1. Introduction
Many researchers have studied the cavitation area prediction of hydrodynamic journal bearing, because the characteristics of the bearings are strongly affected by the gaseous-phase areas. For example, it is known that the gaseous-phase areas of journal bearings expand under starved lubrication conditions; therefore, the stability characteristics under starved lubrication increases, whereas the friction torque decreases. Hashimoto and Ochiai studied the stability characteristics under starved lubrication both theoretically and experimentally [
1]. Furthermore, Ochiai has reported that the gaseous phase has a cooling effect upon the oil film under the starved lubrication condition [
2]. The characteristics of temperature and friction under starved lubrication have been experimentally studied; however, these observations have not been theoretically investigated [
3], because the calculation method for the detailed gaseous-phase area has not yet been proposed. In addition, in actual bearing systems, there are cases where lubrication oil is not able to supply sufficiently and the kind of starved lubrication phenomenon occurs. In the case of strong starved conditions, there is also high possibility to be serious erosion damage on the bearing surfaces. Therefore, it is important to predict the gaseous area.
Generally, the Reynolds equation is used to analyze the journal bearings, and the half-Sommerfeld’s condition or the Swift-Stieber condition have been applied as the simplest method for determining the cavitation areas for the Reynolds equation [
4,
5]. However, these methods ignore the oil films of diverging clearance, because the negative-pressure area is assumed to be the cavitation area. Therefore, the flow-rate conservation does not hold. Coyne and Elrod’s condition is additionally used as a more advanced method; Coyne and Elrod proposed a calculation model that assumes a two phase flow of oil-film rupture to calculate the surface tension between the oil film and cavitation [
6,
7]. The condition is closer to practical bearings than that of the half-Sommerfeld or Swift-Stieber conditions; however, it is impossible to estimate the cavitation area of the entire journal bearing. Therefore, several researchers have proposed boundary condition models that consider the cavitation. For example, Ikeuchi and Mori have analyzed the oil film cavitation areas while using the modified Reynolds equation [
8,
9]. This equation is derived by considering the two-phase flow as an averaged single-phase flow of oil and gas. However, this theoretical model depicts difficulty in converging in the case of a high eccentricity ratio and a starved lubrication condition. Furthermore, Ikeuchi and Mori did not confirm the agreement of their analytical results with the actual cavitation area. However, Boncompine et al. and Hatakenaka et al. have proposed an analytical method by assuming a finger-type cavitation [
10,
11]. However, bearing designers are not able to estimate the variation of the gaseous-phase area when the amount of oil lubricant is changed. Furthermore, some journal bearings contain an oil filler port, which modifies the bearing characteristics [
12,
13]. However, the internal flow of the oil-filler port and the bearing clearance cannot be solved while using the Reynolds equation. Therefore, a different approach is required to analyze the journal bearing while considering the internal flow of an oil-filler port.
Recently, two phase flow computational fluid dynamics (CFD) analysis models that consider vapor pressure and interfacial transport equations have been proposed [
14]. Among other methodologies, the volume of fluid (VOF) method offers good convergence and allows bearings to be evaluated in a relatively short time [
15]. Furthermore, the VOF method can be applied to a journal bearing environment, because it can analyze the flow of slag. To date, analysis utilizing VOF has been utilized to investigate the characteristics of journal bearings while using the reported analytical results. For example, Zhai et al. and Dhande et al. have analyzed the cavitation area of a journal bearing using the VOF method by considering the vapor pressure. Additionally, they have investigated the influence of rotational speed upon the cavitation area [
16,
17]. However, the validity of the calculation method that has been used in this study has not been verified, because the analytical cavitation area has not been compared with that observed in the experiment. On the other hand, Egbers et al. analyzed the journal bearings while using the analytical method that depicted a combination of the Reynolds equation and CFD analysis by considering the two-phase flows [
18]. Furthermore, the cavitation area in the oil-filler port and the opposite load side that was obtained by the analytical method in this study have been compared with that obtained in the experimental results. However, the analytical results cannot precisely produce the cavitation scale and shape, because the influence of the surface tension has not been calculated.
In this situation, the current study visualized the gaseous-phase area on both the bearing surface and the oil-filler port in a small-bore journal bearing under the flooded and starved lubrication conditions. The journal bearing is further analyzed while using a CFD model that is based on the volume of oil, eccentricity ratio, and attitude angle that were obtained from the experiment. Further, the impacts of VOF, surface tension, and vapor pressure of the setting condition were studied, and the analytical and experimental results were compared. Finally, authors considered the influence of surface tension on journal bearing from the Weber number, We.
2. Experiment
Figure 1 depicts the geometry of the test journal bearing that was used in the experiments.
Table 1 lists its major dimensions. The upper part of the test bearing is provided with an oil filler port, allowing for the lubricating oil to flow into the bearing clearance. The diameter of the bearing (
D), clearance of the bearing (
Cr), bearing width (
L), width diameter ratio (
L/D), and the diameter of the oil-filler port (
Dp) are 25.0 mm, 125 μm, 14.5 mm, 0.58, and 8.2 mm, respectively. The test bearing is manufactured while using transparent acryl, which allows for us to observe the formation of the oil film and the generation of the gas.
Figure 2 depicts the schematic of the entire experimental test rig. This rig has a rotor installed at its center and a revolving shaft that is supported by two bearings on its left and righthand sides. The test bearing for visualizations is installed on the right-hand side. The shaft is driven while using a DC motor that is capable of controlling the rotational speed continuously up to a speed of 10,000 rpm. Eddy current type proximity probes are placed in both the horizontal and the vertical direction of the bearing, which allows the journal position to be measured. An oil tank is placed on the top of the bearing, the lubricating oil is supplied through a control valve, and the oil that is leaking from the side end of the bearing is returned to the oil tank while using the pump. The viscosity grade of the lubricant oil is VG22 and the temperature of the supplied oil is fixed to 40 °C. The application of a blue lubricant oil using xylene makes it possible to visualize the region where the gas is being generated.
The oil supply flow rate is measured based on the oil leakage from the side end. A halogen light is used to visualize the gaseous-phase areas on the surface of the bearing, whereas the sheet laser designator is used to visualize the oil film rupture. In the experiment, oil film rupture under the flooded lubrication conditions could not be identified, because the oil leakage from the side end was observed to be large. However, the impact of the interface was large under the starved lubrication conditions, because the gaseous-phase area was observed to broaden. Therefore, in this study, the oil film rupture under a starved lubrication condition has been visualized. In the experiment, the rotational speed, volume of fluid under flooded lubrication, and volume of fluid under starved lubrication are n = 3500 rpm, q = 2.6 cm3/s, and 0.5 cm3/s, respectively.
3. Theory
At present, there are two types of methods for numerically analyzing flows, regardless of the multiphase flow, which are the level set [
19], VOF [
20], front tracking method [
21], and the lattice Boltzmann method [
22] as multiphase flow analysis. The former is a method of treating fluid as a continuum and solving it by discretizing the conservation law of mass, momentum, and energy represented by a partial differential equation. The other is a method of determining the flow by treating the fluid as a collection of virtual particles with a finite number of velocities, and sequentially calculating the collisions and translations that are performed by the virtual particles. The multiphase flow analysis that was developed from the former is the level set, the VOF, and the front tracking, and the latter is the lattice Boltzmann method. In particular, representative examples of the method include the finite-difference method, the finite-element method, and the boundary-element method. Among them, the VOF method can guarantee the volume preservation of fluid at the basic equation stage, and the convergence of gas phase is expressed by volume fraction, and the calculation cost is better when compared with level set method and front tracking method. It is low cost. From the above, in this study, the flow of the journal bearing is analyzed by the VOF method using ANSYS FLUENT 15.0.
3.1. Governing Equation
The Navier–Stokes equation considering the surface tension is used to analyze the journal bearing instead of the Reynolds equation in this study. The mass conservation equation and the momentum equation are shown, as follows;
where
ρ fluid density,
velocity vector,
p fluid pressure,
fluid viscosity,
gravitational force,
σ surface tension,
normal vector,
γ curvature of the boundary surface, and
Dirac’s delta-function.
Here, in the VOF model, using the volume fraction
F, the fluid density and fluid viscosity are expressed, as follows,
where subscription 1 means oil and 2 means gaseous phase.
This CFD analysis can analyze the internal flow of an oil-filler port with bearing clearance simultaneously, because the inertia term is considered in the basic equation.
3.2. Surface Tension and Cavitation Model
The effect of surface tension was considered in this analysis. The continuum surface force (CSF) model that was proposed by Brackbill et al. [
23] was used as the surface tension model implemented in FLUENT out of the box. The last term in Equation (2)
means surface tension. In Brackbill et al.’s CSF model, the effect of surface tension is included as the surface tension term in the Navier–Stokes equation.
In addition, the cavitation model of Schnerr and Sauer [
18] was used. The equation for the volume fraction of steam is as follows.
This cavitation model used the following equation to relate the vapor volume fraction
F to the number of bubbles
per unit volume of liquid and bubble radius.
where
number of bubbles was set as 10
13 in this study.
While considering the vapor pressure, the volume of air that is dissolved in oil expanded and negative pressure was observed in the journal bearing. In this study, the vapor pressure was set to zero. Further, the flow was laminar and the analysis was conducted in a steady state condition.
3.3. Calculation Model
Figure 3 depicts the bearing CFD calculation model. The clearance of the bearing, oil-filler port, and oil supply groove were modeled, and the overall flow in the bearing could be calculated. The calculation model was symmetrically configured against the bearing center. The volume of oil supply, eccentricity ratio, and attitude angle are
q = 2.6 cm
3/s,
ε = 0.54, and
φ = 72.9°, respectively, under flooded lubrication conditions. Further, the volume of oil supply under starved lubrication was
q = 0.5 cm
3/s, the eccentricity ratio was
ε = 0.76, and the attitude angle was
φ = 71.5°. The rotating shaft speed was
n = 3500 rpm. These values were determined based on the experimental results. The computational grid in the minimum clearance contains six layers in the direction of the bearing clearance. The total mesh number under flooded and starved lubrication conditions were 64 × 10
5 and 18 × 10
6, respectively. Previously, we confirmed that mesh size was sufficient for the calculations.
Table 2 shows the calculation conditions. The tension between oil and air was also considered while performing the calculations. The surface tension was set to 0.04 N/m, which was measured while using the du Noüy method (ASTM 971-50).
In this study, to clarify the effects of vapor pressure and surface tension, four types of calculations, which are (i) VOF (negative pressures are allowed), (ii) VOF with vapor pressure, (iii) VOF with surface tension; and, (iv) VOF with vapor pressure and surface tension are conducted.
5. Influence of Surface Tension in Journal Bearing
In the previous section, it was found that the surface tension strongly influenced the gaseous-phase areas in journal bearings under starved lubrication. We considered surface tension influence journal bearing from Weber number
We, which is a dimensionless number, and it is expressed in the following equation:
where
U represents speed and
H represents length.
The Weber number We is a value that is obtained by driving the fluid inertia force by the surface tension. Generally, if We is less than one, the surface tension strongly influences the gas-liquid two-phase flow. This means that the inertial force of the fluid becomes smaller than the surface tension, so the influence of the surface tension becomes remarkable.
First, we examined the surface tension influence on bearing clearance. In this study, it was assumed that the journal center coincided with the bearing center and determined the representative speed
U(=
R(2π
N/60)) and representative length
H.
Figure 14 depicts the relationship between the Weber number
We and the internal flow of bearing clearance. From this figure, the Weber number approximately exceed one in rotating speed 1 m/s, so it is found that surface tension does not have significant influence on flow in the bearing clearance under the analysis condition in the previous section under both lubrication conditions.
However, it has been confirmed by analysis results that the influence of surface tension is significant to gaseous-phase areas under starved lubrications. Therefore, we focused on the Weber number
We of the internal flow of an oil-filler port. In terms of representative speed, it was assumed that the flowing lubricating oil speed in an oil-filler port is uniform and it determined representative speed from supply flow rate.
Figure 15 depicts the relationship between the Weber number
We and the oil flow rate. The continuous line in this figure indicates the value in this study as a reference, whereas several plots indicate the value in other studies of journal bearings under the starved lubrication conditions (Hashimoto and Ochiai [
1], Tanaka [
24], Taura [
25]). Focusing on the continuous line, the Weber number
We drop below 1 approximately under 4 cm
3/s and the influence of surface tension in the internal flow of the oil-filler port become too large to ignore. Moreover, it is found that the Weber number
We is remarkably small in other research of journal bearing under starved lubrications.
From these results, in the case of flooded lubrication conditions, it is considered that the gaseous-phase of analysis condition with surface tension and without surface tension do not greatly differ because the Weber number We of flooded lubrication conditions is larger than the starved lubrication conditions. On the other hand, in the case of starved lubrication condition, the oil flow of the oil-filler port is directly influenced by the surface tension, so it is considered that the gaseous-phase of the upstream region remarkably changed.
6. Conclusions
In this study, gaseous-phase areas in journal bearings under flooded and starved lubrication conditions were analyzed while using two-phase flow CFD analysis. Further, the influence of surface tension and vapor settings on journal-bearing analysis were studied. The analytical results were compared with the experimental results, and the VOF calculation was observed to be in good agreement while considering the vapor pressure and surface tension under both lubrication conditions. Furthermore, under starved lubrication, the gaseous-phase area exhibited the same tendency during the analysis and the experiment, and the analytical results agree rather well with the interface of the oil film and cavitation during oil-film rupture. While using these results, CFD analysis of the two-phase flow of VOF with vapor pressure and surface tension allowed for the calculation of the cavitation area of the journal bearing under flooded and starved lubrication conditions. Moreover, it was found that the oil flow of the oil-filler port was directly influenced by the surface tension, so it is considered that the gaseous phase of the upstream region remarkably changed.