Proceedings of the Institution of Mechanical Engineers,Part J: Journal of Engineering Tribology

Two main effects control the clearances and pressures in rough elastohydrodynamically lubricated contacts. The first is the local attenuation produced by any relative sliding of the surfaces, the second is the clearance variation that is generated in the inlet and which moves through the conjunction at the mean surface velocity – the complementary wave. The first component, which is absent in rolling contacts, is relatively easy to estimate. The second is more complex, particularly in soft contacts, where, for sinusoidal waveforms, it may have a wavelength that differs from the original profile. In addition it will generally decay in amplitude during its transit through the conjunction. The paper explores this component and outlines some of the factors that determine its behaviour with both sinusoidal and more general roughness profiles.

A pull-off force between a sphere and a flat plate is precisely investigated using a newly developed surface force apparatus (SFA). In this system, (1) a pull-off force between a spherical glass probe and a sample plate is measured in vacuum, (2) the probe is directly pulled off from the sample by an electromagnetic force, and (3) the pull-off force and displacement of the probe are measured with ultra-high resolution of 0.4 nN and 0.3 nm, respectively, which are electrical noise levels. The pull-off process, which appears as a part of force curve, is clearly measured by this system, and pull-off force is measured with high reproducibility and accuracy. Pull-off force distributions on flat surfaces of Si, SrTiO₃, glassy carbon, and diamond-like carbon are measured. It is shown that despite the differences between these materials, for all four for them, the distribution strongly depends on the surface roughness, such that the relative standard deviation of pull-off force is proportional to the surface roughness (R_a and R_z). The Greenwood and Williamson (GW) model is then used for the analysis of the pull-off force. Comparing with SFA experiments, it is shown that the pull-off force and separation displacement can be predicted by the GW model and that the pull-off force strongly depends on the standard deviation of surface height, causing the broad distribution of pull-off force at rough surfaces.

A numerical method to generate bifractal surfaces due to a modification of the slope of the power spectral density function in the low- or high-frequency range is proposed. The method has been applied to simulate real surfaces of Ginkgo Biloba leaf scanned at two different magnifications by matching the corresponding experimental power spectral densities. Slight differences have been found in the statistical distributions of the asperity heights and curvatures for the lowest magnification that had marginal influence on the frictionless normal contact response of the surface. For highest magnification, however, the statistics of the simulated numerical surface were quite different from those of the real one, leading also to a significant difference in the normal contact results.

Surface topography is important as it influences contact load-carrying capacity and operational efficiency through generated friction, as well as wear. As a result, a plethora of machining processes and surface finishing techniques have been developed. These processes yield topographies, which are often non-Gaussian, with roughness parameters that alter hierarchically according to their interaction heights. They are also subject to change through processes of rapid initial running-in wear as well as any subsequent gradual wear and embedding. The stochastic nature of the topography makes for complexity of contact mechanics of rough surfaces, which was first addressed by the pioneering work of Greenwood and Williamson, which among other issues is commemorated by this contribution. It is shown that their seminal contribution, based on idealised Gaussian topography and mean representation of asperity geometry should be extended for practical applications where surfaces are often non-Gaussian, requiring the inclusion of surface-specific data which also evolve through process of wear. The paper highlights a process dealing with practical engineering surfaces from laboratory-based testing using a sliding tribometer to accelerated fired engine testing for high performance applications of cross-hatched honed cylinder liners. Such an approach has not hitherto been reported in literature.

Some observations from acoustic emissions recorded during a yield test of a bearing raceway compressed into plasticity using a rolling element are presented. The general objective of the study is to establish whether there is enough evidence of the onset of sub-surface plasticity in the acoustic emissions signature. It is discussed here how acoustic emissions monitoring during compression could indicate the onset of subsurface plasticity as a precursor to damage propagation to the surface. Some comparisons are drawn between the acoustic emissions activity levels and time-frequency response during elastic deformation and at yield loads.

Acoustic emission is an emerging technique for condition monitoring of rolling element bearings and potentially offers advantages for detection of incipient damage at an early stage of failure. Before such a technique can be applied with confidence for health monitoring, it is vital to understand the variation of acoustic emission generation with operating conditions in a healthy bearing. This paper investigates the effects of increased speed and load on the generation of acoustic emission within cylindrical roller bearings, and it was found that the root mean square signal level increased significantly with increasing speed whereas increasing load had a far weaker effect. The AE_RMS value for each experiment was compared with the trend of the Lambda value. The bearing was operating under full film lubrication regime, so it was determined that increases in AE_RMS were not caused by asperity contact. By consideration of trends in frequency energy amplitude, it was determined that excitation of the bearings resonant frequencies were responsible for an increase of energy in the frequency range of 20–60 kHz. The excitation energy at 330 kHz (the acoustic emission sensor’s resonant frequency) increased with load, indicating a link between high-frequency emission and stress at the contact zone. Following characterisation of the bearing under normal operating conditions, an accelerated life test was conducted in order to induce fatigue failure. The frequency response demonstrated that throughout a period of constant wear, the energy amplitude at the bearings resonant frequency increased with time. As the bearing failure became more significant, the energy of the high-frequency components above 100 kHz was spread over a broader frequency range as multiple transient bursts of energy were released simultaneously by fatigue failure of the raceways. This paper demonstrates the potential of acoustic emission to provide an insight into the bearing’s behaviour under normal operation and provide early indication of bearing failure.

Friction tests using ZrO₂ (Y-PSZ: yttria partially stabilized zirconia) pin slid against polymer-like carbon film of bilayer and multilayer structures under H₂ gas environment are conducted. It will be shown that friction coefficients of the level of 0.0001 (friction fade-out) is stably realized by adding alcohol vapors to H₂ gas during run-in stage, then by stepping up the load from 19.8 N to 63.7 N after run-in stage. Four kinds of vapors of alcohol aqueous solutions are tested using bilayer samples, and ethanol-vapored H₂ gas shows the longest friction fade-out duration. Polymer-like carbon/diamond-like carbon multilayer sample shows long-life friction fade-out of 4 h, and it will be shown that the friction trace of 4 h reflects wear process of the first layer of polymer-like carbon and the second layer of diamond-like carbon. ZrO₂ surface is observed by an optical microscope and scanning electron microscopy and measured by surface profiler after friction fade-out test, and it is shown that flat contact area at the central region has many blisters and crimps, and is surrounded by peripheral bumps. It is also shown that the sliding marks are seen only at the top of crimps at the central region. Raman measurements indicate that short-chain carbons are predominant at blisters and ring carbons of small clusters are predominant at bumps. With these observations friction fade-out mechanism is discussed.