CHAPTER 3 DEFECT IDENTIFICATION OF BEARINGS USING VIBRATION SIGNATURES

33 CHAPTER 3 DEFECT IDENTIFICATION OF BEARINGS USING VIBRATION SIGNATURES 3.1 TYPES OF ROLLING ELEMENT BEARING DEFECTS Bearings are normally classified into two major categories, viz., rotating inner race bearings and rotating outer race bearings. In most of the rotating machines, the outer race is fixed and the inner race rotates along with the shaft. In this chapter, such rotating inner race bearing is considered for study. The defects in the rolling element bearings may occur in inner race, outer race, cage and, or rolling element. The typical defects in the bearing are given in Figure 3.1. Rolling-element defect is the most common defect that causes most machinery failure. Figure 3.1 Types of defects in bearings (a) Defect in outer race, (b) Defect in inner race, (c) Defect in cage (d) Defect in rolling element The unique vibration characteristics of each rolling element bearing defect make vibration analysis an effective tool for both early detection and

34 analysis of faults. The specific fault frequencies of the bearing depend on the type of defect, the bearing geometry and the speed of rotation. In this chapter, vibration signatures of good bearings and defective bearings are obtained using the experimental facility created. Next, there is a discussion of how these signatures can be used for defect identification are discussed. The geometry of the rolling element bearing used for testing is shown in Figure 3.2.. Figure 3.2 Geometry of rolling element bearing 3.2 EXPERIMENTAL FACILITY An experimental facility, exhibited in Figure 3.3, has been developed to obtain the vibration signatures of bearings. Figure 3.3 Experimental test rig 1-Accelerometer, 2- Steady rest, 3- Roller bearing, 4- Shaft, 5- Bearing support, 6- Lathe bed, 7- Vibration analyzer, 8- Computer, 9- Chuck,1- Head stock.

35 The experimental facility consists of a precision tool room lathe, shaft and bearing system with necessary instrumentation. The tool room precision lathe is a versatile machine tool used in almost all the manufacturing industries. Such a lathe has been chosen for conducting experiments to achieve variable speed drive; to have exact co-axial set up. Shafts are supported by steady rest which sits on a rigid bed. A mild steel solid shaft of 35 mm diameter and 45 mm length is used between the bearings to accommodate the bearing. Five roller bearings (SKF N37) with 35 mm inner diameter are used for testing. Two bearings are defect-free. The other three bearings contain defects in inner race, outer race and rolling element respectively as shown in Figure 3.1. The defects in the bearings are artificially created using the electrical discharge machining (EDM). The size of the defect created is 1.2 mm width and.8 mm depth on inner race and outer race. The specification of the bearing used in the present study is given in Table 3.1. Table 3.1 Specification of roller bearing Parameter Value Make SKF N37 Number of rollers 11 Outer diameter, mm 8 Inner diameter, mm 35 Pitch diameter, mm 57.5 Roller diameter, mm 11 Contact angle, ß The measuring instrument used for vibration measurement is a Piezoelectric accelerometer (Model:CTC, Type AC 12-A made by Czech Republic, with serial number 6676). Vibration signals are analysed by using a dual channel vibration analyser (ADASH -43 VA3) made by (Czech

36 Republic). The analyser can analyse up to 16 spectral lines. However, 4 spectral lines are sufficient enough to capture necessary features. Beyond 4 spectral, no further improvement in the processed results is obtained. Thus, 4 spectral line setting is used for taking all the readings. For 4 spectral lines settings, the sampling the sampling frequency is 124 Hz and signal length is one second. In each case, four data are captured and the average value of these four data is taken as the output. With 4 spectral line settings, 8192 samples are taken for the time domain analysis. These ranges fall with in the range of data given by most of the earlier researchers, In a nutshell, the following are the instrument and its settings with which all the measurements are carried out. Piezoelectric accelerometer (CTC, Type: AC12-A, S.No. 6676) Dual channel vibration analyzer (ADASH 43 VA3/ Czech Republic make) Number of samples for time domain measurement : 8192 Sampling frequency : 124 Hz Signal length : one second Number of spectral lines : 4 Number of averaging : 4 3.3 EXPERIMENTAL PROCEDURE Two defect-free bearings are tested under no load condition to obtain reference signatures. Before conducting the experiment, the lathe spindle vibrations are measured and verified to check the ovality and misalignment. After correcting misalignment and ovality (if any), the defectfree bearings are carefully fitted into the shaft at specified distance (45 mm). One end of the shaft is firmly fixed in the lathe chuck and the bearings are

37 rigidly supported on the two steady rests shown in Figure 3.3. The setup is run for 15 minutes to stabilize the vibration. An accelerometer having magnetic base is directly mounted over the bearing support to acquire the vibration signals. The accelerometer output signal is directly fed into the dual channel vibration analyzer and is stored as vibration signatures. Time domain and frequency domain signals are acquired at three different speeds. The stored data in the vibration analyzer is retrieved through RS232 cable connected to the computer for further analysis using DDS 27 software. After obtaining necessary signatures with the defect-free bearings, the defect -free bearing at the tailstock end is replaced with defective bearing (one by one) and the experiments are repeated. 3.4 VIBRATION SIGNATURES OF BEARING WITH DEFECT- FREE BEARING Figure 3.4 represents the amplitude of vibrations obtained with time, in the time domain mode and the RMS velocity of the vibrations at various frequencies during the frequency mode. The data shown are obtained at 74 rpm, 115 rpm and 16 rpm, which are the speed settings available at the test facility. In the case of time domain signals, with an increase in time, the amplitudes vary between ±.5 µm, ±.1 µm and ±.13 µm at 74, 115 and 16 rpm respectively. The wave forms are repeated overtime without many deviations. In the frequency domain plots vide Figure 3.4 (b 1 -b 3 ), at all the three speeds, the values of RMS velocities obtained are less than.4 mm/s, which indicate that the bearings are defect-free. This means that the minute imperfections during the manufacturing process give low amplitude vibration

38 signatures which cannot be eliminated. Non-impact type bearing signals, such as pure tone/waviness types, generate only low-frequency signals. Figure 3.5 represents the denoised time domain and envelope spectrum of defect-free bearing. 3.5 VIBRATION SIGNATURES OF BEARING WITH OUTER RACE DEFECT 3.5.1 Fault Frequency Calculation and de-noising the signal In the case of a defective bearing, the interaction of the defect in the rolling element bearing produces pulses of very short duration. This happens when the defect strikes the rotational motion of the system. These pulses excite the natural frequency of the bearing elements, resulting in the increase in the vibration energy at these high frequencies. With a particular defect on a bearing element, an increase in the vibration energy at element rotational frequency may occur. This defect frequency can be calculated from the geometry of the bearings and element rotational speed as given in Table 3.2. For the present bearing geometry (shown in Table 3.1) and the speeds at which the bearings are tested, such fault frequencies are calculated and the values obtained are listed in Table 3.3. For de-noising the raw signal, an adaptive filter is used. It is a filter that self-adjusts its transfer function according to an optimization algorithm driven by an error signal. Because of the complexity of the optimization algorithms, most adaptive filters are digital filters. The adaptive filter uses feedback in the form of an error signal to refine its transfer function to match the changing parameters. The error signal or cost function is the difference between the desired and the estimated signal. Signal-to-noise ratio (SNR) is a measure used in science and engineering to quantify how much a signal has been corrupted by noise. It is defined as the ratio of signal power to the noise

39 power corrupting the signal. A ratio higher than 1:1 indicates more signal than noise. SNR ratio after de-noising obtained are 2.6 or more. (a 1 ) 74 rpm (b 1 ) 74 rpm (a 2 ) 115 rpm (b 2 ) 115 rpm (a 3 ) 16 rpm (b 3 ) 16 rpm Figure 3.4 Vibration signatures of defect-free bearing (a 1 -a 3 : Time wave form; b 1 -b 3 : Frequency spectrum)

4 1.6 Amplitude [ m].5 -.5 RMS velocity [m m/s].4.2-1 Time [ms] Frequency [Hz] (a 1 ) 74 rpm (b 1 ) 74 rpm Am plitude [ m] 1.5 -.5-1 RM S v eloc ity [m m /s ].1.8.6.4.2 Time [ms] Frequency [Hz] (a 2 ) 115 rpm (b 2 ) 115 rpm 1.1 A m plitude [ m].5 -.5 RMS velocity [m m/s].8.6.4.2-1 Time [ms] Frequency [Hz] (a 3 ) 16 rpm (b 3 ) 16 rpm Figure 3.5 Vibration signatures of defect-free bearing (a 1 -a 3 : Denoised time wave form; b 1 -b 3 : Envelope spectrum)

41 Table 3.2 Fault frequency equations (Alfredson and Mathew 1985) Ball-Pass Frequency for Outer-Race Ball-Pass Frequency for Inner-Race BPFO n BD BPFO fr 1 cos 2 PD BPFI n BD BPFI fr 1 cos 2 PD Ball-Rotational Frequency BRF 2 PD BD BPF fr 1 cos BD PD 2 Fundamental Train 1 BD FTF BPF fr 1 cos Frequency 2 PD where n- Number of balls or rollers PD- Pitch diameter of bearing f r - Rotational frequency - Contact angle BD- Rolling element diameter Table 3.3 Fault frequencies at various speeds Speed, rpm Category 74 115 16 BPFO (Hz) 54.56 85.24 118.6 BPFI (Hz) 8.81 125.58 174.72 BRF (Hz) 31.5 48.26 67.15 3.5.2 Time Domain Signal The time domain signals of before de-noising bearing with outer race defect at different speeds are indicated in Figure 3.6 (a 1 -a 3 ). The time wave form indicates the severity of vibrations for defective bearings. The peak to peak vibration amplitude of bearing with outer race defect from time

42 wave form is around ±.15 m at 74 rpm, ±.6 m at 115 rpm and ± 1.1 m at 16 rpm. These amplitudes are comparatively 19 %, 489 % and 775 % higher when compared to the defect-free bearing at 74, 115 and 16 rpm, shown in Figure 3.6 a 1 -a 3 respectively. (a 1 ) 74 rpm (b 1 ) 74 rpm (a 2 ) 115 rpm (b 2 ) 115 rpm (a 3 ) 16 rpm (b 3 ) 16 rpm Figure 3.6 Raw vibration signatures of bearing with outer race defect (a 1 -a 3 : Time wave form; b 1 -b 3 : Frequency spectrum)

43.2.6 Am plitude [µm ].15.1.5 -.5 -.1 RMS Velocity [mm/s].4.2 54,.4 11,.36 279,.21 386,.36 846,.23 -.15 -.2 Time [ms] Ferequency (a 1 ) 74 rpm (b 1 ) 74 rpm RM S velocity [m m /s].4 84,.36 17,.32 765,.31.3.2.1 Frequency [Hz] (a 2 ) 115 rpm (b 2 ) 115 rpm 2.5 Amplitude [µm] 1-1 RMS velocity [mm/s].4.3.2.1 354,.375 236,.358 59,.34-2 Time [ms] Frequency [Hz] (a 3 ) 16 rpm (b 3 ) 16 rpm Figure 3.7 Vibration signatures of bearing with outer race defect (a 1 -a 3 : Denoised time wave form; b 1 -b 3 : Envelope spectrum)

44 3.5.3 Frequency Spectrum Frequency spectrums obtained with the outer race defect are shown in Figure 3.6 (b 1 -b 3 ). It becomes clear that significantly larger RMS velocity values are noticed with the defective bearing when compared to that of defect free bearing shown in Figure 3.4 (b 1 -b 3 ). At 74 rpm, the maximum RMS velocity of defect-free bearing is.26 mm/s whereas that for the outer race defective, it is nearly.69 mm/s at 55 Hz, i.e. about 35 times larger. Similarly, 1 times and 9.5 times larger RMS velocity values at 85 Hz and 118 Hz are obtained when the speeds are 115 rpm and 16 rpm respectively with the defective bearing. Higher amplitude signatures clearly indicate that the bearing is a defective one. Owing to this defect, the amplitudes that are larger can be identified by noticing the peaks of the frequency domain signals. At 74 rpm, from Figure 3.6 b 1, the peaks occur at 55, 11, 165, 22, 275, 33, 385, 44, 495 Hz and so on. From Table 3.3, the values of BPFO at 74 rpm match with the peak amplitude frequency. Thus, the obtained peak value frequency is near the harmonics of the BPFO values as shown in Table 3.4. It is evident that the defect is due to a fault in the outer race. Similar observations are discernible at 115 rpm and 16 rpm as well. Figure 3.7 shows the after de-noising signal and envelope spectrum of the outer race defect bearing. From the envelope spectrum peak amplitudes at 54, 84 and 119 Hz and its harmonics are noted. It falls in line with the raw signal and is very close to the theoretical fault frequency.

45 Table 3.4 Comparison of fault frequencies (outer race defect) Sl. No Type of defect Speed (rpm) Theoretical frequency (Hz) Experimental frequency (Hz) 1 74 54.56 55, 11, 165, 22and 275, Outer race defect 2 115 85.24 85, 17, 255, 34 and 425, (BPFO) 3 16 118.6 118, 236, 354, 472 and 59, 3.6 VIBRATION SIGNATURES OF BEARING WITH INNER RACE DEFECT 3.6.1 Time Domain Signals The time domain signals of before de-noising the bearing with inner race defect at different speeds are shown in Figure 3.8 (a 1 -a 3 ). The time wave form indicates the severity of vibrations for defective bearings. The peak to peak vibration amplitude of bearing, with inner race defect from time wave form, is around.9 m at 74 rpm,.9 m at 115 rpm and 2.2 m at 16 rpm. These amplitudes are 7 %, 35 % and 816 % higher when compared to the defect-free bearing at 74, 115 and 16 rpm respectively. 3.6.2 Frequency Spectrum Frequency spectrums obtained with the inner race defective bearings are illustrated in Figure 3.8 (b 1 -b 3 ). Significantly, larger RMS velocity values are deciphered in defective bearing when compared to RMS velocity of defect free bearing signatures shown in Figure 3.4 (b 1 -b 3 ). At 74 rpm, the maximum RMS velocity of defect-free bearing is.26 mm/s, whereas for the inner race defect is nearly.356 mm/s at 83 Hz, i.e. about 13.69 times larger. Similarly, 38.61 times and 34.1 times larger RMS velocity

46 values at 123 Hz and 176 Hz are obtained at 115 rpm and 16 rpm respectively with the defective bearing. (a 1 ) 74 rpm (b 1 ) 74 rpm (a 2 ) 115 rpm (b 2 ) 115 rpm (a 3 ) 16 rpm (b 3 ) 16 rpm Figure 3.8 Vibration signatures of bearing with inner race defect (a 1 -a 3 : Time wave form; b 1 -b 3 : Frequency spectrum)

47 (a 1 ) 74 rpm (b 1 ) 74 rpm (a 2 ) 115 rpm (b 2 ) 115 rpm (a 3 ) 16 rpm (b 3 ) 16 rpm Figure 3.9 Vibration signatures of bearing with inner race defect (a 1 -a 3 : Time wave form; b 1 -b 3 : Envelope spectrum) A higher amplitude signature clearly indicates that the bearing is a defective one. This is because the amplitudes of larger signatures can be identified by noticing the peaks of the frequency domain signals. At 74 rpm,

48 from Figure 3.8 b 1, the peaks frequency distribution occurs at 82, 161, 241, 323, 397, 482, 81, 979, 12 Hz and so on. From Table 3.3, it is understood that the values of BPFI at 74 rpm match with the peak amplitude frequency. Thus, the obtained peak value frequency is near the harmonics of the BPFI values as vide in Table 3.5. Thus it is evident that the defect is due to outer race. Similar observations are noticed at 115 rpm and 16 rpm as well. Figure 3.9 shows the after de-noising the time signal and envelope spectrum of the inner race defect bearing. From the envelope spectrum peak amplitudes at 83, 123 and 176 Hz and its harmonics are noted. It falls in line with the raw signal and is very close to the theoretical fault frequency. Table 3.5 Comparison of fault frequencies (inner race defect) Sl. No 1 Type of defect Speed (rpm) Theoretical frequency (Hz) 74 8.81 Inner race 2 115 125.58 Defect (BPFI) 3 16 174.72 Experimental frequency (Hz) 82,161,241,323, 397and 482, 123, 252, 374,748, 5, 623, 748 and 878 176, 352, 528, 699 and 876 3.7 VIBRATION SIGNATURES OF BEARING WITH ROLLING ELEMENT DEFECT 3.7.1 Time Domain Signals The time domain signals of bearing before de-noisingwith rolling element defect at different speeds are shown in Figure 3.1 (a 1 -a 3 ). The time wave forms indicate the severity of vibrations due to the defect in the bearing. The peak to peak vibration amplitude of bearing with rolling element defect from time wave form is around.28 m at 74 rpm,.59 m at 115 rpm and

49 2.31 m at 16 rpm. These amplitudes are 18%, 21 % and 862% higher when compared to the defect-free bearing at 74, 115 and 16 rpm respectively. (a 1 ) 74 rpm (b 1 ) 74 rpm (a 2 ) 115 rpm (b 2 ) 115 rpm (a 3 ) 16 rpm (b 3 ) 16 rpm Figure 3.1 Vibration signatures of bearing with rolling element defect (a 1 -a 3 : Time wave form; b 1 -b 3 : Frequency spectrum)

5.2.8 Am plitude [ m ].1 -.1 RMS Velocity [mm/s].6.4.2 62,.5 183,.38 38,.35 427,.32 549,.31 67,.3 -.2 Frequency [Hz] Time [ms] (a 1 ) 74 rpm (b 1 ) 74 rpm.6.4 A m plitude [ m ].4.2 -.2 RM S v eloc ity [m m /s].3.2.1 94,.28 282,.2 376,.23 846,.23 -.4 Time [ms] Frequency [Hz] (a 2 ) 115 rpm (b 2 ) 74 rpm 2.4 Am plitude [µm ] 1-1 RMS velocity [mm/s].3.2.1 131,.23 393,.226 786,.23-2 Time [ms] Frequency [Hz] (a 3 ) 16 rpm (b 3 ) 16 rpm Figure 3.11 Vibration signatures of bearing with rolling element defect (a 1 -a 3 : Time wave form; b 1 -b 3 : Envelope spectrum)

51 3.7.2 Frequency Spectrum Frequency spectrums obtained with the rolling element defective bearings are indicated in Figure 3.1 (b 1 -b 3 ). Significantly larger RMS velocity values are noticed in defective bearing when compared to that of defect-free bearing shown in Figure 3.4(b 1 -b 3 ). At 74 rpm, the maximum RMS velocity of defect-free bearing is.26 mm/s whereas for the defective rolling element it is nearly.675 mm/s at 63 Hz, i.e. about 25.96 times large. Similarly, 8 and 5.89 times larger RMS velocity values at 94Hz and 131 Hz are obtained at 115 rpm and 16 rpm respectively with the defective bearing. Higher amplitudes of the signatures clearly indicate that the bearing is a defective one. Owing to this defect, the amplitudes of larger signatures can be identified by noticing the peaks of the frequency domain signals. At 74 rpm, from Figure 3.6 b 1, the peak frequency occurs at 62, 124, 183, 244, 35, 366, 427, 488, 549, 61, 671, 732 Hz and so on. From Table 3.3, it is seen that the values of BRF at 74 rpm nearly match with the peak amplitude frequency. Thus, peak values of frequency obtained are near the harmonics of the BRF values vide Table 3.6. Thus it is evident that the defect is due to the rolling element. Similar observations are shown up at 115 rpm and 16 rpm as well. Figure 3.11 shows the after de-noising time domain signal and envelope spectrum of the rolling element defect. From the envelope spectrum peak amplitudes at 62, 94 and 131 Hz and its harmonics are observed at 74, 115 and 16 rpm respectively. It falls in line with the raw signal and is very close to the theoretical fault frequency.

52 Table 3.6 Comparison of fault frequencies (rolling element defect) Sl. No 1 Type of defect Speed rpm Theoretical frequency (Hz) 74 62.11 Rolling element 2 115 96.4 defect (BRF) 3 16 134.13 Experimental frequency (Hz) 62, 124, 183, 244, 35and 366, 94, 188, 282, 376 and 47, 131, 262, 393,524, 655, 786, and 917 3.8 CONCLUDING REMARKS In this study, the bearing with defects in outer race, inner race and rolling element defect has been studied. The frequency spectrum and time domain graphs are obtained and drawn for various speeds. The following conclusions are drawn from these experimental results: 1. The time wave form indicated the severity of vibration in defective bearings. Then, the amplitude of the defect free bearing and defective bearing are compared. The amplitude of vibration obtained is 2-7 times larger for the bearing with outer race defect, 3-8 times larger for bearings with inner race and 2-8 times larger for bearings with rolling element defect, when compared to the defect-free bearing. 2. Frequency spectrum helps in identifying the exact nature of the defect in bearing. From Figures 3.5, 3.6 and 3.7 (b 1 - b 3 ) the harmonics of fault frequency corresponding to outer race defect, inner race defect and rolling element defect are noticed. This is a good indication of the presence of the defect at specified bearing elements. These harmonic frequencies are very close to the theoretical fault frequencies.