International Engineering Research Journal Page No 1524-1530

Self Aligned Bearing Fault Detection Using vibration Signals

Analyzed by Spectrum analysis.
Gaikwad Rajendra Dattatraya, Prof. Khaire N.D.
rajendragaikwad355@gmail.com
(M.E. (Design) Student, RSCOE, Tathawade, Pune.)
(Prof. in Mechanical Engineering, RSCOE, Tathawade, Pune.)

ABSTRACT defects plays an important role in quality

Over the past decades there have been many inspection as well as for condition
major advances in the development of fault monitoring of the self aligned ball
diagnosis of deep grove ball bearing and bearing/machine element. This paper
roller bearing had take placed In this paper describes the vibration analysis technique to
an effort is made to study the performance of detect the defects in the self aligned ball
Self aligned ball bearings. Self aligned ball bearing. The Fast Fourier Transform (FFT)
bearings are widely used in industry from detected the frequencies of damage present
home appliances to aerospace industry. during the vibration analysis of a ball
Proper functioning of these machine bearing. This faults and their related
elements is extremely important in order to frequency is simulated with MATLAB
prevent catastrophic damages. It is software.
therefore, important to monitor the KEYWORDS
condition of the bearings and to know the Self aligned ball bearing, signal analysis
severity of the defects before they cause Spectrum analysis, FFT, MATLAB
serious catastrophic damages. Hence, the simulation.
study of vibrations generated by these
1. INTRODUCTION misalignment of the shaft relative to the
Ball bearing is the most basic component housing. Self-aligning ball bearings generate
used in machinery for various engineering less friction than any other type of rolling
applications. Most of the engineering bearing, which enables them to run cooler
applications such as textile industry, even at high speeds.
agricultural industry and food and beverage As ball bearing is most commonly
industry use self aligned ball bearings, used component in machinery, it has
which enable rotary motion of shafts apart received a great attention in the field of
from complex mechanisms in engineering condition monitoring. Even a newly
such as power transmissions, rolling mills manufactured bearing may also generate
and aircraft gas turbines. Self-aligning ball vibration due to components running at high
bearings have two rows of balls and a speeds, heavy dynamic loads and also
common sphered raceway in the outer ring. contact forces which exist between the
The bearings are insensitive to angular bearing components. Bearing defects may be
 International Engineering Research Journal Page No 1524-1530

classified as localized and distributed. The 2. LITERATURE REVIEW

localized defects include cracks, pits and The effect of vibration on perfect bearing
spalls caused by fatigue on rolling surfaces can be considerably reduced by selecting the
[7]. The distributed defects include surface correct preload and number of balls [9]. The
roughness, waviness, and off size rolling vibration monitoring technique is used to
elements. The sources of defects may be due analyze various defects in bearing and it also
to either manufacturing error or abrasive provides early information in case of
wear. Hence, study of vibrations generated progressive defects [8]. Radial vibration
by these defects plays an important role in measurement was used to capture the signals
quality inspection as well as for condition and it was found that defect bearing has a
monitoring of the ball bearing/machinery strong effect on the vibration spectra [10]. In
[2]. In order to prevent bearing failure there case of defect on the fixed ring the
are several techniques in use, such as, oil frequency spectrums generated will appears
analysis, wear debris analysis, vibration at its multiples. If the defect is located on the
analysis and acoustic emission analysis. inner ring or the ball, frequency spectrum is
Among them vibration and acoustic amplitude modulated. The more is the wear,
emission analysis [8] is most commonly higher are the amplitudes of the
accepted techniques due to their ease of components. Low speed fault simulation
application. The time domain and frequency tests were conducted with various defects on
domain analysis [3] are widely accepted for the bearing. This study gives the best
detecting malfunctions in bearings. The frequency bandwidth for early detection of
frequency domain analysis is more useful as bearing defects running at lower speeds
it identifies the exact nature of defect in the [11].
bearings. These frequencies of the ball 3. EXPERIMENTAL SETUP
bearing depend on the bearing The experimental bearing test rig is designed
characteristics and are calculated from the and fabricated to identify the presence of
relations shown below [4]. defects on a radially loaded self aligned ball
Cage defect frequency (FC ) bearing by vibration analysis technique is
[ ] shown in Figure 1. The test rig consists of a
1.5 Hp motor connected with step pulley
Inner race defect frequency (FIRD) which is connected with another step pulley.
[ ] Test bearing is mounted on shaft supported
by two support bearings. Test bearing is
Outer race defect frequency (FORD) loaded with dead weight.
[ ]

Ball defect frequency (FB)

[ ]

Fig. 1 Experimental test rig.

 International Engineering Research Journal Page No 1524-1530

3.2. Bearing type and fault creation

The bearing type used in this study is a
double row self aligned ball bearing with
bearing model 1206 series. The self aligned
ball bearing is shown in Figure 2. 1206 Self
Aligning Ball Bearing has a spherical outer
ring raceway, the balls and cage continue to
rotate, this type of bearings is suitable when
the displacement of the centers around
which the shaft and housing rotate and shaft
deflection are likely to occur. Bearing is
made of Chrome Steel.

Figure 3 Fault created on ball.

Figure 2 Double row self aligned ball

bearing
The details of the ball bearing used in the
vibration analysis is as below, Figure 4 Fault on inner race.
 Bearing: 1206 Ball Bearing 3.3. Experimentation
Experiments were carried on four sets of self
 Type: Self Aligning Bearing
aligned ball bearings, a new bearing and
 Dimensions: 30mm x 62mm x
three defective bearings, i.e., inner race
16mm/Metric
defect, outer race defect and ball defect. The
 Inner Diameter: 30mm
faults are created on outer race, inner race
 Outer Diameter: 62mm
and on one ball by laser machining. Initially
 Width: 16mm
a new self aligned ball bearing was fixed in
 Size: 30 x 62 x 16 mm
the bearing test rig and vibration signals
 Dynamic load rating C: 15,600 N
were captured using FFT analyzer. Then the
 Static load rating Co: 4,650 N
new bearing is replaced with the three
defective bearings for capturing the
vibration signals. Recording of signals were
observed for three running conditions for
each bearing. The central radial load on the
 International Engineering Research Journal Page No 1524-1530

ball bearing was set to 177N, 216N and Table 1 Fault frequency (analytical and
255N. The speed variation is done for experimental)
1090rpm 1820 rpm and 3025rpm.
Speed Theoretical Experimental
4. RESULTS AND DISCUSSIONS (N) outer race Fault freq.
The vibration analysis or condition (rpm) faulty freq. using FFT
monitoring is based on the principle that all BPFo (Hz)
systems produce vibration. When a bearing
1090 125 143
is running properly, the vibrations generated
are very small and generally constant. But, 1809 255 284
due to some of the dynamic processes that 3015 295 306
act in the machine, defects develop causing
the changes in the vibration spectrum.
Firstly, Vibration signals collected in the
form of time domain are converted into
frequency domain by FFT analyzer. Fast MATLAB has used to develop the particular
Fourier Transform (FFT) on each of the four algorithm which makes it easy to find the
bearings. The frequency domain signals and faulty frequencies. The power spectrum
amplitude of vibration at predominant code has also generated in the MATALAB.
frequencies are considered for the analysis. The following steps were used to generate
Vibration signals of a new bearing and the MATLAB program.
defective bearings for a radial load of 177N,  Code for a clean signal
216N and 255N. at 1090rpm 1820 rpm and  Adding noise to clean signal
3025rpm.are shown in Figure 5, 6and 7. The  Calculating and Adding faulty
fundamental frequency theoretically frequencies
calculated for the inner race defect bearing Clean signal are generated using MATLAB
from equation (1) is found to be 143.71Hz, program. Next step was to add noise to the
239.96Hz, and 398.84Hz. The experimental clean signal then small amount of noise has
frequency spectrum of the vibration signals added in clean signal, to add this noise array
for the inner race defective bearing (fir) function and random noise is used. Program
shows higher peaks at 100Hz and 275Hz output clearly shows the clean and noisy
and 300 Hz (Figure 7,8 and 9) compared to signals. MATLB also helps to calculate the
new bearing frequency spectrum peaks. The all faulty frequencies at same time using
frequency spectrum of the vibration signals properties of the bearing. The program has
for the outer race defective bearing shows developed using „for loop‟ properties to
peaks at 125Hz, 255Hz and 295Hz which calculate and add faulty frequencies i.e.
are closely matches with the fault frequency (BPFo, BPFi, BSF and FTF). One of the
of outer race defect from equation (2) is output of MATLAB has shown in Fig. 10
143.71Hz., 284.691Hz and 306.98Hz and 11
 International Engineering Research Journal Page No 1524-1530

Figure 5 Vibration for new bearing at

1090rpm Figure 7 Vibration for inner race defect at
255N and 1090rpm

Figure 6 Vibration for new bearing at

1820rpm Figure 8 Vibration for inner race defect at
255N and 1820rpm

Figure 6 Vibration for new bearing at

3025rpm
Figure 9 Vibration for inner race defect at
255N and 3025rpm
 International Engineering Research Journal Page No 1524-1530

REFERENCES
[1]Purwo Kadarno & Zahari Taha, Tatacipta
Dirgantara, Kimiyuki Mitsui, (2008)
“Vibration Analysis of Defected Ball
Bearing using Finite Element Model
Simulation”, Proceedings of the 9th Asia
Figure 10 Time domen MATLAB output for Pasific Industrial Engineering &
inner race defect at 1820rpm Management Systems Conference,
Indonesia, Vol. 1,pp2832-2840.
[2] Andy C.C. Tan, Katie L, McNickle and
Danieal L. Timms, (2003) “A practical
approach to learning vibration condition
monitoring”. Journal of World
Transactions on Engineering and
Figure 11 Time domen MATLAB output for Technology education, Vol.2, No.2, pp
217-220.
inner race defect at 1820rpm
[3] Zeki Kiral, Hira Karagulle, (2003)
5. CONCLUSIONS “Simulation and analysis of vibration
The vibration response of new and defect signals generated by rolling element
self aligned ball bearing is compared. The bearing with defects”. Journal of
Fast Fourier Transform are performed on Tribology International, 36, pp. 667-678.
each of the four bearings. From the vibration [4] P. N. Botsaris, D. E. Koulouritis, (2007)
data, the amplitude of vibration spectra is “A preliminary estimation of analysis
relatively small for new bearing and defect methods of vibration signals at fault
ball bearing cases, where as vibration diagnosis in ball bearing”. Proceedings
spectra is comparatively larger in both the of the 4th International Conference on
cases of load i.e., 177N, 216N and 255N.for NDT, Chania, Crete-Greece, Vol.1, pp
defects on inner race and outer race. Also 1-6.
from Figure 5 and 7, the values computed [5] N. Tandon and A. Choudhury, (1999) “A
from the frequency domain signals and review of vibration and acoustic
amplitude of vibrations for new and defect measurement methods for the detection
bearings shows the location of the fault and of defects in rolling element bearings”,
severity of the defect. The value shows that Tribology International, Vol. 32, pp469–
as the load increases, the magnitude of 80.
vibration response also increases. [6] Attel Manjunath, D.V.Girish, (2010)
Additionally, the amplitude value for new “Characterization of short carbon fiber
bearing is lies below 5 which is a clear reinforced Polyacetal”, Proc. of
indication that no defects in the bearing, for International conference on Advanced
inner race defect the value lies between 20 Materials, Manufacturing, Management
and 25, it shows the moderate defect in the and Thermal Sciences. Tumkur,
bearing, where as for inner and outer race Karnataka, India, Vol.1, pp79-85.
defects the amplitude value exceeds 25, [7] Sadettin Orahan, Nizami Akturk, Veli
which shows larger indication of damage on celik, (2006) “Vibration monitoring for
inner and outer race ways. Hence amplitude defect dignosis of rolling element
value shows the state of the bearing. bearings as a predictive maintenance
tool: Comprehensive case studies”,
 International Engineering Research Journal Page No 1524-1530

Journal of Non Destructive Testing &

Engineering International. Vol.39, pp
293-298.
[8] M. Amarnath, R Shrinidhi, A.
Ramachnadra and S.B. Kandagal, (2004)
“Prediction of defects in antifriction
bearings using vibration signal analysis”,
Journal of IEI, Vol. 85, pp 88-92.
[9] R. K. Purohit and K. Purohit, (2006)
“Dynamic analysis of ball bearings with
effect of preload and number of balls”,
International Journal of Applied
Mechanics and Engineering, Vol. 11,
No. 1, pp 77-91.
[10] H. Mohamadi Monavar, H. Ahmadi and
S.S. Mohtasebi, (2008) “Prediction of
defects in roller bearings using vibration
signal analysis”, World Applied Science
Journal. Vol.4(1). pp 150-154.
[11] Eric Y.Kim, Andy C.C.Tan, Bo-suk
Yang and Vladis Kosse, (2007)
“Experimental study on condition
monitoring of low speed bearings: Time
domain analysis”, 5th Australasian
Congress on Applied
Mechanics,Brisbane, Australia. pp 1-7.