Engineering Failure Analysis 40 (2014) 97–113
Contents lists available at ScienceDirect
Engineering Failure Analysis
journal homepage: www.elsevier.com/locate/engfailanal
Tribological failure analysis of a heavily-loaded slow speed
hybrid journal bearing
S.M. Muzakkir ⇑, K.P. Lijesh, Harish Hirani
Department of Mechanical Engineering, Indian Institute of Technology Delhi, India
a r t i c l e
i n f o
Article history:
Received 21 December 2013
Received in revised form 7 February 2014
Accepted 13 February 2014
Available online 28 February 2014
Keywords:
Journal bearing
Magnetic bearing
Hybrid bearing
Rotor fracture
a b s t r a c t
In the present work, the feasibility of hybridizing a magnetic arrangement in the conventional journal bearing system has been experimented for the operating conditions of heavy
load and slow speed. A test setup has been developed to perform testing on four types of
bearing arrangements: conventional journal bearing arrangement, cylindrical magnetic
bearing arrangement, circular arc (180°) magnetic bearing arrangement and a hybrid
bearing arrangement. The magnetic levitation force was determined theoretically for these
arrangements to identify reasons for the mechanical contact between rotor and stator
magnets. The results of the experimental investigations in terms of the weight loss (wear),
reduction in the magnetic flux density, acceleration signals, and photographs of worn &
fractured rotor are reported.
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
The operating conditions of heavy load and slow speed are encountered in many applications, like sugar mills, cement-manufacturing plants and steam turbines, where the support bearings operate in mixed-lubrication regime. Under these conditions
the asperity contact occurs and excessive wear of bearing causes failure [1–3]. One way adopted by various researchers is to
reduce friction and wear in the mixed-lubrication regimes using different material combinations, better lubricants & additives
[4] and improving geometry (bearing clearance [5], grooving [6], texturing [7]). Other approach adopted is to hybridize the different bearing technologies, like hydrodynamic bearing and hydrostatic bearing [8], electromagnetic bearing and permanent
magnetic bearings [9], hydrodynamic and permanent magnetic bearing [10,11], etc. It has been investigated earlier that the
hybridization of magnetic bearing and hydrodynamic bearing is feasible for several applications.
In the present work, experiments have been conducted to determine the feasibility of hybridized bearing (magnetic and
journal bearings). The experiments were conducted in four phases. In the first phase experiments were conducted on conventional journal bearings under the operating conditions of heavy load and slow speed. The wear of the bearing was measured as its weight loss after the test. In the second phase the experiments were conducted on cylindrical magnetic bearing
arrangement under the similar operating conditions of heavy load and slow speed and it was observed that the arrangement
is not capable of supporting the dynamic load. In the third phase the experiments were conducted on circular arc (180°)
magnetic bearing arrangement under the similar operating conditions and it was observed that even though the magnetic
arrangement was able to separate the journal in static condition, but rotor experienced vibrations under dynamic conditions
and finally resulted into the breakage of the rotor magnets. Significant wear was also observed on the surface of stator
⇑ Corresponding author. Tel.: +91 98 100 24335.
E-mail addresses: mez108659@mech.iitd.ac.in, smmuzakkir@jmi.ac.in (S.M. Muzakkir).
http://dx.doi.org/10.1016/j.engfailanal.2014.02.016
1350-6307/Ó 2014 Elsevier Ltd. All rights reserved.
98
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
Nomenclature
r1
r2
r3
r4
Br1
Br2
l0
Fr
e
a
b
z
L
g
N
S
K
inner radius of rotor magnet, mm
outer radius of rotor magnet, mm
inner radius of stator magnet, mm
outer radius of stator magnet, mm
magnetic remanence, Tesla
magnetic remanence, Tesla
permeability of free space, H/m
radial force, N
eccentricity ratio
angular variable of rotor, Radian
angular variable of stator, Radian
axial offset, m
length of cylindrical magnet, m
dynamic viscosity, Pa s
journal speed, rpm
Sommerfeld number
specific film thickness
magnet. In the fourth phase experiments were conducted on the hybrid bearing arrangement. Significant wear of both the
rotor and stator was observed in this arrangement with a consequent reduction in the magnetic strength.
The reasons of the bearing arrangement failures were analyzed and the conclusions drawn are reported.
2. Experimental details
The schematic diagram of the experimental setup is shown in Fig. 1a. The provisions have been made in the setup to
conduct experimental studies on four types of bearing arrangements:
1.
2.
3.
4.
Conventional journal bearing.
Cylindrical magnetic bearing arrangement.
Circular sector (180°) magnetic bearing arrangement.
Hybrid bearing arrangement.
These arrangements are shown in Figs. 1b–1e. The main experimental setup, shown in Fig. 1a, consists of a stainless steel
(grade 303) shaft whose one end is free and other is connected, using a spiral coupling, to an induction motor (AC, 3 phase,
1.5 kW). The motor is rigidly mounted on a base plate and is controlled by an ABB frequency drive (IP20/ll open type). The
loading arrangement consists of a horizontal loading platform supported on linear bearing mounted on vertical slide ways of
circular cross section. A maximum of 500 N load may be applied on the bearing. A (shielded) deep groove ball bearing
transfers the load from the loading platform to the shaft. Two accelerometers (type: KS 76C-10, voltage sensitivity:
Detailed View Figure 1(b)
Detailed View Figure 1(c)
Detailed View Figure 1(d)
Detailed View Figure 1(e)
Fig. 1a. Schematic diagram of the main experimental setup.
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
99
Fig. 1b. Journal bearing test setup.
Fig. 1c. Cylindrical magnetic bearing test setup.
Fig. 1d. Circular sector (180°) magnetic bearing test setup.
1.032 mV/m/s2) are connected at the top and side of the bearing housing. The data is acquired using NI USB-4432(24 bit) and
stored in computer using LabVIEW interface. The detailed description of the experimental setups shown in Figs. 1b–1e are
given in the corresponding sections.
The photographs of the experimental setup are shown in Fig. 2a and b.
The description of the experimental setup, operating conditions, experimental results and corresponding discussion for
the four different bearing arrangements are given in the following sections.
100
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
Fig. 1e. Hybrid magnetic bearing test setup.
Loading platform
Vertical slideways
Linear Bearing
Motor
Spiral coupling
Proximity sensor
Lubricant tank
(a)
Accelerometers
(b)
Fig. 2. (a) Left side view of the experimental setup. (b) Right side view of the experimental setup.
Fig. 3a. Journal.
2.1. Experiments on conventional journal bearing
2.1.1. Description of experimental setup
The experimental setup for the conventional bearing arrangement consists of a stainless steel journal fitted at the free end
of the shaft and supported in the bearing as shown in Fig. 3a. The bearing is assembled in a housing, which is mounted rigidly
on the base plate. A lubricant tank is provided with heater and adjustable thermal cut-off switch that maintains the lubricant
at a set temperature. The bearing assembled in the housing is shown in Fig. 3b. The experimental setup is shown in Figs. 1a
and 1b.
2.1.2. Experimental operating conditions
The operating conditions of the conventional journal bearing were decided so as to correspond to that of a heavily loaded
slow speed journal bearing. The Sommerfeld number of one such application, namely a sugar mill bearing, is in the range of
0.0010–0.0026 [12]. The eccentricity ratio corresponding to this Sommerfeld number is 0.99 [13]. It was decided to maintain
the same Sommerfeld number for the test bearing. SM175, a commonly used lubricant in the operating conditions of heavy
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
101
Fig. 3b. Bearing assembled in the housing.
load and low speed, was used in the experiment. It was maintained at a temperature of 40 °C. The dynamic viscosity was
determined to be 0.10025 Pa s at 40 °C. A phosphor bronze bearing with inner diameter 50 mm and length 30 mm was used
for conducting experimental investigations. In order to obtain same Sommerfeld number for the test bearing, a suitable
radial clearance (for the journal bearing) is to be decided considering the dynamic viscosity of lubricant, dimensions of
the test bearing, load and journal speed. A radial clearance of 0.25 mm (0.1% of journal radius) will require a load of
2891 N at a journal speed of 5 rpm. But due to the limitation of the test setup in respect of the maximum applied load
(500 N), the radial clearance was increased to 1.00 mm to obtain a reduced load of 180 N at 5 rpm. However, the shaft could
not be rotated at 5 rpm and it became possible to rotate the journal only at 27 rpm and a load of 373 N corresponding to the
Sommerfeld number of 0.0068. The experiment was thus conducted under these operating conditions for 6 h [14] so as to
obtain measureable wear of the bearing. The wear of the bearing was measured as its weight loss on a precision weighing
balance. The bearings were cleaned using toluene before the weight measurement, both before and after the test.
2.1.3. Results and discussion
The journal bearing (journal diameter: 50 mm, bearing inner diameter: 48 mm, radial clearance: 1 mm, length: 30 mm)
was operated at heavy load and slow speed conditions (load: 373 N, journal speed: 27 rpm, lubricant temperature: 40 °C) for
6 h duration. The weight loss of the bearing after the 6 h test was measured to be 60 mg. This amount to a reduction of
0.033% weight of the bearing and is considered a significant amount of wear which indicates that the journal remained in
contact with the bearing under these operating conditions. The heavy load and slow speed conditions prevented the operation of the journal bearing in the hydrodynamic lubrication regime. The theoretical analysis was then carried out to verify
the operative lubrication regime of the journal bearing. The lift-off speed [15] was determined to be 529 rpm, corresponding
to a specific film thickness parameter, K ¼ 3, which is much higher than 27 rpm at which the experiment was conducted.
This confirms that the journal bearing operated in the mixed lubrication regime.
Since the operation of the journal bearing in mixed lubrication regime results in significant wear, therefore in order to
separate the journal from the bearing, the magnetic (passive) levitation may be employed. The magnetic force (F) between
two magnets is inversely proportional to square of the distance (d) between them and is expressed as F / d12 [16]. It may be
inferred from this expression that an infinite large repulsive magnetic force may be obtained when the two magnets with
same polarity approach towards each other i.e. d ? 0. If this principle is used then the magnetic bearing may be used to
separate the journal from coming in contact with bearing. In order to explore this option and to analyze the feasibility of
the magnetic bearing, the experiments have been conducted on the magnetic bearing arrangements, the details of which
are given in the next sections.
2.2. Experiments on cylindrical magnetic bearing arrangement
2.2.1. Description of experimental setup
The test setup for the cylindrical magnetic bearing arrangement consists of rotor made up of four neodymium magnetic
discs having inner diameter of 10 mm, outer diameter of 48 mm and thickness of 8 mm as shown in Fig. 4a. The rotor is push
fitted on the stainless steel shaft and acts as journal. The bearing stator is made of two cylindrical magnets having inner
diameter of 50 mm, outer diameter of 100 mm and thickness of 15 mm as shown in Fig. 4b. The stator magnets are assembled in the bearing housing as shown in Fig. 1c. The housing is rigidly mounted on the base plate.
2.2.2. Experimental operating conditions
The operating conditions of the experimental setup employing cylindrical magnetic bearing arrangement were kept similar to the one used for the conventional journal bearing setup so as to obtain direct comparison of their performances. The
rotor was subjected to a load of 373 N and rotated at 27 rpm. The radial clearance was taken to be 1 mm.
102
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
Fig. 4a. Rotor.
Fig. 4b. Cylindrical magnet (stator) assembled in the bearing housing.
2.2.3. Results and discussion
The cylindrical magnets have axial magnetization therefore a radial repulsion force is developed between the rotor and
stator. It was observed on the application of the load of 373 N that direct contact occurred between the stator and rotor magnets. The magnetic arrangement was thus unable to levitate the rotor at a load of 373 N. The load was then reduced so as to
determine the maximum load which this magnetic arrangement could carry without allowing contact between the stator
and rotor, however it was found that even at a reduced load of 60 N the rotor and stator magnets remained in contact with
each other and the magnetic arrangement was unable to levitate the rotor as shown in Fig. 5a and b.
To understand the reasons for the contact between rotor and stator, a theoretical estimation of the magnetic levitation
force was carried out. The theoretical radial repulsive force between axially polarized stator and rotor magnets in the
magnetic bearing is determined using the following equation [17].
Fr ¼ A
Z 2p Z 2p Z
0
0
r4
r3
Z
r2
r1
2
1
1
jr 23 j3 jr 24 j3 jr 14 j3
!
ðe þ r 3 cos a r 2 cos bÞr 3 dr3 r 2 dr 2 dadb
ð1Þ
where
h
i1
2 2
r23 ¼ ðL þ zÞ2 þ ðe þ r3 cos a r 2 cos bÞ2 þ ðr3 sin a r 2 sin bÞ
ð2Þ
h
i1
2 2
r24 ¼ z2 þ ðe þ r 3 cos a r 2 cos bÞ2 þ ðr 3 sin a r 2 sin bÞ
ð3Þ
2
1
r14 ¼ ½ðL zÞ2 þ ðe þ r3 cos a r2 cos bÞ2 þ ðr3 sin a r 2 sin bÞ 2
and A ¼
Br1 Br2
4pl0
ð4Þ
ð5Þ
The dimensional parameters, r1 = 5 mm (d1 = 10 mm), r2 = 24 mm (d2 = 48 mm), r3 = 25 mm (d3 = 50 mm), r4 = 50 mm
(d4 = 100 mm), L = 30 mm, z = 1 mm, Br1 = 1.0 T, Br2 = 1.2 T and l0 = 4p 104 H/m as applicable to the present test setup
are shown in Fig. 6.
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
(a) Cylindrical magnetic bearing
arrangement under loaded condition
103
(b) Close up view of contact between
stator and rotor
Fig. 5. (a) Cylindrical magnetic bearing arrangement under loaded condition. (b) Close up view of contact between stator and rotor.
Fig. 6. Dimensional details of cylindrical magnetic bearing arrangement.
Eq. (1) is solved numerically using the Gaussian Quadrature numerical integration technique [18] and the results of repulsive force are plotted in Fig. 7.
It is observed from Fig. 7 that the radial repulsive force increases with the increase in the eccentricity ratio. The maximum
static load that bearing can support is 52.8 N, which is much lower than the applied load of 373 N. Therefore, this magnetic
bearing arrangement would not be able to sustain a load of 373 N.
However, another magnetic configuration which consists of a 180° arc magnet have been proposed and evaluated by
Samanta and Hirani [19]. It was shown to be providing a larger radial force (with minimum axial force) as compared to cylindrical magnetic bearing. The feasibility of employing this magnetic bearing arrangement was then explored and experiments
were conducted on this magnetic bearing configuration, the details of which are described in the next section.
2.3. Experiments on circular 180° arc magnetic bearing arrangement
2.3.1. Description of experimental setup
The test setup for the cylindrical magnetic bearing arrangement consists of rotor made up of four neodymium magnetic
discs having inner diameter of 10 mm, outer diameter of 48 mm and thickness of 8 mm as shown in Fig. 4a. It is supported in
two circular 180°arc magnetic bearing having inner diameter of 50 mm, outer diameter of 100 mm and thickness of 15 mm.
The circular arc (180°) magnets are assembled in the bearing housing which is mounted rigidly on the base plate as shown in
Fig. 8. Two accelerometers were attached on the bearing housing in vertical and horizontal planes to pick the vibration signals along the vertical and horizontal directions.
104
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
Fig. 7. Radial repulsive force of cylindrical bearing.
Fig. 8. Circular arc (180°) magnet in the bearing housing.
2.3.2. Experimental operating conditions
The experiment was planned to be conducted on this bearing arrangement for 6 h duration with similar operating conditions as that of a conventional journal bearing (load 373 N and journal speed 27 rpm).
2.3.3. Results and discussion
It was observed that the rotor remained separated from the stator under static condition with application of 373 N load.
The radial repulsive force was therefore sufficient enough to separate the journal from the bearing under the static condition.
The same was verified by the theoretical estimation of the static load carrying capacity of this bearing arrangement. The theoretical static load carrying capacity or magnetic levitation force or radial repulsive force of this configuration is expressed by
Eq. (6), which is obtained by modifying Eq. (1).
Fr ¼ A
Z
p
2
p2
Z 2p Z
0
r4
r3
Z
r2
r1
2
1
1
jr23 j3 jr24 j3 jr 14 j3
!
ðe þ r3 cos a r 2 cos bÞr 3 dr 3 r 2 dr 2 dadb
ð6Þ
The dimensional parameters, r1 = 5 mm (d1 = 10 mm), r2 = 24 mm (d2 = 48 mm), r3 = 25 mm (d3 = 50 mm), r4 = 50 mm
(d4 = 100 mm), L = 30 mm, Br2 = 1.2 T and l0 = 4p 104 H/m as applicable to the present test setup are shown in Fig. 9.
The Eq. (6) is solved numerically using the Gaussian Quadrature numerical integration technique [18]. The results of
magnetic repulsive force vs eccentric position of rotor are plotted in Fig. 10.
It is observed from Fig. 10 that the maximum static load this bearing arrangement can support is 377.7 N at an eccentricity ratio almost equal to 1.0. This maximum static load is higher than the applied load of 373 N.
The test was then started and it was observed that the bearing operation was smooth in the beginning until 130 min of
the test when suddenly high noise and vibration were observed which continued to increase and it became difficult to
continue the test further. The test was then stopped after 140 min of its operation.
The bearing (rotor and stator) condition was examined after the test. The visible cracks appeared in the rotor magnet as
shown in Fig. 11a. On the disassembly of the rotor it was found that two of the rotor magnets had been fractured into two
105
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
Radial Repulsive Force (N)
Fig. 9. Dimensional details of circular arc (180°) magnetic bearing arrangement.
390
380
370
360
350
340
330
-1
-0.8
-0.6
-0.4
320
-0.2
0
0.2
0.4
0.6
0.8
1
Eccentricity Ratio
Fig. 10. Radial repulsive force of 180° arc magnet.
pieces as shown in Fig. 11b, indicating catastrophic failure of the rotor magnets. The surface of the rotor showed significant
wear indicating contact between rotor and stator as shown in Fig. 11c.
The further investigation of the bearing failure was carried out by conducting the analysis of the acceleration data
acquired using the accelerometers during the experiment. Fig. 12a depicts the acceleration vs time plot in the vertical direction. It is observed from Fig. 12a that the bearing operation was relatively smoother until 130 min when a sudden increase in
the acceleration occurs. This is probably the instant of the fracture of rotor magnets. The acceleration continued to increase
cyclically thereafter until the test was stopped at about 140 min of its operation.
The acceleration peaks in Fig. 12a indicates the instance of rotor hitting/striking the stator. The striking of rotor with the
stator is because of the dynamic unbalance introduced due to errors and variability in manufacturing of the rotor magnets as
it is difficult to manufacture a completely balanced rotor. Moreover, the 180° arc magnet has lower stiffness as compared to
the full cylindrical magnet due to the absence of the magnetic repulsive force in the top half of the bearing [20]. Therefore,
106
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
Fig. 11a. Visible cracks on the rotor magnets.
Fig. 11b. Fractured rotor magnets.
Fig. 11c. Circumferential wear marks on the rotor surface.
the 180° arc magnet does not perform smoothly under dynamic loads even though it has higher static load carrying capacity
as compared to cylindrical magnetic bearing. The rotor unbalance coupled with the low stiffness aggravates the problem of
rotor hitting/striking the stator. The continuous hitting/striking of the rotor with stator causes formation of cracks and final
fracture of the rotor magnets. This also introduces cracks in the stator magnets. The sudden increase in the acceleration after
130 min of operation may be attributed to the fractured rotor magnets. This fracture of rotor magnets significantly reduces
the magnetic strength resulting in continuous contact between rotor and stator resulting in wear.
The FFT plot of Fig. 12a is shown in Fig. 12b. Fig. 12b shows the vibration frequency and amplitude that are induced due to
the instability of the magnetic arrangement including the hitting/striking of the rotor with stator. Fig. 12b shows the initial
operation of the bearing test setup with corresponding sub-synchronous and super-synchronous frequencies of the vibration
signal. The vibration amplitudes are small initially.
On the examination of the stator after the test it was observed that significant wear of the bearing surface has taken place
with a crack in the axial direction as shown in Fig. 13a and b. The stator magnet also suffered fracture at the edges as shown
in Figs. 14a and 14b.
Even though the load applied (373 N) was less than the theoretical estimate of the maximum permissible static load of
377.7 N, the failure of both the stator and rotor indicates that the magnetic arrangement was not able to separate the rotor
107
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
Accelration (m/sec2)
150
100
Fracture of rotor magnet
50
0
-50
122
124
126
128
130
132
134
136
138
Time (min)
Fig. 12a. Acceleration vs time plot showing the instant of rotor magnet fracture.
0.3
X: 0.2109
Y: 0.3028
Amplitude (dB)
0.25
X: 0.457
Y: 0.1815
0.2
0.15
X: 0.6679
Y: 0.1089
0.1
X: 0.3515
Y: 0.08425
0.05
0
0
1
2
3
4
5
Frequency (Hz)
Fig. 12b. FFT of the acceleration signal during the initial operation of the bearing.
(a)
(b)
Fig. 13. (a) Circumferential wear marks on the stator surface with an axial crack. (b) Close up view of stator surface with circumferential wear marks and
axial crack.
from the journal under the dynamic conditions. This failure of the 180° arc magnetic arrangement is catastrophic and needs
to be avoided. The breakage of the stator magnets and significant wear of the rotor magnets suggest the need to obtain solutions to this problem. One possible solution is to introduce lubricant between the stator and rotor that may possibly alleviate
the problem of high wear and rotor magnet fracture by sharing the applied load.
The use of such hybridization has been proved to be feasible [10]. The next section contains the details of the experiment
conducted on a hybrid bearing.
108
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
Fig. 14a. Fracture of stator magnet at the edge.
Fig. 14b. Fracture of stator magnet at the corner.
Fig. 15. Test setup for hybrid bearing arrangement.
2.4. Experiments on hybrid bearing arrangement
2.4.1. Description of experimental setup
A lubricant tank is provided with the test setup for the circular arc (180°) magnetic bearing arrangement and shown in
Fig. 15. The remaining arrangement, however, was similar to the one described in Section 2.1 above and shown in Figs. 1a
and 1e.
109
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
(a)
(b)
Damaged
coating
Wear marks
Magnetic flux density (Tesla)
Fig. 16. (a) Rotor magnets before test, and (b) Rotor magnets after test.
Unused rotor magnet
Used rotor magnet
0.6
0.5
0.4
0.3
0.2
0.1
0
0
45
90
135
180
225
270
315
Location of measurement, θ degree
Fig. 17. Magnetic flux density (T) of rotor before and after test.
Table 1
Magnetic flux density of rotor surface before and after test.
S. no.
1
2
3
4
5
6
7
8
Measurement location (°)
0
45
90
135
180
225
270
315
Magnetic flux density of rotor surface (Tesla)
Before test
After test
0.478
0.497
0.482
0.488
0.468
0.489
0.523
0.510
0.462
0.488
0.457
0.431
0.332
0.422
0.455
0.486
Percentage reduction
3
2
5
13
41
16
15
5
Wear
marks
Fracture at
edges
Fig. 18. Wear and fracture of stator after test.
110
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
Acceleration (m/sec 2)
6
4
2
0
-2
-4
0
2
4
6
8
10
12
14
16
18
20
Time (min)
Fig. 19a. Acceleration vs time plot in vertical direction during beginning of test.
0.4
X: 1.337
Y: 0.3917
Amplitude (dB)
3rd Harmonics
0.3
0.2
2nd Harmonics
X: 0.8911
Y: 0.07716
0.1
X: 0.4455
Y: 0.01931
0
0
1
X: 2.673
Y: 0.1537
X: 4.01
Y: 0.1292
X: 1.621
Y: 0.1079
X: 1.782
Y: 0.09375
2
X: 2.541
Y: 0.03814
3
X: 3.564
Y:
X:0.06312
4.455
Y: 0.04407
4
5
Frequency (Hz)
Fig. 19b. FFT of the acceleration signals (vertical direction) during beginning of test.
2.4.2. Experimental operating conditions
The experiment was conducted on the hybrid bearing arrangement for 6 h duration with similar operating conditions as
that of a conventional journal bearing (load 373 N and journal speed 27 rpm). The lubricant temperature was maintained at
40 °C. The acceleration data of the bearing housing in horizontal and vertical directions were acquired using two
accelerometers.
2.4.3. Results and discussion
The experiment was conducted on hybrid bearing for 6 h duration. No significant vibrations levels were observed during
the test as were observed in the circular arc (180°) magnetic bearing arrangement without the lubricant. The rotor and stator
were examined after the completion of the test. Fig. 16a shows the rotor before the test and Fig. 16b shows the rotor after the
test of 6 h duration. It is observed from the examination of these surfaces that significant wear of the rotor surface has
occurred during the test and the coating has been damaged. This is a cause for concern as the wear of surface indicates that
contact occurred between stator and rotor magnets and the magnetic repulsive force was unable to levitate the rotor under
the dynamic conditions. The wear of the rotor was measured in terms of weight loss. The weight of rotor magnet was measured before and after the test and the weight loss was determined to be 3.48 g which is a significant amount of wear.
The magnetic flux density of the surface of rotor magnet was measured after the test using the Gauss meter (in Tesla) and
was compared with its magnetic flux density before the test. A reduction in the magnetic flux density was observed after the
test as shown in Fig. 17 and tabulated in Table 1. The contact between the stator and rotor magnets during operation with
consequent wear and increase in surface temperature due to sliding is the cause of reduction in magnetic flux density of the
magnets.
The non-uniformity in the magnetic strength introduces instability in the bearing operation [21] and causes the rotor to
hit/strike the stator.
The examination of the bearing surface also indicated significant wear and fracture at the edges as shown in Fig. 18.
The wear of the stator and rotor indicates that the hybrid arrangement has also not been able to maintain the separation
of the rotor from the stator under the dynamic conditions with the consequent wear of both rotor and stator. However a
111
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
0.02
Acceleration (m/sec2)
0.015
0.01
0.005
0
-0.005
-0.01
-0.015
-0.02
0
2
4
6
8
10
12
14
16
18
20
Time (min)
Fig. 20a. Acceleration vs time plot in horizontal direction during initial running.
Acceleration (m/sec2)
10
5
0
-5
340
342
344
346
348
350
352
354
356
358
360
Time (min)
Fig. 20b. Acceleration vs time plot in vertical direction after 340 min of the test.
6
Amplitude (dB)
5
4
X: 1.398
Y: 2.753
3
2
1 0.4251
X:
X: 0.8774
Y: 0.7333
X: 1.755
Y: 0.968
X: 2.802
Y: 1.315
X: 2.632
Y: 0.6564
X: 3.51
Y: 0.7108
Y: 0.2602
0
0
1
2
3
4
X: 4.2
Y: 1.041
X: 4.387
Y: 0.4579
5
Frequency (Hz)
Fig. 20c. Amplitude vs frequency plot in vertical direction near the end of test.
112
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
hybrid bearing caused lesser wear as compared with the circular arc (180°) arrangement as described in the previous section
due to the presence of lubricant. The wear was also not localized as was observed in the case of conventional journal bearing
test. The analysis of the acceleration data obtained during the test indicates a much smoother operation of the hybrid bearing
with reduced level of vibrations highlighting the role of lubricant in damping the vibrations.
Fig. 19a depicts the acquired acceleration data in vertical direction during the beginning of the test. It is observed from
Fig. 19a that the maximum amplitude of acceleration in vertical direction is about 6 m/s2. A FFT of the same data is given in
Fig. 19b to show the harmonics. It is observed from Fig. 19b that first harmonics occurs at 0.4455 Hz corresponding to a rotor
speed of 27 rpm. The peaks represent the harmonics of the rotor frequency.
Fig. 20a depicts the acceleration vs time signal in horizontal direction during the beginning of the test. As may be
observed from Fig. 20a that the maximum acceleration values are much smaller as compared to that in the vertical direction
given in Fig. 19a.This is due to the low stiffness in the top half of the circular arc magnetic bearing arrangement
Fig. 20b depicts the acceleration vs time plot in the vertical direction after the elapse of 340 min during the test. The large
value of the acceleration indicates the hitting/striking of the rotor on stator.
Fig. 20c depicts the FFT of the acceleration signal near the end of the test in vertical direction showing very large
amplitudes of vibration indicating impact of rotor on stator which caused the stator fracture.
It is observed from Fig. 21a that the acceleration values are significantly smaller in horizontal direction near the end of the
test. The FFT of the acceleration signal in horizontal direction near the end of the test is depicted in Fig. 21b.
Fig. 21b indicates very small magnitudes of the vibration amplitudes in the horizontal direction. This indicates that the
effect of striking of rotor on the stator in horizontal direction is not significant.
0.02
0.01
0.005
0
-0.005
-0.01
-0.015
-0.02
340
342
344
346
348
350
352
354
356
358
Time (min)
Fig. 21a. Acceleration vs time plot in horizontal direction after 340 min of the test.
x 10
8
-3
7
X: 0.03246
Y: 0.006842
6
Amplitude (dB)
Acceleration (m/sec2)
0.015
X: 1.264
Y: 0.00475
5
X: 0.05966
Y: 0.005155
4
X: 1.475
Y: 0.002619
3
X: 1.685
Y: 0.002298 X: 2.261
Y: 0.001897
2
1
0
X: 2.529
X: 2.95
Y: 0.0009982
Y: 0.0008007
X: 1.054
X: 0.8432
Y: 0.0006745
X: 0.4212
Y: 0.0003782
Y: 0.0002275
0
1
2
3
X: 4.52
Y: 0.001191
4
5
Frequency (Hz)
Fig. 21b. Amplitude vs frequency plot in horizontal direction near the end of test.
360
S.M. Muzakkir et al. / Engineering Failure Analysis 40 (2014) 97–113
113
3. Conclusions
Based on the experimental observations on the conventional journal bearing and different magnetic bearing arrangements, it is concluded that the cylindrical magnetic bearing arrangement has the minimum static load carrying capacity
and it is not possible to use it for heavy load and slow speed conditions. The circular arc (180°) magnetic arrangement
has higher static load carrying capacity as compared to cylindrical magnetic bearing arrangement and is able to separate
the rotor from the stator under static conditions, however it is unable to maintain the separation under the dynamic conditions resulting in the contact between rotor and stator consequently resulting in significant wear and fracture of the rotor
and stator. It is, therefore, concluded that the static load carrying capacity alone is not the sufficient condition for a magnetic
bearing to operate successfully. A complete dynamic analysis is thus required to determine the dynamic load carrying capacity of the bearing to determine its feasibility. The use of lubricant with the magnetic bearing arrangement (hybrid bearing) is
able to contain the severity of the magnetic bearing failure. It is able to obviate rotor fracture and resulted in reduced wear of
the rotor and stator due to the action of lubricant. Even though the hybrid bearing arrangement with the provision of lubrication also failed to maintain the separation between stator and rotor in dynamic conditions but resulted in less severe wear
as compared to circular arc magnetic arrangement.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
Muzakkir SM, Hirani H, Thakre GD, Tyagi MR. Tribological failure analysis of journal bearings used in sugar mills. Eng Fail Anal 2011;18:2093–103.
Dufrane KF, Kannel JW. Wear of steam turbine journal bearings at low operating speeds. J Lubr Technol 1983;105:313–7.
Singhal S. Sleeve bearing design for slow speed applications in cement plant. Tech Conf Rec IEEE 2008:283–90.
Spikes H. The history and mechanisms of ZDDP. Tribol Lett 2004;17:469–89.
Loxham J. Proposals for the selection of optimum clearances in journal bearings. Tribol Int 1983;16:185–92.
Feng M, Kenjo T. A comparative study of turning and stationary groove journal bearings for HDD spindle motors under transient contact conditions.
IEEE Trans Magn 2007;43:3727–33.
Kumar CP, Menezes PL, Kailas SV. Role of surface texture on friction under boundary lubricated conditions. Tribol Online 2008;3:12–8.
Zhang R-Q, Chang HS. A new type of hydrostatic/hydrodynamic gas journal bearing and its optimization for maximum stability. Tribol Trans
1995;38:589–94.
Eryong H, Kun L. Investigation of axial carrying capacity of radial hybrid magnetic bearing. IEEE Trans Magn 2012;48:38–46.
Hirani H, Samanta P. Hybrid (hydrodynamic + permanent magnetic) journal bearings. Proc Inst Mech Eng Part J J Eng Tribol 2007;221:881–91.
Qingchang T, Wei L, Xuejun X. Investigations on a magnetic–hydrodynamic hybrid thrust bearing. Tribol Int 2002;12:61–6.
Muzakkir SM, Hirani H, Thakre GD. Lubricant for heavily loaded slow-speed journal bearing. Tribol Trans 2013;56:1060–8.
Khonsari MM, Booser ER. Applied tribology: bearing design and lubrication. John Wiley & Sons Ltd.; 2001.
Begelinger A, De Gee AWJ. A study of the effect of radial clearance, contact angle and contact pressure on the wear of boundary-lubricated bearing
bronze. Wear 1982;77:45–56.
Lu X, Khonsari MM. On the lift-off speed in journal bearings. Tribol Lett 2005;20:299–305.
Noar JH, Evans RD. Rare Earth magnets in orthodontics: an overview. Br J Orthod 1999;26:29–37.
Tan Q, Li W, Liu B. Investigations on a permanent magnetic–hydrodynamic hybrid journal bearing. Tribol Int 2002;35:443–8.
Samanta P, Hirani H. A simplified optimization approach for permanent magnetic bearing. Indian J Tribol 2007;2:23–34.
Samanta P, Hirani H. Magnetic bearing configurations: theoretical and experimental studies. IEEE Trans Magn 2008;44:292–300.
Mukhopadhyay SC, Ohji T, Iwahara M. Modeling and control of a new horizontal-shaft hybrid-type magnetic bearing. IEEE Trans Magn 2000;47:100–8.
Ohji T, Mukhopadhyay SC, Iwahara M, Yamada S. Performance of repulsive type magnetic bearing system under nonuniform magnetization of
permanent magnet. IEEE Trans Magn 2000;36:3696–8.