIR COMPARISON TO THE CAPTURED ULTRASONIC SIGNALS

Fig 2: The main procedures of the computer program ULTRAZ for the theoretical display of ultrasonic signals by individual axles.

The computer program was made with object programming, which includes the importance of and the need for the given testing procedure, whereas the procedure order is not important. The procedure is concluded when all procedures are regulated in accordance with the instructions, which the operator should have.
The course of the program operation is shown in Fig. 2 and includes the plotting of the pictorial display. Then follows the reading of the computer command which enables the following procedures:
• to prescribe the shape of the axle or some other machine element;
• to draw the shape of the element to be tested and to check its correctness;
• to input the relevant settings of the test specimen and the ultrasonic probe, which produces a corresponding pictorial display or an echogram;
• to calculate the theoretical sound paths between the individual reflections on the given axle, i.e., machine element;
• to draw sound paths for the given axles, i.e., machine elements, and to present the calculated data, i.e., sound paths, in a graphic and/or table form for the individual types of sound paths A, B, C, D and E.

The computer program works according to a flowchart, which shows the course of individual operations, i.e., procedures.

4. PHYSICAL MODEL OF THE CALCULATION OF SOUND PATHS AS REGARDS THE SHAPE OF THE AXLES, I.E., MACHINE ELEMENTS

A physical model according to which the computer program works and which permits to define the ultrasonic reflections from the edges and calculate the lengths of the sound paths according to input data about the axles, i.e., machine elements, was elaborated. In the simplest case when the longitudinal waves which are directed straight to the edges are taken into account, the longitudinal waves are reflected directly back to the probe. The pulse-echo technique of transmission and receiving of ultrasonic signals at the probe as shown in Fig. 3 was selected. The probe can be moved in circle over the axle front and also closer to or further from the axle centre, which, of course, affects the calculation of the sound paths.

Fig 3: Display of ultrasonic signals of types A and B; LZV - sound path; includes the longitudinal component of ultrasonic waves LL; k - distance from the probe to the edge of the first reflection; h - distance across the radius to the centre of the probe.

Fig. 3 shows ultrasonic signals of types A and B, which are direct longitudinal waves, which reflect from edges on the upper or the lower half of the axle.

5. SCREEN DISPLAY

Fig. 4 shows an example of a screen display for the engine axle, which includes in the upper part the theoretical spots of sound reflection for the individual types of sound paths.

Fig 4: Display of ultrasonic signals of types A, C and E as a function of the sound path at reflection from edge r4.

The left side of Fig. 5 shows the actual signal, while on the right side, there is only a part of the signals left, which deviate from the theoretical, i.e., anticipated, reflections. It was proved with a statistical analysis of the reflected signals which deviate from the anticipated ones that the deviation is important with regard to the theoretical signal from the viewpoint of location of the actual signal and its intensity. That is why the necessary criteria for the acceptance of the axle quality were established or the decision about additional tests that need to be performed is made:

• the first criterion defines the admissible deviation of the sound wave length;
• the second criterion defines the admissible deviation of the signal intensity.

Fig 5: Computer program window with the display of axle shape and the actual captured signal, which is used for further analysis. The right side shows the status after the comparison of the actual signal with the theoretical one, and the verification signal marked ERROR.

The research was limited only to the basic statistical evaluations which were adapted to suit the known types of sound paths. The signal intensity, which is outside the theoretical location of the reflection, being also important for the testing, we have limited ourselves only to the estimation of the signal intensity. If the signal intensity is too great and deviates from the theoretical location, it is necessary to perform an additional testing of that spot with the same method or with a different method or methods, which confirm or refute the given signal deviation. This means that we confirm or refute the existence of a defect for the given axle. We have looked for such a way of processing of the captured and theoretical signals so that a differential signal, which exceeds the allowed threshold can be displayed and then an additional testing of that spot performed. This means that the remaining signals are either irrelevant or critical. In the case of critical intensity of the differential signal, which exceeds the defined threshold value, the criterion of the greatest signal intensity, which requires an additional testing of that spot with the same or a different method, is used.

For the purposes of documenting a subprogram for the display of signals which were actually captured on the real axle, i.e., machine element, was elaborated.

Fig 6: Display of the actual signals with the recording data.

6. CONCLUSION

Engine-axle examinations from the front provide us with only the rough information for the detection of larger cracks present. Due to the transformation of ultrasonic waves from longitudinal into transversal and vice versa one pass over the axle gives three or four signals of different amplitudes and in different places. The signals are recorded as a screen display at fixed amplification or at definition of the height of the last signal at a chosen signal length for the calibration of the device. The parts attached to the axle give a more complicated screen display. Routine work with the given experimental system confirmed that small initial cracks cannot be detected in this way, that is why the proposed method of testing is used only where there are no other options. The method is sufficiently reliable in spite of this limitation because the axles are overdimensioned and can withstand deeper defects, i.e., cracks, before destruction.

As we continue our improvements, we intend to supplement the echogram evaluation criteria and improve the computer program, which will recognise and evaluate the defects from the echograms. We will also have to improve the criteria for defect acceptance, i.e., assessment of the critical defect size, so that there will be a constant system of automatic elimination of axles or other unsuitable vital elements of railway vehicles.

7. REFERENCES

1. Internal Standard of JZS V3.006.
2. Official Communications of Slovenian Railway of Ljubljana, No.: 19, 1979.
3. Krautkrämer, J. H.: Ultrasonic Testing of Materials, Third Edition, Springer - Verlag, New York, Heidelberg, Berlin 1993.
4. JEMEC Viktor, ŠIPEK Mitja : Testing of rail shafts of the series 342/362 with normal ultrasonic probes placed at a certain angle, UDK 620.179.16; 629.423.027.11, Int. Conf. RELIABILITY ON NDT RESULTS NDT '88, Dubrovnik, Croatia, 1988, 189-195 (in Croatien).
5. GERKEN C. in WITKOPP H.: Automatic ultrasonic wheel set testing stand for railway axles with data recording, 10th World Conf. on Non-destructive Testing, Moscow 1982
6. JEMEC Viktor, Grum Janez, Petrišic Joze, Lunar Bojana, Štefotic Boris, Kokolj Franc, Šipek Mitja, Brezan Marija, Pozar Silvester, Kastelec Štefan, Gregoric Anton, Frankic Franc, Zerjav Anton, Beci Anton, Vonca Janez, Manfreda Srecko, Jemec Metka, Breznik Marko, Grabec Igor: Use the nondestructive testing for control of railway bogie, Ministry Of Science And Technology num.: B 42-0477-0361, Ljubljana, 1994 (in Slovene).
7. JEMEC Viktor, PETRIŠIC Joze, LUNAR Bojana: Computer support nondestructive testing on railway axles, The News of Nondestructive association of Slovenia, Ljubljana, vol. 1, no. 3-4, 1994, 14 - 19, (in Slovene).
8. JEMEC Viktor, POLJAK Janko: Automation of Control the Vitalization Elements of Railway Locomotive, Int. Conf. KBR'95 NDT "Inservice Inspection", Pula, Croatia,1995, 67-78
9. GRUM Janez, JEMEC Viktor, BECI Anton: Ultrasonic Testing of Axles and Axle sets of Diesel-Engine Trains, Int. Conf. MATEST'98 - Life extension, Brijuni, Croatia, 1998, 329-339.
10. GRUM Janez, JEMEC Viktor, BECI Anton: Testing of Axles in Wheel and Axle sets of Diesel-Engine Trains, SDA-Proceedings of the first of Slovenian Acoustical Society with International Participation, Portoroz, Slovenia, 1998, 237-244.

© AIPnD , created by NDT.net

|Home|    |Top|