Joel Furustig

Published: 22 January 2015

Semi-deterministic numerical simulations of wear on various scales:

From Chemo-Mechanical Effects to the wear of components in the orbital type hydraulic motor

A costly effect, caused by the relative motion of machine components in contact, is wear. Solid surfaces come into contact when the lubricant film thickness is thin, resulting in locally high pressures and high wear rates. Components and machines will, due to this wear, eventually lose their functionality.

Machines are often lubricated, meaning that there is a sharing of load between lubricant and asperities in the interfaces between machine components. The roughness of the surfaces influences the likelihood that asperities collide.

When asperities collide, protective layers grow on the solid surfaces due to chemical constituents of lubricants. These chemical constituents are called additives and some additives can, under boundary lubrication conditions, interact with surfaces to form the protective layers, also called tribofilms. The formed layers and the formation of layers have previously been studied separately.

Numerical simulations are often used to model phenomena such as wear. In this thesis it is asked whether it is possible to model wear on different length and time scales, and to model the interplay between these length and time scales? Can we model the chemical effects on the nano scale, in combination with contact mechanical effects due to roughness on the micro scale? Can we learn something from numerical simulation of the small changes in roughness during a wear process of a mixed lubrication contact in a component?

Several numerical simulation methods have been used. Contact mechanics problems are treated through the boundary element method. A model based on Arrhenius equation is used to include chemical effects in a chemo-mechanical simulation of the growth of tribofilm. An equation for the lubricant film thickness is combined with contact mechanical behaviour to estimate load sharing. A method to include the effect of wear on the load sharing is included in the model. Archards wear equation is used to simulate wear at several length scales. A model to determine the load distribution between gears in an orbital type hydraulic motor is developed.

Three wear test rigs are utilized. An orbital type hydraulic motor test rig is used to validate wear models for the gear set. A ring-on-ring test is used to experimentally simulate steady state conditions of wear on the topography scale. A sliding reciprocating ball-on-disk test rig is used to evaluate the wear in a contact.

Topography measurements and microscopy are used to quantify and qualitatively locate wear. The methods for topography measurements are stylus profilometry, vertical scanning interferometry and atomic force microscopy. The microscopy techniques which are used to locate and classify wear scars, are confocal microscopy and scanning electron microscopy.

Numerical simulation results are compared with measurements of the wear in a reciprocating ball on disk rig. Wear on the nano-meter scale on a rough surface is simulated by means of numerical methods, and wear depths from the numerical simulations are compared with wear depths measured by Atomic Force Microscopy. A model for tribofilm growth due to chemical reactions on the nano-scale is applied to model a wear process on a surface topography. Wear is numerically investigated in a gear set in an hydraulic motor and the results are compared with results from experimental investigations. A two scale model is applied to the gear set, in which wear effects the geometry of the gears and the small scale surface topography.

In Paper A, ball on disk wear is simulated numerically and by experimental methods. In Paper B, dry contact wear on the topography scale is considered. Roughness features on the scale of 100 nm, were similarly worn in experiment and simulation. In Paper C, a tribochemistry model is constructed in order to evaluate the growth of tribofilms on rough surfaces. The numerical model exhibits similar behaviour as other authors have observed when conducting experiments involving lubricant additives.

A multi scale wear model is applied to study wear in hydraulic motors in Paper D-F. The measured wear depths agree well with the wear depths predicted by numerical simulation. The location of the majority of the wear is the same in the numerical simulations and the experiments. Refinements of the methods to determine force distribution as well as the influence of roughness, improves the predictive capabilities of the multiscale wear models.

Further improvements of multi scale wear modelling are suggested, and some interesting future research topics are suggested.