Vibration Signature Analysis & condition monitoring of Tapered Roller Bearing

Vibration Signature Analysis & condition monitoring of Tapered Roller Bearing

Manoranjan Mahanta

Research Scholar: Department of Mechanical Engineering

Indira Gandhi Institute of Engineering & Technology Sarang, Dhenkanal, Odisha, India

Rabi Narayan Sethi

Assistant Professor: Department of Mechanical Engineering

Indira Gandhi Institute of Engineering & Technology Sarang, Dhenkanal, Odisha, India

AbstractA tapered roller bearing is a rolling element bearing which is used in most of the rotating machinery. These bearing support axial forces as well as radial forces. These are the crucial component in mechanical transmission systems commonly used for moderate speed, heavy duty applications where durability is required. Prognostic maintenance of the bearing plays an important role in maintaining the durability in operation, checking sudden failure, human safety, and cost minimization. Any defects occurring in any one of the components, however small it may be can cause catastrophic damage to the bearing as well as to the entire mechanical system. Therefore, it is important to develop The reliable condition monitoring & fault diagnosis method in preventing malfunctioning of the roller bearings. Vibration signature analysis and signal processing are the most important techniques used nowadays in condition based monitoring of rotating elements. This experimental analysis is focused on establishing a robust signal processing technique from which the behavior of the component can be analyzed easily. Every mechanical component has a characteristic frequency/frequencies. like every human being have a unique signature, every machine has a unique vibration signature. So if something is wrong with the machine, the vibration signature would change. Considering this idea an experiment has been carried out with tapered roller bearing having different depth of defects in different rollers and the corresponding vibration signals have been investigated using the impulse response experiment. Comparing the changes in faulty bearing with the healthy bearing, immediate actions can be taken to change the machinery condition or to replace the faulty bearing, which can avoid further damaging of the bearing parts or system

Keywords: Characteristic frequency, vibration signature, Impulse response.

1. INTRODUCTION

   A tapered roller bearing is most widely used in motor land vehicles and in heavy rotating machinery in industries. Tapered roller bearings use conical rollers which run on conical races. Most roller bearings take radial or axial loads, but Tapered roller bearing supports both radial as well as axial loads and generally can carry higher loads than ball bearings due to its greater contact area. To help ensure long bearing service life, it is most important to determine the condition of machinery and bearing while in operation. Good predictive maintenance will help reduce the machine downtime and decrease overall maintenance cost of the industry. The health of the bearings is critical to the health of the equipment and process. with proper performance monitoring, imminent

   failures can be identified and corrected. Without a good monitoring technique in place and subsequent corrective actions not being taken when required, a single bearing failure can result in full machine failure and countless hours of lost production. Bearing condition monitoring is recognized by three main human senses: sight, sound, and touch. Basic monitoring is generally carried through elemental observations. Also, there are highly sensitive tools that can amplify these observations making them noticeable and recordable. Vibration analysis is the most information-rich method available for bearing failure, to identify overall machine performance and problems. There are many condition monitoring techniques for bearings have been developed by researchers.

   1. Review

      Xiang Tian [1] (2018) developed a fault detection technique using modulation signal bi-spectrum and compared its performance against kurtogram. Since kurtogram is used for narrowband selection, it produces poor detection results. MSB was used to de-noise both stationary and discrete aperiodic noise. The high magnitude feature of MSB also enhances the modulation effects of a bearing fault and can be used to provide an optimal frequency band for fault detection. Yuh tay sheen [2] (2007) has studied impulse response of the defects impacts for a roller bearing. envelope detection was implemented for the vibration signals with amplitude modulation through the linear least square analysis under the assumptions of the stepwise function for the envelope signal. Then the signal processing of impulse response enhancer is applied to eliminate the sidebands around the defect frequencies. Yongbo Li [3] (2017) tried to separate low Q- factor components from the high Q-factor components using tunable Q-factor wavelet transform (TQWT). Characteristic frequency ratio (CFR) was used to optimize the Q-factor of TQWT. Xiuxing Yin et al [4] (2017) have analyzed the non- uniform contact loading distribution due to the combined radial and moment loads. The experimental result demonstrates that the wear and deformation of the bearing raceway can be significantly promoted by non-uniform loading due to sliding motion between asperities of the bearing contact surfaces. Fafa Chen et al [5] (2013) have developed a fault diagnosis model based on multi-kernel support vector machine (MSVM) with chaotic particle swarm optimization (CPSO). By this the optimal parameters for MSVM can be obtained with high accuracy and great

      generalization ability. N Hiremath and D.M Reddy [6] (2016) have studied to asses the surface wear of outer race of the roller bearing using degradation of grease and analysis of vibration signals. Using Fourier transform infrared radiation (FTIR) it was observed that the degradation of grease and kurtosis value is extracted from time domain signals to indicate progressive wear in bearing. M Craig [7] (2009) et al have studied a new condition monitoring technique based on electrostatic charge. It describes the condition monitoring of tapered roller bearing that incorporated electrostatic wear site sensor to identify charge during surface wear and oil line sensors to detect debris in oil line sensors to detect debris in oil scavenging lines. The demographic analysis identified the severity of wear index. I.M Jamdar et al [8] (2016) have developed a model using matrix method of dimensional analysis (MMDA) for investigation of the localized defect in a tapered roller bearing. An analytical model was developed to investigate the non-linear dynamic behavior due to cage run out and varying no of balls results into the reduction in vibration amplitude. Rajesh Kumar [9] (2013) measured the defect width of outer race in tapered roller bearing using discrete wavelet transform (DWT) of the vibration signal. By this technique, defect width over a range of (0.5776- 1.964) mm can be measured successfully. Reanaudin [10] (2010) tried to detect the fault by angular measurement of true instantaneous angular speed.

   2. FAULT DIAGNOSTICS

   A Roller bearing generally contains Rollers, inner race, cage, and outer race. These parts may damage due to ineffective lubrication, heavier loading than anticipated, improper handling or installation, incorrect shaft or housing fits. Fault diagnosis is the process of detecting, isolating, and identifying an impending or incipient failure condition of the affected component. Fault diagnosis can be done in two ways- a model-based and data-driven (Vibration analysis).

2. EXPERIMENTAL

   Fig-1 Instrumented Bearing Test Rig

   The experiment was conducted by rotating the weighted shaft at a constant speed using a motor. First two healthy tapered roller bearings having no defects fitted into the bearing housing and rotated at constant speed. The vibration signals were recorded at different instances and 10 to 20 readings of frequency value and its corresponding spectral unit values were noted down. The no of peaks in the time domain signals was also noted down. After that the faulty bearings were tested one by one having different depth of defects.

   B. Experimental Parameters

   Table 1 and 2 represents bearing specifications and input parameters for the analysis respectively.

   Table-1 Bearing Specifications

   Shaft diameter	30 mm
   Outer diameter	62 mm
   Width	14 mm
   No of Rollers	17
   Dynamic load carrying capacity (c)	43.5 KN
   Static load carrying capacity	48 KN
   Inner and Outer ring material	Steel

   Table- 2 Experimental Parameters

   1. Procedure

      The Experimental setup & the analysis was undertaken at IGIT, Sarang, Dhenkanal. Odisha, India. The set up consists of bearing housing, a bearing shaft with load, motor, pulley, & a rigid base attached to the ground.

      The sensor attached to the top of the bearing housing senses the raw vibration signals. The sensor is connected to the Data acquisition unit, which consists of an accelerometer and a data acquisition device. These devices are connected to the computer. The Labview software installed in the computer gives the impulse response data of the vibration signals in the Time domain. The impulse response curve was then converted into a frequency spectrum using fast Fourier transform analyzers & using some filters the actual signal was produced on the screen.

      Parameter	Value
      Driving Speed	1200 RPM
      Bearing No	4T-30206
      Mean Roller diameter	6.2 mm
      Sampling Rate	50HZ

3. FINDINGS OF THE EXPERIMENT

   When the defect in the rolling element strikes the rotational motion of the bearing, it produces pulses of very short duration which excites the natural frequency of the bearing elements, which results in an increase in the vibration energy at these frequencies.

   The characteristic frequencies were obtained from the frequency spectrum signal for the healthy as well as defective bearing one by one.

   Serial No	Relative depth of defect	Frequency (in Hz)	Relative Frequency
   1	1 (Healthy)	20.846
   2	0.99838797	16.1666667	0.22447152
   3	0.996774194	17.77666665	0.14723848
   4	0.995967742	17.50833333	0.160110653
   5	0.99516129	18.3133333	0.121494133
   6	0.993548387	14.6945	0.295092584

   Serial No	Relative depth of defect	Frequency (in Hz)	Relative Frequency
   1	1 (Healthy)	20.846
   2	0.99838797	16.1666667	0.22447152
   3	0.996774194	17.77666665	0.14723848
   4	0.995967742	17.50833333	0.160110653
   5	0.99516129	18.3133333	0.121494133
   6	0.993548387	14.6945	0.295092584

   Table-3

   Fig- 2 Pictorial view of the different depth of defects in Rollers

   Fig-3 Bearing cage Fig-4 A defective Roller

   1. Graphs

      Here the frequency domain analysis was carried out using Fast Fourier transform algorithms available in the software. unlike in time domain it gives easy and clear results. The main advantage of frequency domain analysis is that the repetitive nature of the vibration signal is displayed as peaks in the frequency spectrum at the frequencies where the repetition occurs. From the graph it can be seen that the characteristic frequency of healthy bearing as 20.84 Hz. As the defect size increases the natural frequency changes accordingly. Hence first noting down the vibration signature of the healthy bearing, the vibration signals of the bearing were checked periodically. The deviation from the original vibration signature indicates that the bearing condition needs to be evaluated. The level of defects can also be interpreted from the vibration signals by looking at the signature graphs.

4. CONCLUSION

In this paper the Vibration signatures of the healthy and defective Tapered Roller bearings were obtained and using the vibration analysis, it can be seen that there is a change in the signature of the Natural frequencies of the roller bearing with an increase in depth of defects of the rollers. By this indication, the fault frequencies can be compared with the healthy bearing frequencies and incipient faults can be avoided by taking necessary actions before it becomes a major defect.

Fig-5 Graph of Relative depth of defect vs Relative Frequency

1. Equations

The Relative Frequencies of the Defective bearings were calculated as-

Relative Frequency ( ) = Where = Relative Frequency

= Frequency of Healthy Bearing

= Frequency of Defective Bearing Relative depth =

REFERENCES

1. Xiang Tian, James Xi Gu, Ibrahim Rehab, Gaballa M Abdalla, Fengshou Gu, A.D Ball, A robust detector for rolling element bearing condition monitoring based on modulation signal bispectrum and its performance against Kurtogram. Mechanical Systems and signal processing 100 (2018).

2. Yuh Tay sheen, An impulse response extracting a method from the modulated signal in a roller bearing, Measurement 40 868-875 (2007).

3. Yongbo Li, Xihui Liang, Minqiang Xu, Wenhu Huang, Early fault feature extraction of roller bearing based on ICD and Tunable Q-factor wavelet transform. Mechanical systems & signal processing 86, 204- 223, (2017).

4. Xiuxing Yin, Haigang Gu, Gang Feng, Analysis of non-uniform abrasion evolution for cylindrical roller bearings. Mechanisms and Machine Theory 117, 148-159 (2017)

5. Fafa Chen, Baoping Tang, Tao sang, Lili, Multi fault diagnosis study on roller bearing based on multi-kernel support vector machine with chaotic particle swarm optimization s0263-2241 (2013).

6. Niranjan Hiremath, D. Mallikarjun Reddy, Experimental studies to asses surface wear using grease degradation, Bearing temp and the statistical parameter of vibration signals in a roller bearing. Materials Today 4, 8370-8377 (2017)

7. M.Craig, T.J Harvey, R.J.K Wood, K. Masuda, M. Kawabata, H.E.G Powerie, Advanced condition monitoring of tapered roller bearings, Part-1, Tribology International 42, 1846-1856, (2009).

8. I M Jamdar & Dr. D.P Vakharia, Model development for investigation for localized defects in Tapered roller bearing using Matrix method of dimensional analysis. Procedia Enginering 144 (2016).

9. Rajesh Kumar & Manpreet Singh, Outer race defect width measurement in Tapered roller bearing using discrete wavelet transform of vibration signal. Measurement 46 , 537-545 (2013).

10. L. Renaudin, F. Bonnardot, O.Musy, J.B Dorey, D. Remond. Natural roller bearing fault detection by angular measurement of true instantaneous angular speed. Mechanical systems and Signal Processing 24, 1998- 2011 (2010).