Exploring CFD applications in rolling bearing lubrication: objectives, challenges, and methodologies – a literature review

Abstract

The efficiency of lubrication in rolling bearings is a key topic in industry and research, particularly for computational fluid dynamics (CFD) analysis. This review critically examines 56 studies on CFD applications in rolling bearing lubrication, identified across three databases, evaluating not only how the literature can be categorised but where its modelling assumptions and validation practices constrain current use. Oil remains the dominant lubricant studied, especially in ball bearings, while grease lubrication is comparatively underrepresented; this gap is shown to stem mainly from limited CFD adoption for grease (7 of 32 grease-related studies) rather than weaker validation within the grease-CFD studies that do exist. Among mesh dynamics strategies, the multiple reference frame (MRF) and rotating moving mesh (RMM) approaches are the most frequently used for roller bearings, but this frequency coincides with commercial-software availability rather than demonstrated accuracy, since quantitative data exist almost exclusively for MRF and no study directly compares the two. The review identifies specific research gaps, including benchmark comparisons of mesh dynamics strategies, transferable grease rheology models, empirical loss correlations derived from existing CFD data, and open validation datasets needed to move CFD from a research tool toward reliable bearing design practice.

1 Introduction

In recent years, the efficiency of lubrication in rolling bearings has become increasingly important in industry and consequently in research (Lugt, 2009). Numerical simulations are becoming crucial in optimising and improving the study of lubricant flows, power losses and stress-related parameters on bearing components. Recently, there has been an increasing use of computational fluid dynamics (CFD) and numerical models for analysing fluid-dynamic behaviour inside bearings. Numerical techniques and mesh dynamics strategies vary widely between studies. Despite numerous works on this topic, a systematic overview of these aspects is lacking: how mesh dynamics are chosen and implemented; which accuracies the works claim; how common is the use of open-source software in this area. The aim of this state of the art is to propose a categorisation of the articles on the subject, with a quantitative and graphical analysis of the numerical strategies adopted in the CFD simulation of bearing lubrication. This work represents the first systematic analysis crossing mesh dynamics, accuracy, rolling element type and use of open-source software in the CFD context for bearings. The structure of this study consists of: item selection methodology, categorisation, data analysis and commentary (aided by graphs and maps), conclusions.

2 Literature search and selection process

The exploration of the relevant articles was conducted in April 2026 using the Scopus, IEEE and Google Scholar databases. The research was made systematically by means of combinations (using Boolean operators) of keywords related to the subject of the study, the fluid and the methodology. This systematic review was performed in accordance to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. The literature selection process is summarized in the PRISMA 2020 flow diagram (Figure 1).

FIGURE 1

The review was structured according to the PICOS framework as follows. The Population (P) consists of rolling bearings (ball and roller types) subjected to oil or grease lubrication under various operating conditions. The Intervention (I) refers to the application of CFD-based numerical simulation methods, encompassing different mesh dynamics strategies (e.g., Multiple Reference Frame, Rotating Moving Mesh, sliding plane, CICSAM) and multiphase flow modelling approaches (e.g., VOF, CLSVOF). The Comparator (C) includes alternative numerical approaches (e.g., FEM, analytical models), different software environments (commercial vs. open-source), and different lubricant types (oil vs. grease). The Outcomes (O) of interest are simulation accuracy, lubricant distribution, churning losses, temperature fields, and hydrodynamic forces on bearing components. The Study designs (S) considered are peer-reviewed journal articles and conference papers presenting original CFD simulations of bearing lubrication, published up to April 2026, with or without experimental validation.

2.1 Subject of the study

As keywords for the subject of the study, Rolling Bearings has been identified. However, in order to avoid neglecting interesting records, firstly the string “Rolling Bearing*” and secondly “Roll* Bearing*” have been used. The * character is used to include all the words having different suffixes.

2.2 Fluid

Considering that the goal of this project is to study grease lubrication, to limit the number of records, the strings “lubricant* flow”, “lubrica* flow”, “grease” have been used. Since the first two strings gave as a result much more papers related to oil as a lubricant fluid, then the third one has been used to better focus on the topic of this research.

2.3 Methodology

To identify contributions focusing on numerical fluid-dynamic simulation approaches, the search strings “numerical method*”“, “CFD” OR “computational fluid dynamics”, “method*”“, “CFD” OR “computational fluid dynamics”, “CFD” OR “computational fluid dynamics”, “analysis” have been selected filter out works focusing on analysis and methodological development.

2.4 Document search on scopus

To synthesize the description of the searching process followed, the abovementioned strings (groups of keywords) are referred with a Group ID, as shown in Table 1. For example, “A” refers to the subject of study, i.e., the string “Rolling Bearing*”“.

Combination of strings have been used to query the database and systematically extracting the most pertinent records. The research has been limited to the title, the keywords and the abstracts. The results of each query have been further screened based on the reading on the titles and/or the full papers.

Table 2 shows the results of the main query combinations used during the document search. As can be seen, in the first iterations, the number of articles found was limited due to excessively restrictive selection criteria. The gradual development of the combinations has made it possible to detect a wider range of contributions relevant to the research topic. The 56 selected records are listed in Table 3.

3 Categories definitions and classification criteria

Each of the 56 retrieved articles was classified against thirteen criteria. The thirteen criteria were not chosen arbitrarily, but reflect four aspects of a CFD study that determine whether its results can be compared with, or transferred to, another study. Lubricant type, EHL treatment, turbulence model, and multiphase interface model describe the physical assumptions built into the simulation, and therefore its expected fidelity. Rolling element type, oil supply mode, and computational domain extent describe the physical scope of the problem being solved, which limits how far a given result can be generalised to other bearing configurations. Mesh dynamics strategy, mesh element type, and simulation type/software describe the numerical implementation, which is where most of the practical differences between studies actually originate. Finally, the investigated losses/parameters and the analysis type (qualitative/quantitative) describe what each study actually reports, which is necessary to assess how comparable or verifiable its accuracy claims are.

Where accuracy is reported in Figures and discussed in Section 4 (>90%, <90%, or not specified), this classification refers specifically to the percentage agreement, as explicitly stated by the study’s authors, between the CFD prediction and an independent experimental measurement of a quantitative output variable (e.g., flow rate, torque, film thickness). Only studies reporting an explicit, author-stated error or agreement percentage against experimental data were classified as “>90%” or “<90%”; studies reporting only qualitative agreement (e.g., visual comparison of flow patterns) or no experimental comparison at all were classified as ““not specified”“. This criterion therefore reflects the validation rigor claimed by each study’s authors rather than an independent reassessment of accuracy by the present reviewers, and should be interpreted accordingly.

3.1 Type of lubrication

This category has been introduced to distinguish articles according to the type of lubricant used, with reference to ‘oil’ or ‘grease’. The classification was made by reading each article’s title and abstract, as well as the materials and methods section and by identifying the type of lubricant used in the experimental tests or considered in the numerical models.

3.1.1 Inclusion criteria

All articles that clearly specify the use of oil as a lubricant, possibly indicating its type or physical and chemical characteristics, or that present experimental data or numerical simulations of oil-lubricated systems were included in the ‘oil’ category. Similarly, articles have been classified as ‘grease’ when they explicitly describe the use of grease, giving details of the type, technical characteristics or application methods, or when they present data obtained from tests on grease-lubricated systems.

3.1.2 Exclusion criteria

Articles that do not specify the type of lubricant used, limiting themselves to generic mentions of ‘lubricant’ or ‘lubrication’ without further details have been excluded from the classification. Studies in which an ideal lubrication condition is assumed without considering the properties of oil or grease have also been excluded. Moreover, review articles only or those reporting literature data without including new measurements or simulations with specific lubricant types have not been included.

3.2 Type of simulation

3.2.1 Experimental

An article was categorized as ‘experimental’ if it presents a study based on physical tests taken in laboratory or in a real operation environment, with the aim of collecting experimental data on bearings’, lubricant’s and correlated systems’ behaviour.

3.2.2 Inclusion criteria

The inclusion criteria based on which an article has been categorized as ‘experimental’ are the ones in which, in the materials and methods section, an experimental setup is described, comprehensive of test benches, sensors, temperature, vibration, pressure or other parameters measurements. Moreover, the articles that explicitly report data obtained from physical tests fall under this categorization, containing images or testing equipment screens and experimental configurations that have been used. Another inclusion criteria regards the presence in the methodology of the declaration of use of specific measurement instruments, e.g., accelerometers, thermocouples, dynamometers or optical and electronic microscopes (SEM). Eventually, the presence of graphs or tables containing real data obtained during measurements has been considered a distinguishing element.

3.2.3 Exclusion criteria

The articles that focus only on running and presenting numerical simulations (e.g., through CFD, FEM or other computational techniques), without any experimental verification to support the results obtained, have been excluded from the ‘experimental’ category. Similarly, the works that employ literature data or data coming from open access databases, in the absence of physical measurements done directly by the authors, have not been considered as experimental. Lastly, the articles that are configured as a review or that report the results from experimental measurements done by other researchers, without conducting new tests, have been excluded.

3.2.4 CFD

The category ‘CFD’ collects the articles in which the lubrication, thermal dissipation phenomena or fluid film behaviour analysis inside the bearing has been conducted through numerical simulations based on computational fluid dynamic models.

3.2.5 Inclusion criteria

An article has been categorized as ‘CFD’ if in the methodology is declared the use of computational codes dedicated to the resolution of the Navier-Stokes equations (or a simplification of these). Moreover, a differentiation has been made by open-source tools, e.g., OpenFOAM, and commercial softwares like ANSYS Fluent, STAR-CCM + or COMSOL Multiphysics. Categorisation is based on the explicit statement in the text or figures, or by deducing the software used from the descriptions of the solvers and models employed. Also included in this category are studies which, although not specifying the calculation code, describe the computational mesh, boundary conditions and typical CFD simulation parameters in detail.

3.2.6 Exclusion criteria

Articles have been excluded from the CFD category if they perform structural or thermal analyses using FEM or limit themselves to analytical or semi-empirical models without implementing a dedicated CFD solver. Studies that report data from the literature or reviews of others’ CFD simulations, without conducting new simulations, have also been excluded.

3.2.7 FEM

Articles have been assigned to the FEM category if they present numerical simulations based on the finite element method for the structural, thermal or tribological analysis of components or materials in the context of lubricated bearings.

3.2.8 Inclusion criteria

Also in this case, the adoption of open-source codes has been distinguished from the use of commercial software such as ANSYS Mechanical or COMSOL. The classification has been carried out by checking the presence of finite element meshes, descriptions of material properties, boundary conditions and applied loads. The lack of supporting experimental data has not been a criterion for exclusion from this category, as long as the FEM approach has been described in sufficient detail to allow numerical reproduction of the study.

3.2.9 Exclusion criteria

Studies that exclusively use CFD models, analytical approaches, or FEM simulations present in reviews by other authors, without developing new simulations, have been excluded from the FEM category. Articles that exclusively report experimental data without any finite element analysis were also excluded from this category.

3.2.10 Analytical

The ‘analytical’ category includes articles in which physical phenomena have been described and analysed exclusively by using analytical or semi-analytical mathematical models that can be solved with standard symbolic or numerical tools (such as MATLAB, Python or Excel).

3.2.11 Inclusion criteria

This classification includes studies that involve balance equations (of mass, energy or momentum), tribological models or thermal formulations, developed on the basis of simplifying assumptions and solved using deterministic approaches. Also in this case, a distinction has been made between implementation using open-source tools and commercial software, if this has been specified. An article has been assigned to this category if it lacks numerical finite element or CFD simulations and if the analyses have been based exclusively on analytical relations or symbolic calculations, eventually compared with experimental or literature data.

3.2.12 Exclusion criteria

Articles that implement numerical finite element or CFD simulations as well as those that limit themselves to present experimental data without any analytical modelling have been excluded from the ‘analytical’ category. Purely compilatory studies or reviews without the development of new models were also excluded.

3.3 Rolling element

This category has been introduced to distinguish articles according to the type of rolling element used in the systems studied, mainly considering two classes: ball bearing and roller bearing. The classification has been based on the analysis of the title, abstract and Materials and Methods section of each article and, when present, on the technical specifications of the experimental configurations or numerical models analysed.

3.3.1 Inclusion criteria

Articles have been included in the ‘ball’ category when they explicitly mention the use of ball bearings within the experimental tests or simulations. The reference may be by means of textual description, technical specification of a test rig, construction diagrams or indication in the results and discussion. Similarly, all articles declaring the use of roller bearings, whether cylindrical, tapered or needle roller, have been classified as ‘roller’. In both cases, a direct reference in the methodology or description of the object of study was considered sufficient.

3.3.2 Exclusion criteria

Articles in which the type of rolling element used is not clearly stated, being limited to generic definitions such as ‘bearing’ without further specification, have been excluded from categorisation. Furthermore, works that do not analyse the behaviour of grease or oil inside bearings, but carry out tests on different testing devices, such as the grease worker, tribometers, or other equipment dedicated to the study of the rheological or tribological properties of lubricants, without involving an actual bearing in the test, have also been excluded.

3.4 Type of oil lubrication

Within the articles examined, the type of oil-based lubricant supply has been distinguished. In particular, the following categories have been considered: starvation, limited lubricant supply, oil splash, oil spray and oil jet. For simplicity and consistency with the literature, starvation and limited lubricant supply conditions have been merged into a single category that identifies situations of insufficient lubricant supply, where the film does not remain constant along the contact due to fluid shortage, whether intentional or due to critical operating conditions. Correspondingly, oil spray and oil jet methods have been grouped into a category that includes forced lubrication techniques through jets or direct atomisation of the lubricant over the contact zone. The oil splash type, on the other hand, has been considered as an autonomous category since it represents a typical condition of systems lubricated by partial immersion or mechanical agitation, without a controlled external supply, and with a lubrication dynamic strongly influenced by the rotation speed of the elements and the geometry of the system. Included in the classification are all articles which clearly specify the lubrication mode used or which, through the system description or pictures, allow the type of lubricant supply to be deduced. Articles that do not state the lubrication mode or that deal with non-lubricated systems or those based on other principles (e.g., compressed air or dry systems) have not been classified.

3.5 Computational domain

In classifying the articles, consideration has been given to the size and extent of the computational domain used for numerical simulations. A distinction has been made between section-type configurations, in which the numerical model reproduces a limited portion of the bearing or system, and full configurations, in which the entire component or rotating system is modelled. The distinction has been made by reading the methodological section of the articles, looking at the images of the computational models or the descriptions of the three-dimensional domains.

Articles in which the computational domain is not clearly described or which carry out theoretical analyses without a true three-dimensional numerical model have been excluded from the classification.

3.6 Model

With regard to the numerical model used to trace the interface between fluid phases, a distinction has been made between approaches based on the Volume of Fluid (VOF) method and those using the Coupled Level Set Volume of Fluid (CLSVOF) variant. The classification has been made by examining the description of the models in detail, usually in the section on the definition of the equations and the handling of the interface tracing between oil, grease and air. Articles that do not specify the method used for interface tracking or that are limited to single-phase models have not been included in the classification.

3.7 Elastohydrodynamic lubrication (EHL)

An important criterion concerns the possible treatment of the elastohydrodynamic effect (EHL) in the lubricating film. All articles that integrate the elastic deformation of rolling elements or contact tracks within the lubrication model, be it numerical, experimental or analytical, have been classified in this category. This has been verified by explicitly searching for the term elastohydrodynamic or associated models in the methodological section or in the equations used. Works dealing with lubrication in the simple hydrodynamic regime without considering the elastic effects of the materials in contact have been excluded from the category.

3.8 Type of turbulence

The articles have been divided according to the turbulence model adopted in the numerical simulations. The categories considered are laminar, k-epsilon and k-omega, based on the description given in the sections on the computational setup. Classification has been carried out in a direct manner through verification of the stated model. Articles not specifying the type of turbulent model or dealing with exclusively stationary or hyperviscous flows without turbulent modelling were not classified.

3.9 Investigation target

3.9.1 Losses

For each study, the types of energy or dissipative losses considered in the analysis have been identified. The items covered in the classification are churning, drag, lubrication, oil losses, friction losses, mechanical losses, power losses, bleed losses, and pressure losses. Each item has been classified according to one or more of these types, depending on the parameters studied or the phenomena investigated, as described in the results section or in the objectives’ definition. Articles that do not provide a clear distinction of the types of losses studied or that only analyse the system in terms of global parameters without detailing the individual dissipative components have been excluded from the classification.

3.9.2 Fluxes, force-related parameters and lubricant properties

For each contribution, the specific object of the technical investigation has been identified. The categories considered include flow, torque, film thickness, deformation, friction, pressure, vibrations, angles, stress, forces, wear, and grease properties. Classification has been done by reading the results section and abstracts, checking which physical or quantitative parameters are actually analysed and discussed. Articles reporting qualitative analyses without numerical data or limiting themselves to general considerations on bearing operation without measuring or simulating specific parameters have not been classified.

3.10 Analysis type

This category has been defined to distinguish studies according to the type of approach used to analyse the data obtained, separating them into qualitative and quantitative studies. To be classified as quantitative, an article must present measurable, calculated or simulated numerical data, accompanied by graphs, tables, or statistical analysis. Studies that employ numerical or experimental models to obtain physical quantities expressed in numerical values are also included in this category. On the other hand, a work is classified as qualitative when it merely provides descriptions, visual observations, diagrams or interpretative images of phenomena, without presenting numerical values or quantitative relationships between variables. Articles that do not report any form of analysis of the results or that limit themselves to a theoretical discussion without data processing have been excluded from both classifications.

3.11 Mesh dynamics

The Mesh Dynamics category collects all the strategies and models used to describe and manage the relative motion between components in the computational domain of CFD simulations. The Multiple Reference Frame (MRF) category has been assigned to articles that apply rotating reference systems stationary with respect to the fixed domain. Rotating Moving Mesh (RMM) identifies cases in which the mesh itself rotates or deforms during the simulation, accurately reproducing the relative motion. The Compressive Interface Capturing Scheme for Arbitrary Meshes (CICSAM) category is reserved for studies that specify the use of this algorithm for capturing the sharp oil-air interface. The Physics Controlled group includes articles in which the motion and mesh management is delegated to the predefined settings of the software, without explicit user control. Sliding Plane includes cases of moving surfaces coupled by sliding interfaces. Dynamic indicates simulations with a dynamic mesh updated at each time step and Relative Rotation Simulated indicates works in which “the relative mesh rotation was simulated by activating the Coriolis and centrifugal force terms in the transport equations of the mathematical model at cells within each specified mesh block. It is then possible to have multiple adjoining rotating blocks, each with its own rotational speed but without actually causing the mesh to rotate.” Articles that do not deal with rotating domains or that simulate stationary systems without relative motion were excluded from the classification.

3.12 Mesh type

This category classifies the type of spatial discretization used in numerical models. Tetrahedral refers to three-dimensional meshes made of tetrahedral elements, generally used for complex geometries. Hexahedral refers to meshes structured with regular cubic or prismatic elements, preferred for their numerical efficiency. Articles using Polyhedral elements use cells with more than four faces, common in open-source codes. Unstructured refers to meshes without ordered patterns, automatically generated for irregular geometries. Meshes specific for ANSYS software, such as Plane77 and Surf151, used for two-dimensional models or contact surfaces, were considered separately. Prismatic refers to meshes with prismatic layers juxtaposed to the wall to correctly resolve the viscous gradient. Finally, Triangular 2D was assigned to articles with two-dimensional meshes composed of triangles. All studies that do not describe the type of mesh used or that are limited to ideal simulations without geometric discretization were excluded.

3.13 Temperature

This category has been introduced to identify articles that directly investigate the effect of temperature on the behaviour of bearings or lubricants. An article has been included in this category if it involves experimental measurements or numerical simulations of temperature within the lubricant system or on mechanical components, or if it explicitly studies how temperature variation affects quantities such as friction, lubricant film thickness, wear or energy losses. Articles reporting temperature only as a boundary condition, without analysing its effects or measuring its trend, or mentioning it marginally in the text without providing values, distributions or trends, have been excluded.

3.14 Other testing machines

This category includes experimental studies that use equipment different from bearing test rigs to test lubricants or tribological components. This includes, for example, articles that involve the use of a grease worker, tribometers, devices for measuring the consistency or stability of lubricant greases, or any laboratory machines other than actual bearing configurations or standardised tribological test rigs. Articles performing tests on lubricants in mixing devices, shear testing, stress tests under controlled pressure or devices for detecting the electrical resistance of greases were included. On the other hand, all studies that do not make use of physical testing machines or that only use numerical simulations and literature reviews, as well as those dealing with tests on bearing test rigs have been excluded.

3.15 Additives

This category identifies articles that study the effect of chemical additives in the lubricant. A paper has been included if it explicitly describes the formulation of the lubricant with specific additives, such as anti-wear, antioxidants, friction modifiers, high-pressure agents or anti-foaming agents, and if it reports experimental tests or numerical simulations aimed at evaluating their effectiveness. Works comparing several lubricating oil or grease formulations differentiated by the presence or absence of additives have also been considered. Articles using commercial lubricants without indicating or analysing the presence of specific additives or focusing on other variables without considering the role of additives in the tribology of the system have been excluded.

Tables 4, 5 report the synthesis of the abovementioned classifications. For the categories ‘temperature’, ‘other testing machines’ and ‘additives’, the last three columns of Table 2 assign to each record a value of 0 if that category has not been treated and one if it has been included and discussed. Tables 4 and 5 report the classification of the 56 included studies (Zhu D. et al., 2022; Peterson et al., 2021; Heng et al., 2021; Adeniyi et al., 2015; Ariffin et al., 2022; Lugt, 2009; Siegle et al., 2025; Xue et al., 2022; Wu et al., 2018; Gurt and Khonsari, 2025; Jiao et al., 2023; Cann and Lubrecht, 2005; Zhang et al., 2015; Zhu W. et al., 2022; Lu et al., 2020; Wang et al., 2017; Aamer et al., 2023; Li et al., 2018; Maccioni et al., 2023a; Lin et al., 2023; Shetty et al., 2024; Maccioni et al., 2023b; Cann, 2007; Osara et al., 2024; Nassef et al., 2022; Maccioni et al., 2025; Wu C. et al., 2023; Chen et al., 2023; Maccioni et al., 2023c; Raju et al., 2013; Wu et al., 2020; Sedlář and Abrahámek, 2023; Fischer et al., 2021; Wang et al., 2024; Hu et al., 2014; Wang et al., 2021; Gao et al., 2025; Li et al., 2025a; Wen and Oshima, 2014; Lugt et al., 2009; Pozzebon et al., 2020; Li et al., 2025b; Wang, 2023; Maccioni et al., 2022; Wei et al., 2024; Wu H. et al., 2023; Wang et al., 2022; Bao et al., 2020; Meijer and Lugt, 2022; Gao et al., 2024; Liang and Ji, 2022; Noda et al., 2020; Wu et al., 2025; Tatzgern et al., 2025; de la Presilla et al., 2025; Benetti et al., 2025).

4 Data analysis and commentary

Figure 2 shows the distribution of the number of records over the years, subdivided according to the type of lubricant used. The trend shows a clear increase of the number of records in the years thanks to the developments in computer science and to the growing interest in improving energy efficiency. It can be noted that, especially in recent years, interest in grease has been increasing considerably. This growing interest is also reflected in recent studies addressing advanced grease formulations and their performance in electric vehicle bearings (Abozeid et al., 2025).

FIGURE 2

Figure 3 aims to show, for each Mesh Dynamics approach, the distribution of articles as a function of accuracy (>90%, <90%, not specified) and the type of rolling element (ball or roller). The size of the bubbles, as well as the number within them, represents the number of records in each combination. For the accuracy, the value −1 means that it has not been made explicit; one means that the accuracy of the results obtained was below 90%, while 2 means that the accuracy is over 90%. Figure 3 is intended to show how each approach performs. The MRF approach seems to show a good level of applicability for roller bearings since it has been applied mostly in case studies that involved these latter. On the other hand, the Relative rotation simulated, the sliding plane and the CICSAM approaches have been used only in studies involving ball bearings, with variable outcomes in terms of accuracy. Finally, it is worth noting that the RMM approach, despite being used in a quite large number of works, does not show any accuracy outcome.

FIGURE 3

With reference to Figure 4, it is interesting to note that although most studies use commercial software, the RMM approach is only implemented with open-source software. MRF is the most widely used approach, moreover only with commercial software, followed by RMM and sliding plane. The choice between VOF and CLSVOF as the multiphase interface-tracking method is seldom discussed explicitly in the reviewed studies. VOF is by far the most commonly adopted approach, reflecting both its widespread implementation in commercial and open-source CFD solvers and its relatively low computational cost compared with CLSVOF. By contrast, CLSVOF, which combines the volume-of-fluid formulation with a level-set function to improve interface sharpness and curvature accuracy, appears only in a limited number of studies, generally where detailed interface dynamics (e.g., thin-film breakup or bubble entrainment) are central to the research objectives. However, no study in the reviewed literature performs a direct quantitative comparison between VOF and CLSVOF for the same bearing configuration and operating conditions. Consequently, the accuracy gained by adopting CLSVOF, and the operating regimes in which its additional computational cost is justified, remain unclear. This reflects the same evidence gap identified for MRF and RMM: numerical methods are generally selected on the basis of established practice and software capabilities rather than through systematic comparisons demonstrating a case-specific accuracy advantage.

FIGURE 4

Figures 5, 6 show the distribution of the records worldwide. The most active country seems to be China, but also Italy and the Netherlands are productive. An analysis made on institutions shows that most of the studies done by China come from 3 principal cities: Xi’an (4 articles), Beijing and Nanjing (5 articles each). The 6 records from Italy are all produced by the Free University of Bolzano/Bozen, while all the 4 records from the Netherlands have as principal author Prof. P. Lugt.

FIGURE 5

FIGURE 6

Examining Tables 4, 5 together, rather than considering each criterion in isolation, reveals a clearer pattern than lubricant type alone suggests. All 24 oil-lubricated records in this review employ CFD, whereas only 7 of the 32 grease-related records (22%) do so; the remaining grease studies rely on experimental characterisation, FEM, or analytical modelling without a fluid-dynamic simulation component. Among the studies that do combine grease with CFD, quantitative accuracy is reported in 4 of 7 cases (57%), compared with 6 of 24 oil-lubricated CFD studies (25%). The available evidence therefore suggests that the principal difference between oil and grease in the reviewed literature lies in the adoption of CFD itself, rather than in the validation practices of the grease-CFD studies that have been reported. A further asymmetry concerns software usage: all six validated oil-CFD studies employ commercial software, typically together with an explicitly stated mesh dynamics strategy (MRF, RMM, or sliding plane), whereas the four validated grease-CFD studies are evenly divided between commercial and open-source tools and do not report an explicit mesh dynamics strategy. This indicates that mesh dynamics have received considerably less explicit attention in grease-CFD studies than in the oil-lubricated literature, rather than simply reflecting differences in validation practice.

5 Discussion

A systematic literature review has been performed on the Scopus database. The search has been carried out taking into consideration the subject of study, fluid properties and methodology. A total of 56 articles have been analysed and found to be relevant. Of the different mesh dynamics approaches, MRF and RMM have been found to be the most suitable as well as the most used for roller bearings. Concerning the RMM approach, no accuracy data have been found, whereas MRF provides results that are more than 90% accurate. Most of the studies were performed with commercial software. The most prolific country is China, followed by Italy. The most active cities are Beijing, Nanjing and Bolzano.

6 Open challenges and future research directions

The analysis of the reviewed literature reveals several specific unresolved challenges that future research should address. First, grease rheology remains the most critical modelling bottleneck: only a handful of studies (e.g., (de la Presilla et al., 2025)) have attempted to implement non-Newtonian constitutive models such as Herschel-Bulkley within CFD solvers, and none of the reviewed works provides a validated, transferable grease model applicable across different bearing types or operating conditions. The viscoelastic behaviour of grease, its thixotropy, and its degradation over time are largely neglected in current CFD frameworks. Physically, grease behaves as a two-phase colloidal system in which a fibrous or lamellar thickener matrix retains the base oil at rest. Below a critical shear stress (the yield stress), this matrix resists deformation and the material behaves essentially as a soft viscoelastic solid; once the applied stress exceeds this threshold, the matrix breaks down locally and the material flows as a shear-thinning fluid, with base oil migrating from the thickener structure—a process commonly termed bleed. This stress-induced solid-to-liquid transition is one of the key characteristics distinguishing grease from lubricating oil and motivates the use of yield-stress constitutive models, most commonly the Bingham plastic and, more accurately, the Herschel-Bulkley model, which additionally captures the shear-thinning behaviour observed above the yield point (e.g., (Benetti et al., 2025)). However, even Herschel-Bulkley-based formulations do not capture the thixotropic component of grease behaviour—the time-dependent recovery of the thickener microstructure after shear ceases—which governs how quickly a starved or sheared contact re-establishes its lubricating film. A more physically representative constitutive framework would therefore combine a yield-stress formulation with an explicit structural parameter describing thickener breakdown and recovery over time. To the best of our knowledge, such an approach was not identified in the rolling-bearing CFD literature reviewed in this work and therefore represents a promising direction for future model development. Second, starvation and lubricant replenishment are insufficiently covered: only a minority of studies model partial-film or starved conditions, despite their industrial relevance for energy efficiency and bearing life. Third, the RMM approach—despite being the second most used mesh dynamics strategy—shows no reported accuracy data in any of the reviewed works, leaving its quantitative reliability unassessed. Future studies should systematically compare RMM and MRF on the same benchmark geometry to establish clear applicability criteria. Fourth, the evidence base for grease-lubricated CFD remains small in absolute terms: only 7 of the 32 reviewed grease-related studies employ CFD at all, limiting how much can be concluded about validation practice specifically. Expanding this small sample, rather than improving validation rigor within it, is the more pressing need. The development of open, standardised validation benchmarks—combining PIV measurements and high-speed imaging—would help attract more CFD-based grease research toward a shared, comparable evidence base.Finally, the growing adoption of open-source tools such as OpenFOAM offers an opportunity to develop community-maintained, grease-specific solver extensions, which could accelerate progress in this area while reducing dependency on commercial licensing.

A further gap concerns the link between CFD research and the empirical formulas actually used in bearing design. Power losses in rolling bearings are usually split into a load-dependent term, due to elastohydrodynamic friction in the loaded contact, and a load-independent term, due to churning and windage, which depends on speed, viscosity, immersion level, and geometry rather than on load. Palmgren’s original correlations, and the catalogue-based calculation methods still used across the bearing industry, remain the standard tool for estimating the load-independent term, since they require no simulation. Most of the CFD studies reviewed here predict churning losses accurately for one specific case but do not turn their results into a general formula that could be reused for a different geometry without repeating the simulation. The CFD data already produced across the reviewed literature, particularly for oil, could be used to derive updated loss correlations of this kind. For grease, where no such correlation currently exists, this is an open and practically relevant research direction, complementary to the higher-fidelity modelling challenges discussed above.

Uncertainty quantification and digital-twin frameworks are entirely absent from the reviewed literature, and machine learning appears in only one of the 56 studies ((Li et al., 2025b), a GA-BP neural network used to predict scuffing failure in grease-lubricated ball bearings), applied independently of any CFD simulation rather than as a surrogate or reduced-order model built on CFD data. These approaches are increasingly used in adjacent areas of fluid dynamics and are directly relevant to the challenges identified above. Uncertainty quantification would be particularly valuable for grease-lubricated simulations, where rheological parameters are often estimated rather than directly measured, and propagating this uncertainty through the CFD model would give a more honest picture of prediction confidence than a single deterministic result. Machine-learning surrogate models, trained on the CFD datasets already produced across the reviewed literature, could substantially reduce the computational cost of the parametric studies needed to build the generalisable loss correlations discussed above, and could in principle support real-time or near-real-time digital-twin applications for bearing condition monitoring, an area with clear industrial relevance that remains entirely unexplored in the CFD-specific literature reviewed here.

7 Conclusion

This review has examined 56 studies on CFD applications in rolling bearing lubrication, with the aim of identifying not only general trends but also specific, actionable research gaps. Oil-lubricated systems, particularly ball bearings, remain the most extensively investigated configuration, supported by established mesh dynamics strategies. In contrast, grease lubrication remains comparatively underdeveloped, with CFD being adopted far less frequently for grease than for oil (7 of 32 grease-related studies, compared with all 24 oil-lubricated studies). Among the grease-CFD studies that do exist, quantitative validation is reported with a frequency comparable to that observed in the oil-lubricated CFD literature; the principal limitation therefore appears to be the small number of available CFD studies rather than the validation practices adopted within them. Grease’s non-Newtonian and thixotropic rheology is represented in only a handful of CFD studies overall, and evidence of a CFD model that has been systematically validated and demonstrated to be transferable across different bearing types remains lacking.

Beyond this overall finding, the review identifies five key research gaps that currently limit the development of robust and reusable CFD methodologies for rolling bearing lubrication. First, the comparative performance of the leading mesh dynamics strategies (MRF and RMM) for roller bearings remains unclear, since their reported use is strongly influenced by software availability and no systematic benchmark under equivalent operating conditions has been identified in the reviewed literature. Second, despite the growing availability of CFD datasets, relatively few studies derive simplified empirical correlations for load-independent power losses, particularly for grease-lubricated bearings, limiting the transfer of CFD results into engineering design practice. Third, the non-Newtonian, thixotropic rheology of grease remains only partially represented in current CFD models, with no validated formulation shown to transfer across bearing types or operating conditions. Fourth, starvation and lubricant replenishment dynamics—directly relevant to energy efficiency and bearing life—are addressed by only a minority of the reviewed studies, particularly for grease lubrication. Finally, the evidence base for grease-lubricated CFD remains limited in absolute terms, restricting the extent to which validation practices, modelling strategies, and model transferability can currently be assessed. Expanding this evidence base, together with the development of shared benchmark datasets, would enable more objective comparisons between different modelling approaches and provide a stronger foundation for future model validation.

Among these challenges, the limited number of CFD studies addressing grease lubrication represents one of the most significant barriers to broader industrial adoption. Grease-lubricated bearings are increasingly employed in maintenance-free and energy-efficient applications across sectors such as automotive, wind energy, and industrial machinery, yet the available CFD evidence remains too limited to establish modelling approaches that are both broadly validated and demonstrably transferable across different bearing configurations. Addressing this limitation will require not only the high-fidelity multiphysics developments discussed in Section 6, but also a broader body of systematically validated CFD studies, reduced-order engineering models, and shared validation frameworks capable of translating advanced simulations into practical bearing design tools.

References

• 1

  AamerS.SadeghiF.MeinelA. (2023). Cylindrical roller bearing cage pocket lubrication. Tribol. Int.188, 108851. 10.1016/j.triboint.2023.108851

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  AbozeidA. A.YoussefM. M.MegahedT. F.El-HelalyM.PapeF.NassefM. G. A. (2025). Tribo-electric performance of nano-enhanced palm oil-based glycerol grease for electric vehicle bearings. Lubricants13, 354. 10.3390/lubricants13080354

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  AdeniyiA. A.MorvanH. P.SimmonsK. A. (2015). “A multiphase computational study of oil-air flow within the bearing sector of aeroengines,” in Volume 5C: Heat Transfer (Montreal, Quebec, Canada: American Society of Mechanical Engineers), V05CT15A024. 10.1115/GT2015-43496

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  AriffinN. A. A. M.AngH. H.HamdanS. H.AhmadH. A. I. A.ChongW. W. F. (2022). A numerical algorithm for grease-lubricated point contact at different regimes of lubrication. Jurnal Tribologi35, 50–67.

  • Google Scholar
  • View reference in article

• 5

  BaoH.WangC.LuF. (2020). Temperature impact analysis of star gear bearing inner ring based on under-race lubrication passage. J. Mech. Sci. Technol.34 (12), 5271–5278. 10.1007/s12206-020-1128-x

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  BenettiM.MaccioniL.JuckelandtK.PlogmannM.ConcliF. (2025). “Rheological calibration and CFD simulation of grease behaviour using the herschel-bulkley model,” in Proceedings of the 11th World Congress on Mechanical, Chemical, and Material Engineering (MCM'25) (Paris, France). 10.11159/icmie25.167

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  CannP. M. (2007). Grease lubrication of rolling element bearings — role of the grease thickener. Lubr. Sci.19 (3), 183–196. 10.1002/ls.39

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  CannP. M.LubrechtA. A. (2005). “Bearing performance limits with grease lubrication,” in World Tribology Congress III (Washington, D.C., USA: ASMEDC), 2, 35–36. 10.1115/WTC2005-63205

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  ChenH.LiangH.WangW.ZhangS. (2023). Investigation on the oil transfer behaviors and the air-oil interfacial flow patterns in a ball bearing under different capillary conditions. Friction11 (2), 228–245. 10.1007/s40544-021-0592-3

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  de la PresillaR.Calderon-SalmeronG.LecknerJ.KitamuraR.SatoK.SasakiS.et al (2025). Lubricant design for oscillating rolling bearings: greases, ionic liquids, and friction torque. Tribol. Int.210, 110721. 10.1016/j.triboint.2025.110721

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  FischerD.Von GoeldelS.JacobsG.StratmannA. (2021). Numerical investigation of effects on replenishment in rolling point contacts using CFD simulations. Tribol. Int.157, 106858. 10.1016/j.triboint.2021.106858

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  GaoS.HouX.MaC.YangY.LiZ.YinR.et al (2024). Transient simulation analysis of needle roller bearing in oil jet lubrication and planetary gearbox lubrication conditions based on computational fluid dynamics. Lubricants12 (2), 39. 10.3390/lubricants12020039

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  GaoW.LiY.LiC.XuY.LiuZ. (2025). Numerical prediction of drag force on spherical elements inside high-speed ball bearing with under-race lubrication. Mech. Syst. Signal Process.224, 112024. 10.1016/j.ymssp.2024.112024

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  GurtA.KhonsariM. (2025). Analysis of grease mechanical degradation in standard equipment. Tribol. Int.202, 110308. 10.1016/j.triboint.2024.110308

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  HengY. J.YusofN. F. M.YenL. A.HamidS. A.YusofN. N. M. (2021). A contact characteristic of roller bearing with palm oil-based grease lubrication. ARFMTS89 (2), 139–149. 10.37934/arfmts.89.2.139149

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  HuJ.WuW.WuM.YuanS. (2014). Numerical investigation of the air–oil two-phase flow inside an oil-jet lubricated ball bearing. Int. J. Heat Mass Transf.68, 85–93. 10.1016/j.ijheatmasstransfer.2013.09.013

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  JiaoB.BianQ.WangX.ZhaoC.ChenM.ZhangX. (2023). Analysis of high-speed Angular Ball Bearing Lubrication Based on bi-directional fluid-solid Coupling, Archive of Mechanical Engineering. 531–551. 10.24425/ame.2023.148698

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  LiY.YangZ.ChenF.ZhaoJ. (2018). Effect of air inlet flow rate on flow uniformity under oil-air lubrication, ILT70 (2), 282–289. 10.1108/ILT-12-2016-0296

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  LiJ.AnL.HuangY.LiZ.AnB.GuanB.et al (2025a). Numerical thermal analysis of greased rolling bearing considering surface topography and plastic deformation. Lubr. Sci.37 (2), 129–141. 10.1002/ls.1730

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  LiJ.LiZ.AnB.LiR. (2025b). Prediction of scuffing failure in grease-lubricated ball bearing based on GA-BP neural network. Eng. Fail. Anal.170, 109331. 10.1016/j.engfailanal.2025.109331

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  LiangX.JiH. (2022). Tribological properties of lubricating grease additives made of silica and silicon carbide nanomaterials, integrated ferroelectrics. 225 (1), 212–224. 10.1080/10584587.2021.1911243

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  LinC.-L.PozzebonM.SokolowskiK. A.MeehanP. A. (2023). Experimental investigation on rolling contact wear in grease lubricated spherical roller bearings using microcomputed tomography (μCT). Wear534–535, 205121. 10.1016/j.wear.2023.205121

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  LuF.WangM.PanW.BaoH.GeW. (2020). CFD-based investigation of lubrication and temperature characteristics of an intermediate gearbox with splash lubrication. Appl. Sci.11 (1), 352–dic. 10.3390/app11010352

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  LugtP. M. (2009). A review on grease lubrication in rolling bearings. Tribol. Trans.52 (4), 470–480. 10.1080/10402000802687940

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  LugtP. M.VelickovS.TrippJ. H. (2009). On the chaotic behavior of grease lubrication in rolling bearings. Tribol. Trans.52 (5), 581–590. 10.1080/10402000902825713

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  MaccioniL.ChernorayV. G.MastroneM. N.BohnertC.ConcliF. (2022). Study of the impact of aeration on the lubricant behavior in a tapered roller bearing: innovative numerical modelling and validation via particle image velocimetry. Tribol. Int.165, 107301. 10.1016/j.triboint.2021.107301

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  MaccioniL.ConcliF. (2023a). “Estimation of hydraulic power losses in a double-row tapered roller bearing via computational fluid dynamics,” in Towards a Smart, Resilient and Sustainable Industry. Editors BorgianniY.MattD. T.MolinaroM.Orzese G.diA. c. (Cham: Springer Nature Switzerland), 745, 655–666. 10.1007/978-3-031-38274-1_55

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  MaccioniL.ChernorayV. G.ConcliF. (2023b). Fluxes in a full-flooded lubricated tapered roller bearing: Particle image velocimetry measurements and computational fluid dynamics simulations. Tribol. Int.188, 108824. 10.1016/j.triboint.2023.108824

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  MaccioniL.RüthL.KochO.ConcliF. (2023c). Load-independent power losses of fully flooded lubricated tapered roller bearings: numerical and experimental investigation of the effect of operating temperature and housing wall distances. Tribol. Trans.66 (6), 1078–1094. 10.1080/10402004.2023.2254957

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  MaccioniL.ChernorayV. G.ConcliF. (2025). Investigating lubricant behavior in a partially flooded tapered roller bearing: validation of a multiphase CFD solver for aerated oil sump via particle image velocimetry studies and high-speed camera acquisitions. Tribol. Int.201, 110274. 10.1016/j.triboint.2024.110274

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  MeijerR. J.LugtP. M. (2022). The grease worker and its applicability to study mechanical aging of lubricating greases for rolling bearings. Tribol. Trans.65 (1), 32–45. 10.1080/10402004.2021.1979151

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  NassefM. G. A.SolimanM.NassefB. G.DahaM. A.NassefG. A. (2022). Impact of graphene nano-additives to lithium grease on the dynamic and tribological behavior of rolling bearings. Lubricants10 (2), 29. 10.3390/lubricants10020029

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  NodaT.ShibasakiK.MiyataS.TaniguchiM. (2020). X-Ray CT imaging of grease behavior in ball bearing and numerical validation of multi-phase flows simulation. Tribol. Online15 (1), 36–44. 10.2474/trol.15.36

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  OsaraJ. A.ChatraS.LugtP. M. (2024). Grease material properties from first principles thermodynamics. Lubr. Sci.36 (1), 36–50. 10.1002/ls.1672

  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  PetersonW.RussellT.SadeghiF.BerhanM. T.StackeL.-E.StåhlJ. (2021). A CFD investigation of lubricant flow in deep groove ball bearings. Tribol. Int.154, 106735. 10.1016/j.triboint.2020.106735

  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  PozzebonM. L.LinC.-L.MeehanP. A. (2020). On the modeling of wear in grease-lubricated spherical roller bearings. Tribol. Trans.63 (5), 806–819. 10.1080/10402004.2020.1743400

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  RajuK. S.VeettilM. P.RayS.ShiF. (2013). “Needle roller bearing lubricant flow CFD simulations,” in Presentato Al Symposium on International Automotive Technology 2013. 10.4271/2013-26-0041

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  SedlářM.AbrahámekP. (2023). Numerical analysis of oil lubrication and cooling of roller thrust bearing in high-performance mixed-flow pump. Energies16 (19), 6890. 10.3390/en16196890

  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  ShettyP.MeijerR. J.OsaraJ. A.PasaribuR.LugtP. M. (2024). Film thickness in grease-lubricated deep groove ball bearings—A master curve. Tribol. Trans.67 (5), 1057–1068. 10.1080/10402004.2024.2398239

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  SiegleL.HilligerA.HerbstS.EngelmannF. (2025). Analysing the impact of solid contaminants on grease viscosity and the theoretical temperature performance of rolling bearings. J. Loss Prev. Process Industries94, 105562. 10.1016/j.jlp.2025.105562

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  TatzgernF.SchirruM.SchandlM.SchneidhoferC.VargaM. (2025). Tribo-acoustic tachometer: a novel ultrasonic device for the simultaneous operando measurement of grease film thickness and cage rotational speed in rolling element bearings. Tribol. Int.212, 110930. 10.1016/j.triboint.2025.110930

  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  WangQ. (2023). “Research on the lubricating grease distribution patterns of the beam pumping unit in the oil and petrochemical industry,” in 2023 3rd International Conference on Electrical Engineering and Control Science (IC2ECS) (Hangzhou, China:: IEEE), 732–735. 10.1109/IC2ECS60824.2023.10493762

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  WangK.LiuX.WuX.ZhuZ. (2017). “Condition monitoring on grease lubrication of rolling bearing using AE technology,” in 9th International Conference on Modelling, Identification and Control (ICMIC) (Kunming: IEEE), 595–599. 10.1109/ICMIC.2017.8321525

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  WangB.WangZ.SunC. (2021). Numerical investigation of the heat-fluid characteristic inside high-speed angular contact ball bearing lubricated with grease. IJE34 (5). 10.5829/ije.2021.34.05b.26

  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  WangL.PanY.JiangZ.ZhangY.FuX.MenX. (2022). “Temperature field analysis of cylindrical roller bearing based on rheological properties of grease,” in 2022 the 14th International Conference on Computer Modeling and Simulation (Chongqing China: ACM), 115–120. 10.1145/3547578.3547596

  • CrossRef
  • Google Scholar
  • View reference in article

• 46

  WangZ.WangF.DuanH. (2024). Numerical investigation of oil–air flow inside tapered roller bearings with oil Bath lubrication. JAFM17 (1). 10.47176/jafm.17.1.1944

  • CrossRef
  • Google Scholar
  • View reference in article

• 47

  WeiF.LiuH.LiuY. (2024). Study on atomization mechanism of oil injection lubrication for rolling bearing based on stratified method. Lubricants12 (10), 357. 10.3390/lubricants12100357

  • CrossRef
  • Google Scholar
  • View reference in article

• 48

  WenY.OshimaS. (2014). Oil flow simulation based on CFD for reducing agitation torque of ball bearings. SAE Int. J. Passeng. Cars - Mech. Syst.7 (4), 1385–1391. 10.4271/2014-01-2850

  • CrossRef
  • Google Scholar
  • View reference in article

• 49

  WuZ.-H.XuY.-Q.DengS.-E. (2018). Analysis of dynamic characteristics of grease-lubricated tapered roller bearings. Shock Vib.2018 (1), 7183042. 10.1155/2018/7183042

  • CrossRef
  • Google Scholar
  • View reference in article

• 50

  WuZ. H.XuY. Q.ChenZ. Y. (2020). Numerical analysis of grease film characteristics in tapered roller bearing subject to shaft deflection. IJE33 (7). 10.5829/ije.2020.33.07a.28

  • CrossRef
  • Google Scholar
  • View reference in article

• 51

  WuC.HongY.NiJ.TealP. D.YaoL.LiX. (2023). Investigation of mixed hBN/Al2O3 nanoparticles as additives on grease performance in rolling bearing under limited lubricant supply. Colloids Surfaces A Physicochem. Eng. Aspects659, 130811. 10.1016/j.colsurfa.2022.130811

  • CrossRef
  • Google Scholar
  • View reference in article

• 52

  WuH.ZhangS.SunH.WangJ.ZhangH. (2023). “Study on grease elasto-hydrodynamic lubrication of deep groove ball bearings,” in Recent Advances in Applied Mechanics and Mechanical Engineering. Editors YadavS.KumarH.WanM.AroraP. K.YusofY.diA. C. (Singapore: Springer Nature Singapore), 477–486. 10.1007/978-981-99-2375-5_47

  • CrossRef
  • Google Scholar
  • View reference in article

• 53

  WuC.ZhengC.ZhaoH.GhazaliM. J.WeiY.ZhuZ. (2025). Enhancement of thrust ball bearings performance through biomimetic composite textures: a comparative analysis under grease-starved lubrication condition. Tribol. Int.204, 110483. 10.1016/j.triboint.2024.110483

  • CrossRef
  • Google Scholar
  • View reference in article

• 54

  XueK.WangX.HuangH. (2022). Analysis of cage wear in rolling bearing greased. Tribol. Online17 (4), 291–299. 10.2474/trol.17.291

  • CrossRef
  • Google Scholar
  • View reference in article

• 55

  ZhangR.WeiC.WuW.YuanS. (2015). “CFD investigation on the influence of jet velocity of oil-jet lubricated ball bearing on the characteristics of lubrication flow field,” in 2015 International Conference on Fluid Power and Mechatronics (FPM) (Harbin, China: IEEE), 1324–1328. 10.1109/FPM.2015.7337326

  • CrossRef
  • Google Scholar
  • View reference in article

• 56

  ZhuD.ChenG.LiY.ZhangC.WangL. (2022). A 3D CFD simulation of oil spray-collection and delivery process in an aeroengine inter-shaft bearing. Chin. J. Aeronautics35 (9), 366–378. 10.1016/j.cja.2021.07.003

  • CrossRef
  • Google Scholar
  • View reference in article

• 57

  ZhuW.ZhuR.TangX.LuF.BaiX.WuX.et al (2022). CFD-based analysis of oil and gas two-phase flow characteristics in double-row tapered roller bearings with different rib structures. Appl. Sci.12 (3), 1156. 10.3390/app12031156

  • CrossRef
  • Google Scholar
  • View reference in article