The effects of external loading on low displacement wear rates of unlubricated steels

Highlights

• An experiment was designed to find the effects of external loads on fretting wear

• Tensile, compressive and torsional loads were applied to crossed cylinder contacts

• Wear rate was found to be insensitive to all forms of external loading considered.

Abstract

Whilst contact loading is known to affect wear, the general stress field is rarely considered. Steel fretting (and low amplitude reciprocating) wear contacts typically develop through a transient regime and wear by multiple mechanisms. It is expected that if these mechanisms were controlled by plasticity or fatigue, the wear rate would be altered by external stresses. Whether or not these stresses must be accounted for is an important consideration. This paper assesses the sensitivity of wear rate to external stresses, experimentally. An apparatus was designed to apply external loads to fretting wear contacts. The wear rates throughout the tests were insensitive to changes in external load, indicating that wear models need only model stresses due to contact loading, and external loads can be disregarded.

Introduction

Wear problems are often overcome through the use of large bearing surfaces to lower pressure, and through the use of lubricants to lower the coefficient of friction. Fretting wear is a challenge in mechanical design as it can occur between components which are in nominally static contact, preventing lubrication. The range of amplitudes within which fretting wear mechanisms operate is not universally agreed, however a common figure is around $\text{300}\text{µm}$ [1]. This absolute displacement is not important, but rather the displacement relative to the contact size, as this determines the ease of debris egress. Often, wearing components will be subject not only to contact loads but also to external loading. A clarification of this distinction is given in Fig. 1. Classical wear models such as the Archard law [2], acknowledge the influence of normal load in a wearing contact, but not the effects of the more general stress field, in particular the impact of stress acting parallel with the free surface. Fig. 1 shows the loads required to define a partially stuck and incomplete contact. The contacts studied here are fully sliding albeit of small displacement, so the components ${\sigma }_A$ and ${\sigma }_B$ will not have an effect on the tractions as they do in contacts under partially stuck conditions [3]. The question addressed here, is whether the additional stress in the wearing material, resulting from these loads, has any effect on the wear rate.

Damage in metal components is usually attributed to one of the following:

• Plastic flow

• Brittle fracture

• Chemical change/corrosion

If the dominant wear mechanisms were plasticity driven, then it is reasonable to assume that wear rate ought to be controlled, not by the contact pressure alone, but by a yield parameter. So if we assume that von-Mises’ criterion applies, the non-traction component of stress, for example (the direct stresses parallel with the free surface) ought also to contribute to the wear rate. Further, if the wear particle nucleation and early development were crack related, then surface tension ought also to have a strong effect on rates of wear, just as it would if it were causing the growth of fatigue cracks [4]. We conclude that if either of these mechanisms is responsible, fundamentally, for the generation of wear particles, then tension (or compression) parallel with the free surface might be expected to change wear rate. These remarks apply whether the tension is present along the direction of sliding, or transversely, or both. This hypothesis is explored experimentally in this paper.

Stresses parallel with the free surface might originate either from externally applied loads, or from the presence of self-equilibrating residual stresses. These might arise because of manufacturing processes, or because the surface has been subjected to some surface treatment which gives rise to residual stresses (peening or electro-plating, for example). Lin et al. [5] showed that residual compressive stresses (700 MPa) in treated steel reduced abrasive wear. Much of the previous work, such as that cited here and that reviewed in [5], on the effects of residual stress on wear rate have focused on the wear resistance of coatings where clearly, the material being worn is also changed. Vierneusel et al. [6] explored the links between residual stresses and coating wear in molybdenum disulphide, observing significant reduction in wear in specimens with coatings with a large compressive residual stress state.

Naga et al. [7] conducted tests with a contact geometry and loading similar to that studied in this paper but with brass, concluding that the wear rates of brass were accelerated by the presence of static tensile loads. Upon review of this data, any effect is small and inconsistent, with no data given for variability. Ho et al. [8] studied the effect of residual stresses imposed by shot peening on wear rates of steels, concluding that the residual stresses created by the wear mechanisms themselves often exceed the magnitude of any pre-existing stress, therefore ‘washing them out’. Ho measured wear induced axial stresses of 600–800 MPa, and shear stresses of up to 125 MPa but negligible stress perpendicular to the surface. He argued that only if the contact stresses induced were below the magnitude of the pre-existing stress state, could the pre-stressing influence the wear rate.

Owing to the number of compounding effects known to influence wear rates in metals, it is infeasible to quantify empirically the effects of changes in any single variable relative to all others. This paper simply explores whether external loading (stress due to sources other than contact loading) in a component has any significant effect on wear rate for fretting steel–steel contacts, so that it can be identified as a potential modelling requirement. The type of contact studied here was a dry (unlubricated), steel-on-steel contact as these have been thoroughly studied in the literature due to their obvious industrial prevalence. Initially, contacts of this nature wear quickly by severe mechanisms such as abrasion or adhesion, after which the wear rate slows as more mild mechanisms such as oxidation become dominant [9], [10]. In the experiments to be described, measurements of worn volume are taken cycle by cycle, so the wear rates resulting from the various mechanisms can be measured independently (as opposed to tests in which one measurement is taken at the end, and so the ‘rate’ calculated is an average). Consequently, the influence of external loading can be seen for each of these wear regimes (or mechanisms).

In this paper we describe the results of an experiment in which constant external loads are imposed on the specimen being worn. Experiments have been designed to apply either axial stress or torsion to cylindrical specimens. Tests in which axial pre-stresses are applied explored the possibility that the wear mechanisms are crack driven. This is based on the assumption that changes in near surface direct stress will substantially impact crack growth, as discussed in detail by Susmel et al. [4]. Torsional pre-stress experiments were performed to find the response of wear rates to more complex in-surface states of stress. Torsion was used as it allowed for application of high shear stresses in the surface of the material, as well as higher von Mises stresses than could be achieved with the axial loading method.