Volume 22 Issue 2 Article 6

COMBINED EFFECTS OF MAGNETIC FIELD AND SURFACE ROUGHNESS ON

LONG JOURNAL BEARING LUBRICATED WITH FERROFLUID
Tze-Chi Hsu
Department of Mechanical Engineering, Yuan-Ze Unversity, Chung-Li, Taiwan, R.O.C

Jing-Hong Chen
Department of Mechanical Engineering, Yuan-Ze Unversity, Chung-Li, Taiwan, R.O.C, [email protected]

Hsin-Lu Chiang
Department of Mechanical Engineering, Taoyuan Innovation Institute of Technology, Chung-Li, Taiwan, R.O.C.

Tsu-Liang Chou
Department of Mechanical Engineering, Taoyuan Innovation Institute of Technology, Chung-Li, Taiwan, R.O.C.

Follow this and additional works at: https://jmstt.ntou.edu.tw/journal

Part of the Engineering Commons

Recommended Citation
Hsu, Tze-Chi; Chen, Jing-Hong; Chiang, Hsin-Lu; and Chou, Tsu-Liang (2014) "COMBINED EFFECTS OF MAGNETIC
FIELD AND SURFACE ROUGHNESS ON LONG JOURNAL BEARING LUBRICATED WITH FERROFLUID," Journal of Marine
Science and Technology: Vol. 22: Iss. 2, Article 6.
DOI: 10.6119/JMST-013-0207-4
Available at: https://jmstt.ntou.edu.tw/journal/vol22/iss2/6

This Research Article is brought to you for free and open access by Journal of Marine Science and Technology. It has been
accepted for inclusion in Journal of Marine Science and Technology by an authorized editor of Journal of Marine Science and
Technology.
154 Journal of Marine Science and Technology, Vol. 22, No. 2, pp. 154-162 (2014 )
DOI: 10.6119/JMST-013-0207-4

COMBINED EFFECTS OF MAGNETIC FIELD

AND SURFACE ROUGHNESS ON LONG
JOURNAL BEARING LUBRICATED WITH
FERROFLUID

Tze-Chi Hsu1, Jing-Hong Chen1, Hsin-Lu Chiang2, and Tsu-Liang Chou2

Key words: ferrofluid, long journal bearing, stochastic surface tude angle of the bearings, while decreasing the modified
roughness. friction coefficient in cases of high eccentricity. However, the
model used to determine the applied magnetic field must be
selected with care to avoid causing side leakage due to ex-
ABSTRACT
cessive film pressure. Other similar studies include those by
This study investigated the influence of ferrofluids on the Montazeri [12], and Huang et al. [8].
lubrication performance of long journal bearings under the The above studies all assumed that the surfaces within
combined effects of stochastic surface roughness and a mag- bearings are entirely smooth. Christensen [4] and Chris-
netic field generated by an infinitely long wire. According to tensen and Tonder [5] developed stochastic Reynolds equa-
our results, placing an infinitely long wire magnetic field at an tions to describe average pressure, addressed the issue of
appropriate distance from the center of the bearing can sup- transverse and longitudinal roughness, and determined the
press side leakage in long journal bearings, thereby extending influence of roughness on the performance of bearings.
the life of the bearings. Under a higher power-law index and Chiang et al. [1-3] derived a generalized stochastic Rey-
induced magnetic force, the introduction of transverse rough- nolds equation based on Christensen’s stochastic model
ness can enhance film pressure and load capacity, while re- and the Stokes microcontinuum theory [20]. Through nu-
ducing the attitude angle and modified friction coefficient. merical simulation, the performance of lubrication in bear-
The introduction of longitudinal roughness has the opposite ings under the combined influence of couple stresses and
effect. We believe that these findings provide a valuable ref- surface roughness was determined in order to characterize
erence for the design of bearings in the future. the dynamic squeeze film. Hsu et al. [7] applied Christen-
sen’s stochastic roughness theory to the problem of two
parallel circular disks, the results of which demonstrate that
I. INTRODUCTION
surface roughness could improve the behavior of squeeze
Ferrofluids are stable colloidal liquids comprising ferro- films. Magnetic particles are often trapped in the valleys
magnetic particles suspended within a carrier fluid. Special associated with surface roughness; therefore, Chu et al. [6]
attention must be paid to the non-Newtonian characteristics of and Osman et al. [16, 17] used the power-law flow model
ferrofluids under various operating environments [6, 9-11, 14, as the rheology model for ferrofluid characteristics. None-
18-20, 22]. Nada and Osman [13] and Osman et al. [15-17] theless, previous studies have tended to focus on the impact
investigated the influence of ferrofluid lubricants on the op- of surface roughness or ferrofluids, without considering the
erational characteristics of journal bearings, using specifically influence of magnetic fields. The objective of the current
designed magnetic field models. Their results indicate that study was to investigate the combined effects of surface
increasing the power law index of ferrofluids and the intensity roughness and ferrofluid characteristics as well as magnetic
of the magnetic field can enhance the load capacity and atti- fields.
This study combined Christensen’s stochastic surface rough-
ness model with a magnetic field generated using an infinitely
Paper submitted 10/16/12; revised 01/18/13; accepted 02/07/13. Author for long wire and investigated how this would affect bearings
correspondence: Jing-Hong Chen (e-mail: [email protected]). lubricated with ferrofluids. It is hoped that this comprehensive
1
Department of Mechanical Engineering, Yuan-Ze Unversity, Chung-Li,
Taiwan, R.O.C.
research on the lubrication performance of hydrodynamic
2
Department of Mechanical Engineering, Taoyuan Innovation Institute of bearings will provide a valuable reference for the design of
Technology, Chung-Li, Taiwan, R.O.C. bearing in the future.
 T.-C. Hsu et al.: Combined Effects of Magnetic Field and Surface Roughness on Long Journal Bearing Lubricated with Ferrofluid 155

y
where I is the strength of the current passing through the wire
Wπ /2
L→ ∞ and K = R0/R. The optimum ψ, as given by Tarapov [21] is
ω
π/2.
θ
z
According to Christensen’s theory, using the expected
ψ
h values derived from Eq. (1), the stochastic modified Reynolds
e

R
U equation for the journal bearing with rough surface can be
R0 g wire
ϕ
Infinit
ely lon x
written as:
W0
Infinitely long wire
W
∂   h n+2  ∂p  ∂  n + 2 ∂p  n ∂E ( h)
E 
∂x   n
  + E h
∂ ∂ 
(∂ 
)
 = 6m0 ( Rω ) ∂x
h
 
x z z
δ = δ1 + δ 2
hm
σ 2 = σ 12 + σ 22
δ1 ∂   h n+2   ∂ n+2
δ2
+ E   f MX  + ( E (h ) f MZ ) (5)
A. Transverse B. Longitudinal ∂x   n   ∂z
Roughness Roughness

Fig. 1. The frame structure of the transmitted bit stream.

where the expectancy operator E(·) is defined by

II. THEORETICAL ANALYSIS E (⋅) = ∫ (⋅) f (δ )dδ (6)

−∞

Fig. 1 depicts the configuration of a long bearing with a

journal radius R rotating at angular speed ω, with the lubrica- and f(δ) is the probability density distribution for the stochas-
tion of the interior bearings provided by ferrofluid. A mag- tic variables. As most rough surfaces in the field of engi-
netic field is produced by a current passing through an infi- neering are Gaussian in nature, for the sake of simplicity, we
nitely long wire, which is displaced at a distance (R0) greater used a polynomial form for integration (instead of a Gaussian
than the radius of the bearing. The wire is placed at angle distribution function), as shown in the following:
ψ from the centerline of the bearing.
In accordance with the Navier-Stokes equation, magnetic  35 2
 (c − δ 2 )3 if − c ≤ δ ≤ c
force was treated as an external body force and the modified f (δ ) =  32c 7 (7)
Reynolds equation was then obtained as [13]  0 elsewhere

∂  h n+2 ∂p  ∂  n+2 ∂p  ∂h ∂  h n+2  where c is the half total range of random film thickness vari-
 + h  = 6m0 ( Rω ) n +  f MX 
∂x  n ∂x  ∂z  ∂z  ∂x ∂x  n  able and, the function terminates at c = ±3σ, where σ is the
standard deviation.
In this study, local film geometry of the lubricant is treated
∂ n+ 2
+ (h f MZ ) (1) as a stationary, ergodic, stochastic process with zero mean,
∂z written as follows:
where n is the power-low index, p is local film pressure, and
h = hm ( x, z ) + δ ( x, z , ξ ) (8)
m0 is the viscosity constant. The fMX and fMZ respectively
represent magnetic force in the circumferential and axial di-
rections, written as follows: where hm represents the nominal smooth part of the film ge-
ometry according to the x and z coordinates, while δ (x, z, ξ) is
∂hM the part due to the surface asperities measured from the
f MX = µ0 X M hM (2) nominal level and is regarded as a randomly varying quantity
∂x
of zero mean. Using the circumferential coordinate, film thick-
ness hm can also be expressed as
∂hM
f MZ = µ0 X M hM (3)
∂z hm = C (1 + ε cos θ ) (9)
The magnetic field hM of the infinitely wire long [13] is
represented by where C is the radial clearance, and ε is eccentricity ratio,
ε = e/C .
The surface feature of one-dimensional longitudinal rough-
I
hM (θ ) = (1 + K 2 − 2 K cos{ψ − θ }) −0.5 (4) ness is assumed to have the form of long narrow ridges and
2π R valleys running in the direction of rotation. Thus, the thick-
156 Journal of Marine Science and Technology, Vol. 22, No. 2 (2014 )

ness of the lubricating film listed in Eq. (8) can be expressed as 1 n 2 2

G1 ( H , Λ, n) = H n+ 2 + H Λ (n + 3n + 2) (20)
a function of the form 18

h = hm ( x) + δ ( x, ξ ) (10) 1 Λ 2  n 2 + 5n + 6 
G2 ( H , Λ, n) = n+ 2 + (21)
H 18 ⋅ H n+ 4
in which Eq. (5), dealing with longitudinal roughness, can be
reduced to and the intensity of the dimensionless magnetic field of Eq.
(4) can be written as
∂   h n+2  ∂ p  ∂  1 ∂p n ∂E ( h)
E   +  n+ 2  = 6m0 ( Rω ) H M (θ ) = (1 + K 2 − 2 K cos{ψ − θ }) −0.5 (22)
∂x   n  ∂x  ∂z  E (1/ h ) ∂z  ∂x

∂   hn+2   ∂  1  where
+ E   f MX  +  n+2
f MZ  (11)
∂x   n   ∂z  E (1/ h )  H M = hM hM 0 (23)

where p is mean film pressure. and

To simplify analysis, the magnetic field coefficient can be
defined as hM 0 = I 2π R (24)
n +1
hM2 0 µ0 X M  C  One-dimensional transverse roughness is assumed to have
αm =
m0ω n  R 
(12)
the form of long narrow ridges and valleys running in the
z direction. Therefore, the thickness of the lubricant film can
while dimensionless mean film pressure P and surface rough- be expressed as a function of the form:
ness parameter Λ, are written as
h = hm ( x) + δ ( z , ξ ) (25)
n +1
p C
P= (13)
m0ω n  R 
Eq. (15) for transverse roughness can be derived as

c ∂  ∂P  n ∂  ∂P 
Λ= (14)
∂θ G2 ( H , Λ, n) ∂θ  + 4λ 2 ∂Z G1 ( H , Λ, n) ∂Z 
C    

∂ ∂  ∂H M 
The non-dimensional modified Reynolds equation for lon- = 6n G.3 ( H , Λ, n) + α m  G2 ( H , Λ, n) H M
gitudinal roughness can then be expressed as follows: ∂θ ∂θ  ∂θ 

nα m ∂  ∂H M 
∂  ∂P  n ∂  ∂P  +  G1 ( H , Λ, n) H M (26)
G1 ( H , Λ, n) ∂θ  + 4λ 2 ∂Z G2 ( H , Λ, n) ∂Z  4λ ∂Z 
2
∂Z 
∂θ    
∂H ∂  ∂H M  where
= 6n + αm  G1 ( H , Λ, n) H M
∂θ ∂θ  ∂θ 
18 H 2 + Λ 2 [ n 2 + 5n + 6]
nα m ∂  ∂H M  G3 ( H , Λ, n) = H ⋅ (27)
+ G2 ( H , Λ, n) H M (15) 18 H 2 + Λ 2 [ n 2 + 7 n + 12]
4λ ∂Z 
2  ∂Z 
Because λ >>5 in the approximation of a long journal
where bearing, variations in axial pressure can be disregarded in
favor of circumferential variations. Therefore, Eqs. (15) and
hm (26) are reduced to
H= (16)
C
Z = z/L (17) ∂  ∂P 
∂θ G1 ( H , Λ, n) ∂θ  = 6n( −ε ⋅ sin θ )
 
L
λ= (18)
2R ∂  ∂H M (θ ) 
+ αm G1 ( H , Λ, n) H M (θ ) ∂θ  Longitudinal (28)
θ = x/R (19) ∂θ  
 T.-C. Hsu et al.: Combined Effects of Magnetic Field and Surface Roughness on Long Journal Bearing Lubricated with Ferrofluid 157

∂  ∂P  ∂ where
G2 ( H , Λ, n)  = 6n G3 ( H , Λ, n)
∂θ  ∂θ  ∂θ
1 Λ [ n 2 + 5n + 6 ]
2

∂  ∂H M (θ )  G4 ( H , Λ, n) = + (36)
+ αm H n+ 2 18 ⋅ H n+ 4
∂θ G2 ( H , Λ, n) ⋅ H M (θ ) ⋅ ∂θ  Transverse (29)
 
18 H 2 + Λ 2 (n 2 + 5n + 6)
The boundary conditions for film pressure are as follows: G5 ( H , Λ, n) = (37)
18 H 2 + Λ 2 (n 2 + 7 n + 12) θ
2

P(θ ) = 0 at θ = 0 (30)
1
R(ψ , θ , K ) = ⋅  H M2 (θ ) − H M2 (0)  (38)
dP
= 0 at θ = θ 2 (31) 2 
dθ
(1 + K 2 − 2 K cos{ψ − θ })−0.5 ⋅ K ⋅ sin(ψ − θ )
By applying boundary conditions (30) and (31), the non- Q(ψ , θ , K ) = (39)
(1 + K 2 − 2 K cos{ψ − θ }) −1.5
dimensional mean film pressure and θ 2 can be obtained by
integrating Eqs. (28) and (29) as
By integrating the non-dimensional mean film pressure
P(θ ) = acting on the journal bearing, the dimensionless load compo-
nents parallel (W0) and perpendicular (Wπ /2) to the centerline
 θ (cos θ − cos θ )
6 ⋅ ε ⋅ n ⋅ ∫ 0 G ( H , Λ, n) ⋅ dθ + α m ⋅ R(ψ , θ , K )
2 can be obtained by the following:
 1 θ
 θ 1 θ2
1

 − α m ⋅ [G1 ( H , Λ, n) ⋅ Q(ψ , θ , K )] θ ⋅ ∫ 0 ⋅ dθ W0 = −2∫ ∫ 2 P(θ )cosθ dZdθ (40)

 2 G1 ( H , Λ, n) θ 0 0

 Longitudinal (32)
 θ2
1
 Wπ / 2 = 2 ∫ ∫ 2
P(θ )sin θ dZdθ (41)
 θ
0 0

6n ⋅ ∫ 0 G3 ( H , Λ, n) θ − G3 ( H , Λ, n) θ2  ⋅ G4 ( H , Λ, n) ⋅ dθ
 Thereafter, dimensionless load capacity W and attitude
 + α m ⋅ R(ψ ,θ , K ) − α m ⋅ [G2 ( H , Λ, n) ⋅ Q(ψ ,θ , K ) ]
 θ2 angle ϕ can be evaluated from the above, as follows:
 θ
 ⋅ ∫ 0 G4 ( H , Λ, n) ⋅ dθ Transverse (33) 1
 W = W0 2 + Wπ / 2 2  2
(42)

cosθ 2 = ϕ = tan −1 Wπ / 2 W0  (43)

 θ2 cosθ
 6 ⋅ ε ⋅ n ⋅ ∫ 0 G ( H , Λ, n) ⋅ dθ + α m ⋅ [ R (ψ , θ , K ) ]
 1 θ Shear stress on the moving surface in the direction of rota-
 θ2 1 tion can also be calculated, such that the modified friction
 6 ⋅ε ⋅ n ⋅ ∫ ⋅ dθ
0 G ( H , Λ, n) coefficient f R/C is determined by
 1 θ
 θ2 1
 α m ⋅ [G1 ( H , Λ, n) ⋅ Q(ψ , θ , K )] θ ⋅ ∫ ⋅ dθ n
∂P(θ )
1
G ( H , Λ , n) θ 2π 1  
2∫ 2 ∫
0
 − 2
1 ⋅ 1+ γ ⋅
n 
− γ ⋅ α m ⋅ Q(ψ , θ , K  dθ dZ
 θ2 1 R
=
0 0 H  ∂θ 
 6 ⋅ε ⋅ n ⋅ ∫ ⋅ dθ f
C W
0 G ( H , Λ, n)
 1 θ
 (44)
 Longitudinal (34)
 where

{ θ
( ) }
 6n ⋅ 2 G3 ( H , Λ, n) ⋅ G4 ( H , Λ, n) ⋅ dθ + α m ⋅ R(ψ , θ , K )
 ∫0 θ
(1 + ε cosθ ) n+1
 θ2 γ= (45)
 6n ⋅ θ 2 ⋅ ε ⋅ ∫ G5 ( H , Λ, n) ⋅ G4 ( H , Λ, n) ⋅ dθ 2n
0
 θ
 α m ⋅ [G1 ( H , Λ, n) ⋅ Q(ψ , θ , K )] θ ⋅ ∫ G4 ( H , Λ, n) ⋅ dθ 1
 − 2 0
− III. RESULTS AND DISCUSSION
 θ2
ε
 6 n ⋅ θ 2 ⋅ ε ⋅ ∫0 5 G ( H , Λ , n ) ⋅ G4 ( H , Λ , n ) ⋅ d θ
This study investigated the combined influence of surface

 Transverse (35) roughness patterns and a magnetic field produced by an infi-
158 Journal of Marine Science and Technology, Vol. 22, No. 2 (2014 )

70 4
αm = 0
αm = 1
60
αm = 3
αm = 5 3 αm = 5
50

40
2 αm = 0

P
P

30

20
1
Chiang (2004), Smooth, Λ = 0
10 Smooth, Λ = 0,
Longitudinal, Λ = 0.3
Transverse, Λ = 0.3
0 0
0 20 40 60 80 100 120 140 160 180 200 220 0 20 40 60 80 100 120 140 160 180 200 220
θ θ
Fig. 2. Numerical analysis of αm influence on dimensionless pressure Fig. 4. Comparison between current simulations and those reported by
(Z = 0, n = 1, ε = 0.4 , Λ = 0, λ = 5, K = 1.2). Chiang [15] (Z = 0, n = 1, ε = 0.4, λ = 5, K = 3).

65 example, dimensionless pressure increased steadily from θ = 0

K = 1.2 to the maximum value and then swiftly dropped to zero. This
60
K = 1.8 is considered by many scholars to be the ideal curve.
55 K = 2.4 Fig. 4 compares the simulation results of this study and
50 K=3
those of Chiang [2]. Under a magnetic force of 0, our nu-
45
merical results were similar to those obtained by Chiang,
40
demonstrating that an increase in the magnetic field contrib-
P

35
uted to an increase in dimensionless pressure. Using αm = 5
30 as an example, compared to a smooth bearing surface, trans-
25 verse roughness increased Pmax whereas longitudinal rough-
20 ness decreased Pmax.
15 Although the viscosity of ferrofluids changes only slightly
10 in the Schliomis model, a number of studies [6, 16, 17] have
5 shown that viscosity can be affected by differences in the
0 weight ratio of ferromagnetic particles within the fluid.
0 20 40 60 80 100 120 140 160 180 200 220 Therefore, the impact of various power law indices on di-
θ mensionless pressure distribution under the influence of sur-
Fig. 3. Numerical analysis of K influence on dimensionless pressure (Z = face roughness is shown in Fig. 5. Observation revealed that
0, n = 1, ε = 0.4, Λ = 0, λ = 5, αm = 5). with a fixed power law index, surface roughness has a sig-
nificant influence on the distribution of dimensionless pres-
sure. Adopting a smooth surface as the standard, transverse
nitely long wire on the distribution of hydrodynamic pressure roughness increases Pmax, while longitudinal roughness de-
in long journal bearings lubricated with ferrofluids. Fig. 2 creases Pmax. With an increase in the power law index, using
depicts the impact of various magnetic field coefficients on transverse roughness as an example, the shear thickening fluid
dimensionless pressure distribution under smooth surface con- (n = 1.5) has the highest Pmax, followed by Newtonian fluid
ditions. The coefficient of the magnetic field in this study was (n = 1.0) and shear thinning fluid (n = 0.7).
4 λ2 times that of the coefficient used by Osman [12] (i.e., Table 1 illustrates the combined effects of surface rough-
αm = 4 λ2α). Clearly, Pmax increases with the coefficient of the ness and a magnetic field according to variations in Pmax.
magnetic field. However, excessively high and concentrated When the magnetic field coefficient equaled zero, the maxi-
dimensionless pressure was found to cause side leakage, mum dimensionless pressure increased to 14.37%, in con-
thereby reducing the operational life of the bearing. In seeking junction with an increase in transverse roughness from Λ = 0
to address this defect, we discovered that the relative location to Λ = 0.3. In contrast, an increase in longitudinal roughness
of the metal wire has an absolute impact on the distribution of slightly decreased the maximum dimensionless pressure,
dimensionless pressure (Fig. 3). When K was increased from compared to the smooth case. The current study assumed that
1.2 to 3, Pmax decreased significantly. Taking K = 3 as an the long journal bearings had a length-to-diameter ratio equal
 T.-C. Hsu et al.: Combined Effects of Magnetic Field and Surface Roughness on Long Journal Bearing Lubricated with Ferrofluid 159

Table 1. Pmax as a function of surface roughness various αm(n = 1, ε = 0.6, λ = 5, K = 3).

Magnetic Magnetic Magnetic
Sureface field Pmax, Λ − Pmax,α = 0, Λ= 0 field Pmax, Λ − Pmax,α = 0, Λ= 0 field Pmax, Λ − Pmax,α = 0, Λ= 0
m
× 100% m
× 100% m
× 100%
roughness coefficient Pmax,α m = 0, Λ= 0 coefficient Pmax,α m = 0, Λ= 0 coefficient Pmax,α m = 0, Λ= 0

αm = 0 αm = 2 αm = 5
Pmax Pmax Pmax
Λ = 0.1 5.531 -5.52% 5.648 -3.52% 5.879 0.43%
Longitudinal
Λ = 0.3 5.816 -0.6% 5.996 2.43% 6.182 5.60%
Smooth Λ=0 5.854 0% 6.053 3.40% 6.324 8.03%
Λ = 0.1 6.095 4.12% 6.296 7.55% 6.576 12.33%
Transverse
Λ = 0.3 6.695 14.37% 7.092 21.15% 7.364 25.79%

14 250
Smooth
Longitudinal, Λ = 0.3 245
12 Transverse, Λ = 0.3
240
10 n = 0.7
235
n=1
8 230
n = 1.5
θ2
P

6 225
n = 1.0 n = 1.5
220
4
n = 0.7
215
2 Smooth, Λ = 0
210 Longitudinal, Λ = 0.3
Transverse, Λ = 0.3
0 205
0 20 40 60 80 100 120 140 160 180 200 220 0.1 0.2 0.3 0.4 0.5 0.6
θ ε
Fig. 5. Relationship between dimensionless pressure P and θ under the Fig. 6. Relationship between zero pressure gradient angle θ2 and ε under
influence of various power law indices and surface roughness (ε = various power law indices and surface roughness (λ = 5, αm = 5,
0.6, λ = 5, αm = 5, K = 3). K = 3).

to 5; therefore, the axial variations in pressure were neglected indices. Obviously, when the journal bearing is operated using
in favor of those in the circumferential direction. As a result, the same power law index, θ 2 reduces as the eccentricity ratio
the transverse grooves generated stronger hydrodynamic lift, ε increases. Taking shear thickening fluid (n = 1.5) as an
resulting in higher maximum dimensionless pressure; The example, transverse roughness reduced θ 2 by approximately
influence of a magnetic field on maximum dimensionless 33° following an increase in eccentricity from 0.1 to 0.6.
pressure is presented in Table 1. Because the surface was With the same roughness pattern, lower power law indices
assumed to be smooth, the maximum dimensionless pressure lead to higher θ 2. Thus, shear thinning fluids (n = 0.7) produce
increased to 8.03% when the magnetic field coefficient was higher θ 2.
increased from 0 to 5. Finally, combining the influence of The control circuit adjusts the timing alignment between
surface roughness and magnetic field resulted in a maximum the locally generated spreading signal and the received signal.
dimensionless pressure of 25.79%, in the case of Λ = 0.3 and When the timing alignment is correct, the correlator output
αm = 5. It is clear that with an increase in the magnetic field, will reach its maximal value. Therefore, the code acquisition
the influence of surface roughness is far more pronounced can be attained by examining the peak locations of the corre-
than that of a magnetic field and the coupled effects become lator output signal.
more significant. To ensure strong hydrodynamic lift within Fig. 7 exhibits the relationship between dimensionless load
the journal bearing and prevent side leakage caused by ex- W and eccentricity ratio ε under various magnetic field coef-
cessive hydrodynamic pressure, an appropriate balance must ficients. Clearly, dimensionless load is directly proportional
be attained between these two factors. to eccentricity ratio. The increase in dimensionless load re-
Fig. 6 illustrates the relationship between θ 2 and the ec- sulting from an increase in the magnetic field coefficient is
centricity ratio ε under the influence of various power law more apparent under conditions of low eccentricity ratio.
160 Journal of Marine Science and Technology, Vol. 22, No. 2 (2014 )

10 1.6
Smooth, Λ = 0 Smooth, Λ = 0
Longitudinal, Λ = 0.3
9 Longitudinal, Λ = 0.3
Transverse, Λ = 0.3
Transverse, Λ = 0.3
8 1.4

7 ε = 0.3

6 1.2

f R/C
W

4 αm = 5 1
ε = 0.6
3

2 αm = 0 0.8

1
0.1 0.2 0.3 0.4 0.5 0.6 0 1 2 3 4 5
ε αm

Fig. 7. Relationship between dimensionless load and eccentricity ratio Fig. 9. Relationship between modified friction coefficient and magnetic
under the influence of various magnetic fields and surface rough- field coefficient under various eccentricity ratios and surface
ness (n = 1, λ = 5, αm = 5, K = 3). roughness (n = 1, λ = 5, K = 3).

75 1

72

69
0.9
αm = 0
66
n = 0.7
63
f R/C

n = 1.0
ϕ

60 0.8 αm = 5
n = 1.5
57

54
Smooth Smooth
0.7
51 Longitudinal, Λ = 0.3 Longitudinal
Transverse, Λ = 0.3 Transverse
48
0 0.1 0.2 0.3 0.4
0.1 0.2 0.3 0.4 0.5 0.6
ε Λ
Fig. 8. Relationship between attitude angle and eccentricity ratio under Fig. 10. Relationship between modified friction coefficient and surface
the influence of various power law indices and surface roughness roughness under the influence of various magnetic field coeffi-
(λ = 5, αm = 5, K = 3). cients and surface roughness ( n = 1, ε = 0.6, λ = 5, K = 3)

Moreover, surface roughness patterns also have an influence eccentricity increased from 0.1 to 0.6. With the same rough-
on dimensionless load. Taking αm = 5 as an example, trans- ness pattern, lower power law indices lead to higher attitude
verse roughness increases dimensionless load (compared with angles. Thus, shear thinning fluids (n = 0.7) produce higher
smooth bearing surfaces), whereas longitudinal roughness attitude angles.
slightly decreases dimensionless load at a fixed eccentricity Fig. 9 illustrates the relationship between the modified
ratio. friction coefficient and the magnetic field coefficient under
Fig. 8 illustrates the relationship between attitude angle ϕ various eccentricity ratios. Taking ε = 0.3 as an example, the
and eccentricity ratio ε, under the influence of various power modified friction coefficient shows a distinctively decreasing
law indices. Clearly, the attitude angle was markedly reduced trend as the magnetic field coefficient increases. However, in
with an increase in eccentricity ratio. With a fixed power the case of a high eccentricity ratio (ε = 0.6), only a slight
law index, the influence of roughness patterns on attitude change is observed. Transverse roughness reduces the modified
angle grew with an increase in eccentricity ratio. Taking shear friction coefficient, whereas longitudinal roughness slightly
thickening fluid (n = 1.5) as an example, longitudinal rough- increases the modified friction coefficient.
ness reduced the attitude angle by approximately 21° when Fig. 10 illustrates the relationship between the modified
 T.-C. Hsu et al.: Combined Effects of Magnetic Field and Surface Roughness on Long Journal Bearing Lubricated with Ferrofluid 161

friction coefficient and roughness parameters under the x, y, z rectangular coordinates

influence of variations in the magnetic field. Using αm = 5 as Z dimensionless coordinate in the z-direction, Z = z/L
an example, the modified friction coefficient associated with αm magnetic field coefficient, α m = (hM2 0 µ0 X m / m0ω n )
transverse roughness decreases as the roughness parameter
increases; however, the modified friction coefficient associated (C R) n+1
with longitudinal roughness increases. ε eccentricity ratio, ε = e/C
δ random part of film geometry
θ circumferential coordinate, x = Rθ
IV. CONCLUSION
θ2 zero-pressure gradient angle
This study investigated the influence of lubricants con- λ length-to-diameter, λ = L / 2 R
taining suspended ferromagnetic particles on the lubrication Λ roughness parameter, Λ = c/C
characteristics of long journal bearings under the combined µ0 permeability of free space of air, µ0 = 4π × 10-7
influence of a magnetic field and various surface roughness AT/m
patterns. Using Christensen’s stochastic surface roughness ω angular speed
model and an infinitely long wire magnetic field model, we
derived a Reynolds equation for long journal bearing ap- REFERENCES
proximation. The results obtained in other studies support the
1. Chiang, H. L., Lin, J. R., and Hsu, C. H., “Linear stability analysis of
theoretical models outlined in this study and demonstrate the
a rough short journal bearing lubricated with non-Newtonian fluids,”
reliability of the proposed numerical analysis. Tribology Letters, Vol. 17, No. 4, pp. 867-877 (2004).
According to our results, the combined effect of surface 2. Chiang, H. L., Lin, J. R., and Hsu, C. H., “Lubrication performance of
roughness and magnetic forces within bearings are significant finite journal bearings considering effect of couple stresses and surface
and cannot be disregarded. Compared to bearings with roughness,” Tribology International, Vol. 37, pp. 297-307 (2004).
3. Chiang, H. L., Lin, J. R., Hsu, C. H., and Chang, Y. P., “Lubrication per-
smooth surfaces, the introduction of transverse roughness and
formance of long journal bearings considering effects of couple stresses
a magnetic field can increase the film pressure and load ca- and surface roughness,” Journal of the Chinese Institute of Engineers, Vol.
pacity, while decreasing the attitude angle and modified fric- 27, No. 2, pp. 287-292 (2004).
tion coefficient. Longitudinal roughness has the opposite 4. Christensen, H., “Stochastic models for hydrodynamic lubrication of
effect. The aforementioned trends become increasingly sig- rough surface,” Proceedings of the Institution of Mechanical Engineers,
Vol. 55, pp. 1013-1025 (1969-70).
nificant with an increase in the power-law index and eccen-
5. Christensen, H. and Tonder, K., “The hydrodynamic lubrication of rough
tricity ratio. We believe that these simulation results will bearing surface of finite width,” ASME Journal of Lubrication Tech-
provide a valuable reference for the design of bearings in the nology, Vol. 93, pp. 324-330 (1971).
future. 6. Chu, L. M., Li, W. L., Hsu, H. C., and Chang, Y. P., “Magneto-
elastohydrodynamic lubrication of circular contacts at impact loading,”
Tribology International, Vol. 42, pp. 333-339 (2009).
NOMENCLATURE 7. Hsu, C. H., Lai, C., Lu, R. F., and Lin, J. R., “Combined effects of surface
roughness and rotating inertia on the squeeze film characteristics of
c one half of the total range of the random film thick- parallel circular disks,” Journal of Marine Science and Technology, Vol.
ness variable 17, No. 1, pp. 60-66 (2009).
C radial clearance 8. Huang, W., Shen, C., and Wang, X., “Study on static supporting capacity
e eccentricity, e = ε C and tribological performance of ferrofluid,” Tribology Transactions, Vol.
fmx magnetic forces in x-direction (circumferential d 52, pp. 717-723 (2009).
9. Katiyar, A., Singh, A. N., Shukla, P., and Nandi, T., “Rheological of mag-
direction) netic nanofluids containing spherical nanoparticles of Fe-Ni,” Powder
fmz magnetic forces in z-direction (axial direction) Technology, Vol. 224, pp. 86-89 (2012).
hM magnetic field intensity 10. Lin, J. R., “Derivation of ferrofluid lubrication equation of cylindrical
hM 0 characteristic value of magnetic field intensity squeeze films with convective fluid inertia forces and application to cir-
HM dimensionless magnetic field intensity, HM = hM/hM0 cular disks,” Tribology International, Vol. 49, pp. 110-115 (2012).
11. Lin, J. R., Liang, L. J., and Chu, L. M., “Effects of non-Newtonian mi-
h thickness of lubricant film cropolar fluids on the squeeze-film characteristics between a sphere and a
hm, H nominal smooth part of the film thickness, H = hm/C = plate surface,” Proceedings of the Institution of Mechanical Engineers,
1 + ε cosθ Part J: Journal of Engineering Tribology, Vol. 224, pp. 825-832 (2010).
K distance ratio parameter K = R0/R 12. Montazeri, H., “Numerical analysis of hydrodynamic journal bearings
L length of the bearing lubricated with ferrofliud,” IMechE, Part J: Journal of Engineering Tri-
bology, Vol. 222, pp. 51-60 (2008) .
m0 viscosity constant 13. Nada, G. S. and Osman, T. A., “Static performance of finite hydrody-
n power-law index namic journal bearings lubricated by magnetic fluids with couple
pm induced magnetic pressure stresses,” Tribology Letters, Vol. 27, pp. 261-268 (2007).
p,P mean film pressure, P = ( p / m0ω n )(C / R ) n +1 14. Oladeinde, M. H. and Edokpia, R. O., “Performance modeling of infi-
nitely wide exponentially shaped slider bearing lubricated with couple
R radius of the journal stress fluids,” Journal of Engineering and Applied Science, Vol. 5, No. 8,
Xm susceptibility of ferrofluid pp. 26-31 (2010).
162 Journal of Marine Science and Technology, Vol. 22, No. 2 (2014 )

15. Osman, T. A., Nada, G. S., and Safar, Z. S., “Static and dynamic charac- ance of a Wet Friction Clutch Material,” Wear, Vol. 181-183, No. 2, pp.
teristics of magnetized journal bearings lubricated with ferrofluid,” Tri- 850-855 (1995).
bology International, Vol. 34, pp. 369-380 (2001). 19. Sihuan, Y., Huang, W., and Wang, X., “Orientation effects of micro-
16. Osman, T. A., Nada, G. S., and Safar, Z. S., “Effect of using current- grooves on sliding surface,” Tribology International, Vol. 44, pp. 1047-
carrying-wire models in the design of hydrodynamic journal bearings 1054 (2011).
lubricates with ferrofluid,” Tribology Letters, Vol. 11, pp. 61-70 (2001). 20. Stokes, V. K., “Couple stress in fluids,” The Physics of Fluids, Vol. 9, pp.
17. Osman, T. A., Nada, G. S., and Safar, Z. S., “Different Magnetic models 1709-1715 (1966).
in the design of hydrodynamic journal bearings lubricated with non- 21. Tarapov, I. E., Magnetohydrodynamics, Vol. 8, p. 444 (1972).
Newtonian ferrofluid,” Tribology Letters, Vol. 14, No. 3, pp. 211-223 22. Wang, X. L., Zhu, K. Q., and Wen, S. Z., “On the performance of dy-
(2003). namically loaded journal bearings lubricated with couple stress fluids,”
18. Scott, W. and Suntiwattana, P., “Effects of Oil additives on the perform- Tribology International, Vol. 35, pp. 185-191 (2002).