EDUCATIONAL SET-UP FOR VIBRATION MEASUREMENTS IN VEHICLE GEARBOXES

10.2478/v10138-012-0002-z

EDUCATIONAL SET-UP FOR VIBRATION MEASUREMENTS IN VEHICLE GEARBOXES GABRIELA ACHTENOVÁ Czech Technical University in Prague, Prague, Czech Republic E-mail: [email protected]

MOHAMED EL MORSY Helwan University, Cairo, Egypt E-mail: [email protected]

SHRNUTÍ V následujícím článku je představen zkušební stav vybudovaný za účelem diagnostiky chyb v převodovce pomocí měření vibrací na převodovce. Stav je využívaný zejména studenty automobilové specializace. Otevřený měřicí stav sestává ze dvou dynamometrů a je připraven pro upevnění různých typů převodovek. Jako testovací převodovka byla vybrána nejčastější typ převodovky pro příčnou zástavbu hnací ústrojí vpředu a pohon předních kol. Pro měření byly k dispozici dvě naprosto totožné a nové převodovky. Diferenciál byl blokován. Výkon byl odebírán pouze z jedné výstupní příruby. V jedné z převodovek byly uměle vyrobeny následující dvě závady: • Na vnitřním kroužku ložiska vstupního hřídele umístěného ze strany spojky byla vyfrézována podélná drážka o šířce 1 mm. • Důlky představující poškozením pittingem byly vybroušeny do boku zubu IV. rychlostního stupně. Upínací podstavec umožňuje rychlou výměnu převodovky. Při studentském měření je nejdříve jsou měřeny a analyzovány vibrace pomocí FFT, řádové a obálkové analýzy na nové převodovce. Poté je převodovka zaměněna za převodovku s uměle vytvořeným poškozením a za stejných podmínek jsou měřeny a analyzovány vibrace. Studenti mají identifikovat přítomnost a typ závad v převodovce.

KLÍČOVÁ SLOVA: DETEKCE PITTINGU, DIAGNOSTIKA ZÁVAD LOŽISEK, DIAGNOSTIKA POMOCÍ MĚŘENÍ VIBRACÍ ABSTRACT In this paper we present a set-up specifically prepared for an educational application. The aim of the set-up is to build an educational tool for students of automotive engineering. The open test stand we used is equipped with two dynamometers and can accommodate various automotive gearboxes. The most common type of gearbox intended for transversal powertrain placement in small and midsize vehicles was chosen as the test gearbox. At our disposal we had two identical and completely new gearboxes provided by a car manufacturer. The differential of the gearbox was blocked. The total power is taken-off from one side only. In one gearbox two artificial faults were introduced: • An orthogonal placed groove on the inner race with the initial width of 1 mm. The damaged bearing is a roller bearing located on the gearbox input shaft – on the clutch side. • Pitting defect ground on the teeth sides of the IVth speed. The mounting base of the test stand allows easy changing of the gearboxes. During the student training, the new undamaged gearbox is first measured. After completion of signal measurement and analysing by means of FFT, Order and Envelope analysis, the gearbox is changed for the damaged one. After measurements under the same conditions, the same type of analysis of the measured signal is performed. The students have to identify the type of damage(s) present in the damaged gearbox.

KEYWORDS: PITTING DETECTION, BEARING FAULT DETECTION, VIBRODIAGNOSTICS

1. INTRODUCTION

The investigation of diagnostics of mechanical machines with aid of vibration measurements is described in many papers; McFadden [1] has described a series of seeded fault tests performed on the double-helical gears used in a marine

gearbox and presents analyses of the vibration signals measured at two locations on the gearbox for each of five faults; Dalpiaz and Rivola [2] have assessed and compared the effectiveness and reliability of different vibration

Educational Set-up for Vibration Measurements in Vehicle Gearboxes Gabriela Achtenová, Mohamed El Morsy

MECCA 01 2012 PAGE 7

2. DESCRIPTION OF THE TEST SET-UP

analysis techniques such as amplitude probability density for fault detection and diagnostics in cam mechanisms used in high performance automatic packaging machines; Meltzer and Nguyen [3] have studied the effectiveness of the continuous wavelet transform in vibro-acoustical diagnostics of gearboxes operating under non-stationary rotational speed, which has been illustrated by means of a program test for fault diagnosis on helical spur gears. The set-up conceived and described in this paper is mainly intended for educational and scientific purposes. For educational purposes the Automotive Engineering Masters Degree students learn how to determine various gearbox faults with the help of vibration measurements. For the scientific purposes we are investigating: • Analysis of fault propagation with help of different methods (Crest factor, Kurtosis factor, etc). • Detection of combined faults (fault on bearings and pitting faults). • The influence of a change in torque or speed of rotation on the magnitude of fault factors. • The influence of the accelerometer position on the gearbox casing on the magnitude of fault factors. The results of these investigations are described in this article.

n1

The measurements are conducted on an open loop test bed consisting of two dynamometric machines, the gearbox to be tested, and torque flanges for torque measurement. On the input – as a motor – is used a three-phase dynamometer KS 56-4. On the output side is an eddy current water cooled dynamometer 2VD 110 / 6 as a brake. To obtain a wide range of rotation speeds we use a planetary two speed overdrive unit after the input dynamometer. The planetary gear set can function as direct with ratio = 1 or overdrive (ratio = 0.4). The shift of the overdrive unit can only occur in steady state. In our case, we used the ratio = 1 in the planetary gearbox. When the gearbox for transversal disposition of drivetrain is used – in the case of the investigated gearbox – the differential has to be locked. The power flows through one output shaft only. The mounting base for the gearbox is easy to manipulate. Therefore, the gearboxes can easily be changed, which is useful for educational purposes, where the faultless gearbox is switched for the faulty one during student experimental exercises, and the measurements are compared.

3. INVESTIGATED GEARBOX

The gearbox used for our measurements is the type most commonly used in current small and mid-size passenger cars with transversally mounted powertrain and front wheel drive:

M1

M2

tw

tpl

n2 2VD 110/6 (brake)

KS 56 – 4B-1 160 kW/3000 min (motor)

toil

n planetary overdrive i=1; 0,4

a2

a1

B&K LAN XI 3050

FIGURE 1: Diagram of the open loop test bed used for investigation of faults in an automotive gearbox with the aid of vibration measurements OBRÁZEK 1: Schéma otevřeného zkušebního stavu pro zjištění závad automobilové převodovky pomocí měření



FIGURE 3: Detail of the bearing inner race damage. A = 0,6 to 0,7 mm, B = 0,6 mm OBRÁZEK 3: Detail poškození vnitřního kroužku ložiska.

FIGURE 4: Detail of the pitting damaged produced on one tooth only. OBRÁZEK 4: Detail poškození pittingem vyfrézovaném na jednom zubu.

a two-shaft, five speed gearbox with final drive gear and front wheel differential. The internal arrangement of gears, shafts and bearings is depicted in Figure 2. Two faults are artificially introduced into the gearbox: • Groove on the inner race of a bearing mounted on the input shaft in the clutch casing (the faulty bearing is highlighted in the Figure). • Pitting on the free rotating wheel of the IVth speed. This fault is combined with the bearing fault. The following two Figures depict the artificially fabricated faults. In the case of the bearing a groove was orthogonally ground on

the inner race of the roller bearing on the input shaft. The width of the groove varies from 0.6 to 0.7 mm. The diameter of the rollers is 4 mm, so a groove of these dimensions produces a significant impact on the rollers – which will be demonstrated in section 4, where the results of vibration measurements are discussed. The free rotating wheel of the IVth speed has damage on one tooth only. The mesh side of the tooth was damaged by grinding three pits. The pits have a total surface area of 4.894 mm2, which equals 8% of the total surface of the tooth mesh side.

5-1 J5

2-1

1-1

J3

Z-1

L4

4-1

3-1

SP3-4

J4

L1 Input

SP5

4. RESULTS OF VIBRATION MEASUREMENTS

This section is split into three parts. The first deals separately with the case of the damaged bearing only. The measurements were performed when IIIrd speed was engaged. The second covers the case of the IVth speed fault in combination with the bearing fault. The last part is dedicated to the determination of the gearbox faults at different accelerometer positions.

Z-Z L6

J2

J1

L2

SZ1 Output

5-2

Z-2 1-2

2-2

4-2

3-2

K1

SP1-2 Differential K2

FIGURE 2: Diagram of the investigated five-speed automotive gearbox. The highlighted gear wheel and bearing are the two faulty elements OBRÁZEK 2: Schéma vyšetřované pětistupňové převodovky. Zvýrazněné ozubené kolo a ložisko jsou poškozené členy.

4.1. BEARING FAULT Figure 5 and Figure 6 show the frequency domain and the time history of the synthesized signal, 20 kHz baseband and 125 ms, of the faulty bearing in third speed (2000 rpm) and input shaft load 50Nm, and output speed 370 rpm at output shaft load 310 Nm. The peaks in the time domain are repeated at approximately 30 ms (33.3 Hz), which matches the input shaft speed of 2000 rpm (33.33 Hz). In the frequency domain there are harmonics at high frequencies that indicated a fault in the gearbox bearing, but from which it is difficult to determine the bearing and faulty element in that bearing. There are some steps that can determine these unknowns; the first one is to check the time domain to determine the revolution speed of the bearing shaft (33.3 Hz) and the second is to make an envelope in the frequency domain to determine the faulty element of the bearing.



4 Acceleration (m/sec2)

Acceleration (m/sec2)

400 200 0 -200

3 2 1

Period between impacts 0

-400 0

20m

40m

60m 80m Time (sec)

100m

FIGURE 5: Time domain for faulty bearing (inner race fault) in IIIrd speed. OBRÁZEK 5: Časová oblast záznamu měření vibrací pro poškozený vnitřní kroužek ložiska při zařazeném 3. rychlostním stupni.

Shaft speed

2

Inner race frequency

1.6 1.2 800m

0

1.2k

Input speed: 2000 rpm Input load 20, 50 Nm respectively

19 .06

Shaft speed (33.3 Hz)

400 Frequency [Hz]

600

800

FIGURE 8: A high-resolution frequency domain envelope for faulty bearing (inner race fault) in IIIrd speed. OBRÁZEK 8: Výřez nízkofrekvenčního záznamu z obrázku 7.

Input speed: 1500, 2000, 2500 rpm respectively Input load 50 Nm

25 .55

20k

Inner race frequency 355 Hz

200

1.4k

FIGURE 7: Frequency domain envelope for faulty bearing (inner race fault) in IIIrd speed. OBRÁZEK 7: Obálka záznamu frekvenční analýzy pro měření poruchy vnitřního kroužku ložiska při 3. zařazeném stupni.

16k

800m

0 600 800 1k Frequency [Hz]

12k Frequency [Hz]

1.2

400m

400

8k

1.6

400m

200

4k

FIGURE 6: Frequency domain for faulty bearing (inner race fault) in IIIrd speed. OBRÁZEK 6: Frekvenční analýza záznamu z obrázku 5



2

0

120m

59 .18

19 .58

16 .04

9.37

25 .55

10 .21 19 .06 11.98

Kurtosis values

RMS values

Crest factor

FIGURE 9: Comparisons between parameters that are used in fault detection for a faulty bearing at low load and differing loads in IIIrd speed. OBRÁZEK 9: Porovnání jednotlivých parametrů používaných pro detekci závad převodovek v závislosti na zatížení.

15 .62

Kurtosis values

14 .51 8.75 10.21 8.5

RMS values

Crest factor

FIGURE 10: Comparisons between parameters that are used in fault detection for a faulty bearing at low load and differing speeds in IIIrd speed. OBRÁZEK 10: Porovnání parametrů používaných pro detekci závad převodovek v závislosti na otáčkách vstupního hřídele.



2

Shaft speed = 33.3 Hz Acceleration (m/sec2)


80 40 0 -40 -80

1.2 800m 400m 0

0

100m

200m 300m Time [s]

400m

0

FIGURE 11: Time domain for faulty gear (Pitting of tooth) in IVth speed OBRÁZEK 11: Časová oblast záznamu vibrací pro kombinaci poruch na ložisku a pitting na IV. rychlostním stupni

6

2

Shaft speed (33.3 Hz)

4

Harmonic

2

1k

2k

3k Frequency [Hz]

4k

5k

6k

FIGURE 12: Frequency domain for faulty gear (Pitting of tooth) in IVth speed OBRÁZEK 12: Záznam z obrázku 11 po použití FFT



Shaft speed = 33.3 Hz Mesh frequency = 1.367 kHz

1.6

Shaft speed

Inner race frequency

1.6 1.2 800m 400m 0

0 0

200

400

600 800 Frequency [Hz]

1k

1.2k

1.4k

1.6k

100

200

300 400 Frequency [Hz]

500

600

FIGURE 13: Frequency domain after application of envelope analysis for faulty bearing (inner race notch) in IVth speed. OBRÁZEK 13: Obálka hodnot signálu ve frekvenční oblasti pro případ kombinace poruchy na ložisku a pitting na IV. rychlostním stupni.

FIGURE 14: High resolution frequency domain after application of envelope analysis for faulty bearing (inner race notch) in IVth speed. OBRÁZEK 14: Výřez nízkofrekvenční oblastí z obrázku 13.

Figure 8 shows a high-resolution spectrum of the envelope. Figure 7 shows four harmonics of the BPFI = 355 Hz, all amplitudes modulated by the shaft speed = 2000 rpm = 33.3 Hz accounting for the load variation. The modulation of the BPFI may be perceived as spoiling the clean 355 Hz, BPFI, spectrum, but on the other hand it is very informative. In a real situation where bearing frequencies would not be known in advance, this spectrum is the signature of a rotating race fault, most often the inner race. The Kurtosis, Crest and RMS factors are calculated; the equations for computing all mentioned factors are listed e.g. in [2]. The Kurtosis factor is often used for determination of faults in a gear or bearing in the gearbox. The following graphs show the change in the factors with change of load and change of rotational speed. The values are calculated from the measurements acquired using an accelerometer placed inside the clutch cover “on” the faulty bearing (position “A”, Figure 18).

4.2. COMBINED FAULT PITTING AND BEARING Figure 11 and Figure 12 show the time domain and frequency domain signal for the faulty gear (41 teeth) at 0.5 sec and 6.4 Hz respectively and IVth speed at input speed = 2000 rpm (33.33 Hz) and input load 50 Nm, and output speed = 515 rpm and output load = 230 Nm. To obtain information about faults in the bearing, the envelope analysis is used again. This highly accurate diagnosis can provide information about a bearing fault even when the signal is affected by strong noise – as in our case with the faulty gear. Figure 13 and Figure 14 show the envelope in the frequency domain for the faulty bearing (inner race) in IVth speed 2000 rpm and input load 50 Nm. These figures show harmonics in addition to the shaft speed. It is clear that fault detection and classification (faulty gear and faulty bearing) can be achieved using vibration analysis. Kurtosis is a parameter that is sensitive to the signal peaks and is well adapted to the impulse nature of the simulating forces generated by component damage. Over time the defect becomes



Input speed: 2000 rpm Input load 20, 50, 90 Nm respectively

9.09

9.91

8.19

8.09

5.83 6.12

7.65

5.71

4.54

Kurtosis values

RMS values

Crest factor

FIGURE 15: Comparison of parameters used in gear fault detection at different load in IVth speed. OBRÁZEK 15: Porovnání parametrů používaných pro detekci závad převodovek v závislosti na různém zatížení


10.80 9.27

9.09

8.09 6.82 5.83

5.50

Kurtosis values

5.85

5.27

RMS values

Crest factor

FIGURE 16: Comparison of parameters used in fault gear fault detection at intermediate load and different speeds in IVth speed. OBRÁZEK 16: Porovnání parametrů používaných pro detekci závad převodovek v závislosti na otáčkách vstupního hřídele


8.19

7.9 7.97

6.29 6.48 5.71 5.06

4.54 2.97

Kurtosis values

RMS values

Crest factor

FIGURE 17: Comparison of parameters used in gear fault detection at low load and different speeds in IVth speed. OBRÁZEK 17: Porovnání parametrů používaných pro detekci závad převodovek v závislosti na otáčkách vstupního hřídele při malé zátěži

greater or more defects propagate in the gearbox in such manner that the impulsive signal changes to a continuous one (the first signal does not expire before the second one is produced [1]), the Kurtosis index diminishes to a value of around 3; i.e. to the value of a “not damaged” signal. This explains why the Kurtosis index as well as the RMS and Crest factor are lower for the IVth speed, than for the IIIrd speed, where only the bearing fault occurred. Figure 15 shows the influence of input load change on Kurtosis values at input shaft IVth speed (2000 rpm = 33.33 Hz) in IVth speed, and input load change between 20, 50, 90 Nm with output shaft loads of 250, 230 and 365 Nm respectively. It is clear that Kurtosis values increase proportionally with increasing input load. It is important to avoid high loads when a fault is present to prevent accelerated progress of the fault and sudden breakdown. Figure 16 shows the change of Kurtosis values against the input shaft speed in IVth speed (1500, 2000, 2500 rpm), input load 50 Nm with output shaft loads of 135, 150 and 250 Nm respectively. It is important to note that it is not sensitive to speed, but rather load. Figure 15, Figure 16 and Figure 17 show a comparison between some parameters that are used in gear fault detection at different loads and speeds. It is clear from the above figures that the RMS values are a sensitive and good indicator of gearbox faults at both different loads and speeds. On the other hand, the RMS does not indicate the magnitude of fault progress. The Kurtosis values are used as the best indicator of a fault at different loads because its values also indicate the fault progress, which may be useful for predictive maintenance, but it should be taken into consideration that it is not sensitive to speed. 4.3. FAULT DETECTION WITH DIFFERENT ACCELEROMETER POSITIONS When measuring in real situations, we never know whether a fault is in the gearbox or its position. For building a monitoring system which will indicate the fault propagation in a gearbox often only one or at most two accelerometers are used. The accelerometers should be placed such that wherever the fault is located, detection of the fault is possible. In our case both faults are concentrated near the clutch cover. In the following study we tried four different accelerometer positions (see Figure 18): A –on the faulty bearing housing inside the clutch cover, B – on top of the gearbox housing near the mounting point with the clutch cover. In our case near to the faulty bearing and gear, C – on the gearbox housing above the bearing shield, D – on the gearbox housing above the tapper roller bearing of the output shaft. For every position we determined the Kurtosis index, RMS and Crest value. The following graphs show the results for the IVth



FIGURE 19: Comparison of results of Kurtosis index, RMS and Crest factor for different bearing positions in IVth gear speed. Input speed = 2000 rpm and differing loads. OBRÁZEK 19: Porovnání parametrů používaných pro detekci závad převodovek v závislosti na zátěži a umístění akcelerometrů. Kombinovaná závada na ložisku a pitting ozubení. FIGURE 18: Accelerometers positions. OBRÁZEK 18: Umístění akcelerometrů

and IIIrd gear speed. With Kurtosis and Crest factors we are only able to indicate the fault when the accelerometer is placed close to the faulty object(s). For any other position, even for position “B” (which was unexpected) we obtain a Kurtosis value of around 3, even when considering just the faulty bearing – as demonstrated in Figure 21. The only factor that increases in magnitude with increasing load, or with increasing speed of rotation for any placement of accelerometer on the gearbox casing is the RMS value. However, as stated before, the RMS magnitude itself does not indicate the fault occurrence and/or its propagation. To construct the monitoring system we should first measure the undamaged gearbox to have a reference value for RMS. Due to the fact that position “C” in the case of the IVth speed has a similar RMS magnitude to that measured from the accelerometer in position “A”, we propose the accelerometer placement in position “C”. This position lies on the most rigid part of the gearbox housing, where bearings of input shaft and lay shaft are located, and therefore we expect to obtain a good response for bearing and gear faults. Future measurements should confirm this assumption.

FIGURE 20: Comparison of results of Kurtosis index, RMS and Crest factor for different bearing positions in IVth gear speed. Input load = 50 Nm and different input speeds of rotation. OBRÁZEK 20: Porovnání parametrů používaných pro detekci závad převodovek v závislosti na otáčkách vstupního hřídele a umístění akcelerometrů. Kombinovaná závada na ložisku a pitting ozubení.

5. CONCLUSIONS

• The FFT technique and the high order statistics of RMS and Kurtosis reflect the vibration responses of the gearbox. This can be an effective tool for planning predictive maintenance and avoiding a sudden breakdown. • This study shows that high order statistics of the RMS and Kurtosis techniques seem quite effective in detecting combined faults (fault on bearing and pitting

FIGURE 21: Comparison of results of Kurtosis index, RMS and Crest factor for different bearing positions in IIIrd gear speed. Input load = 50 Nm and different input speeds of rotation. OBRÁZEK 21: Porovnání parametrů používaných pro detekci závad převodovek v závislosti na otáčkách vstupního hřídele a umístění akcelerometrů. Závada pouze na vnitřním kroužku ložiska.



fault). The technique excels in extracting transients, which are often indicators of incipient damage and fault propagation in geared systems. While a certain amount of data processing is necessary, the time and frequency domain offer great detail concerning the meshing conditions of two gears. The high order statistics of RMS and Kurtosis are shown to reflect features associated with the presence of damaged teeth to yield a more positive assessment of tooth condition. • This study shows the influence of change of torque or speed of rotation on the magnitude of fault factors and its effect on fault propagation • Envelope Analysis for Diagnostics of Local Faults in Rolling Element Bearings of raw signal can be an effective way to obtain great detail about a bearing fault.

ACKNOWLEDGEMENTS

This research has been realized using the support of EU Regional Development Fund in OP R&D for Innovations (OP VaVpI) and Ministry for Education, Czech Republic, project # CZ.1.05/2.1.00/03.0125 Acquisition of Technology for Vehicle Center of Sustainable Mobility. Further this research has been realized using the support of Technological Agency, Czech Republic, programme Centres of Competence, project # TE01020020 Josef Božek Competence Centre for Automotive Industry. These supports are gratefully acknowledged.

REFERENCES

[1] Mc Fadden P. D. (2000). Detection of Gear Faults by Decomposition of Matched Differences of Vibration Signals. Mechanical Systems and Signal Processing 14 (5), pp. 805 – 817 [2] Dalpiaz, G. & Rivola, A. (1997). Condition Monitoring and Diagnostics In Automatic Machines: Comparison of Vibration Analysis Techniques. Mechanical Systems and Signal Processing 11 (1) , pp. 53 – 73, 1997. [3] Meltzer G. & Nguyen Phong Di. (2004). Fault Diagnosis in Gears Operating under Non-Stationary Rotational Speed using Polar Wavelet Amplitude Maps. Mechanical Systems and Signal Processing 18, pp. 985 – 992. [4] de Lorenzo F. & Calabro M. (2007). Kurtosis: A Statistical Approach to Identify Defect in Rolling Bearings. In: Proceedings of 2nd International Conference on Marine Research and Transportation, Naples 2007, Session A, pp. 17 – 24. [5] Abouel-seoud S.A., Elmorsy M.S. & Dyab E.S. (2011). Robust Prognostics Concept for Gearbox with Artificially Induced Gear Crack Utilizing Acoustic Emission. Energy and Environment Research 1 (1), 2011. [6] Tůma J. (1997). Zpracování signal získaných z mechanických systémů užitím FFT. Sdělovací technika, Praha. ISBN 80-901936-1-7. [7] Tomeh E. (2003). Identifying Motor Vehicles Mechanical Defects by Vibrodiagnostics Methods. MECCA 1 (1), ISSN 1214-0821, pp 16 – 22.



EDUCATIONAL SET-UP FOR VIBRATION MEASUREMENTS IN VEHICLE GEARBOXES

Recommend Documents