Gear Pitting Assessment Using Vibration Signal Analysis

Gear Pitting Assessment Using Vibration Signal Analysis Zoltan Iosif KORKA Romania, z.korka@uem.ro Aurel BARA Romania, bara@uem.ro Bogdan CLAVAC Romania, clavac@uem.ro Lidia FILIP Romania, l.filip@uem.ro Abstract: - Gears are common machine elements of transmission systems, being used to transfer power. When gears are operating near their maximum load capacity, the gear teeth are subjected to very high contact pressures, leading to contact fatigue failures such as pitting, scoring, spalling or scuffing, even in adequate lubrication conditions. Currently, three methods to detect faults in gear systems are approached: analysis of oil/ wear particle, vibration analysis and acoustic signal analysis. In this study, experimental investigations are carried out to assess gear pitting using vibration signal analysis. In this respect, four pinions with different pitting grades where investigated on a single helical gearbox integrated in an openenergy test rig. These results provide a better understanding of the consequences of pitting evolution in gear systems, providing information for the assessment of gear fault stage. Keywords: - gear teeth, pitting, signal, vibration analysis 1. INTRODUCTION The current trend in the construction of gear systems, regarding speed and power increasing, favors the development of damages on the gear teeth surface, such as pitting, scoring, spalling or scuffing. These deteriorations at the level of the active surfaces of the teeth lead to increased vibration and noise levels in the operation of gear transmissions. Therefore, in the overall effort to reduce pollution produced by vibration and noise in the industrial environments, [1], [2], it is essential to seize the techniques for recognizing these defects. Pitting is the most common gear surface failure, being related to lubricated gears. As stated in the American Society for Metals (ASM) handbook [3], Pitting occurs when fatigue cracks are initiated on the tooth surface or just below the surface. Usually, pits are the result of surface cracks caused by metalto metal contact of asperities or defects due to low lubricant film thickness. High-speed gears with smooth surfaces and good film thickness may experience pitting due to subsurface cracks. These cracks may start at inclusions in the gear materials, which act as stress concentrations, and propagate below and parallel to the tooth surface. Pits are formed when these cracks break true the tooth surface and cause material separation. When several pits join, a larger pit (or spall) is formed. Pitting can also be caused by foreign particle contamination of the lubricant. These particles create surface stress concentration points that reduce lubricant film thickness and promote pitting. In the last two decades, vibration signal analysis has been employed for condition monitoring and fault detection in gearboxes. As in other areas [4], different techniques have been developed, including signal demodulation [5-7], analysis in time and frequency domains [8-10] or cepstrum techniques [11-13]. The present study is concerned with the investigation of the vibration signal of a gear transmission showing various pitting grades (from new ones to gears with severe pitting) and driving at different speeds. 2. PROBLEM FORMULATION Our aim is to assess the most common gear surface failure, i.e. pitting, using vibration signal analysis. RJAV vol XIV issue 1/2017 44 ISSN 1584-7284

For this purpose, a test stand with open energy flow circuit, also employed for previous researches [14], [15] and shown in Figure 1, was used. Figure 1. View of the test stand For the drive of the gearbox, an electric motor with variable speed was used, since a gear pump (produced by Kracht, Germany) was employed for the breaking system. The technical data of the single helical gearbox are shown in Table 1. In order to isolate the vibrations transmitted from the electrical motor and from the pump respectively, couplings with rubber strips were used. Besides a good torsional vibration damping, they also have the advantage of a smooth assembling, as well as do not require a rigorous alignment of the coupled elements. For adjusting the speed of the electrical motor, a torque flange type T 10 FS fitted to the input shaft was utilized, the speed values being read on the acquisition module MP 60, both devices being made by HBM, Germany. Table 1. Gearbox technical data Performance data Maximum power P [kw] 2,5 Ratio i 2,529 Range of input speed n 1 [rpm] 1.000-1.500 Main geometrical data of the gears Centre distance A [mm] 125 Number of teeth z 1 /z 2 17/43 Helix angle β [ ] 9 Module m n [mm] 4 Gear width b [mm] 40 Normal pressure angle α 0 [ ] 20 Addendum mod. coeff. x 1 / x 2 0,485/ 0,477 Reference diameters d 1 /d 2 [mm] 68,85/ 174,14 For vibration measurements an accelerometer of type 8772, produced by Kistler, has been used. It was installed on the housing, above the high speed shaft, on the electrical motor side. The acquisition board, consisting of a NI (National Instruments) chassis cdaq-9172 and the signal acquisition module NI 9234, was used to send binary encoding date to a laptop programmed to run a specially designed application in LabView software, for fine processing of the vibration signal. Test pinions with four different conditions were investigated: new pinion, pinions with slight pitting, moderate pitting and severe pitting. In order to mimic the different pitting stages, artificial grooves have been created on new manufactured pinions, along the pitch line, which had a diameter of 3 mm and a depth of about 0,5 mm. The four pinions with different failure conditions are shown in Figure 2 (anew pinion; b- pinion with slight pitting; c- pinion with moderate pitting; d- pinion with severe pitting). d. Figure 2. Pinions with different pitting grades RJAV vol XIV issue 1/2017 45 ISSN 1584-7284

The procedure was organized in following steps: - the pinion without pitting was mounted into the gearbox; - the test stand was operated at three different input speeds: 1000, 1250 and 1500 [rpm]; - the vibration signals were collected, stored and processed; - the pinion was replaced with the one having slight pitting, followed by the last previously described two steps; - the experimental measurements were continued as described, with the other two pinions having moderate and severe pitting. 3. RESULTS AND DISCUSSIONS Table 2. Time representation of vibration acceleration Table 2 presents the results regarding the vibration acceleration for the 3 operating speeds, and for the four pinions with various pitting stages respectively. As it can be seen, the diagrams provide information on the increase of the global vibration level, indicating changes in operation. For a better understanding of the phenomenon, Figure 3 provides the shape evolution of the vibration measured at the investigated operating speeds and pitting stages. n 1 = 1.000 rpm n 1 = 1.250 rpm n 1 = 1.500 rpm New pinion Pinion with slight pitting Pinion with moderate pitting d. Pinion with severe pitting RJAV vol XIV issue 1/2017 46 ISSN 1584-7284

15 10 5 0 1,000 1,250 n1 [rpm] 1,500 Without pitting Figure 3. Vibration acceleration at different operating speeds and pinion conditions It was observed that the increase of the input speed from 1000 to 1500 rpm doubles the amplitudes of the vibration. Furthermore, a visible detachment of the acceleration variation curves for the pinions with pitting was noticed. As the representations presented in Figure 2 are on a time scale, they indicate only a vibration amplitude increase, without being able to specify the responsible cause for this increase. Therefore, it is necessary to use a frequency representation of the vibration spectrum. In this way, it can be seen which frequency is responsible for a certain acceleration peak, thus deducting the causative source of the vibration increase. Thereupon, Figures 4, 5, 6 and 7 show the frequency spectrum at the three investigated operating speeds and for the four pitting stages. Figure 4. Frequency spectrums for the pinion without pitting, on: a) 1000 rpm, b)1250 rpm Figure 5. Frequency spectrums for the pinion with slight pitting, on: a) 1000 rpm, b)1250 rpm RJAV vol XIV issue 1/2017 47 ISSN 1584-7284

Figure 6. Frequency spectrums for the pinion with moderate pitting, on: a) 1000 rpm, b)1250 rpm Figure 7. Frequency spectrums for the pinion with severe pitting, on: a) 1000 rpm, b)1250 rpm Value of dominant harmonic [mm/s2] In the upper figures, the harmonics giving the maximum amplitude in the frequency spectrum were marked with yellow. Table 3 provides a summary of the data presented in the Figures 4, 5, 6 and 7. Furthermore, Figure 8 depicts the variation of the dominant harmonics for the investigated speeds, respective pitting stages. Table 3. Values of dominant harmonics at different operating speeds and pinion conditions n1 [rpm] 1.000 1.250 1.500 Pinion without pitting 0,003 0 0,007 Pinion with slight pitting 0,028 0,048 0,190 Pinion with moderate pitting 0,040 0,080 0,980 Pinion with severe pitting 0,120 0,145 1,350 RJAV vol XIV issue 1/2017 1.5 1 0.5 0 1,000 Without pitting 1,250 n1 [rpm] 1,500 Slight pitting Figure 8. Variation of dominant harmonics with operating speed and pinion condition 48 ISSN 1584-7284

Analyzing the data presented in Figs 4 to 8, and in Table 3 respectively, it can be determined: a) For the pinion without pitting, the measurements have shown dominant frequencies which don t belong to the teeth engagement frequencies or their harmonics. These frequencies are associated to the rotational frequency of other gearbox components, such as bearings and couplings or to imbalances and misalignments. b) Usually, the gear pitting could be highlighted at the higher (second or third) harmonics of the teeth engagement frequencies. c) The frequency spectrum shows sidebands around the engagement frequencies and their harmonics. These sidebands are becoming less visible as the wearing grade (pitting) of teeth increases. d) The increase of the operating speed and the wearing grade increases the vibration amplitudes at the engagement frequencies and their harmonics. 4. CONCLUSIONS Vibration signals collected from gear systems contain information regarding the gear deterioration progress, since the tooth wear process is having a significant effect on the gear dynamics. This paper has assessed different wearing grades of a single helical gearbox by using vibration signal analysis. It was assessed that with the increase of the pitting grade of the gears, the vibration amplitudes at the engagement frequencies and their harmonics are increasing too. Finally, vibration signal analysis is an important tool for the early detection of faults (when they are just starting to occur), being widely used for condition monitoring, breakdown avoidance and maintenance planning of machineries. ACKNOWLEDGMENTS The authors would like to acknowledge the cooperation with the gearbox factory Resita Reductoare si Regenerabile S.A. (RRR), where the pinions used in the present research were manufactured. REFERENCES [1] P. Bratu, N. Dragan, Innovative Composite Systems for Vibration and Mechanical Shocks Damping. Numerical Nonlinear Modelling, Proceedings of the 5th International Vilnius Conference Knowledge-based technologies and OR methodologies for strategic decisions of sustainable development (KORSD-2009), September 30- October 3, 2009, Vilnius, Lithuania, pp. 357-362 [2] O. Vasile, N. Dragan, Innovative Sound Insulation and Absorption Modular Systems for Public Works Equipment, Proceedings of the 5th International Vilnius Conference Knowledge-based technologies and OR methodologies for strategic decisions of sustainable development (KORSD- 2009), September 30- October 3, 2009, Vilnius, Lithuania, pp. 404-409 [3] P.J. Blau, ASM Handbook, Volume 18- Friction, lubrication and wear technology, ASM International, 1992 [4] G.R. Gillich, N. Maia, I.C. Mituletu, M. Tufoi, V. Iancu, Z. Korka, A new Approach for Severity Estimation of Transversal Cracks in Multi-Layered Beams, Latin American Journal of Solids and Structures, Vol.13, No.8, 2016, pp. 1526-1544 [5] P.D. McFadden, Detecting fatigue cracks in gears by amplitude and phase demodulation of the meshing vibration, ASME, Journal of Vibration, Acoustics, Stress, and Reliability in Design, Vol. 108, 1986, pp. 165 170 [6] J.E Nicks, G. Krishnappa, Gear fault detection using modulation analysis technique, Proceedings of the Second International Conference on Gearbox Noise, Vibration, and Diagnostics, Institute of Mechanical Engineers, UK, 1995, pp. 81 89 [7] W.Wang, Early detection of gear tooth cracking using the resonance demodulation technique, Mechanical Systems and Signal Processing Vol. 15 No. 5, 2001, pp. 887 903 [8] X.W. Chen, Z.P. Feng, Iterative generalized time-frequency reassignment for planetary gearbox fault diagnosis under nonstationary conditions, Mechanical Systems and Signal Processing, Vol. 80, 2016, pp. 429-444 [9] C. Xu, C. Wang, W. Liu, Nonstationary Vibration Signal Analysis Using Wavelet-Based Time-Frequency Filter and Wigner-Ville Distribution, Journal of Vibration and Acoustics-Transactions of the ASME, Vol. 138, No. 5, 2016, article number: 051009 [10] X.W. Chen, Z.P. Feng, Time-frequency Demodulation Analysis for Gearbox Fault Diagnosis under Nonstationary Conditions, IEEE International Conference on Prognostics and Health Management (ICPHM), 20-22 June, 2016, Ottawa, Canada [11] X.H. Liu, Y. Zhang, B.B. Wang, X.D. Li, W. Zhang, Research on Order Cepstrum Method for Gear Cracking Fault Diagnosis, International Conference on Mechanical Engineering and Automation (ICMEA 2015), 7-8 August, 2015, Chengdu, P.R. of China, pp. 93-97 [12] S. Ziaran, R. Darula, Determination of the State of Wear of High Contact Ratio Gear Sets by Means of Spectrum and Cepstrum Analysis, Journal of Vibration and Acoustics- Transactions of the ASME, Vol. 135, No. 2, 2013, article number: 0210008 [13] Y. Guo, J. Han, Analysis about coordinate of cepstrum and its application in gearbox fault diagnosis, Applied Mechanics and Materials, Vols. 397-400, 2013, pp. 2219-2222 [14] Z.I. Korka, G.R. Gillich, I.C. Mituletu, M. Tufoi, Gearboxes Noise Reduction by Applying a Fluoropolymer Coating Procedure, Environmental Engineering and Management Journal, Vol.14, No.6, 2015, pp. 1433-1439 [15] Z.I. Korka, V. Cojocaru, N. Gillich, C.O. Miclosina, Comparative assessment of noise reduction in helical gearboxes, Journal of Vibration Engineering & Technologies, Vol. 5, No. 3, June 2017, pp. 263-268 RJAV vol XIV issue 1/2017 49 ISSN 1584-7284