Philadelphia University

Faculty of Engineering
Mechanical Engineering Department

Submitted as a report for machine design part 2 course

Methods and approaches to solve

Reynold hydrodynamic equation

Using MATLAB and ANSYS fluent software

Submitted By: Supervisor

Zine Al Abdeen Baraa Shaheen Dr. Mohammed Gohazeh
ABSTRACT
Hydrodynamic journal bearings have been widely used to support high speed
rotating machinery such as turbines and compressors because of their superior
durability and load-carrying capacity. Therefore, the bearings are essential to
machine elements for enhancing the quality of the rotating machinery. In this
dissertation, I declare the importance and advantages of journal bearing and have
an overview of the Reynold hydrostatic equation and applying Boundary Value
finite-difference Solution to solve the second-order nonlinear differential
equation using numerical iterations approach with MATLAB software.
Analytical Solution and Semi-Analytical Solution of 1D transient Reynolds
Equation have been mentioned to investigate the Pressure distribution, pressure
gradient and the comparison between transient and steady loads on journal
bearings also numerical solution using Ansys fluent has been applied to
investigate the most critical parameters that influence the performance of the
journal bearing
 Table of contents
ABSTRACT .......................................................................................................... 2
INTRODUCTION ................................................................................................ 6
1.1 Overview .................................................................................................. 6
1.2 WHEN TO USE FLUID FILM BEARINGS........................................... 7
1.3 BASIC CONCEPTS ................................................................................. 7
1.4 Bearing Performance ................................................................................ 9
1.5 Differential Equation for Hydrodynamic Lubrication Theory ............... 11
Literature review ................................................................................................. 12
Reynold solving methods .................................................................................... 13
2.1 Boundary Value Solution Explicit Finite Difference Method ............... 13
2.2 Analytical Solution of 1D Reynolds Equation ....................................... 16
2.3 Semi-Analytical Solution of 1D Transient Reynolds Equation (Grubin's
Approximation) ............................................................................................... 18
2.4 Generalization of Reynolds Equation ....................................................... 19
2.5 Reynold equation with pressure viscosity relation ................................... 22
Numerical CFD analysis ..................................................................................... 23
Ansys fluent .................................................................................................... 24
Static pressure ............................................................................................. 24
Dynamic pressure........................................................................................ 25
Pressure coefficient ..................................................................................... 25
Acoustic power ........................................................................................... 26
Total pressure .............................................................................................. 26
Turbulence energy ...................................................................................... 27
CONCLUSION ................................................................................................... 28
Appendix ............................................................................................................. 29
A1: Boundary Value Solution Explicit Finite Difference Method ................. 29
A2: Analytical Solution of 1D Reynolds Equation ........................................ 30
A3: Semi-Analytical Solution of 1D Transient Reynolds Equation (Grubin's
Approximation) ............................................................................................... 34
A4 Generalization of Reynolds Equation ....................................................... 39
REFERENCES.................................................................................................... 41
 List of figures
Figure 1 ................................................................................................................. 6
Figure 2 ................................................................................................................. 7
Figure 3 ................................................................................................................. 8
Figure 4: Hydrodynamic Pressure Profile - Low Load and High load ................. 9
Figure 5: Plain Bearing Eccentricity versus Speed with a Constant Load ......... 10
Figure 6: Effect of Load and Speed on Plain Journal Eccentricity..................... 10
Figure 7: l/d vs. mxx,myy ................................................................................. 15
Figure 8:laminar and turbulent............................................................................ 15
Figure 9:Reynold number ................................................................................... 16
Figure 10: Pressure distribution .......................................................................... 17
Figure 11: Pressure gradient ............................................................................... 17
Figure 12: Transient Load comparison ............................................................... 18
Figure 13:Pressure distribution with 20*20nodes .............................................. 21
Figure 14:Pressure distribution with 30*30 nodes ............................................. 21
Figure 15:Pressure distribution with tolerance of 0.1% ..................................... 21
Figure 16:Pressure distribution with tolerance of 1% ........................................ 21
Figure 17:Pressure distribution with tolerance of 0.01% ................................... 22
Figure 18:Pressure distribution with tolerance of 0.1% ..................................... 22
Figure 19:journal bearing .................................................................................... 24
Figure 20:static pressure ..................................................................................... 24
Figure 21:Dynamic pressure ............................................................................... 25
Figure 22:pressure coefficient............................................................................. 25
Figure 23:Acoustic .............................................................................................. 26
Figure 24:Total pressure ..................................................................................... 26

Table 1 ................................................................................................................. 23
INTRODUCTION

1.1 Overview

Journal bearings are one of the most common types of hydrodynamic bearings.
Their primary purpose is to support a rotating shaft. They are used in various
subsystems in engines and power trains, for example, for support of
both crankshaft and camshaft. They are also used in the rocker shaft of rocker-
arm valve train systems.[1]
Hydrodynamic journal bearings have been widely used to support high speed
rotating machinery such as turbines and compressors because of their superior
durability and load-carrying capacity. Therefore, the bearings are essential to
machine elements for enhancing the quality of the rotating machinery. As the
performance characteristics of high-speed, hydrodynamic journal bearings
operated in both laminar and turbulent flow regimes are governed by several
bearing parameters, the bearing designers usually try to select the design variables
within constraints by a trial and error method using many design charts obtained
from the characteristic bearing analysis. However, this approach only guarantees
acceptable solutions, and it does not necessarily produce optimum solutions.
Moreover, even in the case that the bearing designers can get the optimum
solutions successfully by such an approach, a considerable amount of working
time and cost will be needed to complete the optimum design of high-speed
journal bearings.[2]

Figure 1
1.2 WHEN TO USE FLUID FILM BEARINGS

There are applications where anti-friction bearings are the best choice.
Commonly, smaller motors, pumps, and blowers use rolling element bearings.
Papermill rolls often use large specialized spherical roller bearings. Anti-friction
bearings are best for these applications. However, once the size of a pump (or fan
or motor) gets large enough and fast enough, a gray area is entered. Here still find
rolling element bearings used successfully, but as speeds increase and
temperatures rise, rotor dynamics often become a concern, and critical speeds are
encountered. That is when damping is required, and fluid film bearings become
increasingly necessary. That turbomachinery designers (and users) should
consider using fluid film bearings if running above 3,000 RPM or the machine
exceeds 500 HP, at 1,000 HP and up, all machines except exceptional cases
should be on journal bearings specifically designed for that service. There are
exceptions, of course, and the decision where to apply what type of bearing is
ultimately done for every machine individually based on sound engineering
practice and experience. Unfortunately, this decision is sometimes based on
economics, which keeps maintenance engineers and consultants employed.[3]

1.3 BASIC CONCEPTS

To illustrate the basic nomenclature, geometry, and introduce the ideas of how
fluid film bearings work, the most straightforward bearing called an everyday
journal bearing examined. Figure 1 is a photograph of a plain bearing. A steel
base material is overlaid with a babbitt material and bored to a circular diameter
equal to the shaft diameter plus the desired clearance. Scallops are cut at the split
line to admit oil. Figure 2 is a computer model of this same bearing.

Figure 2
At zero speed, the shaft rests on the bearing at the bottom dead center. As soon
as shaft rotation begins, the shaft "lifts off" on a layer of oil. In fluid film bearings,
lubrication is required between a pair of surfaces with relative motion between
them. There is always a convergent wedge developed that is formed due to the
relative surface speeds and the lubricant viscosity to carry the applied load. An
oil pressure film develops with equal and opposite force vectors to the applied
load. One surface drags the lubricant, usually an oil, into a converging gap. As
the space available in this gap decreases, the fluid develops a pressure gradient or
pressure hill. As the fluid leaves the gap, the high pressure helps expel it out the
other side. A simple diagram of this is shown in figure 3.

Figure 3
In this example, the load is supported by a high-pressure oil region, as shown in
figure 5. Each line in the pressure profile represents an oil pressure vector at the
centreline of the bearing. The sum of the vertical components adds up to the
applied load, and the horizontal components cancel out for equilibrium. Oil inlet
ports are placed in areas of minimum pressure. The pressure profile can also be
examined in a three-dimensional format, as shown in figure 4 (low load) and (for
a highly loaded bearing).
 Figure 4: Hydrodynamic Pressure Profile - Low Load and High load

These figures show the bearing "unwrapped" with the top half on the left and the
loaded bottom section on the right. It is important to note several things about the
hydrodynamic pressure profiles. First, the peak pressure is significantly higher
than the specific load (W/LD). Secondly, the pressure at the margins always
returns to the boundary condition, which is usually atmospheric pressure. The
unloaded top half, even though it is cavitated, is substantially at atmospheric
pressure, which is why no cavitation damage ever occurs in this type of bearing.

1.4 Bearing Performance

While the stiffness and damping provided by a journal bearing are crucial, other
design factors must be considered to understand how bearings work. For example,
if the eccentricity is too high, there is a risk of metal-to-metal contact and higher
dynamic loads being imparted to the babbitt causing premature fatigue. If the
eccentricity is too low (the journal is nearly centered), then the machine could
more easily become unstable. Eccentricity is a function of both speed and load.
Figure 5 indicates that, with a constant load, as speed increases, the eccentricity
decreases.
 Figure 5: Plain Bearing Eccentricity versus Speed with a Constant Load

By combining the effect of speed and load on the bearing eccentricity, figure 6
tells the complete story. At low speeds b the eccentricity ratio, e/C is high. At
low loads, the eccentricity is high.

Figure 6: Effect of Load and Speed on Plain Journal Eccentricity

1.5 Differential Equation for Hydrodynamic Lubrication
Theory

The two-dimensional classical theory of fluid mechanics lubrication was 1st

given by O. Reynolds. In 1886, within the wake of a classical experiment by
Beauchamp Tower, he developed an equation celebrated as Reynolds Equation.
The formation and underlying mechanism of the fluid film were analyzed by that
experiment by taking some assumptions that the film thickness is exceptionally
tiny as compared to the axial and longitudinal dimensions of fluid film. If the
lubricator layer is to transmit pressure between the shaft and, therefore, the
bearing, the layer should have a variable thickness. Osborne Reynolds himself
derived "Generalized Reynolds Equation," which depends on density, film
thickness, surface, and transverse velocities. The equation derived initially by
Reynolds was restricted to incompressible fluids. Thus it had been developed
generally enough to incorporate the results of compressibility and dynamic
loading and was aforementioned to be Generalized Reynolds Equation. So, the
ultimate form of Generalized Reynolds Equation was as given:
 h3 dp d  h3 dp d (  h)
( )+ ( ) = 6(U1 − U 2 ) + 12 V
 dx dz  dz dx

Where ρ is that the density, μ is that the viciousness, h is that the film thickness
of the fluid film, U1, and U2 are the surface velocities, and V is that the general
velocity. Within the equation (1), the term was because of the bearing velocities
on the lubricator film and depended on whether or not the bearing surfaces have
angular or translational velocities. In contrast, the term was because of the relative
speed of bearing surfaces within the direction traditional to the fluid film. In most
cases, the bearing is stationary and solely the runner in thrust bearings, and also,
the shaft within the journal bearings are moving, therefore U1=U and U2=0.
Currently, the ultimate equation for incompressible lubricants was based by
Reynolds is as given:
d  h3 dp d  h3 dp d (  h)
( )+ ( ) = 6(U) + 12 V
dx  dx dz  dz dx

Where U is that the sliding velocity, V0 is the motion of the journal center.[4]
LITERATURE REVIEW
Going back to 1959 (L. N. TAO et al.)[5] investigated the Load carrying
capacity, Coefficient of friction, and Pressure distribution using Heun's equation
and Separation of variables method. (Juliana Javorova et al. 2018)[6] presented
on their paper precisely, step by step, with necessary explanation obtaining of a
modified Reynolds equation for lubrication of finite length journal bearing
lubricated with non-Newtonian Rabinowitsch fluid. Starting from the momentum
and continuity equations, and based on the cubic stress constitutive equation, the
velocity components for the two-dimensional flow of a Rabinowitsch fluid are
obtained. For this purpose, differentiation and integration techniques are applied,
.and they found that The performances of journal bearings lubricated with a non-
Newtonian Rabinowitsch fluid can be compared with the case of Newtonian
lubricant through the variation of the non-Newtonian parameter, i.e., the
nonlinear factor . The future analyses can be performed with combinations of
other effects in lubrication performances of journal bearings as the influence of
fluid inertia, elastic deformations effects, surface roughness.
(Gustavo G.Vignoloabc et al. 2011)[8] declared an analytical tool to perform
finite journal bearing calculations is introduced. The novelty of the proposed
regular perturbation method lies in that both the pressure and the Ocvirk number,
are expanded. This idea is supported by the fact that the Ocvirk number is a
dimensionless mean pressure and that the order of magnitude of the
pressure varies with aspect ratio. The square of the aspect ratio is then used
as the perturbation parameter. Pressure and shear stress, as well as the
Ocvirk number and the friction coefficient (total friction force over load-
carrying capacity), were calculated and analyzed as a function of azimuthal
position applies. Aspect and eccentricity ratios.they also mentioned that,
Consequently, the treatment of the Ocvirk number as an expansible variable
produces noticeable improvements in the analytical description of the flow
fields of journal bearings and even more remarkable ones in the calculations of
load-carrying capacity and friction coefficient.
(SaeidDousti et al2016)[9] mentioned the preliminary results attest to the
prominent turbulence effects in pressure and velocity profiles and comparatively
a lesser convective inertia contribution. Convective inertia effects also boost the
load capacity and shift the journal position to a lower eccentricity ratio and higher
attitude angle than what the classical theory predicts, depending on the magnitude
of Re*. The turbulence effect is, in general, more pronounced on static
characteristics by increasing the load capacity similar to the convective inertia
effect.
REYNOLD SOLVING METHODS

2.1 Boundary Value Solution Explicit Finite Difference

Method

Boundary value problems arise in several branches of physics, as any physical

differential equation will have them. Problems involving the wave equation, such
as the determination of normal modes, are often stated as boundary value
problems. A broad class of significant boundary value problems is the Sturm–
Liouville problems. The analysis of these problems involves
the eigenfunctions of a differential operator.
Boundary value problems are similar to initial value problems. A boundary value
problem has conditions specified at the extremes ("boundaries") of the
independent variable in the equation. In contrast, an initial value problem has all
of the conditions specified at the same value of the independent variable (and that
value is at the lower boundary of the domain, thus the term "initial" value).
A boundary value is a data value that corresponds to a minimum or maximum
input, internal, or output value specified for a system or component.[10]
Simply second-order nonlinear differential equation can be solved using an
explicit finite difference method as shown:
2 y dy  2 y yi − 2 yi +1 + yi + 2
+ t ( ) 2 − y = te2t , y(0) = 1, y(1) = e 
u 2
dt t 2 h2

 y yi +1 − yi
 t = ih
t h

yi + yi + 2 h h2 h3
yi +1 = + i ( yi +1 − yi ) 2 − yi − i e 2ih
2 2 2 2
The results of the iterative described as shown: (MATLAB code in A1)

Figure 7: l/d vs. mxx,myy

Figure 8:laminar and turbulent

 Figure 9:Reynold number

2.2 Analytical Solution of 1D Reynolds Equation

The derivation that is presented follows from five assumptions that are valid in
many lubrication applications. The five assumptions are:
• The fluid is assumed to be Newtonian, where stress is directly proportional
to the strain rate.
• Inertia effects and body forces are assumed to be negligible compared to
the viscous terms.
• The variation of pressure across the film is assumed to be negligible.
• Flow is laminar.
• The curvature effects are negligible, implying that the thickness of the
lubricant film is much smaller than the length or width of the bearing and
allowing the use of a Cartesian coordinate system.
In the case of the 1-D Reynolds equation, several analytical or semi-analytical
solutions are available. In 1916 Martin obtained a closed-form solution for
minimum film thickness and pressure for a rigid cylinder and plane geometry.
This solution is not accurate for the cases when the elastic deformation of the
surfaces contributes considerably to the film thickness. In 1949, Grubin obtained
an approximate solution for the so-called elastohydrodynamic lubrication (EHL)
line contact problem, where he combined both elastic deformation and lubricant
hydrodynamic flow. In this solution, it was assumed that the pressure profile
follows the Hertz solution. The model is, therefore, accurate at high loads when
the hydrodynamic pressure tends to be close to the Hertz contact pressure. The
code below shows the methodology of solving the equation to investigate the
pressure distribution and pressure gradient profiles.[11] (MATLAB code in A2)
The radius Ambient Average Width or Viscosity Cavitation
of curvature Pressure sliding length[m] [Pa.s] pressure
[m] [Pa] speed [Pa]

0.1680 10^5 0.33/2 0.0112 0.007 -50*P0

Figure 10: Pressure distribution

Figure 11: Pressure gradient

2.3 Semi-Analytical Solution of 1D Transient Reynolds
Equation (Grubin's Approximation)

The classical Ertel–Grubin solution for the inlet of EHL line contacts has been
revisited. The scheme has been modified to account for the inlet elastic
deformation via linear fracture mechanics. Inlet waviness can be modeled in
steady and transient conditions. The same fracture mechanics scheme has been
applied to the outlet of the contact to model the exit bump and to calculate the
minimum film thickness in the contact. For smooth surfaces, a modified transient
Ertel–Grubin scheme has been used to model the sudden halting motion problem.
The Ertel–Grubin scheme has also been used to calculate shear thinning in
isothermal and thermal conditions with two rheological models Eyring and
GLSS.[12] (MATLAB code in A3)

Figure 12: Transient Load comparison

2.4 Generalization of Reynolds Equation

Various of generalized Reynolds equations were derived from weakening the

assumptions used to derive the classical form. For example, compressible, non-
Newtonian lubricant behavior can be considered. Reynolds equation is used to
predict the thickness of the lubricant film, but also to predict friction developed
by the lubricant on the surfaces. Since many tribological contacts operate in the
highly loaded regime and thin films, the shear rates can be very high (in the order
of 107-109). Many of the typical lubricants start to behave non-Newtonian in the
contact conditions, and therefore, Reynolds equation was generalized to the case
of non-Newtonian lubricant.
Another generalization includes slip boundary conditions. This form of the
Reynolds equation is used to calculate film thicknesses and friction in textured
surfaces or surfaces with high slip.
Then the Finite Difference Method to Solve Reynolds Equation is as shown in
details:

d 3 dp X 2 d 3 dp C dh
(h )+ 2 (h )= (1)
dx dx Z dz dz X dx
d 3 dp h3 .p + h3i −0.5 j .pi −1 j − (h3i + 0.5 j + h3i −0.5 j ).pi. j
Where (h ) = i + 0.5 j i +1 j (2)
dx dx (x) 2

X 2 d 3 dp h3i + 0.5 j .pi +1 j + h3i −0.5 j .pi −1 j − (h3i + 0.5 j + h3i −0.5 j ).pi. j
And ( h ) = (3)
Z 2 dz dz (z )2

The boundary conditions are pressure equal to zero at node zero

p i. j = 0

pi. j = Ai. j .pi. j +1 + Bi. j .pi. j −1 + Ci. j .pi. j +1 + Bi. j .pi. j −1 + Ei. j (4)

By rearranging the previous 1,2,3 equations

X 2 x 2 3 C x
(h3i + 0.5 j .pi +1 j + h3i −0.5 j .pi −1 j − (h3i + 0.5 j + h3i −0.5 j ).pi. j ) + ( ) .h i. j (pi. j +1 + pi. j −1 − 2 pi. j ) = (h i +1 j − hi −1 j )
Z z
2
X 2
(5)
Then by assuming film thickness is only function od x the nodal non-
dimensional pressure is:
 
 3 3 
h i + 0.5 j h i −0.5 j
pi. j =  .p + .p 
i −1 j +
 3 X 2
x 
i +1 j
 X 2
x  
  h i + 0.5 j + h i −0.5 j + 2 2 ( ) .h i. j  + + 
3 2 3 3 3 2 3
 i + 0.5 j
h h i − 0.5 j 2 ( ) .h i. j 
 Z z   Z 2 z  
X 2 x 2 3
( ) h i. j (h i +1 j − hi −1 j )
x C
 3
Z 2 z
X 2 x 2 3 
( p i −1 j + p i +1 j ) −
2 X  3 X 2 x 2 3 
 i + 0.5 j
h + h 3
i − 0.5 j + 2 ( ) .h i. j   i + 0.5 j
h + h 3
i − 0.5 j + 2 ( ) .h i. j 
 Z 2 z   Z 2 z 

(6)
x
=1
z

By solving equation (4) and (6) j has been omitted from all h values
 
 3 3 
pi. j =  h i + 0.5
.pi +1 j +
h i −0.5
.pi −1 j  +
 3 X x 2 3
2
  X x 2 3
2
 
  h i + 0.5 + h i −0.5 + 2 2 ( ) .h i.   h i + 0.5 + h i −0.5 + 2 2 ( ) .h i.  
3 3 3

 Z z   Z z   (7)
X x 2 3
2
( ) h i. (h i +1 − hi −1 )
x C
 3
Z 2 z
X 2 x 2 3 
( p i −1 j + p i +1 j ) −
2 X  3 X 2 x 2 3 
 h i + 0.5 + h i −0.5 + 2 2 ( ) .h i.   h i + 0.5 + h i −0.5 + 2 2 ( ) .h i. 
3 3

 Z z   Z z 

By implementing the previous equation in MATLAB, as shown in Appendix

A4, the pressure profile is:
 Figure 13:Pressure distribution with Figure 14:Pressure distribution with 30*30
20*20nodes nodes

As shown in figure 13, 14 there are issues with the pressure distribution with
20*20 and30*30 number of nodes so by increasing the number of nodes to
50*50 the results will be as shown in figure15 and16 for a tolerance of 1%
and0.1%

Figure 16:Pressure distribution with tolerance of Figure 15:Pressure distribution with

1% tolerance of 0.1%
Figure 18:Pressure distribution with tolerance of Figure 17:Pressure distribution with tolerance of
0.1% 0.01%

The suitable tolerance for the iteration is 0.01% percentage of error, as shown in
figure 17, where the pressure distribution profile is in appropriate condition.

2.5 Reynold equation with pressure viscosity relation

d 3 dp d 1 dp d
(h )+ ( ) = 6U h
dx dx dz  dz dx

 =  e p

d  h3 dp  3 d  1 dp  d 
 p +h  p  = 6U  h 
dx  e dx  dz  e dz   dx 

By assume:
1 − e− p dq dp
q= ; = e − p
 dx dx

d  3 dq  3 d  dq  d 
h +h   = 6U  h 
dx  dx  dz  dz   dx 

The pressure results are shown in table1:

Table 1
Pressure Multiplier
1.e5 1.0010
1.e6 1.0101
1.e7 1.1052
1.e8 2.7183
1.e9 22.026

NUMERICAL CFD ANALYSIS

K.P. Gertzos et al. (2008) [13] also mentioned the performance characteristics
of a hydrodynamic journal bearing lubricated with a Bingham fluid are derived
through three-dimensional computational fluid dynamics (3-D CFD) analysis.
The FLUENT software package is used to calculate the hydrodynamic balance of
the journal using the so-called "dynamic mesh" technique. The results obtained
from the developed 3-D CFD model are found to be in excellent agreement with
experimental and analytical data from previous investigations on Bingham fluids.
Additionally, they mentioned that Recently, following the progress in computer
technology, many researchers began to use commercial computational fluid
dynamics (CFD) programs in their investigations. The main advantage of CFD
code is that it uses the full Navier–Stokes equations and provides a solution to the
flow problem at the end of their research they conclude that at high value of
relative eccentricity, a core is formed and adheres to a small region of the journal.
As the value of eccentricity increases, the solid on the bearing separates into two
or three parts, and a hollow core between these parts is observed. M. Deligant et
al. (2011) [14] study on their research a 3D CFD model to compute power friction
losses due to journal bearings. Computations were carried out for various oil
entrance temperatures and rotational speeds. Results are presented and discussed,
making comparisons with some sets of experiments carried out in the CNAM
laboratory using a special turbocharger test rig equipped with a torque meter. M.
Deligant, P et al. (2009) [15], decided to solve the Reynolds equation by
antidifference method using the Gauss-Seidel iterative method in the case of
turbocharger bearings. P. Podevin et al. (2005) [16] studied gave good results,
but the friction power was still overestimated compared to experiments. That was
due to the isothermal assumption which neglects viscous heating and
modification of the viscosity. That is the reason for developing the 3D CFD
model, which is presented in their paper.
 Ansys fluent

By drawing a simple journal bearing with a diameter of 25 mm, as shown

below:

Figure 19:journal bearing

By assuming a rotation of 1500rpm with oil lubricant the following parameters
investigated
Static pressure
with maximum static pressure 452 pascal

Figure 20:static pressure

Dynamic pressure

The dynamic pressure of 625 pascals at the interior of the bearing

Figure 21:Dynamic pressure

Pressure coefficient
With a maximum value of 452

Figure 22:pressure coefficient

Acoustic power
The maximum power of acoustic is 15.8dB

Figure 23:Acoustic
Total pressure
The total pressure is 745 as a maximum value and 1.97 pascals as the minimum
value

Figure 24:Total pressure

Turbulence energy
Some turbulence has been noticed with the highest energy of 7 Kjoule
CONCLUSION
The journal bearing is one of the most critical applications of hydrodynamic
lubricant devices, Reynold equation has the primary role in investigating the
pressure gradient and pressure distribution in the journal bearing. A simple
boundary condition solution has been motioned firstly, but it has a limit to 30
elements. The 1D analytical solution has also been declared and the pressure
distribution and gradient plotted and investigated. Also, the comparison
between the transient and steady-state has been motioned.
After that general finite-difference solution has been applied, and the pressure
gradient profile plotted and has the most appropriate shape with 50*50 element
mesh and 0.001%tolerance.
Ansys fluent has been used to apply numerical analysis to the journal bearing,
static pressure, dynamic pressure, velocity, turbulence effect, and acoustics
power have been taken in consideration and declared at the final section of the
report
APPENDIX

A1: Boundary Value Solution Explicit Finite Difference

Method
function FDM()
nn=38.0;%number of elements and nn+1 is the number of nodes
lngth=1.0; %length of the domain
he=lngth/nn; %increment
t=[0.0:he:lngth]; %nodal points in domain
AC=0.0001;%For Accuracy(tolerance)
U=zeros(nn+1);%to initialize the solver
U(1)=1.0; U(nn+1)=exp(1); %Boundary conditions
c=true;
while(c)
[U1]=equation(U,nn,he);%calling the equation function
c=false;
for i=1:nn+1
if (abs(U(i)-U1(i))>AC)%to check the value with tolerance
c=true;
break;
end
end

U=U1;
end
disp('Hence solution=:');

U2(1:nn+1)=U

t=[0:he:lngth];
figure(1)
plot(t,U2,'o',t,exp(t),'-')
hold on
end

function[U1]=equation(U,nn,he)
for i=1:nn-1
U1(i+1)=((U(i)+U(i+2))/2)+i*(he/2)*((U(i+1)-U(i))^2)-
he*(he/2)*U(i)-i*(he/2)*he*he*exp(2*i*he);
end
U1(1)=U(1); %To keep the boundary conditions fixed
U1(nn+1)=U(nn+1);
end
A2: Analytical Solution of 1D Reynolds Equation
function Analytical_Reynolds()
global P0;
R = 0.2; % m, Radius of curvature
P0 = 10^5; % Pa, ambient pressure
Um = 0.4/2; %%% m/s U1 = 2piR => Um = U1/2 avergae slding speed
B = 0.0112; %%% m, width or length of the contact
mu = 0.007; % viscosity %%% Pa*s
Wl =41.24; % Newtons
Wll = Wl/(B) % Newtons/m
Pc = -50*P0;
Pc = P0; %%% cavitation pressure, Pa
global alpha;
%%%%% left_border parameter shows the %
%%%%% starting point of ksi to plot
global left_border;
global step_ksi;
left_border = -pi/2.15;
step_ksi = 0.0001;
%%%% dimensionless parameters %
global A;
global A1;
global Wld;
global Pcd;

A = R*P0/(12*mu*Um);
Wld = Wll/(R*P0);
Pcd = (Pc - P0)/P0;
[ksic fval exitflag]=
fminsearch(@find_ksic,0.5,optimset('MaxFunEvals',15000,'MaxIter',120
00));

if exitflag~=1
display('Algorithm did not converge');
end
%%%% calculate pressure for given parameters %%%%
[P x a xc ksi ksic w] = calc_p(ksic);
%
%%%% calculate pressure without cavitation with %
%%%% same parameter a %%%%%%%%%%%%%%%%%%%%%%%%%%%
[Prey xrey] = calc_p_reynolds(a);

%%%% calculate load %%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%% in dimensions %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
w_t = sum(P.*step_ksi.*sqrt(2*a)./cos(ksi).^2);
w_calc_dimens = w*R*P0

%%%% calculate approximate pressure for case of %

%%%% large load %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
[Pwl xwl awl xcwl ksiwl ksicwl wwl] = calc_p_wl();

%%%% calculate approximate pressure for case of %

%%%% light load %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
[Pl xl al xcl , ~, ksicl wl] = calc_p_pc();
%%%% pressure gradient distribution %%%%%%%%%%%%%%%%%%%%%%
c1 = -1/cos(ksic)^2;
dpdx = (cos(ksi).^2 +
c1.*cos(ksi).^4)*sqrt(2*a)/(A*a^2).*cos(ksi).^2./sqrt(2*a);
h3 = figure;
plot(x,dpdx,'m','LineWidth',2);
xlabel('Xc, dimensionless','fontsize',14);set(gca,'FontSize', 14);
ylabel('dP/dx, dimensionless','fontsize',14);set(gca,'FontSize',
14);
title('Pressure Gradient','fontsize',14);
grid on;
%%%% pressure distribution with and without cavitation
%%%%%%%%%%%%%%%%%
h3 = figure;
plot(x.*R,P.*P0,'r','LineWidth',2);
xlabel('Xc, m');
ylabel('P, Pa');
%title(strcat('Pressure. a =
',num2str(a*R),'m','.Wl=',num2str(Wll),'N/m'),'fontsize',14);
grid on;
hold on;
set(gca,'FontSize',14)
title('Pressure Distribution');
%%%% first equation of the system (from load balance)%
function W = load_equation(ksic)
c1 = -1/cos(ksic)^2;
W = 0.5*(1+3/4*c1)*(1+(ksic+pi/2)*tan(ksic)) + 1/8;
%%%% second equation of the system (from Reynolds equation) %%
function P = pc_equation(ksic)
c1 = -1/cos(ksic)^2;
P = 1/8*sin(2*ksic)*(1+3/2*c1)+1/2*(1+3/4*c1)*(ksic+pi/2);
%%%% equation for ksic (for numerical solution)
function goal_val = find_a(a)
global Pcd;
global A1;
global A;
global Wld;
global alpha;
global P0;
alphap = alpha*P0;

A1 = A;
ksic = 0.4436;
dksi = (pi/2+ksic)/100;
ksi = [-pi/2+0.1:dksi:ksic];
c1 = -1/cos(ksic)^2;
k = 3*pi/16*c1+pi/4;
fksi = 1/4.*sin(2.*ksi) + ksi./2 + c1.*(1/32.*sin(4.*ksi) +
1/4.*sin(2.*ksi) + 3.*ksi./8) + k;

Pmu = sqrt(2*a)/(a^2*A1).*fksi -1/alphap;

integral = sum((sqrt(2*a)./cos(ksi).^2.*log(-alphap.*Pmu)).*dksi);
if a<=10^-10
goal_val = 10^3;
else
goal_val = abs(-Wld*alphap - integral);
end
%%%% equation for ksic (for numerical solution)
function goal_val = find_ksic(ksic)
global Pcd;
global A;
global Wld;
f1 = load_equation(ksic);
f2 = pc_equation(ksic);
goal_val = abs(2*Pcd/(sqrt(A)*Wld^(3/2)) - f2/(f1)^(3/2));

%%%% Pressure function

function [P x a xc ksi ksic w] = calc_p(ksic)
global A;
global Wld;
global Pcd;
global left_border;
global step_ksi;

f1 = load_equation(ksic);
a = 2*f1/(A*Wld);

c1 = -1/cos(ksic)^2;
k = 3*pi/16*c1+pi/4;

ksi = left_border:step_ksi:ksic;

P = sqrt(2*a)/(a^2*A)*(1/4.*sin(2.*ksi) + ksi./2 +
c1.*(1/32.*sin(4.*ksi) + 1/4.*sin(2.*ksi) + 3.*ksi./8) + k);

x = sqrt(2*a).*tan(ksi);

xc = sqrt(2*a)*tan(ksic);

w = 2*load_equation(ksic)/(a*A);

%%%% Approximate value of pressure in case of infinite load %%%

function [P x a xc ksi ksic w] = calc_p_wl()
global A;
global Wld;
global Pcd;
global left_border;
global step_ksi;

ksic = 0.4436;
f1 = load_equation(ksic);
a = 2*f1/(A*Wld);

c1 = -1/cos(ksic)^2;
k = 3*pi/16*c1+pi/4;

ksi = left_border:step_ksi:ksic;

P = sqrt(2*a)/(a^2*A)*(1/4.*sin(2.*ksi) +ksi./2 +
c1.*(1/32.*sin(4.*ksi) + 1/4.*sin(2.*ksi) +3.*ksi./8) + k);
x = sqrt(2*a).*tan(ksi);

xc = sqrt(2*a)*tan(ksic);

w = 2*load_equation(ksic)/(a*A);

%%%% Approximate value of pressure in case of light load %%%

function [P x a xc ksi ksic w] = calc_p_pc()
global A;
global Wld;
global Pcd;
global left_border;
global step_ksi;

ksic = 0.6334;
f2 = pc_equation(ksic);
a = (sqrt(2)*f2/(A*Pcd))^(2/3);

c1 = -1/cos(ksic)^2;
k = 3*pi/16*c1+pi/4;

ksi = left_border:step_ksi:ksic;

P = sqrt(2*a)/(a^2*A)*(1/4.*sin(2.*ksi) +ksi./2 +
c1.*(1/32.*sin(4.*ksi) + 1/4.*sin(2.*ksi) +3.*ksi./8) + k);

x = sqrt(2*a).*tan(ksi);

xc = sqrt(2*a)*tan(ksic);

w = 2*load_equation(ksic)/(a*A);

%%%% Pure Reynolds solution (no cavitation)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [P x] = calc_p_reynolds(a)
global A;
global left_border;
global step_ksi;

c1 = -4/3;

ksi = left_border:step_ksi:-left_border;

P = sqrt(2*a)/(a^2*A)*(1/4.*sin(2.*ksi) + ksi./2 +
c1.*(1/32.*sin(4.*ksi) + 1/4.*sin(2.*ksi) + 3.*ksi./8));

x = sqrt(2*a).*tan(ksi);
A3: Semi-Analytical Solution of 1D Transient Reynolds
Equation (Grubin's Approximation)
function [AmplDiff, PhaseShift] = transient_analytical_grubin(w)
%%%% this file works on semi-analytical transient approximate
solution
Nt = 1200; %% number of time steps
R = 2*10^-3; % Radius, m
B = 6*10^-3; % width of the cylinder, m
P0 = 10^5; % ambient pressure, Pa
Um = 0.1; %%% m/s, average sliding speed, (U1+U2)/2
mu = 0.34; % viscosity, Pa s
Wl = 2.5; % Load, Newtons
Wll = (2.5+3.75)/B; % Newtons/m
El=231*10^9/2; %%% GPa, reduced Elastic modulus
alpha = 2.8*10^-9; %%% 1/ Pa, pressure viscosity coefficient
alpha = alpha*P0; %% make it dimensionless
% R = 0.01; % Radius
% Um = 6.28; %%% m/s U1 = 2piR => Um = U1/2
% mu = 0.4; %1 ;%* 10^-3; %%% Pa*s
% Wl = 2.5; % Newtons
% Wll = 1.15*10^6; % Newtons/m
% El=300*10^9/2; %%% vmoih oboznacheniyah, 2/El = 1-nu2/E1+.. , a u
nego 1/El=...
% %%% => chtobi moi uravneniya ispol'zovat', nado ego El podelit' na
2.
% alpha = 2.5*10^-8;
% %alpha = 1.4*10^-8/5; %%% to compare with VIchards, fig.5
% %alpha = 1.4*10^-8*5; %%% to compare with VIchards, fig.5
% alpha = alpha*P0; %% make it dimensionless

if nargin==0
w = 341; %% Hz, frequency
end
if nargin==1
plot_flag = 0;
else
plot_flag = 1;

A = R*P0/(12*mu*Um);
N = 1/A;
kappa = w*2*pi*R*sqrt(mu/(Um*Wll));
ke = w/Um*sqrt(Wll*R/El);
Wd = Wll/(El*R);
kp = Wd^1.5/(alpha*N);

S = w*R/Um;

display(strcat('Kappa=',num2str(kappa)));
Number_of_waves = 7;
t1 = 0; %% in seconds
tend = Number_of_waves/w; %% in seconds
dt = (tend-t1)/Nt; %% dimensional
t = [t1:dt:tend].*w; %% dimensionless
dt = dt.*w; %% dimensionless

Wlld = Wll/(P0*R);
Am = 0.25*Wll;
F0 = Wll/(El*R);
Am = 3.75/B;
Am = Am/(El*R);
F = F0 + Am.*sin(2*pi.*t);
dF = Am.*2*pi.*cos(2.*pi.*t);

H0 = 4.8936.*mu*Um*R./(Wll);
H0 = H0/R;

ke = w/Um*sqrt(Wll*R/El);
z = [0.0001:0.01:1];
f = @(H0)GrubinsStationary(H0,F0,A,alpha);
[H0 fval] = fzero(f,[10^-26 1]);
h0t =
(2.3664*mu*Um*alpha/P0*sqrt(R)/((Wll/(2*El))^(1/8)))^(8/11); %%%
reshenie Grubina, stacionarnoe
b = H0/F0;
[RHS Gp Gs Gw] = calc_rhs(H0,F0,A,alpha,0,S)
[Gpa Gsa Gwa] = calc_approximate_Gi(b*2);

idxi = 6610; %%% for plots

H0 = H0;
H0T = H0;
H0M = H0;
XT = {};
H0S = H0;
H0SG = h0t/R;
H0SKR = 1;
for i = 2:length(t)
tc = t(i);
%%%%%% solve system of equations
flag = 0;
k = 1;
H0NP=H0;
while flag==0
%dt = 0;

H0N = RungeKutta4(F(i),A,S,H0,dt,alpha,dF(i)); %%% zdes'

doljno bit' H0 - s predidushego shaga.
dH0 = (H0N-H0)./dt;

error = abs(H0N-H0NP)./H0NP;
if error<10e-3 || k>100
flag = 1;
end
if k>100
display(strcat('Warning! Time dependent value did not
converge:',num2str(i)));
end
 H0NP = H0N;
k = k + 1;
end
H0 = H0N;
f = @(H0)GrubinsStationary(H0,F(i),A,alpha);

H0T = [H0T H0];

[H0s fval] = fzero(f,[10^-26 1]);
H0sg =
(2.3664*mu*Um*alpha/P0*sqrt(R)/((F(i)*El*R/(2*El))^(1/8)))^(8/11);
H0sg=H0sg/R;
H0SG = [H0SG H0sg];
H0S = [H0S H0s];
H0SKR = [H0SKR H0S(end)/H0M];
end

%%%% find the amplitude of film thickness

AmplDiff = max(H0T) - min(H0T);
%%%%%%%%%%%%%%%
tt = Number_of_waves-1;
idx = find(t>=tt);
temph0t = H0T(idx);
temph0s = H0S(idx);
[rm rt] = max(temph0t);
[rm rs] = max(temph0s);
ph_shift = rt-rs;
delta_phase = ph_shift*2*pi/length(idx)*180/pi;
%%% calculate the phase shift between h0 transient and stationary
% [delta_phase percentage] = find_phase_shift(H0T,H0S);
PhaseShift = delta_phase

% kappa = w*R*sqrt(mu/(U*P0));
if plot_flag==1
figure
plot(t./w,H0T,'LineWidth',3);
hold on;
plot(t./w,H0S,'r','LineWidth',3);
xlabel('Time,s','FontSize',14);
ylabel('H0/R','FontSize',14);
legend('Transient','Stationary');
grid on;
xlim([t1 tend]);
title('Transient Load: Comparison','FontSize',14);
set(gca,'FontSize',14);
end
end

function H0 = RungeKutta4(F,A,S,H0,dt,alpha,dF)
w0 = H0;
k_1 = dt*calc_rhs(w0,F,A,alpha,dF,S);
k_2 = dt*calc_rhs(w0+k_1/2,F,A,alpha,dF,S);
k_3 = dt*calc_rhs(w0+k_2/2,F,A,alpha,dF,S);
k_4 = dt*calc_rhs(w0+k_3,F,A,alpha,dF,S);
w1 = w0 + (1/6)*(k_1+2*k_2+2*k_3+k_4);
w0 = w1;
H0 = w1; %%% H0 after one time step dt;
end

function [RHS Gp Gs Gw] = calc_rhs(H0,F,A,alpha,dF,S)

z = [0.0001:0.001:1];
f = @(z)calc_Gp(H0,F,z);
Gp = integral(f,z(1),z(end));
f = @(z)calc_Gs(H0,F,z);
Gs = integral(f,z(1),z(end));
f = @(z)calc_Gw(H0,F,z);
Gw = integral(f,z(1),z(end));
t1 = A*F.^(3/2)/(alpha*sqrt(2));
t2 = S*sqrt(2/F);
t3 = dF.*Gw;
t4 = Gs+F.^3/(2.*H0.^3);

RHS = -((t1-Gp)/t2-t3)/t4;
RHS = (t3-(t1-Gp)/t2)/t4;

end
function Gp = calc_Gp(H0,Wl,z)
delta = (H0.*z.^2/Wl + sqrt(1-z.^2) - z.^2.*log((1+sqrt(1-
z.^2))./z)).^3;
Gp = z.^2.*(sqrt(1-z.^2)-z.^2.*log((1+sqrt(1-z.^2))./z))./delta;
end
function Gs = calc_Gs(H0,Wl,z)
delta = (H0.*z.^2/Wl + sqrt(1-z.^2) - z.^2.*log((1+sqrt(1-
z.^2))./z)).^3;
Gs = z.^3./delta;
end
function Gw = calc_Gw(H0,Wl,z)
delta = (H0.*z.^2/Wl + sqrt(1-z.^2) - z.^2.*log((1+sqrt(1-
z.^2))./z)).^3;
Gw = z.^3.*(log((1+sqrt(1-z.^2))./z) - sqrt(1-z.^2))./delta;
end
function [delta_phase percentage] = find_phase_shift(H0T,H0S)

C2FFT = fft(H0T);
C2PWR = abs(C2FFT);
[maxv idxmax] = max(C2PWR(2:end));

f = @(phase)fit_phase(H0T,idxmax,phase);
x_phase2 = fminbnd(f,-pi,pi);
A1 = (max(H0T)-min(H0T))/2;
x = [1:length(H0T)];
C21 = A1.*sin(2*pi/length(H0T)*idxmax.*x+x_phase2)+mean(H0T);
x_phase3 = x_phase2;
if x_phase2<0
x_phase2 = x_phase2 + pi; %%% bring the angle to format from 0
to 2*pi
end

C1FFT = fft(H0S);
C1PWR = abs(C1FFT);
[maxv idxmax] = max(C1PWR(2:end));
f = @(phase)fit_phase(H0S,idxmax,phase);
x_phase1 = fminbnd(f,-pi,pi);
A1 = (max(H0S)-min(H0S))/2;
x = [1:length(H0S)];
C11 = A1.*sin(2*pi/length(H0S)*idxmax.*x+x_phase1)+mean(H0S);
x_phase4 = x_phase1;
if x_phase1<0
x_phase1 = x_phase1 + pi; %%% bring the angle to format from 0
to 2*pi
end
C111 =
A1.*sin(2*pi/length(H0S)*idxmax.*x+x_phase4+(abs(abs(x_phase4)-
abs(x_phase3))))+mean(H0S);
COF_CLEAN = C11./H0T;
COF_MOVED = C111./H0T;
TT = (COF_CLEAN - COF_MOVED);
[mv mi]=max(TT);
percentage = (COF_CLEAN(mi) - COF_MOVED(mi))./COF_CLEAN(mi)*100;
delta_phase = abs(x_phase1-x_phase2)*180/pi;

end
function tol = fit_phase(C1,idxmax,phase)
x = [1:length(C1)];
A1 = (max(C1)-min(C1))/2;
C11 = A1.*sin(2*pi/length(C1)*idxmax.*x+phase)+mean(C1);

tol = sum(abs(C11-C1));

end
function [Gpa Gsa Gwa] = calc_approximate_Gi(b)
if b<6.66
alphap = 0.394;
betap = 1.387;
alphas = 2.0965;
betas = 2.247;
else
alphap = 0.4747;
betap = 1.476;
alphas = 1.378;
betas = 2.0496;
end
if b<1
alphaw = 0.181;
betaw = 1.408;
elseif b>1 && b <20
alphaw = 0.1883;
betaw = 1.52;
elseif b>20 && b <220
alphaw = 0.2948;
betaw = 1.667;
else
alphaw = 0.5135;
betaw = 1.781;
end
Gpa = alphap*b.^-betap; %%% exponent approximation from Vichard
Gsa = alphas*b.^-betas; %%% exponent approximation from Vichard
Gwa = alphaw*b.^-betaw; %%% exponent approximation from Vichard
end
function RHS = GrubinsStationary(H0,F,A,alpha)
z = [0.0001:0.001:1];
f = @(z)calc_Gp(H0,F,z);
Gp = integral(f,z(1),z(end));
% b = H0/F;
% [Gp Gsa Gwa] = calc_approximate_Gi(b*2);
t1 = A*F.^(3/2)/(alpha*sqrt(2));
RHS = t1-Gp;
end

A4 Generalization of Reynolds Equation

N=60;
M=60;
C=0.03;
Z1=25;
X1=10;
delxbar=1/N;
delzbar=1/M;
const1=X1*X1/(Z1*Z1);
ITER=1000;
for i=1:N+1
for J=1:M+1
p(i,J)=0;
end
end
sum(1)=0;
for K=1:ITER
sumiJ=0;
for i=2:N
X(i)=1/N*(i-1);
h=2/3*(2-X(i));
hm=2/3*(2-X(i)-0.5*delxbar);
hp=2/3*(2-X(i)+0.5*delxbar);
hm1=2/3*(2-X(i)-delxbar);
hp1=2/3*(2-X(i)+delxbar);

cubh=h*h*h;
cubhm=hm*hm*hm;
cubhp=hp*hp*hp;
const2=(cubhp+cubhm+2*const1*cubh);
A=const1*cubh/const2;
CA=cubhp/const2;
D=cubhm/const2;
E=-0.5*delxbar*(hp1-hm1)*C/(const2*X1);
for J=2:M
Z(J)=1/M*(J-1);
p(i,J)=A*p(i,J+1)+A*p(i,J-1)+CA*p(i+1,J)+D*p(i-1,J)-E;
sumiJ=sumiJ+p(i,J);
end
 end
sum(K+1)=sumiJ;
percentage=abs(sum(K+1)-sum(K)/abs(sum(K+1)));
if percentage<0.001
break
end
end
y=K
surf(p);
REFERENCES
1) S. Balakrishnan... H. Rahnejat, in Tribology and Dynamics of Engine and Powertrain,
2010
2) Hiromu Hashimoto, in Tribology Series, 1998
3) Understanding Journal Bearings Malcolm E. Leader, P.E. Applied Machinery
Dynamics Co. Durango, Colorado
4) The solution of Reynolds Equation for the Short Journal Bearings Rotation System
Dr. Mohammad Miyan Associate Professor, Department of Mathematics, Shia
P.G.College, Lucknow, India -226020, International Journal of Pure and Applied
Physics. ISSN 0973-1776 Volume 13, Number 1 (2017), pp. 117-124
5) GENERAL SOLUTION OF REYNOLDS EQUATION FOR A JOURNAL
BEARING OF FINITE WIDTH Author(s): L. N. TAO Source: Quarterly of Applied
Mathematics, Vol. 17, No. 2 (JULY, 1959), pp. 129-136 Published by Brown
University Stable URL: http://www.jstor.org/stable/43634920 Accessed: 09-07-2016
06:10 UTC
6) On the modified Reynolds equation for journal bearings in a case of non-Newtonian
Rabinowitsch fluid model Juliana Javorova1,*, and Jordanka Angelova2 1UCTM,
Dept. of Applied Mechanics, 8 Kliment Ohridski Blvd., 1756 Sofia, Bulgaria
2UCTM, Dept. of Mathematics, 8 Kliment Ohridski Blvd., 1756 Sofia, Bulgaria
7) an Extended Reynold Equation Applicable to High Reduced Reynolds Operation of
Journal Bearings
8) The approximate analytical solution to Reynolds equation for finite length journal
bearing author links open overlay panel Gustavo G.Vignoloabc Daniel O.BarilácLidia
M.Quinzania
9) An extended Reynold equation applicable to high reduced Reynolds number operation
of journal bearing author links open overlay panel Saeid Dousti Paul Allaire Timothy
DimondJianming Cao
10) Daniel Zwillinger (12 May 2014). Handbook of Differential Equations. Elsevier
Science. Pp. 536–. ISBN 978-1-4832-2096-3
11) Daniel Zwillinger (12 May 2014). Handbook of Differential Equations. Elsevier
Science. Pp. 536–. ISBN 978-1-4832-2096-3.
12) Accuracy and Grid Convergence of the Numerical Solution of the Energy Equation in
Fluid Film Lubrication: Application to the 1D Slider, Sun Zhang 1,2, Mohamed-
Amine Hassini 1 and Mihai Arghir 2,
13) K.P.Gertzos P.G. Nikolakopoulos C.A. Papadopoulos. CFD analysis of journal
bearing hydrodynamic lubrication by Bingham lubricant, Volume 41, Issue
12, December 2008, Pages 1190-1204.
14) Michael Deligant Pierre Podevin Georges Descombes. CFD model for turbocharger
journal bearing performances, DOI: 10.1016/j.applthermaleng.2010.10.030.
15) M. Deligant, P. Podevin, F. Vidal, W. Tyminski, S. Guilain, H. Lahjaily, 3D thermal
steady-state CFD analysis of power friction losses in a turbocharger's journal bearing
and comparison with finite difference method and experimentation, 12th EAEC,
Bratislava 2009.
16) P. Podevin, G. Descombes, V. Hara, A. Clenci, C. Zaharia, Performances of
Turbochargers at Low Speeds. EAEC, Belgrade, 2005.