Printer Friendly

Psychoacoustic analysis of gear noise with gear faults.

ABSTRACT

Gear drives are widely used in the transmission of many types of vehicles and various gear faults were reported to have different effects on the performance of transmission systems. The psychoacoustics metrics, which are used to represent the human hearing property, are objective indicators of product sound quality performance. Therefore, psychoacoustic analysis of gear noise with gear faults needs to be conducted. In this paper, different types of gear faults are summarized, and two of them, including wear and misalignment, are studied separately in the psychoacoustic analysis of the synthesized noise signal of an example gearbox. The gear noise spectra for the cases with different gear faults are synthesized based on the findings of previous publications, where it shows that the two gear faults can either increase the amplitude at the harmonics of the gear mesh frequency or cause the sideband responses. Five psychoacoustics metrics, including loudness, sharpness, tonality, spectral centroid and Kurtosis, are analyzed for the cases with various gear faults. The analysis results indicate that the psychoacoustic analysis provides more gear faults information, and it can be used in gear health monitoring and troubleshooting. In addition, the psychoacoustic metrics can also be used as design targets in the early stage of the gear design.

CITATION: Guo, D., Shi, Q., and Yi, P., "Psychoacoustic Analysis of Gear Noise with Gear Faults," SAE Int. J. Passeng. Cars - Mech. Syst. 9(2):2016.

INTRODUCTION

Helical gear drives are widely used in the transmission of automotive, wind turbine, helicopters and other gear reducers. One of the main reasons of the transmission system breakdown was reported to be the gear failure [1], and a lot of efforts have been applied to the gear fault detection and diagnosis in the past few decades [2, 3, 4, 5, 6, 7, 8, 9, 10]. Various signal processing methods were developed and proved to be effective tools for gear condition monitoring and troubleshooting, including the time-domain [2, 3], frequency-domain [4, 5, 6] and time-frequency domain approaches [7, 8, 9, 10]. However, these signal processing methods mostly focus on using the objective metrics to detect gear faults, such as the magnitude variations of the noise and vibration signals [3, 8, 10].

Psychoacoustic analysis of noise becomes more and more important, and some standard psychoacoustics metrics, developed in the past few decades, are widely used in the automotive industry to evaluate the sound quality of an entire car or its components [11, 12, 13, 14]. This is mainly because that the calculated psychoacoustic metrics are the direct measurements of human perceptions of the noise [15, 16]. On the other hand, there are only a few publications talking about the psychoacoustic analysis of the gear noise. Brecher, et.al [17, 18] investigated the effect of the system geometry parameters and the gear tooth profile on the human perception of the gear noise. Kim, et. al [19] developed a new tonality psychoacoustic metric to evaluate the sound quality of a special axle-gear whine noise, which is a non-stationary signal. Later on, they also correlated the subjective and objective evaluation results for the same axle-gear whine noise through using the Artificial Neural Network (ANN) method [20, 21]. Becker [22] developed an objective method for the analysis of the invehicle gear whine noise, which was shown to be an effective approach for evaluating the gear whine noise by performing a statistical analysis of the physical data. However, most of these psychoacoustic analyses of the gear noise were performed for the sound quality evaluation purpose, and there is no publication dealing with the diagnosis of gear faults using the psychoacoustic analysis of the gear noise, to the best knowledge of the authors.

Therefore, one primary purpose of this study is to diagnose gear faults through performing psychoacoustic analysis of the noise radiated by geared system with different type of gear faults. To do so, different types of gear faults are summarized, and two of them, including wear and misalignment, are studied separately in the psychoacoustic analysis of the synthesized noise signal of an example helical gear box. The gear noise spectra for the cases with different gear faults are synthesized based on the findings of previous publications, where it shows that the three gear faults can either increase the amplitude at the harmonics of the gear mesh frequency or cause the sideband responses. Five subjective metrics, including loudness, sharpness, tonality, spectral centroid and Kurtosis, are analyzed for the cases with various gear faults. The analysis results indicate that the psychoacoustic analysis provides more information for detecting gear faults, and it can be used in gear health monitoring and troubleshooting.

EFFECT OF GEAR FAULTS ON THE GEAR NOISE

The effects of the gear faults on the gear dynamics are widely studied, and a large number of publications can be found [29, 30, 31, 32, 33]. As a comparison, to the best knowledge of the authors, there is very few published paper directly studying the effect of the gear faults on the gear noise [24]. As a primary work, the noise signals for the cases with different gear faults are synthesized based on the previous findings about the effects of gear faults on the gear dynamics. In the following parts of this section, various gear faults are summarized and the effects of different gear faults on the gear dynamics are introduced to give the basis of synthesizing noise signals.

Summary of Various Gear Faults

Gear faults are more likely to occur in the modern geared transmission systems. This is mainly because that the geared transmission systems are designed to be lightweight, and usually work under very high speed and load conditions. Different gear faults are categorized into two groups in the ISO standard depending on how the gear faults occur [23]. Among all the different gear faults, the gear tooth wear and misalignment are two of the most frequent occurred gear faults. Wear is the gear fault which uniformly distributed over the gear tooth surface, and it is normally caused by the sliding motion between gear teeth. Misalignment is space errors, which come from the manufacture and assembly process. Therefore, in this study, psychoacoustic analyses of the noise generated by gears with wear and misalignment are performed independently.

Effect of Gear Faults on the Gear Dynamics

Gear dynamics has obtained extensive research and numerous gear models are established to study the dynamic behavior of all kinds of gears in the past decades. Among these models, the lumped parameter model of the contacting gear pair is mostly used. In this model, both pinion and gear are lumped as the mass of inertia that are coupled by mesh parameters such as mesh stiffness, damping and transmission error [24]. The occurrence of gear faults will directly affect those mesh parameters, hence the dynamic model for gear system with various gear faults can be simply formulated by changing these mesh parameters.

Dynamics of gears with different gear faults have already been widely studied and reported in previous literatures. For gear pairs with worn teeth, typically, the shaft frequency response, the gear mesh frequency and its harmonic responses can be observed in the vibration spectrum of the housing, which radiates noise into the environment. As wear increases, the magnitude of the responses at those harmonics increases gradually [29, 30]. For gear pairs with the eccentricity fault, the third gear mesh harmonic response is higher than its second harmonic response, and the responses having the gear body natural frequencies can also be observed in the spectrum [30]. Also, as eccentricity increase, the magnitudes of these frequency components increase gradually. For gear pairs with misalignment errors, the second gear mesh harmonic response is higher than the third harmonic response, and the amplitude of sidebands increases as the misalignment increases [30].

Psychoacoustic Parameters

Psychoacoustic metrics are widely used to represent human perceptions of various sounds, and actually some of the most used ones, for example the loudness and sharpness, have already been standardized for routine uses. Besides those standardized metrics, there are some other psychoacoustic metrics including the tonality, spectral centroid and Kurtosis, which are also very useful in revealing the features of the human perception of sound. In this study, five metrics are used to analyze gear noise to detect gear faults.

CASE STUDY

A one-stage helical gearbox is studied as an example to evaluate the capability of the psychoacoustic analysis for detecting gear faults, as shown in Fig.1. The parameters of the gear pair used in this model are listed in Table 1. To make an effective comparison of the cases with different gear faults, the gearbox for all the psychoacoustic analysis cases are assumed to operate in the same steady-state condition with the gear shaft rotating at 1200 round per minute (rpm). Hence, the basic frequency components, such as the gear shaft rotating frequency, fundamental gear mesh frequency and its harmonics, are all constant, and they are given in Table.2.The spectrum of the noise radiated by the helical gearbox with health gears is given as the baseline in Fig.2.

Case1: Wear

The responses at the shaft rotating frequency, gear mesh frequency (GMF) and its harmonic frequencies can be observed in the spectrum of the noise radiated by a gearbox with gear tooth wear. As the gear tooth wear develops, the magnitudes of the second and third gear mesh harmonic responses increase much more significantly as compared with the increase of the magnitudes of the noise at the shaft rotating frequency and the first gear mesh frequency [29, 30]. Based on these previous findings, the noise radiated by the example helical gearbox with increasing levels of gear tooth wear is generated and labeled from A to F. In all the synthesized noise, the magnitudes of the second and third gear mesh harmonic responses increase gradually as the gear tooth wear develops, while the magnitudes of the noise at the shaft rotating frequency and the first gear mesh frequency are kept the same. The magnitudes of the noise at different frequencies for all the cases are given in Table 3.

All the five psychoacoustic metrics introduced in the previous section were calculated, and the total metric results are shown in Fig.3. It can be easily observed that the kurtosis, total loudness, spectral centroid and sharpness of the noise increase gradually as the gear tooth wear increases. This is mainly because that the magnitudes of the high frequency components of the noise increase gradually as the wear develops. Meanwhile, it can be seen that the tonality of the noise firstly decreases to its minimum at Case D and then increases as the gear tooth wear develops. This is due to the masking effect, which will be discussed in detail later in this section. To sum up, the simultaneously increasing results of kurtosis, total loudness, spectral centroid and sharpness of a noise is one clue of detecting gear wear and the firstly decrease then increase trend of tonality is another clue of wear troubleshooting.

The specific loudness results for the cases with different levels of gear tooth wear are shown in Fig.4, where two peaks of the specific loudness can be seen.

The first peak of the specific loudness occurs at the critical band 11 for all the cases, which is contributed by the responses at the first gear mesh frequency of 1640 Hz [15]. Right before the first peak, approximately from critical band 7 to 10, step increases of the specific loudness can also be observed, while there is no signal, which can contribute to the specific loudness in these critical bands, can be found from the noise spectra. This is because that the specific loudness in the critical band 11 has a significant loudness perception effect on the neighboring critical bands. The second peak occurs at the critical band 16, which is contributed by responses at the second gear mesh harmonic frequency of 3280 Hz. As the gear tooth wear develops, both the magnitude of the second peak of the specific loudness and the critical bands it spans increase gradually. The shaft rotating frequency and the third gear mesh harmonic frequency have very little effect on the loudness results because their magnitudes are relatively low comparing with the first two gear mesh harmonics. To sum up, if the specific loudness curve has two space separated peaks and the second peak increase obviously, it indicate that wear happens.

As mentioned previously, the tonality of the noise firstly decrease and then increase as the gear tooth wear develops, which is different from the changes of all the other four parameters shown in Fig.3.The tonality of a noise is controlled by the characteristics of its tonal components, including the frequencies and the amplitude differences of the tonal contents, and the number of tones contained in a critical band. For the helical gearbox examined in this study, the 1st gear mesh frequency of 1640Hz and the 2nd gear mesh frequency of 3280Hz are followed into two different critical bands. Besides, the magnitude of the 3rd gear mesh harmonic response is much lower than the magnitudes of the first two gear mesh harmonic responses, and it has less effect on the tonality results. Therefore, the frequencies and amplitude difference between the first two harmonic responses are the main factors that define the tonality of the noise radiated by the example gearbox. The tonality of the noise for the cases from A to F is given again in Fig.5.

For the noise radiated by the gearbox without gear tooth wear, its tonality is mainly determined by the response at the first gear mesh frequency. As the gear tooth wear develops (cases from A to F), the magnitude of the second gear mesh harmonic response increases, which results in a stronger interaction between the first and second gear mesh harmonic responses. For the case D, the magnitude difference between the 1st and 2nd harmonic responses is 12 dB that results in the minimum tonality, and for cases after the Case D, the tonality is mainly determined by the second gear mesh harmonic response. As previously mentioned, the particular trend of tonality results is also a clue of identifying wear.

The wear characteristics were detailed described in different aspects by metrics results. Through combining all the information provided by the results of total loudness, sharpness, kurtosis, spectral centroid, specific loudness and tonality, the wear can be detected.

Case2: Misalignment

Similarly as the gear tooth wear cases, the responses at the shaft rotating frequency, gear mesh frequency and its harmonic frequencies will appear in the gear noise spectrum, if the misalignment occurs. Besides, the sideband (SB) responses will also appear in the spectrum of the noise generated by misaligned gears. The magnitude of the response at the second gear mesh harmonic frequency was reported to be the highest as compared with the magnitudes of other frequency components [30]. Based on this finding, the magnitudes of the synthesized gear noise at the second gear mesh harmonic frequency and its sideband responses are assumed to increase as the misalignment error increases. Meanwhile, the magnitudes of all other frequency components are assumed to be the same for cases with different levels of misalignment. The synthesized noise radiated by the example helical gearbox with increasing levels of misalignment is generated and labeled from A to G. The magnitudes of the noise at different frequencies for all the cases are given in Table 4, and the noise spectrum of the Case C is given in Fig.6.

The psychoacoustic metrics were calculated for all the cases, and the total metrics results are given in Fig.7.It can be easily seen that, from normal case to the Case G, the spectral centroid and sharpness firstly increase significantly and then level off slightly. The total loudness increases gradually as the misalignment increases, while the kurtosis increases at the beginning and then keeps almost the same. From the normal case to the Case A, the tonality decreases slightly, and then it increases gradually from signal A to G. Similarly, these particular results trends of spectral centroid, sharpness, total loudness, kurtosis and tonality are clues of detecting misalignment.

Specific loudness results for all the cases are shown in Figure

It can be observed that the first specific loudness peaks of the normal case and the cases with misalignments occur at different critical bands. The first specific loudness peak of the normal case occurs at the critical band 11, which is contributed by the responses at the gear mesh frequency of 1640 Hz. However, the first specific loudness peaks of the cases with misalignments occur at the critical band 15. This is because the second gear mesh harmonic response is the primary component for the cases with misalignments, and it actually dominates the loudness perception. It can also be seen that, as the misalignment error increases, the magnitude of the second specific loudness peak also increases gradually. This is mainly because the magnitudes of both the second gear mesh harmonic response and its sideband responses increase as the misalignment error increases. Correspondingly, the loudness of the noise generated by misaligned gears is higher than the noise generated by healthy gears. In addition, right below the critical band 15, step increased specific loudness can be seen, while there is no such frequency components in the spectrum relating to these critical band ranges, and the spectrum lost this information in view of loudness perception. To sum up, the shift of specific loudness peak to higher critical band range and the increasing specific loudness peak are clues of diagnosing misalignment.

CONCLUSIONS

In this study, psychoacoustic analyses of synthesized noise signals of an example helical gearbox with different gear faults were performed to detect the gear faults. The analyses reveal that psychoacoustics metrics, including loudness, sharpness, tonality, spectral centroid and Kurtosis, can be used for diagnosing gear faults of the studied helical gear box. Total psychoacoustics metrics results can evaluate the overall sound properties and assist gear design optimization by making balance between different metrics. From all the metrics examined, loudness and tonality are the best indicators of the gear faults of this particular helical gear box studied.

However, all the noise signals of the helical gearbox are synthesized based on the conclusions drawn in the previous publications rather than predicted or measured, and for all the cases studied a single constant gear rotating speed (1200 rpm) was assumed. Future work will build structural-acoustic models to predict the noise generated by gearbox with different types of gear faults, and to take into account the realistic operating conditions of gearboxes.

REFERENCES

[1.] Randall R.B., Vibration Based on Condition Monitoring: John Wiley and Sons Publishers, 2010.

[2.] Bechhoefer E., Kingsley M., "A Review of Time Synchronous Average Algorithms," Annual Conference of the Prognostics and Health Management Society, 2009.

[3.] Li Y, Gao Y, Guo J, et al. Gear Fault Diagnosis Based on Time Synchronous Average and Envelope Analysis[J]. Applied Mechanics and Materials, 303: 502-505, 2013.

[4.] Cheng J, Yu D, Tang J, et al, "Application of frequency family separation method based upon EMD and local Hilbert energy spectrum method to gear fault diagnosis," Mechanism and Machine Theory, 43(6): 712-723, 2008.

[5.] Wang W, Wong A K, "Some new signal processing approaches for gear fault diagnosis," Proceedings of the Fifth International Symposium on Signal Processing and Its Applications. IEEE, 2: 587-590, 1999.

[6.] Tian X, Lin J, Fyfe K R, et al, "Gearbox fault diagnosis using independent component analysis in the frequency domain and wavelet filtering," International Conference on Acoustics, Speech, and Signal Processing. IEEE, 2: II-245-8, 2003.

[7.] Yu D, Yang Y, Cheng J, "Application of time-frequency entropy method based on Hilbert-Huang transform to gear fault diagnosis," Measurement, 40(9): 823-830, 2007.

[8.] Wang W J, McFadden P D, "Early detection of gear failure by vibration analysis--ii. Interpretation of the time-frequency distribution using image processing techniques," Mechanical Systems and Signal Processing, 7(3): 205-215, 1993.

[9.] Lin J, Qu L, "Feature extraction based on Morlet wavelet and its application for mechanical fault diagnosis," Journal of sound and vibration, 234(1): 135-148, 2000.

[10.] Cheng J, Yang Y, Yu D, "The envelope order spectrum based on generalized demodulation time-frequency analysis and its application to gear fault diagnosis," Mechanical systems and signal processing, 24(2): 508-521. 2010.

[11.] Genuit, Klaus, "The sound quality of vehicle interior noise: a challenge for the NVH-engineers," International journal of vehicle noise and vibration, 1. (1): 158-168,2004.

[12.] Ingham, R., Otto, N., and McCollum, T., "Sound Quality Metric for Diesel Engines," SAE Technical Paper 1999-01-1819, 1999, doi:10.4271/1999-01-1819.

[13.] Wang X, Subic A, Ren H, et al, "Identification and assessment of vehicle seat adjuster sound quality," International Journal of Vehicle Noise and Vibration, 10(1): 51-63, 2014.

[14.] Jennings P A, Dunne G, Williams R, et al, "Tools and techniques for understanding the fundamentals of automotive sound quality," Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 224(10): 1263-1278, 2010.

[15.] Fastl H, Zwicker E, Psychoacoustics: Facts and models: Springer, 2006.

[16.] Moore BCJ, An introduction to the Psychology of hearing Bingley: Emerald, 2012.

[17.] AGMA Technical paper, "Systematic Approach for the Psychoacoustic Analysis of Dynamic Gear Noise Excitation," (American Gear Manufacturers Association, Alexandria, VA, 2012).

[18.] Brecher C, Gorgels C, Carl C, et al, "Benefit of Psychoacoustic Analyzing Methods for Gear Noise Investigation," Gear Technology, 28(5).,2011.

[19.] Kim E Y, Shin T J, Lee S K, "New tonality design for non-stationary signal and its application to sound quality for gear whine sound," Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 227(3): 311-322,2013.

[20.] Lee H H, Kim H W, Lee S K, "Sound Quality Evaluation for the Axle Gear Noise in the vehicle," Journal of the Acoustical Society of America, 123(5): 3260-3260,2008.

[21.] Lee H H, Kim S J, Lee S K, "Design of new sound metric and its application for quantification of an axle gear whine sound by utilizing artificial neural network," Journal of mechanical science and technology, 23(4): 1182-1193,2009.

[22.] Becker, S. and Yu, S., "Objective Noise Rating of Gear Whine," SAE Technical Paper 1999-01-1720, 1999, doi:10.4271/1999-01-1720.

[23.] ISO 10825:1995(E/F), Gears - Wear and damage to gear teeth - Terminology

[24.] Smith J D, Gear noise and vibration: CRC Press, 2003.

[25.] DIN 45631, Berechnung des Lautstarkepegels und der Lautheit aus dem Gerauschspektrum{Calculation of volume and loudness based on the noise spectrum}, DIN, Berlin, Germany, 1991

[26.] DIN 45692 Messtechnische Simulation der Horempfindung Scharfe, {Measurement based simulation of sharpness perception}, DIN, Berlin, Germany, 2009

[27.] Terhard, E., Stoll, G. and Seewann,M., "Algorithm for extraction of pitch and pitch salience from complex tonal signals,"Journal of the Acoustical Society of America, 71(3):679-688, 1982,

[28.] Aures,W., Berechnungsverfahren fur denWohlklang beliebiger Schallsignale--gehorbezogene Schallanalyse,{Calculation method for euphony of arbitrary signals - perception related noise analysis}Doctoral Thesis, TU Munchen, 1984

[29.] Ding, Huali, "Dynamic Wear Models for Gear Systems," PhD Dissertation, The Ohio State University, 2007.

[30.] Taylor J I, The Gear Analysis Handbook: A Practical Guide for Solving Vibration Problems in Gears: Vibration Consultants, 2000.

[31.] Chen Z, Shao Y. Dynamic simulation of spur gear with tooth root crack propagating along tooth width and crack depth [J]. Engineering Failure Analysis, 2011, 18(8): 2149-2164.

[32.] Ma R, Chen Y, Cao Q. Research on dynamics and fault mechanism of spur gear pair with spalling defect [J]. Journal of Sound and Vibration, 2012, 331(9): 2097-2109.

[33.] Ma R, Chen Y. Research on the dynamic mechanism of the gear system with local crack and spalling failure [J]. Engineering Failure Analysis, 2012, 26: 12-20.

Dong Guo and Quan Shi

Chongqing University of Technology

Peng Yi

Chongqing Academy of Science and Tech

Table 1. Parameters of a helical gearbox

Descriptions         gear   pinion

Tooth number         82     35
Module(mm)            2.75   2.75
Helical angle(deg)   10.00  10.00
Pressure angle(deg)  20.00  20.00
Teeth width(mm)      43.00  45.00

Table 2. Frequency components under 1200 rpm

Gear shaft rotating frequency(Hz)      20
Pinion shaft rotating frequency(Hz)    47
Gear mesh frequency(Hz)              1640
2nd harmonic(Hz)                     3280
3rd harmonic(Hz)                     4920

Table 3. Information of wear signals

GMF  Normal(dB)   A     B     C     D     E    F
                 (dB)  (dB)  (dB)  (dB)  (dB)  (dB)

1st  73.2        73.2  73.2  73.2  73.2  73.2  73.2
2nd  33.2        45.2  52.3  57.3  61.2  64.3  67
3rd   5.6        12.6  17.6  21.5  24.7  27.3  29.7

Table 4. Information of misalignment signals

     Normal   A     B     C     D     E     F     G
     (dB)    (dB)  (dB)  (dB)  (dB)  (dB)  (dB)  (dB)

1st  73.2    35.4  35.4  35.4  35.4  35.4  35.4  35.4
GMF
2nd  33.2    75.4  77.4  79.2  80.9  82.4  83.8  85
GMF
3rd   5.6    39.6  39.6  39.6  39.6  39.6  39.6  39.6
GMF
1st   0      14    14    14    14    14    14    14
SB
2nd   0      34    35    36    36.7  37.5  38    38.8
SB
3rd   0      16    16    16    16    16    16    16
SB
COPYRIGHT 2016 SAE International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2016 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Guo, Dong; Shi, Quan; Yi, Peng
Publication:SAE International Journal of Passenger Cars - Mechanical Systems
Article Type:Report
Date:Jun 1, 2016
Words:4213
Previous Article:Long-haul truck sleeper heating load reduction package for rest period idling.
Next Article:Experimental transfer path contribution study with the projected operational forces estimated from the responses.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters