Graphical Abstract Figure
Graphical Abstract Figure
Close modal

Abstract

Rolling element bearings are an integral component of electric vehicles, supporting radial and axial loads in powertrain components such as electric motor shafts and wheel bearings. Fast-switching inverters enable precise, variable control of motor performance at the cost of possible stray current leakage into mechanical components. These currents naturally seek to cross the insulating fluid film in rolling element bearings. In doing so, a destructive discharge or arc may form and cause irreversible damage to metallic bearing surfaces. A unique contribution of the work is that it provides a method to use the statistical height distribution to predict the likelihood of electrical breakdown and discharging. To predict film thickness it uses a closed-form elasto-hydrodynamic lubrication (EHL) models to present a semi-analytical model of this discharging phenomenon. Existing EHL models are modified for mixed lubrication and electrical contacts by incorporating a solid rough surface asperity contact model and a flow factor modified lubrication model. The model accounts for transient effects and considers changes in speed and other parameters during operation. The resulting model predicts the likelihood of surface damage and electrical properties of the bearings through the statistical asperity height above a critical value calculation. The damaged regions predicted by the model are in qualitative agreement with the experimental tests.

1 Introduction

Despite boasting entirely dissimilar powertrain architectures from their internal combustion predecessors, electric vehicles still rely heavily on existing mechanical technologies, such as rolling element bearings and gears. As the usage of electric vehicles continues to grow, it becomes increasingly important to improve component reliability and optimize performance. Electric vehicle (EV) power train systems convert chemical potential energy stored in high-voltage DC batteries into an appropriate alternating current (AC) form to drive an electric motor shaft. Three-phase inverters, like pulse-width modulation inverters, facilitate variable speed performance of electric motors. Their stepwise output aims to replicate perfect sinusoidal conditions with high-frequency switching rates, but subsequent common mode voltages drive unwanted currents through unintended, conductive mechanical pathways [1]. Additionally, increased switching rates within modern PWM inverters magnify the presence and the effect of shaft voltages created by uneven magnetic pole distributions from asymmetric stator windings or manufacturing tolerances [2].

The complex electrical environments characteristic of EVs and hybrid-electric vehicles (HEV) can contribute to premature electric motor failures, 40% of which are bearing related [3]. Robust lubricants protect surfaces of bearing components from excessive wear. In many lubrication systems, a relatively nonconductive mineral or synthetic hydrocarbon oil or grease creates a dielectric barrier between conductive metallic surfaces. As leakage currents naturally seek to find a grounding location across bearing components, the lubricating film can store charge, like an electrical capacitor.

When the oil film thickness exceeds a certain threshold, it enters an insulating state, leading to an infinite interface impedance. However, when the film thickness is smaller than the critical value, the interface current experiences a sudden surge because of the high enough electrical field across the oil film, causing it to breakdown and initiate discharging. It should be noted that it is difficult to confirm for certain that discharge did occur as an electrical arc. Once the charged voltage exceeds the dielectric strength of the lubricant, the potential can rapidly discharge and arc across the surfaces, destabilizing and decomposing the oil in its path. The discharge phenomenon leads to the conversion of kinetic energy from the ions and electrons into heat energy and eventually it elevates the contact temperatures exceeding the material's melting point [4]. These thermal effects can result in irreversible damage to the surfaces in the form of frosting, fluting, or pitting [2,510]. As a result of fluting or pitting, the fused ferrous metals get removed in the form of black spheres creating craters at the surface. Furthermore, the liberated electrically induced ferrous particles, exposed to intense heat, have contributed to the degradation and oxidation of the grease. Consequently, this results in a loss of the grease's purity. These forms of mechanical damage alter tribological contacts and increase surface roughness, heavily contributing to vibrations and bearing noise [11].

EV powertrains can incorporate electrically insulating bearing surfaces or electrically conductive lubricants to mitigate the adverse effects of electro-discharge across bearing surfaces [11]. Regardless of component configuration, specific lubricants, speeds, loads, and other performance aspects of system design can be tailored to mitigate or avoid electrical damage. If manufacturers can predict the likelihood and circumstances for electrical damage based on select parameters, future designs can avoid problematic lubricants and tribological conditions susceptible to damage. Due to this, researchers are assembling models to predict the occurrence of damage [10,12,13], such as in the work by Schneider et al. which reviews models on the full rolling element bearing and transient electrical models based on the elasto-hydrodynamic lubrication (EHL) film thickness [14]. Some previous models employ an electrical circuit analysis [15] but are only loosely tied to tribological principles. A more recent work sought to improve the modeling of capacitance in bearings to improve prediction capabilities [16]. However, more work is still required for electrified rolling contacts to include roughness, start–stop motion, and grease lubrication.

Due to the rapid expansion of EV production, work on electrified mechanical contacts is still establishing itself. Despite this, the area of electrical connectors has been well researched. Likewise, there is a growing usage of lubricants in electrical connectors aimed at minimizing corrosion and fretting [17,18]. However, issues arise when lubricated conditions or the lubricant alters conductivity for powered and electrical connectors. Contact resistance can increase due to vibrations within the lubricant film, and typical extreme pressure (EP) additives can chemically form a nonconductive solid film on normally conductive surfaces [19]. Rather than risking the formation of a conductivity-altering lubricant film, conductive nanoparticles can be incorporated as a lubricant additive [19,20] to reduce friction and wear while maintaining optimal contact resistance. Similarly, ionic lubricants possess inherently higher electrical conductivity and can be incorporated to achieve comparable effects [21].

The model presented in this work predicts the damage caused by discharging across the surfaces and compares results to experimental results from reciprocating rolling contact measurements. During reciprocation, the film thickness varies as the rotational speed of the rolling element changes. From a rheological perspective of the lubricant, it is essential to acknowledge the significance of the squeeze film phenomenon, particularly caused by abrupt changes in rolling motion acceleration. As the squeezing process progresses, the film thickness gradually decreases toward a steady-state film thickness [22]. Changes in film thickness are not considered within steady-state EHL and require an additional squeeze film effect to account for changing fluid pressure resulting from lubricant squeezed out from contacting surfaces. Similarly, this approach describes a transient EHL problem.

In addition, grease is employed as the lubricant in this work. Grease is a lubricant that behaves as a soft solid until enough shear or pressure release the liquid lubricant held within it by degrading the thickener [23] or pushing it out of the areas of minimum film thickness. Grease is created by adding a thickener, such as lithium, calcium, aluminum complexes, or polyurea, to oil. In this work, the thickener is polyurea, which is often used in EV and electric motor bearings (for additional properties, see Ref. [24]). In EHL, the thickener is often assumed to not influence lubricant rheology in the contact. However, that is only sometimes the case and the base oil viscosity may be due to the lack of widely accepted alternative method [24]. Grease can be pushed out of contact and not allow for replenishment, termed starvation [23,25]. Thickener can enter EHL film and effectively increase viscosity (semisolid layer). This increase in film thickness has been observed at low [26,27] and medium speeds [28].

Electrically induced bearing damage is a growing research area, and numerical modeling is still advancing. The discharging phenomena demand that any approach bridges the disciplines of electrical conduction, heat transfer, chemistry, and multiphase mechanics. The transient nature is also essential to adequately capture the timing of the discharge phenomena. The research, therefore, proposes a simplified transient model of a lubricated electrified rolling bearing contact, which also considers grease lubrication and surface roughness.

2 Numerical Methodology

A robust and efficient model is needed to predict the probability of surface damage resulting from electrical currents discharging across the lubricant film of rolling element bearings. In a rolling element contact, the fluid film develops when motion starts. The pressure generated during film formation is often enough to cause significant deflection of contacting surfaces. This lubrication scenario, often referred to as EHL, has been theoretically and experimentally characterized by many works [29]. When an EHL film develops, the electrical conduction between the metal raceway and the rolling element decreases. Combining EHL and rough surface contact models can produce an approximate transient model of the bearing contact. When considering primary forces at play, it is evident that the applied external force, the fluid lifting force (load support), and the solid rough surface contact force will usually not balance. Therefore, the rolling element bearing will accelerate toward or away from the raceway. This acceleration is used by Newton's method to predict the velocity and position of the bearing. The velocity normal to the surface will also cause a squeeze film mechanism between the surfaces. The details of the theories implemented to include these mechanisms are described in more detail in the following sections. The assembled model can then make predictions of film thickness and electrical performance for different materials, finishes, and operating conditions.

2.1 Force Imbalance.

In an assembled rolling bearing, each element is loaded on one side by the inner race and on the other by the outer race. In the current work, one side of the races is presumably loaded with an external force, FE, while the opposite side only considers the contact and lubrication mechanisms. The sliding and normal motion pressurizes the lubricant that exerts a lifting force of FL. In addition, when the film thickness is small, solid contact could occur between the surfaces and exert a force of Fc. This arrangement closely mimics a rolling ball on disk or a flat experimental test. This results in forces acting in the vertical, z, direction. Summing for forces results in the following Newtonian prediction for acceleration in the z-direction.
(1)

The Newton forward matching time scheme then employs this acceleration to predict the velocity, z˙, and the position, z. Various values were evaluated to find the appropriate time-step for convergence.

Figure 1 shows a schematic of a single spherical rolling element contact. Initially, the film thickness is ho. As the time marching moves forward, the film thickness updates with the motion of the ball, as shown in the following equation:
(2)

2.2 Lubricating Film Thickness.

The equations deduced from numerical EHL simulations by Hamrock and Dowson [30] are used to back-calculate the vertical lifting force of the lubricant film. Although the film thickness is not uniform in an EHL contact, the central film thickness for this model is assumed to be the film thickness, h, in Eq. (2). For spherical EHL contacts, the central film thickness is governed by Eq. (3) for a spherical contact [31].
(3)
The nondimensional parameters used in Eq. (3) are as follows:
(4a)
(4b)
(4c)
The force from EHL is predicted from Eq. (3), which can be solved as Eq. (5):
(5)
Under modeled conditions, the speed and load on the rolling element bearing can change with time, subsequently changing the film thickness. As defined by the Stribeck curve, mixed lubrication exists if the film thickness decreases enough. In the mixed lubrication regime, the roughness can influence the film's flow by obstructing and redirecting it. Patir and Cheng [32,33] derived the flow factor method and a modified form of the Reynolds equation to include the influence of roughness. Flow factors ϕx and ϕy account for the effect of surface roughness redirecting or obstructing fluid flow patterns in the x- and y-direction, respectively. Patir and Cheng provide predictive equations for ϕx and ϕy. If the surfaces are considered isotropic, then ϕx = ϕy. With the flow factors equal, the viscosity and flow factor terms are combined into one effective viscosity term outlined in Eq. (6):
(6)
The effective viscosity is then applied to Eq. (4a) to account for roughness. Assuming isotropic roughness, the Patir and Cheng flow factor is equivalent to:
(7)
To consider the flattening of the surface roughness under the high fluid pressure, the methodology proposed by Ref. [34] and later refined by Ref. [35] for 3d roughness is used. This is critical as it has also been shown that the EHL fluid pressure significantly reduces the surface roughness so that flow factors cannot be used directly [36]. The employed equations forms are adapted from amplitude reduction given by Ref. [35], where Ad is the deformed asperity amplitude and Ai is the initial asperity amplitude:
(8)
where the parameter is given by
(9)
where the parameters M and L are given as follows:
(10)
(11)
The wavelength can be approximately related to the asperity density based on Ref. [37] by
(12)

These reduced roughness parameters are then employed in the flow factors (Eq. (7)) and other places roughness parameters are considered in the model. However, this method would surely deviate from reality or a deterministic, although time consuming, calculation.

Since this model will be compared to tests using ISO 100 NLGI 2 grade polyurea mineral greases, an effective viscosity will be used for µ. As noted previously, at low speeds, the film thickness of EHL contacts with grease is much larger than the base oil alone [26,27]. Measurements made in an Anton-Paar MCR 302e rheometer on the apparent viscosity of the grease for varying shear rates will be used to predict the viscosity, µ. These values are similar to those reported by Sisko [38] and are much larger than the base oil viscosity at room temperature (∼0.1 Pa·s). The effective shear rate in the EHL contact is approximated by the equation given in Ref. [39]. Then the Carreau model has been shown to be a reasonable fit to the grease viscosity behavior [40] and is given as follows [41]:
(13)

Equation (13) is fit to the viscosity measurements, which finds values of γ˙c=0.0017s1, μo=1.78×105Pas, and n = 0.44. The fit is shown with the rheological data presented in Fig. 2 and differs from the data by an average of 7.7%.

Fig. 2
Measured viscosity and the fit equation prediction used as the apparent viscosity in the EHL model
Fig. 2
Measured viscosity and the fit equation prediction used as the apparent viscosity in the EHL model
Close modal
Due to the transient nature of the operating conditions, pressure changes generated from varying film thicknesses are modeled with respect to time by employing the analytical solution to the circular flat squeeze film [31]. As shown in Eq. (14), the Hertz contact solution is incorporated into the analytical solution to predict the radius of the squeeze film, a.
(14)
The resulting forces from the squeeze film and EHL equations are summed to establish the total load support of the fluid (FL).
(15)

2.3 Rough Surface Contact.

In addition to roughness influencing lubricant flow, comparatively tall peaks, known as asperities, create unwanted mechanical contacts despite the average film thickness remaining positive between surfaces. A well-established stochastic model of rough surface contact is employed to account for the contact of these peaks. Since the pressures in EHL contacts can be high, other methods, such as deterministic approaches and multiscale models, are available [42,43]. However, these alternatives can be too computationally expensive and increase complexity when modifying force predictions as a function of film thickness. It is common practice to use statistical models in mixed lubrication models [4451]. There are many versions of the stochastic rough surface contact models [52,53], but the one employed in Eq. (16) assumes a Gaussian distribution and isolated spherical elastic-plastically deforming asperities [54].
(16)
In Eq. (16), P¯ represents the single asperity force predicted by a chosen elastic–plastic contact model. A well-established spherical model is selected in this work [54]. However, other contact geometries, such as sinusoidal [53], also exist and can serve as the topic for the future work. The McCool methodology calculates all roughness parameters from a rough surface contact [55]. Within the height distribution, the rms asperity roughness, or peak roughness, should be used and is related to the rms roughness by [56]
(17)
The solid contact of the asperities can provide a pathway for electrical conduction between the surfaces. If the film thickness greatly exceeds the roughness, asperities are relatively isolated across the surface profile. Therefore, they do not conduct a significant portion of the electrical load. However, if the load or speed is low enough, the film thickness will decrease, bringing more asperities into contact. Equation (18) of the statistical model can also predict electrical spreading resistance or the inverse of conductance.
(18)

In the previous equation, the alleviation factor [57] is given by Ψ=(1Ar/An)3/2, and the real area of contact predicted by the statistical model is Ar. The alleviation factor accounts for the adjacent asperities sharing the electrical conduction once their numbers grow, and they become less isolated under higher contact forces.

In addition to the contact or spreading resistance, the effective film resistance is calculated from resistivity of the fluid, Hertz contact area, and film thickness:
(19)

Note that the total contact resistance is the reciprocal of the sum of the reciprocals of Rs and Rfilm since they are in electrical parallel.

Fig. 1
Schematic of example pseudo-steady-state system
Fig. 1
Schematic of example pseudo-steady-state system
Close modal

2.4 Statistical Prediction of Electrically Induced Damage Between Rough Surfaces.

The dielectric voltage, Vd, is predicted once the asperity model and average film thickness are established. The dielectric voltage is the electrical potential that an insulating material can withstand before breaking down. Commonly referred to as breakdown voltage, the dielectric voltage signifies when the material becomes plasma after breaking electron bonds. The plasma phase change also initiates the discharge phenomena (and perhaps arcing), where the electrical current can cross the once-insulating material. The dielectric voltage, Vd, is calculated as follows:
(20)
where Sd is the dielectric breakdown strength and is approximately 10–15 MV/m [58] for mineral oil. If an electric potential across an oil film exceeds Vd, the model predicts fluid breakdown, allowing an electrical discharge across the film as plasma. The plasma's temperature at atmospheric pressure can be 10,000 K. At temperatures this high, the metal surfaces near the plasma can be melted or ablated.
Since the film thickness between the bearing and raceway is not constant and varies due to roughness and deformation, discharging is believed to initiate on the highest points of the surface (i.e., the asperities). There are two reasons for this: the dielectric voltage is lowest at these points, and the charge tends to build up at these peaks as well [59]. Pitting damage has also been observed more on the peaks or ridges in testing performed that will later be used to compare to the model (see Fig. 3). A critical film thickness, or surface separation, can be calculated from Eq. (20) for an applied voltage (Vapplied) with
(21)
Fig. 3
Scanning electron microscopic image of the damaged area near the end of a track from an electrified rolling element test with polyurea mineral oil grease
Fig. 3
Scanning electron microscopic image of the damaged area near the end of a track from an electrified rolling element test with polyurea mineral oil grease
Close modal
Then, at any moment when the film thickness is h, the asperity height that would result in the surfaces being closer than the critical surface separation would be given by
(22)
This concept is schematically shown in Fig. 4 (for clarity, it is shown as the equivalent of a combined rough surface opposing a smooth surface). Next, the concepts used in the Greenwood and Williamson model for a statistical rough surface can be applied here. Discharging can occur whenever an asperity is taller than dcrit. Therefore, the number of asperities tall enough to reduce the film thickness past the critical value is given by:
(23)
Fig. 4
Schematic depicting the critical height of asperities to be susceptible to discharging
Fig. 4
Schematic depicting the critical height of asperities to be susceptible to discharging
Close modal
Assuming a Gaussian distribution for the surface heights, then the predicted number of asperities tall enough to reduce the gap below the critical value for electrical breakdown can be solved as follows:
(24)
where erf is the error function. Then substituting in Eqs. (20) and (21) results in
(25)

Note that other height distributions could also be implemented into Eq. (23), but a closed-form equation might not be obtainable. This qualitative prediction given by Eqs. (23)(25) merely suggests how many locations and instances the film will electrically break down, causing surface damage. In reality, the mechanisms at work are more complicated, but this simplified approach is adopted to facilitate a simple prediction. The resulting damage prediction should be proportional to the observed surface damage.

3 Results

Predictions of surface damage generated by the model were compared to the observed surface damage in a reciprocating rolling ball on disk measurement to evaluate the effectiveness of the proposed model (see Ref. [9] for all details of the experiment). A schematic and photo of this setup are shown in Fig. 5. A rolling ball contact is electrified by a DC power supply at 0.5 Amps as shown. A multimeter is used to measure the voltage drop across the contact, similar to 4-wire resistance measurement. The circuits are isolated from the rest of the test rig using polymer washers and 3D-printed parts.

Fig. 5
Electrical schematic and photograph of the test setup
Fig. 5
Electrical schematic and photograph of the test setup
Close modal

The experimental testing was conducted in triplicate to confirm experimental repeatability. These experiments evaluated mineral oil-based polyurea grease (NLGI Grade 2, ISO 100) without any additives under electrified conditions. The ball is 9.53 mm (3/8 in.) diameter and loaded to 50 N. An electric current is applied across the contact during operation and will be discussed in more detail later. To mimic the reciprocating motion carried out during testing, a representative function for velocity is inputted into the model. These tests produce a small wear track that might be considered “frosting” of the surface (see Fig. 6). The surface wear from this short 1.5 h test is difficult to measure via mass loss or profilometry. Therefore, this study has used scanning electron microscopy to observe the surface pitting, as shown in Fig. 3. It should be noted that the severe pitting surface damage is observed to be concentrated near the ends of the wear groove. In the middle of the wear track, the surfaces do not have significant damage, are slightly polished, and contain debris redeposited from the pitting [9]. For the first case (case 1) considered, a sawtooth velocity function is employed, matching the velocity in the experiment (see Fig. 7).

Fig. 6
A 1 cm long electrically induced wear track
Fig. 6
A 1 cm long electrically induced wear track
Close modal
Fig. 7
The velocity provide for case 1
Fig. 7
The velocity provide for case 1
Close modal

The Southwest Research Institute tested the grease sample's dielectric breakdown strength using Japanese Industrial Standard (JIS) C2101 to be approximately 6.4 × 106 V/m, which is in the expected range of recent literature [58,60,61]. The previously mentioned method for predicting the probable number of asperities susceptible to breakdown and discharging is employed (see Eq. (25)). In these locations, the grease film is expected to degrade, enabling plasma formation to discharge across. The comparison experiment involved lubricating 52,100 bearing steel samples under electrified conditions with ISO 100 mineral oil-based polyurea NLGI 2 grease. Table 1 contains other relevant experimental conditions for modeling purposes.

Table 1

Assumed geometry and material properties for EHL model

R4.7625 mm
E200 GPa
ν0.3
Sy500 MPa
η1.55 × 1011 m−2
Σ1.98 × 10−6 m
rasp2.56 × 10−7 m
μo1.78×105Pas
γc˙0.0017s1
N0.44
α20/GPa
ρsteel14.3 10−8 Ω·m
ρoil1012 Ω·m
Sd6.4 × 106 V/m
F50 N
R4.7625 mm
E200 GPa
ν0.3
Sy500 MPa
η1.55 × 1011 m−2
Σ1.98 × 10−6 m
rasp2.56 × 10−7 m
μo1.78×105Pas
γc˙0.0017s1
N0.44
α20/GPa
ρsteel14.3 10−8 Ω·m
ρoil1012 Ω·m
Sd6.4 × 106 V/m
F50 N

Following the outlined methodology, film thickness as a function of time is predicted, as shown in Fig. 8. The film thickness is directly related to the rolling speed's increase and decrease. The squeeze film effect and rough surface contact decrease the magnitude of the film thickness fluctuation to a minimal amount for the current case. However, this relatively small change significantly affects the predicted contact resistance (Fig. 9). The contact resistance fluctuates by many orders of magnitude as a film of grease is built up during sliding and then squeezed out and compressed when the velocity slows and changes direction. At larger film thicknesses, the contact resistance is very high because there is a large film of grease between the surfaces. The plateau observed at these high film thicknesses is due to the film resistance dominating (Eq. (19)), and it changes linearly with film thickness.

Fig. 8
Numerically predicted film thickness for case 1
Fig. 8
Numerically predicted film thickness for case 1
Close modal
Fig. 9
Numerically predicted contact resistance
Fig. 9
Numerically predicted contact resistance
Close modal

It should be noted that at the times when the electrical resistance is high in the contact, it is difficult for the electrical current (electrons) to flow across the lubricant film. This allows charge to build and discharge. This build and discharge are also governed in time by the capacitance properties of the film and are equally important for predicting the performance of the bearing [1416]. Using the same general framework presented in this work, the calculation of capacitance is straightforward.

Since the average film thickness changes with time, so does the average dielectric voltage of the film. The predicted dielectric voltage is shown in Fig. 10, but clearly it never falls below the average applied voltage of 1.15 V. This voltage is obtained by measuring the voltage drop across the rolling element in the test. A sample voltage reading during testing is shown in Fig. 11. The voltage does not follow the predicted variations in resistance over this range. This is due to the limiting breakdown voltage of the lubricant film. Therefore, it is more helpful to show it over a smaller time scale. Changes in the voltage can be observed over a smaller time scale that could be the discharges. Note that 31 V is applied to the entire circuit in the test setup, but other resistances account for most of that voltage. The heating element that is in series has a resistance of 58Ω and at 0.5 Amps contributes to most of the voltage drop. Again, it is shown that the voltage across the contact is much smaller than the dielectric voltage predicted by the average film thickness, but this calculation does not yet account for the surface roughness.

Fig. 10
Numerically predicted dielectric voltage
Fig. 10
Numerically predicted dielectric voltage
Close modal
Fig. 11
Measured voltage drop across a rolling element contact
Fig. 11
Measured voltage drop across a rolling element contact
Close modal

Since the dielectric voltage predicted by the average film thickness is not low enough to allow for discharging, the model also numerically predicts the number of asperities larger than the critical height (Eq. (25)). These asperities may facilitate dielectric breakdown below the applied voltage (1.15 V). These predictions are shown in Fig. 12 for different locations along the track of the rolling ball. The model predicts electrical erosion along the outer ends of the wear track, which is observed in the form of pits during experimental tests. As demonstrated by the images, little to no pitting damage appeared in the track's mid-section, where the velocity was highest. However, raised surfaces in the track's mid-section indicate a deposition of material removed during the pit formation process. All three repetitions of the test confirmed this trend.

Fig. 12
Numerically predicted number of asperities susceptible to discharging and the damage experimentally observed in these locations
Fig. 12
Numerically predicted number of asperities susceptible to discharging and the damage experimentally observed in these locations
Close modal

Next, the numerical model was used to predict the damage for two other cases. Case 2 is for a full rotation before the rolling direction is switched and the velocity changes direction. These results are obtained by using the velocity profile given in Fig. 13. The velocity reaches a nearly constant value of 0.01 m/s in between the direction changes. The effective track length of one rotation is 0.14 m. The predicted film thickness is shown in Fig. 14, and the number of asperities above the critical height is shown in Fig. 15. The number of asperities is nearly an order of magnitude smaller than predicted for case 1. Little to no damage is observed for case 2. Therefore, the model is in qualitative agreement with experimental observations. This is also repeated for a third case (case 3), where the disk is rotated ten times at a speed of 0.01 m/s before the direction is changed. This also predicts little to no damage and agrees with experimental observations.

Fig. 13
Provided velocity profile for case 2
Fig. 13
Provided velocity profile for case 2
Close modal
Fig. 14
The predicted film thickness as a function of time for case 2
Fig. 14
The predicted film thickness as a function of time for case 2
Close modal
Fig. 15
The predicted asperities of the critical height susceptible for dielectric breakdown and the film thickness over time for case 2 (one full rotation before changing direction). Inset is an example scanning electron microscopic image of the surface at the end of one of the tracks.
Fig. 15
The predicted asperities of the critical height susceptible for dielectric breakdown and the film thickness over time for case 2 (one full rotation before changing direction). Inset is an example scanning electron microscopic image of the surface at the end of one of the tracks.
Close modal

4 Discussion

The theory outlined in the current work suggests a “sweet spot” or finite range of film thickness that is more conducive to the electrical discharging and perhaps arcing phenomena. When the film thickness is relatively large, the electric voltage cannot overcome the dielectric voltage and insulative strength of the lubricating film. In contrast, when the film thickness is minimal, the solid asperity contacts provide a direct pathway for the electric current. In this case, solid contacts facilitate the flow of charge across the bearing surfaces without discharging or arcing across the lubricant. Although beneficial for electrical performance under electrified conditions, the case of a very thin lubricating film introduces the possibility of increased mechanical wear, such as adhesion and abrasion, between the surfaces.

The proposed model still makes many assumptions that may oversimplify the real situation. For instance, it does not consider how the electrical field influences on fluid and material properties and behavior [62,63]. Many of the properties in the model, such as dielectric strength and conductivity, are also dependent on changes in pressure and temperature, which is neglected by the model. This work does not consider how the relatively high pressures in the EHL contact will change the properties of the solids and fluids (except for viscosity, which is included in the employed EHL theory). The current work does not consider how wear of the surfaces will change the roughness and change the contact resistance. There are also oxides and tribo-films that will influence the contact resistance. In addition, it does not consider how electrical damage to the grease would alter its behavior during the test, or how surface damage would also influence the film thickness. In future iterations of the model, the influence of various lubricant base oils and additives could also be considered. Additives often increase conductivity and can influence degradation mechanism, such as nanoparticles [9].

5 Conclusion

As electrical powertrain architectures continue to replace traditional combustion precursors, the effect of electrified conditions on bearing technologies remains paramount. Until now, limited modeling work has quantified the damaging effect of electrical conduction across lubricated bearing surfaces. The model presented in this work predicts a simple mixed lubrication numerical model of the electrically induced damage of a rolling element contact. The established EHL equation for spherical rolling contact is modified with an effective viscosity and flow factor term and includes a statistical model of elastic–plastic rough surface contact. A new statistical method is used to consider how roughness can locally cause the surface to be close enough for the film to dielectrically break down and allow for discharging and perhaps arcing. The model predicted surface pitting of electrically induced bearing damage at locations and during conditions similar to that observed in testing.

Acknowledgment

The authors thank for the support of the National Lubricating Grease Institute (NLGI) Research Grant. The authors also thank Carlos Sanchez and Peter Lee of the Southwest Research Institute for measuring the dielectric properties of the grease, Chad Chichester of Dupont for the guidance throughout the project, Paul Slade for his insights into the microscale arcing process, and Anton-Paar and Bruker for their support via equipment.

Conflict of Interest

There are no conflicts of interest.

Data Availability Statement

The datasets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request.

Nomenclature

a =

elastic contact radius of rolling element predicted by Hertz model

d =

distance between the mean of the surface asperities or peaks

h =

film thickness, separation of mean surface height

m =

mass of rolling element assembly

t =

time

u =

entrainment velocity

z =

location of rolling element

E =

elastic modulus

G =

EHL normalized elasticity term

L =

Moes dimensionless material parameter

M =

Moes dimensionless load parameter

N =

Carreau model constant

R =

radius of rolling element

U =

EHL normalized velocity term

W =

EHL normalized force term

P¯ =

single asperity contact force

dcrit =

asperity height susceptible to breakdown or discharging

h1 =

film thickness at time, t1

h2 =

film thickness at time, t2

hcrit =

film thickness which is susceptible to discharging

ho =

initial film thickness

rasp =

single asperity contact radius

Ad =

deformed asperity amplitude

Ahertz =

Hertz area of contact

Ai =

initial or undeformed asperity amplitude

An =

nominal or apparent contact area

Fc =

rough surface contact force on rolling element

Fehl =

lubricant hydrodynamic force predicted by EHL

FE =

externally applied force on the rolling element

FL =

lubricant hydrodynamic force from sliding and squeeze film effects

FSF =

fluid squeeze film force

Nca =

number of asperities above critical height for discharging

Rc =

electrical spreading contact resistance

Rfilm =

electrical lubricant film resistance

Rpeak =

asperity radius of curvature

Sd =

dielectric breakdown strength

Vapplied =

voltage applied across the contact in model

Vd =

dielectric breakdown voltage

Vm =

voltage measured during test across the contact

E′ =

equivalent modulus of elasticity: 1E=(1ν12)E1+(1ν22)E2

γ˙ =

shear rate

γ˙c =

critical shear rate for Carreau equation

η =

areal asperity density

λ =

wavelength of the asperities

μ =

dynamic fluid viscosity

μeff =

effective dynamic fluid viscosity

μo =

base dynamic fluid viscosity for Carreau equation

ν =

Poisson's ratio

ρoil =

electrical resistivity of oil

ρsteel =

electrical resistivity of steel

σ =

combined RMS surface roughness

σd =

deformed RMS surface roughness

σi =

initial or undeformed RMS surface roughness

σs =

combined RMS asperity height

ϕx =

flow factor for modified Reynolds equation in x direction

ϕy =

flow factor for modified Reynolds equation in y-direction

Φ =

asperity height distribution

Ψ =

alleviation factor for rough surface conduction

=

dimensionless asperity flattening parameter

References

1.
Whittle
,
M.
,
Trevelyan
,
J.
, and
Tavner
,
P. J.
,
2013
, “
Bearing Currents in Wind Turbine Generators
,”
J. Renewable Sustainable Energy
,
5
(
5
), p.
053128
.
2.
He
,
F.
,
Xie
,
G.
, and
Luo
,
J.
,
2020
, “
Electrical Bearing Failures in Electric Vehicles
,”
Friction
,
8
(
1
), pp.
4
28
.
3.
Chen
,
Y.
,
Jha
,
S.
,
Raut
,
A.
,
Zhang
,
W.
, and
Liang
,
H.
,
2020
, “
Performance Characteristics of Lubricants in Electric and Hybrid Vehicles: A Review of Current and Future Needs
,”
Front. Mech. Eng.
,
6
, p.
571464
.
4.
Lin
,
C.-M.
,
Chiou
,
Y.-C.
, and
Lee
,
R.-T.
,
2001
, “
Pitting Mechanism on Lubricated Surface of Babbitt Alloy/Bearing Steel Pair Under AC Electric Field
,”
Wear
,
249
(
1–2
), pp.
132
141
.
5.
Farfan-Cabrera
,
L. I.
,
Erdemir
,
A.
,
Cao-Romero-Gallegos
,
J. A.
,
Alam
,
I.
, and
Lee
,
S.
,
2022
, “
Electrification Effects on Dry and Lubricated Sliding Wear of Bearing Steel Interfaces
,”
Wear
,
516
, p.
204592
.
6.
Hemanth
,
G.
, and
Suresha
,
B.
,
2021
, “
Hybrid and Electric Vehicle Tribology: A Review
,”
Surf. Topogr.: Metrol. Prop.
,
9
(
4
), p.
043001
.
7.
Prasad
,
S.
, and
Krishnanunni
,
S.
,
2020
, “
Review on Analysis of Failures Modes in the Electric Vehicles Due to Electric Bearings
,”
Int. Res. J. Eng. Technol.
,
7
(
12
), pp.
1722
1725
.
8.
Raadnui
,
S.
, and
Kleesuwan
,
S.
,
2011
, “
Electrical Pitting Wear Debris Analysis of Grease-Lubricated Rolling Element Bearings
,”
Wear
,
271
(
9–10
), pp.
1707
1718
.
9.
Bond
,
S.
,
Jackson
,
R. L.
, and
Mills
,
G.
,
2024
, “
The Influence of Various Grease Compositions and Silver Nanoparticle Additives on Electrically Induced Rolling-Element Bearing Damage
,”
Friction
,
12
(
4
), pp.
796
811
.
10.
Prashad
,
H.
,
2002
, “
Diagnosis of Rolling-Element Bearings Failure by Localized Electrical Current Between Track Surfaces of Races and Rolling-Elements
,”
ASME J. Tribol.
,
124
(
3
), pp.
468
473
.
11.
Suzumura
,
J.
,
2016
, “
Prevention of Electrical Pitting on Rolling Bearings by Electrically Conductive Grease
,”
Q. Rep. RTRI
,
57
(
1
), pp.
42
47
.
12.
Jackson
,
R. L.
,
2024
, “Modeling Electrical Resistance of Lubricated Contacts,”
Electric Vehicle Tribology
,
L. I.
Farfan-Cabrera
and
A.
Erdemir
, eds.,
Elsevier
,
New York
, pp.
207
224
.
13.
Jackson
,
R. L.
, and
Angadi
,
S.
,
2023
, “
Electrical Contact During a Rolling Vibratory Motion Considering Mixed Lubrication
,”
ASME J. Tribol.
,
145
(
8
), p.
082201
.
14.
Schneider
,
V.
,
Behrendt
,
C.
,
Höltje
,
P.
,
Cornel
,
D.
,
Becker-Dombrowsky
,
F. M.
,
Puchtler
,
S.
,
Gutiérrez Guzmán
,
F.
,
Ponick
,
B.
,
Jacobs
,
G.
, and
Kirchner
,
E.
,
2022
, “
Electrical Bearing Damage, A Problem in the Nano – and Macro-Range
,”
Lubricants
,
10
(
8
), p.
194
.
15.
Busse
,
D.
,
Erdman
,
J.
,
Kerkman
,
R. J.
,
Schlegel
,
D.
, and
Skibinski
,
G.
,
1997
, “
System Electrical Parameters and Their Effects on Bearing Currents
,”
IEEE Trans. Ind. Appl.
,
33
(
2
), pp.
577
584
.
16.
Gonda
,
A.
,
Paulus
,
S.
,
Graf
,
S.
,
Koch
,
O.
,
Götz
,
S.
, and
Sauer
,
B.
,
2024
, “
Basic Experimental and Numerical Investigations to Improve the Modeling of the Electrical Capacitance of Rolling Bearings
,”
Tribol. Int.
,
193
, p.
109354
.
17.
Sawa
,
K.
,
Watanabe
,
Y.
, and
Ueno
,
T.
, “
Effect of Lubricant on Sliding Conditions in Au-Plated Slip-Ring System for Small Electric Power Transfer
,”
2014 IEEE 60th Holm Conference on Electrical Contacts (Holm)
,
New Orleans, LA
,
Oct. 12–15
.
18.
Sawa
,
K.
,
Takemasa
,
Y.
,
Watanabe
,
Y.
,
Ueno
,
T.
, and
Yamanoi
,
M.
, “
Fluctuation Components of Contact Voltage at AgPd Brush and Au-Plated Slip-Ring System With Lubricant
,”
2015 IEEE 61st Holm Conference on Electrical Contacts (Holm)
,
San Diego, CA
,
Oct. 11–14
.
19.
Jackson
,
R. L.
,
Coker
,
A. B.
,
Tucker
,
Z.
,
Hossain
,
M. S.
, and
Mills
,
G.
,
2019
, “
An Investigation of Silver Nanoparticle Laden Lubricants for Electrical Contacts
,”
IEEE Trans. Compon. Packag. Manuf. Technol.
,
9
(
2
), pp.
193
200
.
20.
Crilly
,
L.
,
Jackson
,
R. L.
,
Mills
,
G.
,
Bond
,
S.
, and
Bhargava
,
S.
,
2022
, “
An Exploration of the Friction, Wear, and Electrical Effects of Nanoparticle Enhanced and Conventional Lubricants
,”
IEEE Trans. Compon. Packag. Manuf. Technol.
,
12
(
11
), pp.
1757
1770
.
21.
Cao
,
Z.
,
Xia
,
Y.
,
Liu
,
L.
, and
Feng
,
X.
,
2019
, “
Study on the Conductive and Tribological Properties of Copper Sliding Electrical Contacts Lubricated by Ionic Liquids
,”
Tribol. Int.
,
130
, pp.
27
35
.
22.
Chu
,
L.-M.
,
Lin
,
J.-R.
,
Chang
,
Y.-P.
, and
Li
,
W.-L.
,
2013
, “
Effects of Surface Forces and Surface Roughness on Squeeze Thin Film of Elastohydrodynamic Lubricated Spherical Conjunction
,”
Lubr. Sci.
,
25
(
1
), pp.
11
28
.
23.
Chevalier
,
F.
,
Lubrecht
,
A. A.
,
Cann
,
P. M. E.
,
Colin
,
F.
, and
Dalmaz
,
G.
,
1998
, “
Film Thickness in Starved EHL Point Contacts
,”
ASME J. Tribol.
,
120
(
1
), pp.
126
133
.
24.
Sanchez Garrido
,
D.
,
Leventini
,
S.
, and
Martini
,
A.
,
2021
, “
Effect of Temperature and Surface Roughness on the Tribological Behavior of Electric Motor Greases for Hybrid Bearing Materials
,”
Lubricants
,
9
(
6
), p.
59
.
25.
Du
,
H.
,
Xie
,
W.
,
Wang
,
J.
, and
Venner
,
C. H.
,
2025
, “
Rapid Failure of Lubricated Contacts With Grease Under ZEV Condition
,”
ASME J. Tribol.
,
147
(
1
), p.
014301
.
26.
Kanazawa
,
Y.
,
Sayles
,
R. S.
, and
Kadiric
,
A.
,
2017
, “
Film Formation and Friction in Grease Lubricated Rolling-Sliding Non-Conformal Contacts
,”
Tribol. Int.
,
109
, pp.
505
518
.
27.
Zapletal
,
T.
,
Sperka
,
P.
,
Krupka
,
I.
, and
Hartl
,
M.
,
2020
, “
On the Relation Between Friction Increase and Grease Thickener Entraining on a Border of Mixed EHL Lubrication
,”
Lubricants
,
8
(
2
), p.
12
.
28.
Cyriac
,
F.
,
Lugt
,
P. M.
,
Bosman
,
R.
,
Padberg
,
C. J.
, and
Venner
,
C. H.
,
2016
, “
Effect of Thickener Particle Geometry and Concentration on the Grease EHL Film Thickness at Medium Speeds
,”
Tribol. Lett.
,
61
(
1
), pp.
1
13
.
29.
Lubrecht
,
A.
,
Venner
,
C. H.
, and
Colin
,
F.
,
2009
, “
Film Thickness Calculation in Elasto-Hydrodynamic Lubricated Line and Elliptical Contacts: The Dowson, Higginson, Hamrock Contribution
,”
Proc. Inst. Mech. Eng. Part J J. Eng. Tribol.
,
223
(
3
), pp.
511
515
.
30.
Hamrock
,
B. J.
, and
Dowson
,
D.
,
1976
, “
Isothermal Elastohydrodynamic Lubrication of Point Contacts: Part II
Ellipticity Parameter Results
,”
J. Lubr. Techcol.
,
98
(
3
), pp.
375
381
.
31.
Hamrock
,
B. J.
,
Schmid
,
B. J.
, and
Jacobson
,
B. O.
,
2004
,
Fundamentals of Fluid Film Lubrication
,
CRC Press
,
Boca Raton, FL
.
32.
Patir
,
N.
, and
Cheng
,
H. S.
,
1978
, “
An Average Flow Model for Determining Effects of Three-Dimensional Roughness on Partial Hydrodynamic Lubrication
,”
ASME J. Tribol.
,
100
(
1
), pp.
12
17
.
33.
Patir
,
N.
, and
Cheng
,
H. S.
,
1979
, “
Application of Average Flow Model to Lubrication Between Rough Sliding Surfaces
,”
ASME J. Tribol.
,
101
(
2
), pp.
220
230
.
34.
Lubrecht
,
A.
,
Graille
,
D.
,
Venner
,
C.
, and
Greenwood
,
J.
,
1998
, “
Waviness Amplitude Reduction in EHL Line Contacts Under Rolling-Sliding
,”
ASME J. Tribol.
,
120
(
4
), pp.
705
709
.
35.
Venner
,
C. H.
, and
Lubrecht
,
A. A.
,
2000
, “
Multigrid Techniques: A Fast and Efficient Method for the Numerical Simulation of Elastohydrodynamically Lubricated Point Contact Problems
,”
Proc. Inst. Mech. Eng. Part J J. Eng. Tribol.
,
214
(
1
), pp.
43
62
.
36.
Venner
,
C. H.
, and
ten Napel
,
W. E.
,
1992
, “
Surface Roughness Effects in an EHL Line Contact
,”
ASME J. Tribol.
,
114
(
3
), pp.
616
622
.
37.
Jackson
,
R. L.
,
Malucci
,
R. D.
,
Angadi
,
S.
, and
Polchow
,
J. R.
, “
A Simplified Model of Multiscale Electrical Contact Resistance and Comparison to Existing Closed Form Models
,”
2009 Proceedings of the 55th IEEE Holm Conference on Electrical Contacts
,
Vancouver, Canada
,
Sept. 14–16
,
IEEE
, pp.
28
35
.
38.
Sisko
,
A.
,
1958
, “
The Flow of Lubricating Greases
,”
Ind. Eng. Chem.
,
50
(
12
), pp.
1789
1792
.
39.
Greenwood
,
J.
, and
Kauzlarich
,
J.
,
1998
, “
Elastohydrodynamic Film Thickness for Shear-Thinning Lubricants
,”
Proc. Inst. Mech. Eng. Part J J. Eng. Tribol.
,
212
(
3
), pp.
179
191
.
40.
Madiedo
,
J. M.
,
Franco
,
J. M.
,
Valencia
,
C. n.
, and
Gallegos
,
C. s.
,
2000
, “
Modeling of the Non-Linear Rheological Behavior of a Lubricating Grease at Low-Shear Rates
,”
ASME J. Tribol.
,
122
(
3
), pp.
590
596
.
41.
Carreau
,
P. J.
,
1972
, “
Rheological Equations From Molecular Network Theories
,”
Trans. Soc. Rheol.
,
16
(
1
), pp.
99
127
. .
42.
An
,
B.
,
Wang
,
X.
,
Xu
,
Y.
, and
Jackson
,
R. L.
,
2019
, “
Deterministic Elastic-Plastic Modelling of Rough Surface Contact Including Spectral Interpolation and Comparison to Theoretical Models
,”
Tribol. Int.
,
135
, pp.
246
258
.
43.
Jackson
,
R. L.
,
Crandall
,
E. R.
, and
Bozack
,
M. J.
,
2015
, “
Rough Surface Electrical Contact Resistance Considering Scale Dependent Properties and Quantum Effects
,”
J. Appl. Phys.
,
117
(
19
), p.
195101
.
44.
Liu
,
Z.
, and
Zhang
,
L.
,
2020
, “
A Review of Failure Modes, Condition Monitoring and Fault Diagnosis Methods for Large-Scale Wind Turbine Bearings
,”
Measurement
,
149
, p.
107002
.
45.
Jackson
,
R. L.
, and
Green
,
I.
,
2006
, “
The Behavior of Thrust Washer Bearings Considering Mixed Lubrication and Asperity Contact
,”
Tribol. Trans.
,
49
(
2
), pp.
233
247
.
46.
Ruan
,
B.
,
Salant
,
R. F.
, and
Green
,
I.
,
1997
, “
A Mixed Lubrication Model of Liquid/Gas Mechanical Face Seals
,”
Tribol. Trans.
,
40
(
4
), pp.
647
657
.
47.
Varney
,
P.
, and
Green
,
I.
,
2017
, “
Impact Phenomena in a Noncontacting Mechanical Face Seal
,”
ASME J. Tribol.
,
139
(
2
), p.
022201
.
48.
Lebeck
,
A.
,
1999
, “
Mixed Lubrication in Mechanical Face Seals With Plain Faces
,”
Proc. Inst. Mech. Eng. Part J J. Eng. Tribol.
,
213
(
3
), pp.
163
175
.
49.
Polycarpou
,
A. A.
, and
Etsion
,
I.
,
1998
, “
Static Sealing Performance of gas Mechanical Seals Including Surface Roughness and Rarefaction Effects
,”
Tribol. Trans.
,
41
(
4
), pp.
531
536
.
50.
Masjedi
,
M.
, and
Khonsari
,
M.
,
2014
, “
Theoretical and Experimental Investigation of Traction Coefficient in Line-Contact EHL of Rough Surfaces
,”
Tribol. Int.
,
70
, pp.
179
189
.
51.
Davis
,
C. L.
,
Sadeghi
,
F.
, and
Krousgrill
,
C. M.
,
1999
, “
A Simplified Approach to Modeling Thermal Effects in Wet Clutch Engagement: Analytical and Experimental Comparison
,”
ASME J. Tribol.
,
122
(
1
), pp.
110
118
.
52.
Greenwood
,
J. A.
, and
Williamson
,
J. B. P.
,
1966
, “
Contact of Nominally Flat Surfaces
,”
Proc. R. Soc. Lond. A
,
295
(
1442
), pp.
300
319
.
53.
Chu
,
N. R.
,
Jackson
,
R. L.
,
Wang
,
X.
,
Gangopadhyay
,
A.
, and
Ghaednia
,
H.
,
2021
, “
Evaluating Elastic-Plastic Wavy and Spherical Asperity-Based Statistical and Multi-Scale Rough Surface Contact Models With Deterministic Results
,”
Materials
,
14
(
14
), p.
3864
.
54.
Jackson
,
R. L.
, and
Green
,
I.
,
2006
, “
A Statistical Model of Elasto-Plastic Asperity Contact Between Rough Surfaces
,”
Trib. Int.
,
39
(
9
), pp.
906
914
.
55.
McCool
,
J. I.
,
1987
, “
Relating Profile Instrument Measurements to the Functional Performance of Rough Surfaces
,”
ASME J. Tribol.
,
109
(
2
), pp.
264
270
.
56.
Front
,
I.
,
1990
, “
The Effects of Closing Force and Surface Roughness on Leakage in Radial Face Seals
,”
MS thesis
,
Technion, Israel Institute of Technology
,
Haifa, Israel
.
57.
Cooper
,
M.
,
Mikic
,
B.
, and
Yovanovich
,
M.
,
1969
, “
Thermal Contact Conductance
,”
Int. J. Heat Mass Transfer
,
12
(
3
), pp.
279
300
.
58.
Haynes
,
W. M.
,
2016
,
CRC Handbook of Chemistry and Physics
,
CRC Press
,
Boca Raton, FL
.
59.
Morris
,
S. A.
,
Leighton
,
M.
, and
Morris
,
N. J.
,
2022
, “
Electrical Field Strength in Rough Infinite Line Contact Elastohydrodynamic Conjunctions
,”
Lubricants
,
10
(
5
), p.
87
.
60.
Farfan-Cabrera
,
L. I.
,
Erdemir
,
A.
,
Cao-Romero-Gallegos
,
J. A.
, and
Aguilar-Rosas
,
O. A.
,
2024
, “Electrified Tribotesting of Lubricants and Materials Used in Electric Vehicle Drivelines,”
Electric Vehicle Tribology
,
L. I.
Farfan-Cabrera
and
A.
Erdemir
, eds.,
Elsevier
,
New York
, pp.
265
276
.
61.
Joshi
,
A.
, and
Blennow
,
J.
, “
Electrical Characterization of Bearing Lubricants
,”
Proceedings of 2014 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP)
,
Des Moines, IA
.
Oct. 19–22
, pp.
586
589
.
62.
Lee
,
P. M.
,
Sanchez
,
C.
,
Frazier
,
C.
,
Velasquez
,
A.
, and
Kostan
,
T.
,
2023
, “
Tribological Evaluation of Electric Vehicle Driveline Lubricants in an Electrified Environment
,”
Front. Mech. Eng.
,
9
, p.
1215352
.
63.
Holweger
,
W.
,
Bobbio
,
L.
,
Mo
,
Z.
,
Fliege
,
J.
,
Goerlach
,
B.
, and
Simon
,
B.
,
2023
, “
A Validated Computational Study of Lubricants Under White Etching Crack Conditions Exposed to Electrical Fields
,”
Lubricants
,
11
(
2
), p.
45
.