Tribology Transactions
ISSN: 1040-2004 (Print) 1547-397X (Online) Journal homepage: [Link]
Experimental Study of the Influence of Scratches
on the Performance of a Two-Lobe Journal Bearing
Célia Giraudeau, Jean Bouyer, Michel Fillon, Mathieu Hélène & Jérôme
Beaurain
To cite this article: Célia Giraudeau, Jean Bouyer, Michel Fillon, Mathieu Hélène & Jérôme
Beaurain (2016): Experimental Study of the Influence of Scratches on the Performance of a
Two-Lobe Journal Bearing, Tribology Transactions, DOI: 10.1080/10402004.2016.1238528
To link to this article: [Link]
Accepted author version posted online: 03
Oct 2016.
Published online: 03 Oct 2016.
Submit your article to this journal
Article views: 32
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
[Link]
Download by: [Gazi Universitesi] Date: 04 January 2017, At: 12:17
TRIBOLOGY TRANSACTIONS
[Link]
Experimental Study of the Influence of Scratches on the Performance of a Two-Lobe
Journal Bearing
C
elia Giraudeaua,b, Jean Bouyera, Michel Fillona, Mathieu Heleneb, and Jero
^me Beaurainb
a
Pprime Institute, CNRS–Universite de Poitiers–ISAE-ENSMA, Futuroscope Chasseneuil, France; bEDF Lab Paris–Saclay, Palaiseau, France
ABSTRACT ARTICLE HISTORY
During maintenance operations of steam turbines, a common issue is the degradation of the supporting Received 11 July 2016
and guiding components for the rotating shafts due to the presence of scratches. To examine this issue, a Accepted 14 September 2016
numerical software program has been developed. In order to validate and complete the elaborated KEYWORDS
modeling, experimental tests have been carried out on the Pprime Institute test bench. A preloaded two- Journal bearing; scratches;
lobe journal bearing (diameter: 100 mm, length: 68.4 mm, preload: 0.524) lubricated with an ISO VG 46 thermal effects; pressure;
mineral oil has been tested. Scratches were machined on the shaft at two axial locations and at different experiments
depths. Pressures and temperatures have been measured for several rotational speeds and applied loads.
Seven configurations (one unscratched surface and six with a scratch) have been tested in order to
evaluate the influence of a scratch on the performance of the bearing. It is found that the pressure profile
is significantly affected by the presence of a scratch, whereas the temperature field is weakly modified.
Introduction
measured pressure and temperature distributions using a plain
Since the end of the 19th century, extensive research on hydro- fully flooded transparent journal bearing operating under
dynamic lubrication of journal bearings has been carried out. steady-state conditions. That also enabled a photographic
Numerous numerical studies have been conducted on many investigation of the film breakdown in the diverging zone. Sub-
aspects, such as cavitation, thermal effects, wear, unexpected atmospheric pressures in the diverging zone are confirmed and
operating conditions, etc. Experiments have always been experimental results are close to theoretical ones obtained by
needed to validate the theoretical works. Of the various studies the model proposed by Floberg (8), (9). Recently, Cristea, et al.
found in the literature, Tower (1), who was the first to show the (12) presented full experimental pressure and temperature field
increase in pressure in the lubricant film during operation, is measurements in a circumferential journal bearing. Their mea-
notable. He also published pressure measurements carried out surement method allows identification of the cavitated regions
at the film–bushing interface (Tower (2)). During the first half at subatmospheric pressures. They also pointed out that the
of the 20th century, McKee and McKee (3), McKee, et al. (4), maximum temperature occurs in the sharp pressure drop
McKee and White (5), and McKee (6) also studied various con- region. More information on the cavitation formation and
figurations of journal bearings under varying operating param- modeling for fluid film bearings can be found in a recent review
eters. They were the first to perform measurements of the film paper presented by Braun and Hannon (13).
pressure distribution. During the same period, Kingsbury (7) With the increase in power plant lifetime and the increase in
analyzed the thermal effects in a journal bearing. He made the- operating conditions, electric operators are faced with new
oretical assumptions, still often in use nowadays, such as the problems. One of these problems is the apparition of scratches
radial heat flux continuity throughout film layers and metal on the rotating shaft that can create melting of the Babbitt or
interface boundaries or laminar flow of the lubricant. Floberg direct contact between the solid parts. They are often due to
(8), (9) analyzed another important phenomenon in lubricated debris in the lubricant oil that is trapped in the Babbitt. To the
bearings; that is, cavitation. He proposed a cavitation model authors’ knowledge, no experimental study has been done to
predicting subcavity pressures and taking into account oil characterize the influence of a scratch on a journal bearing
streamers. Tanaka and Someya (10) experimentally analyzed behavior. Moreover, numerical studies found in the literature
cavitation by measuring oil film pressure distributions in a on this subject are scarce. Dobrica and Fillon (14) proposed a
journal bearing. Pressures were recorded at the mid-plane of numerical work on the evolution of pressure as a function of
the bearing for several supply pressures under steady-state scratch depth. Taking radial clearance C as a reference to
operating conditions. They showed that, in the divergent zone, describe the scratch depth, they demonstrated that a C/2 value
the film pressure is constant. Braun and Hendricks (11) also only weakly affects the bearing performance, whereas a 2C
CONTACT Jean Bouyer [Link]@[Link]
Color versions of one or more of the figures in the article can be found online at [Link]/utrb.
Review led by Benyebka Bou-Said
© 2016 Society of Tribologists and Lubrication Engineers
2 C. GIRAUDEAU ET AL.
value (or higher) significantly affects the bearing performance. Table 1. Bearing and lubricant parameters.
The main differences are to be noted near or inside the scratch. Bearing diameter Dn 100 mm
Helene, et al. (15) also presented a numerical investigation of Shaft diameter d 99.772 mm
the influence of a scratch on the pressure and temperature Bearing length L 68.4 mm
Bearing thickness H 20 mm
fields. They showed that the pressure decreased as a function of Horizontal radial clearance C 143 mm
scratch depth until rupture of the lubricant film occurred. Bra- Vertical radial clearance Cb 68 mm
nagan (16) focused on the impact of scratch location. He con- Lubricant — ISO VG 46
Lubricant supply temperature Ts 43 C
cluded that the load-carrying capacity is most significantly Lubricant supply pressure Ps 0.17 MPa
affected by a centered scratch, because the maximum pressure Lubricant dynamic viscosity (40 C) m40 C 0.0416 Pa.s
location is changed. This result confirmed the work of Helene, Lubricant dynamic viscosity (60 C) m60 C 0.0191 Pa.s
et al. (15), who obtained a maximum pressure located at § L/6
for this configuration.
Therefore, the present study aims to provide experimental secondary planar hydrostatic bearing placed between the pneu-
measurements of both local parameters, such as pressure and matic cylinder and the spherical hydrostatic bearing. The first
temperature fields, and global parameters, such as the friction one enabled a slight ring lateral displacement relative to the
torque and flow rate. The first part will present the test rig and fixed shaft. The second one enabled correct surface mating by
bearing equipment and the second part presents the experi- unconstraining the rotational degrees of freedom. This system
mental procedure. The third part presents experimental results, prevents bearing misalignment. Two diametrically opposite
which will be discussed and compared to previous numerical pins connected to a torque meter were placed on the ring hous-
findings in the literature. The last part will be devoted to con- ing to prevent its rotation. Details concerning the torque mea-
clusions and perspectives. surement system have been presented by Bouyer and Fillon
(18). The bearing was supplied by an ISO VG 46 oil whose
supply pressure was regulated with a precision proportional–
Test rig, bearing arrangement, and instrumentation
integral–derivative controller at 0.17 MPa relative to the atmo-
All of the experimental results presented in this work were per- sphere. A thermocouple located upstream of the bearing supply
formed on the journal bearing rig presented in Fig. 1. The test holes enabled supply temperature regulation. For the current
rig is located at Pprime Institute. It was first designed at INSA tests, the supply temperature was maintained at 43 § 0.5 C.
of Lyon and then adapted and built by Ferron, et al. (17). It has The evolution of this temperature is similar to a 15-min sine-
been continuously modified and upgraded over the years to be wave periodic profile. Because the data acquisition is performed
able to carry out state-of-the-art experiments. only at a 43 C feeding temperature, the consequence of this
The test rig consists of a tapered end shaft carried on high- type of profile is that the period between two data acquisitions
precision rolling element ball bearings. A 21-kW DC motor is the one of the profile. The resulting supply flow was moni-
allows rotation of the shaft within §1% accuracy of the nomi- tored by an electric pulse gear flow rate meter.
nal speed up to 3,500 rpm. An electronically controlled pneu- The geometrical parameters of the journal bearing and the
matic cylinder allows imposing a maximum vertical load of lubricant properties are presented in Table 1. Two identical
10,000 N with §20 N uncertainty. Alignment between the ring 99.772 mm § 2 mm diameter steel shafts have been used for
and the shaft in operation is ensured by a primary spherical the current tests. The measured roundness of the shafts after
hydrostatic bearing supporting the ring housing and by a mounting on the spindle did not exceed § 4 mm. In order to
Figure 1. Schematic representation of the test rig.
TRIBOLOGY TRANSACTIONS 3
Figure 2. 3D model of two-lobe journal bearing.
confirm the similarity of the shafts, both have been tested under and the front sections of the bearing. The horizontal axis is the
the same operating conditions without any defects. The results circumferential coordinate, where 180 is at the bottom of the
provided by those two tests were very close, with a maximum bearing. Yellow rectangles identify the supply grooves, placed
discrepancy of §0.5% for any parameter. As a consequence, at § 90 from the load direction. One can note that the lower
the results obtained on both shafts were not distinguished. The lobe, which is the loaded lobe, is understandably the most
tested bearing is composed of two identical lobes with a 0.524 equipped one with 10 thermocouples and 14 pressure sensors
preload ratio. It has a diameter of 100 mm with 68 mm § 4 mm placed both axially and circumferentially at strategic locations.
vertical radial clearance (Cb) and 143 mm § 4 mm horizontal Dashed lines give the positions of the scratch that are machined
radial clearance (C). The thickness of the bushing was 20 mm on the shaft at either 7L/12 or L/3.
(radial). After being instrumented, the two-lobe bearing is inserted in
Holes were drilled in the bushing in order to permit pres- a special ring housing (200-mm outer diameter) before being
sure and temperature measurements, as can be seen on Fig. 2. paired with the steel shaft. All tubes, both oil supply and pres-
Pressures were measured with 19 wet-type strain-gauge pres- sure measurement, are then connected. Finally, thermocouples
sure sensors. Several ranges of pressure sensors were chosen are plugged in and the torque meter is linked to the ring hous-
to correctly fit a two-dimensional pressure field obtained from ing. Figure 4 shows a general view of the mounted bearing.
calculations as presented in Table 2. This technique reduces
the uncertainty of the measurements, because the range of
each probe is closer to the value to be measured (error range
Experimental procedure and test parameters
is §0.3% of the full range). An additional absolute pressure
sensor continuously recorded the atmospheric pressure, allow- Before performing any tests, the test rig was heated at least for
ing all pressures to be converted into absolute values. Thus, 2 h to ensure a steady-state thermal regime. Then, operating
the pressures presented in the Results and Discussion section parameters were tuned and, once thermal equilibrium was
are absolute. Table 2 gives the values of uncertainties associ- reached, data were recorded. Data were acquired through a
ated with each sensor. Temperatures were measured with 20 series of 100 values of each probe signal at a 10 Hz frequency.
K-type thermocouples (0.5 mm diameter) with an error range The average from each acquisition thus yielded a single value
of §0.5 K. Thirteen thermocouples were located in the per probe. Each test was repeated at least three times at differ-
mid-plane; the others were close to the bearing ends. Other ent moments and the final value of the considered operating
thermocouples were placed on the rig to provide additional condition was obtained for each recorded datum by averaging
information such as protecting cage temperature, proximity the results of the three acquisitions.
probe temperature, and supply temperature. The operating parameters were varied in terms of speed and
Figure 3 presents the spatial repartition of pressure and tem- load. The applied load varied between 6,000 and 10,000 N with
perature probes. The vertical axis gives the axial coordinate, a 1,000-N step. The rotational speed of the shaft varied between
where the top and bottom extrema are, respectively, the back 500 and 3,500 rpm with a 750-rpm step. A low-speed test has
been carried out for each imposed load at 100 rpm. As stated in
the previous section, prior to performing tests with scratches, a
Table 2. Pressure sensor uncertainties. series of reference tests was performed on both intact shafts
Pressure sensor range (MPa) Accuracy (MPa) Sensor number with the operating conditions cited above. After this series, the
shafts were machined to introduce the scratches. The scratch
0.4 §0.0012 P1, P2, P11, P14, P15, Patm
1.6 §0.0048 P5, P6, P10, P22 characteristics are defined by three parameters: width, depth,
2.5 §0.0075 P13, P21, P23 and location (Fig. 5). It should be noted that the scratch width
6 §0.018 P7, P9, P20 remained constant for all tests at a value of 2 mm. Indeed, the
10 §0.03 P17, P18, P19
25 §0.075 P8
L/D ratio was equal to 0.68 for the experimental model and 0.8
for the real bearings used at Electricite De France (EDF), and
4 C. GIRAUDEAU ET AL.
Figure 3. Pressure, temperature measurement, and scratch locations at the film–bushing interface.
Figure 4. General view of the test bearing.
the 2-mm scratch width used for this study is in accordance (130 mm), and 2C (260 mm). Due to machining uncertainties,
with the real ones observed on the EDF’s machinery. these values were not the exact ones applied on the shafts but
Seven configurations were thus sequentially tested: a refer- they were of the same order of magnitude. The scratches were
ence case without any defect on the shaft surface and six realized sequentially going from the shallower one toward the
others with a scratch. The difference between the two tested deeper one. After machining, scratches were systematically
shafts is the location of the scratch: one at 7L/12 and the other measured with a 3D optical profilometer. Figure 6 shows an
one at L/3, as represented in Fig. 3. The depth was also varied example of the measurement realized on the shaft at L/3 for
as a function of the horizontal diameter clearance in accor- the deepest case. Table 3 provides the results of these meas-
dance with the work of Dobrica and Fillon (14). Indeed, three urements on each scratch.
theoretical depths were chosen for the study: C/2 (65 mm), C
Results and discussion
The first part of this section will be devoted to presenting local
results of pressure and temperature measurements with varying
speed and load for the scratch located at 7L/12. The second part
will give similar results for the scratch located at L/3. After pre-
senting a comparison of axial pressure profiles between the two
locations of the scratch, the last part will present the results for
the global parameters such as friction torque and oil flow rate.
It should be noted that in order to make the results more read-
able, lines have been added to the graphs but only the points
Figure 5. View of a scratched shaft. represent experimental data.
TRIBOLOGY TRANSACTIONS 5
Figure 6. Example of a scratch depth measurement.
Based on the work of Dobrica and Fillon (14), the three dif- other regions of the bearing. This is due to the fact that, as
ferent selected depths should show different impacts on the shown by Helene, et al. (15), a pressure drop is created in front
performance of the bearing. In Figs. 7 to 10, pressure profiles at of the scratch, thus significantly modifying the shape of the
an imposed load of 10,000 N are plotted for four different pressure field (Fig. 12).
speeds (100, 500, 2,000, and 3,500 rpm) in the bearing mid- Figures 13 and 14 show the evolution of the temperature
plane for several scratch depths. It can be observed that the profile in the bearing mid-plane for the highest rotational speed
maximum pressure decreases with an increase in speed. It is and the highest applied radial load. In Fig. 13, which shows the
known that higher speeds decrease shaft eccentricity, which temperature in the mid-plane of the lower lobe (highest loaded
induces a higher minimum film thickness, consequently lobe), it can be noticed that the temperature globally increases
decreasing the maximum pressure. In Fig. 7, the shaft rotational with the scratch depth and maintains the same shape. Only the
speed is very low and the effect of scratch depth on the pressure case with the deepest scratch (green dots) has slight higher inlet
profile is weak. Only the presence of a scratch affects the pres- and lower outlet temperatures.
sure field because the film thickness is very small for such oper- Figure 14 shows the same results for the upper lobe. Here, all
ating conditions; therefore, the depth of the scratches, which is profiles are similar; the scratch weakly affects the temperature
higher than the minimum film thickness, induces a high geom- on this lobe. In addition, temperatures are lower than the ones
etry discontinuity in the lubricating film. With increasing speed obtained for the lower lobe, which is in accordance with the
(Fig. 8), the scratch depth becomes a significant parameter. theory.
When the film thickness is higher, the influence of C/2 on pres- The next part of this section will be devoted to the same
sure is lower and the values tend toward those in the case with- parameters obtained with a different location of the scratch,
out a scratch. At higher speeds (Figs. 9 and 10), the effect of which is L/3. Again, the results for temperature and pressure
scratch depth is again more visible. In fact, the influence of the fields will be presented under varying operating conditions.
scratch on the pressure field is directly correlated to the film Figure 15 shows the pressures in the bearing mid-plane of
thickness: the scratch does affect the pressure field if its depth the lower lobe at 3,500 rpm and 10,000 N. The pressure is simi-
has the same order of magnitude as the minimum film thick- lar regardless of scratch depth for all cases with a scratch
ness. For high eccentricity, all depths influence the pressure located at L/3. This result differs from that observed for 7L/12
field, whereas for low eccentricities, only the deeper scratch has scratch cases; indeed, this is due to the fact that the scratch is
a significant influence. further from the bearing mid-plane and the impact of the
Figure 11 represents the evolution of axial pressure at high scratch is thus less visible.
speed and high load (3,500 rpm and 10,000 N). The results pre- In the schematic representation of the bearing in Fig. 3, one
sented here are for three pressure probes placed at a 185 angu- of the pressure probes is located at the L/3 axial coordinate and
lar coordinate (P8, P17, and P18) and one at 195 angular the 195 angular coordinate, P19. This probe is very important
coordinate (P19, left-hand side of the figure). In comparison to because it faces the scratch, providing the possibility to have
the nominal case without a scratch, it can be noted that the the pressure in front of the scratch. Figure 16 shows the evolu-
maximum measured pressure moves from the bearing mid- tion of its value for all depth cases as a function of imposed
plane toward the center of the largest zone defined by the static load at 500 rpm. Firstly, it can be noted that the pressure
scratch. Moreover, the maximum measured pressure increases increases with load as expected. Secondly, the pressure
with scratch depth and, conversely, the pressure decreases in decreases when the scratch becomes deeper, confirming that
the scratch causes a drop in the pressure field in front of the
Table 3. Description of the tested cases and their scratch characteristics. scratch, as shown by Helene, et al. (15). At 8,000 N, for exam-
ple, the pressure decreases from 4 MPa for the case without a
Case without a scratch on the rotating shaft
Scratch at 7L/12 Scratch at L/3 defect to 2.7, 1.6, and 0.32 MPa for the C/2, C, and 2C scratches,
respectively. Because the eccentricity is high for this low-speed
54 § 5 mm 52 § 5 mm
100 § 5 mm 120 § 5 mm
case, the effect of the scratch depth is obvious: a small scratch
274 § 5 mm 243 § 5 mm depth (C/2) decreases the pressure by 24 to 38% from low load
to high load, an intermediate scratch depth (C) decreases the
6 C. GIRAUDEAU ET AL.
Figure 7. Mid-plane pressure for lower lobe at 100 rpm and 10,000 N.
Figure 8. Mid-plane pressure for lower lobe at 500 rpm and 10,000 N.
Figure 9. Mid-plane pressure for lower lobe at 2,000 rpm and 10,000 N.
TRIBOLOGY TRANSACTIONS 7
Figure 10. Mid-plane pressure for lower lobe at 3,500 rpm and 10,000 N.
Figure 11. Axial pressure at 185 and 195 of the lower lobe at 3,500 rpm and 10,000 N.
pressure from 50 to 69%, and a large scratch depth (2C)
decreases the pressure from 82 to 99%. Moreover, for the C and
2C scratch cases, because the depth is larger than the minimum
film thickness, the presence of the scratch always induces a
drop in pressure, which remains almost constant regardless of
load.
Figure 17 represents the evolution of the pressure measured
for the same probe at 10,000 N as a function of speed. Effects of
eccentricity on pressure are clear. As expected for the nominal
case, the pressure decreases with an increase in speed due to
the increase in film thickness. Conversely, for the scratch cases
the pressure increases with speed. This is due to the fact that,
because the film thickness is increased with speed, the scratch
effect is weaker because the ratio between the scratch depth
and the minimum film thickness decreases.
Next, the probe located at a 185 angular coordinate and 2L/
3 axial coordinate, P17 (Fig. 3), will be studied. It is almost sym-
metric with the previous one from the bearing mid-plane. As in
Fig. 16, the evolution of pressure at 500 rpm as a function of
Figure 12. Example of pressure field with one scratch from Helene, et al. (15). imposed load is plotted in Fig. 18. Firstly, the values obtained
8 C. GIRAUDEAU ET AL.
Figure 13. Mid-plane temperature for the lower lobe at 3,500 rpm and 10,000 N.
Figure 14. Mid-plane temperature for the upper lobe at 3,500 rpm and 10,000 N.
Figure 15. Mid-plane pressure for lower lobe at 3,500 rpm and 10,000 N.
TRIBOLOGY TRANSACTIONS 9
Figure 16. Evolution of pressure at 195 angular coordinate and L/3 axial coordinate (P19) at 500 rpm versus load.
for the unscratched case are very close to those obtained in the temperatures as for the other location (Figs. 13 and 14) is
Fig. 16. Secondly, the pressure for the scratched cases shows an observed. The temperature globally increases with scratch
opposite behavior to that in Fig. 16; that is, it increases with depth and maintains the same shape but less significantly than
scratch depth. Indeed, the pressure measured by this probe, in the other location case. Figure 21 shows the same results for
which is in the middle of the largest region created by the the upper lobe. Again, all profiles are similar; that is, the tem-
scratch, is close to the maximum pressure. As seen in Figs. 16 peratures are lower than those obtained for the lower lobe,
and 17, the scratch causes a drop of pressure at L/3 and thus which is in accordance with the theory.
the maximum pressure logically increases in the bearing to Figures 22 and 23 give an overall view of the evolution of the
compensate the imposed load. The deepest scratch has a more maximum measured temperature in the mid-plane of the bear-
significant effect than the others for all applied loads. ing, respectively, at 3,500 rpm with varying load and at
The same probe values are plotted in Fig. 19 as a function of 10,000 N with varying speed. It should be noted that the loca-
speed for an imposed load of 10,000 N. The maximum mea- tion of the maximum measured temperature obviously changes
sured pressure obviously decreases with speed. However, it can with the operating conditions. This evolution is not presented
be noticed that the higher the rotational speed of the shaft, the in the figures, but the following observations can be made: at
less influence the scratch has on the pressure. Again, this can high speeds, the maximum temperature location is constant at
be directly correlated to the shaft eccentricity. a 200 angular coordinate, but at lower speeds, it varies between
As shown for the other scratch location, temperatures in the 180 and 200 , also depending on the scratch depth. In Fig. 22,
mid-plane of the bearing are presented in Figs. 20 and 21 for it can be noticed that the presence of a scratch does not modify
the lower and upper lobes. respectively. The same behavior of the slope of the temperature variation. Nevertheless, the 2C
Figure 17. Evolution of pressure at 195 angular coordinate and L/3 axial coordinate (P19) at 10,000 N versus speed.
10 C. GIRAUDEAU ET AL.
Figure 18. Evolution of pressure at 185 angular coordinate and 2L/3 axial coordinate (P17) at 500 rpm versus load.
Figure 19. Evolution of pressure at 185 angular coordinate and 2L/3 axial coordinate (P17) at 10,000 N as a function of speed.
Figure 20. Mid-plane temperature for lower lobe at 3,500 rpm and 10,000 N.
TRIBOLOGY TRANSACTIONS 11
Figure 21. Mid-plane temperature for upper lobe at 3,500 rpm and 10,000 N.
Figure 22. Evolution of maximum measured temperature at 3,500 rpm as a function of load.
Figure 23. Evolution of maximum measured temperature at 10,000 N as a function of speed.
12 C. GIRAUDEAU ET AL.
Figure 24. Axial pressure at 185 and 195 of the lower lobe at 3,500 rpm and 10,000 N.
scratch depth induces a global increase of the maximum mea- different: in fact, the deep scratch transforms the bearing in a
sured temperature of 2 K. The shallower scratches also lead to circumferential groove journal bearing with two lands sepa-
an increase in this temperature for highly loaded cases. rated by a groove. Indeed, because the load is constant, the
As could be expected, the maximum measured temperature maximum pressure occurs on the largest land.
increases quasilinearly with speed under a constant load The last part of the study is dedicated to more global parame-
(Fig. 23) for all cases. The first two configurations of scratches ters, which are the friction torque and oil flow rate. The results
do not modify the maximum measured temperatures regardless presented herein are only for cases without a scratch (blue dia-
of the operating conditions. As stated for Fig. 22, only the monds) and with the deepest scratch located at L/3 (orange
deeper scratch leads to a slight increase in temperature at squares). In Fig. 25, in order to observe the influence of both
higher speeds (respectively 1.3 and 2.5 K at 2,750 and 3,500 speed and load, two speeds are presented (500 rpm with plain
rpm). lines and 3,500 rpm with dashed lines) as a function of the
Figure 24 allows analysis of the influence of scratch location imposed load. Because the friction torque is mainly related to
on pressure. The axial repartition of pressure at 3,500 rpm and the rotational speed, it obviously increases with speed for both
10,000 N is presented for sensors P17 and P19 for both locations cases. The presence of the scratch induces an increase in friction
of the deepest scratch (2C). As shown by Branagan (16), the torque, which is more significant at a low speed. Indeed, at
maximum measured pressure obviously depends on the scratch 500 rpm, the friction torque globally increases by 25% regardless
position, which is located on the larger part of the bearing with of the applied load. At high speed, this effect is weaker, with an
the scratch as a boundary. The pressure profiles are completely increase varying between 0.9% at 6,000 N and 6% at 10,000 N.
Figure 25. Evolution of friction torque vs. load at two rotational speeds; that is, 500 and 3,500 rpm.
TRIBOLOGY TRANSACTIONS 13
Figure 26. Evolution of oil flow rate vs. load at two rotational speeds; that is, 500 and 3,500 rpm.
The same comparison has been done for the oil flow rate in Acknowledgement
Fig. 26. It can be noted that the presence of the scratch weakly
The authors thank M. Beau for machining high-precision scratches on the
impacts the evolution of the oil flow rate. The oil flow rate mea- shafts.
sured in the case with a scratch is slightly lower due to a lower
axial oil flow. Nevertheless, the fact that this difference is weak
is due to our operating conditions, for which a constant feeding Funding
pressure is imposed. The authors thank ANRT for their financial support.
Conclusions and perspectives
References
This experimental study enables a better understanding of the
(1) Tower, B. (1883), “First Report on Friction Experiments (Friction of
influence of rotor scratches on the performance of a two-lobe Lubricated Bearings),” Proceedings of the Institution of Mechanical
journal bearing. Both pressure and temperature measurements Engineers, 34, pp 632–659.
were performed under laboratory conditions and the results of (2) Tower, B. (1885), “Second Report on Friction Experiments,” Proceed-
this study lead to the following conclusions: ings of the Institution of Mechanical Engineers, 36, pp 58–70.
The influence of the presence of a scratch on the pressure (3) McKee, S. A., and McKee, T. R. (1932), “Pressure Distribution
in Oil Film of Journal Bearings,” ASME Transactions, 54, pp
field is directly correlated to the film thickness: the scratch 149–165.
does affect the pressure field if its depth is of the same (4) McKee, S. A., White, H. S., and Swindells, J. F. (1948), “Measure-
order of magnitude as the minimum film thickness. ments of Combined Frictional and Thermal Behavior in Journal-
The scratch creates a pressure drop that significantly Bearing Lubrication,” ASME Transactions, 70, pp 409–418.
modifies the shape of the pressure field. This effect is (5) McKee, S. A., and White, H. S. (1950), “Oil Holes and Grooves in
Plain Journal Bearings,” ASME Transactions, 72, pp 1025–1034.
more visible when the eccentricity is high; that is, the (6) McKee, S. A. (1952), “Oil Flow in Plain Journal Bearings,” ASME
rotational speed is low and the applied load is high. Transactions, 74, pp 841–848.
The temperatures are weakly affected by the presence of a (7) Kingsbury, A. (1933), “Heat Effects in Lubricating Films,” Mechani-
scratch. Nevertheless, it can be noticed that temperatures cal Engineering, 55, pp 685–688.
are slightly higher for the deepest scratch at high rota- (8) Floberg, L. (1965), “On Hydrodynamic Lubrication with Special
Reference to Sub-Cavity Pressures and Number of Streamers in
tional speed. Cavitation Regions,” Acta Polytechnica Scandinavica, 19, pp
The presence of a scratch does not affect the oil flow rate 1–35.
but increases the friction torque, depending on the rota- (9) Floberg, L. (1968), “Sub-Cavity Pressures and Number of Oil Stream-
tional speed, from 0.9 to 25%. ers in Cavitation Regions with Special Reference to the Infinite Jour-
This study has thus permitted validation of some of the nal Bearing,” Acta Polytechnica Scandinavica, 37, pp 1–36.
(10) Tanaka, T., and Someya, T. (1972), “Investigation into Oil Film Pres-
observations made in the numerical works of Dobrica and Fil- sure Distributions in Journal Bearings,” Proceedings of the Institution
lon (14), Helene, et al. (15), and Branagan (16). It would be of Mechanical Engineers, pp 20–25.
interesting to compare these experimental results with other (11) Braun, M. J., and Hendricks, R. C. (1984), “An Experimental Investi-
results obtained with a “real” case, close to what happens on gation of the Vaporous/Gaseous Cavity Characteristics in an Eccen-
industrial machines; that is, with several scratches on the shaft tric Journal Bearing,” ASLE Transactions, 27(1), pp 1–14.
(12) Cristea, A. F., Bouyer, J., Fillon, M., and Pascovici, M. D. (2011),
surface. It would also be interesting to study higher rotational “Pressure and Temperature Field Measurements of a Lightly Loaded
speeds in order to study higher thermal effects and thus to con- Circumferential Groove Journal Bearing,” Tribology Transactions, 54
firm the tendencies observed here. (5), pp 806–823.
14 C. GIRAUDEAU ET AL.
(13) Braun, M. J., and Hannon, W. M. (2010), “Cavitation Formation and (16) Branagan, L. A. (2006), “Toward a Quantitative Analysis of Journal
Modelling for Fluid Film Bearings: A Review,” Journal of Engineering and Bearing Scratches,” 5th EDF & LMS Poitiers Workshop, Futuro-
Tribology, 224(9), pp 839–863. scope, October 5, Futuroscope, France.
(14) Dobrica, M. B., and Fillon, M. (2012), “Performance Degradation (17) Ferron, J., Fr^ene, J., and Boncompain, R. (1983), “A Study of the ther-
in Scratched Journal Bearings,” Tribology International, 51, mohydrodynamic Performance of a Plain Bearing. Comparison
pp 1–10. between Theory and Experiments,” Journal of Lubrication Technol-
(15) Helene, M., Beaurain, J., and Raud, X. (2013), “Impact of Scratches in ogy, 105(3), pp 422–428.
Tilting Pad Journal Bearings—Influence of the Geometrical Charac- (18) Bouyer, J., and Fillon, M. (2002), “An Experimental Analysis of
teristics of Scratches,” 12th EDF/Pprime Workshop, Futuroscope, Sep- Misalignment Effects on Hydrodynamic Plain Journal Bearings
tember 17–18, Futuroscope, France. Performances,” Journal of Tribology, 124(2), pp 313–319.