Friction in highly pressurized lubricants and its relation to thermo-physical properties

Abstract: The objective of this work was to increase the understanding of which lubricant properties control transient friction, and during what environmental conditions friction is minimized. The experiments have been conducted during conditions relevant to an elastohydrodynamic contact, with pressures extending up to 2.5 GPa, initial temperatures between 20-80°C in the lubricant and loading durations of approximately 200-400 ms. A major revision of an existing work on the relation between pressure and dilatation, for mineral- and synthetic oils, using a modified split Hopkinson pressure bar was undertaken. A quadratic expression with lubricant specific constants, describing the pressure-dilatation relationship without an asymptotic behavior, was presented. A thermal study of isothermal and adiabatic effects on the result was added to the report. One significant observation was that the difference between adiabatic- and isothermal dynamic behavior of the compressed lubricants was small. The proposed idea was that, in practical calculations, it is unnecessary to distinguish between adiabatic and isothermal material data for dynamic calculations. Time-dependent processes on a molecular level must explain the difference found. A new type of experimental set-up was developed, utilizing the flexural response to an exerted load and dynamic beam theory, enabling simultaneous measurement of transverse and normal forces, and thus friction coefficient, as a function of time or pressure. The reproducibility of the load exertion was found to be low but the accuracy of the measurements were encouraging. Differences in friction histories between lubricated and dry contacts were found. Improvement of the developed friction coefficient measurement device was made to enhance reproducibility and accuracy. The experimental set-up was in many ways inverted and a new theory was developed, encompassing both rotational inertia and shear deformation in the energy of motion consideration. Reproducibility between experiments, along with accuracy, went from low to excellent. Five different lubricants were tested and frictional data, as a function of time, were presented. Several phenomena concerning friction were observed and discussed. Two interesting observations were that: Friction increases with increasing pressure during each load cycle (and decreases again as the loading is removed). Friction decreases with increasing maximum pressure of the loading pulse. This apparent contradiction inspired further work in the field and indications suggested that the solution might be embedded in the thermodynamic properties of the lubricants. Enhanced environmental control, added to the set-up, enabled initial temperature increase up to 60°C above ambient. Focus on thermodynamic properties was rewarded with a number of observations and a suitable explanation to the previously described contradiction was formed. It was found that increased coefficient of heat conduction lowered the friction coefficient, as did an increase in specific heat capacity. Both effects strive to reduce the temperature in the contact (by different means) but the tendency is clear: colder contact temperature, lower friction. Further support through the finding that increased initial temperature correspondingly increased friction was acknowledged. The main conclusion was that solid friction is lower than fluid friction during the prevailing test conditions and that a numerical analysis is considered necessary. Frictional data for 12 different lubricants and model hydrocarbons with different physical- and thermodynamic properties were presented. A numerical study of the thermal, shear stress and velocity gradient profiles along with friction coefficients for a paraffinic mineral oil, subjected to a transient pressure pulse in-between two infinite, rigid plates with relative motion was undertaken. Two rheological models were used: one nonlinear viscous and one nonlinear viscoelastic (Ree-Eyring). It was concluded that a Ree-Eyring rheological model together with pressure and temperature dependent viscosity and heat conduction can be used to qualitatively simulate transient rheology experiments. A localized temperature increase occurs in the lubricant film. It is associated with a viscosity decrease and a velocity gradient increase and a “slip plane” is formed. The slip plane will be more pronounced as the pressure pulse peak becomes higher until eventually all motion is localized to that region. The observed contradiction in friction is not viscosity related. The simulation indicated that the friction coefficient increased with increasing pressure, for the rheological model used, in total contradiction to measured data on a real EHL contact. It was consequently concluded that that the frictional properties of the localized shear bands, occurring in the solidified lubricant, were the origin of the detected anomaly and refined rheological models are necessary to bring crisp clarity to the matter.