Advancing modelling of interfacial processes in boundary lubrication (P13)
Our goal is to investigate friction between DLC coated surfaces in boundary lubrication regime (BL). For BL regime there is no valid theoretical model which would enable us to predict friction coefficient since there are many variables that influence it. Predicting friction in BL regime is becoming more and more important in the construction of high quality mechanical devices so we shall focus on conditions that are important for everyday life. By choosing two surfaces (steel, DLC), minimal number of lubricants and test parameters we will make an attempt to give semi empirical friction equation in BL regime.
1. University of Belgrade – Serbia, Faculty of Physics, Applied Physics and Informatics.
Faculty of Mechanical Engineering
Laboratory for Tribology and Surface Nanotechnology
University of Ljubljana (Slovenia)
For more information contact Kosta;
Supervisors: Dr. Mitjan Kalin
Reports are availbale here
Research in friction is as old as industry itself. Frist attempt to systematically approach friction was made by Da Vinci around 1500. First models were established by Amontons and Coulomb, with about two hundred years in between. This points to the fact how complicated friction phenomena are and how difficult it was for scientist to approach it. First reliable reciprocating test was carried out by Hatchet in 1854. Stribeck published his results in 1901 and proved coefficient of friction dependence of speed. Later on, Stribeck curve was used to distinct between lubrication regimes.
From Stribeck curve, three lubrication regimes are: boundary lubrication (BL), mixed lubrication (ML) and elasto-hydrodynamic lubrication (EHL). Main distinction between the three is the amount of lubricant inside the contact that determines the separation of two surfaces and therefore defines the coefficient of friction.
Tribology has progressed fast in the research of ML and EHL regime; however, knowledge about BL regime is still scarce. One of the main reasons for it is that due to the many asperity – asperity interactions and use of different oil additives there is an overlap of many different mechanical, chemical and physical effects. It is still unclear how all of this occurrences (physical, physicochemical processes) influence the coefficient of friction, which of them are most influential and which of them have no, or limited, influence. As if this is not enough, it is also unclear how the coefficient of friction is dependent of contact parameters (i.e., contact pressure, surface roughness, sliding speed and temperature). There are studies that deal with one or two of these parameters but none considers more comprehensive set of parameters.
Modern tribology is focusing on tribochemistry; how oil and oil additives react with the surface (steel or DLC) and how this reaction contributes to the friction and wear. Much research has been performed with different additives (anti-wear, friction modifiers, extreme pressure, etc.) and the focus has been placed on the mechanism of the reaction, i.e. how the specific additive works. Understandably, these investigations are performed at fixed test parameters, so as a result we may get proposal of the mechanism for certain reaction but we do not know clearly if this mechanism is valid when our test parameters change and weather this change affects friction and how much.
Furthermore, from the performance point of view it is clear that effectiveness of additives depends from their adsorption on the surface. Extensive research has been undertaken in the recent years to determine how additives are adsorbed on different surfaces and what mechanism govern the adsorption process. While it is clear that adsorption will be one of the main determinants of the coefficient of friction the relationship between the adsorption process and the coefficient of friction is not yet defined. Reasons for this gap in knowledge is similar as the one mentioned previously when friction mechanisms were discussed. Studies are mostly done with single parameter set without variation and although detailed analysis of adsorption mechanism is thorough and irreplaceable it does not provide answer about connection of adsorption and performance.
Surpassingly, there are plenty of statistical tools available to help solving problems of huge number of influencing parameters we are facing in tribology today but these tools are seldomly used. Design of experiment (DoE) methods provide us with scientifically justified methods for many parameter variations in the single experimental matrix so we can obtain better insight into the behavior of the friction coefficient within certain range of quite a lot of parameters. Other statistical procedures (regression algorithms) provide us with empirical equations that would be able to predict friction coefficient from the considered test parameters.
In the frame of this project we have investigated how Hertzian pressure, sliding speed, temperature and initial surface roughness influence the coefficient of friction od DLC with additives in boundary lubrication regime. With the help of the DoE methodology we have systematically varied above stated parameters and obtained knowledge about how each of the chosen parameters influence friction. DoE methods have helped us to understand not only the influence of separate parameters but their interactions as well. The goal was not set to understand lubrication and tribochemical mechanism, which is a typical approach, but to estimate absolute and relative effects of different parameters on the coefficient of friction in a broad engineering range. From the assessment of the coefficient of friction attempt was made to assess the level of adsorption of investigated additives on tested surfaces. Finally, regression algorithm was applied and empirical equation containing investigated parameters was obtained. From the obtained equation we have created friction maps for the chosen materials and oils.
For the purposes of this project following contact parameters were chosen: Hertzian contact pressure, sliding speed, surface roughness and temperature. Experimental matrix that contains these four contact parameters has been designed according to the split plot design method with the test temperature being designated as hard-to-change factor.
Steel-steel and steel-DLC contacts were tested with five different model oils: base oil, base oil with w.t. 1% of antiwear additive, base oil with w.t. 1% of antiwear and friction modifier additive, model fully formulated oil and commercially available fully formulated oil. Tribotests were performed on oscillating tribometer. Median coefficient of friction was taken for further analysis with elastic net regression algorithm.
Elastic net algorithm provided us with the information about the correlation of studied contact parameters and produced empirical equation, composed out of studied parameters and their interactions, which was further used for prediction of the friction coefficient and creation of friction maps.
Results and contribution to the ENTICE project
Main results we have obtained during the project:
Obtained results contribute to the understanding of boundary lubrication regime as well as to provide good basis for further discussion on the mechanisms of adsorption and tribochemical reactions.