Mixed Lubrication: Wear Particles Size and Friction Evolution

Wear Model
Reprinted with permission from [1].

Wear in tribological contacts results in generation of wear particles of various sizes and shapes and these particles impact the performance of the mechanical devices. These particles may create additional mechanical damage or act as catalysts and adversely affect the lubricating properties of lubricants. The size of the generated wear particles determines the severity of the impact.

The size of particles also determines the evolution of friction coefficient during initial stage of wear – running-in. In mixed lubrication regime the change of roughness profile is closely related to the wear particle generation. In turn, the evolution of surface roughness determines the friction coefficient. Therefore, the knowledge of the wear particles size evolution allows to predict the evolution of the friction coefficient.

In Surface Technology and Tribology group of University of Twente we developed a model to predict the size of wear particles formed in lubricated contacts and it was applied to calculate the evolution of the friction coefficient during running-in. Typical evolution of friction coefficient with wear during running-in is shown in Fig. 1.

 COF Evolution Running-In
Fig. 1. Friction coefficient at 0.05 m/s sliding speed a) measured curve b) calculated curve. Reprinted with permission from [1].
As shown in the Fig. 1, the typical friction process starts with high friction coefficient (Initial) and after initial phase stabilizes at a lower – (Run-In) level. The theoretical model was validated using experimental data by comparison of initial and run-in values of coefficient of friction. The results are shown in Fig. 2.

COF Evolution Running-In Validation
Fig.2. Initial and Run In friction coefficient curves, experiment vs. calculation. Reprinted with permission from [1].

Prediction of the initial friction coefficient by the model is in a reasonable agreement with the experimental data for both the initial and  run-in friction coefficient. It can be noticed that for the sliding speeds of 0.03 and 0.05 m/s, the model underestimates the friction. This is ascribed to the roughening of the surface due to abrasive wear of particles re-entering the contact, an effect that is not included in the model.
The largest change in the mean coefficient of friction due to running-in is achieved in the mixed lubrication part of the friction (Stribeck) curve  (0.03–0.05 m/s). Obviously, the effect of running-in was absent in the EHL regime, 0.2–0.4 m/s. In boundary lubrication the effect of running-in was less pronounced. After all, the fraction of the contact separated by the lubricant is already very small and will not change much by wear.

The model was also validated by the measurement of wear particles size using Dynamic Light Scattering and a good agreement was observed. Developed model can be used to optimize the design of the mechanical components. As discussed in the original work, the hardness of the material can be chosen to obtained a minimum friction coefficient after running-in.

For further details see “Generation of wear particles and running-in in mixed lubricated sliding contacts”, Aydar Akchurina, Rob Bosman, Piet M. Lugt, Tribology International 110 (2017) 201–208, http://doi.org/10.1016/j.triboint.2017.02.019.

[1] “Generation of wear particles and running-in in mixed lubricated sliding contacts”, Aydar Akchurina, Rob Bosman, Piet M. Lugt, Tribology International 110 (2017) 201–208, http://doi.org/10.1016/j.triboint.2017.02.019.

Save

tribonet
About tribonet 172 Articles
Administration of the project

5 Comments

  1. “In the current paper, wear particles formed in boundary-lubricated steel on steel and steel on brass sliding contacts were studied….”

    “Centrifugation [was] necessary to collect most of the particles from the lubricant and… to eliminate unwanted contaminants…. [S]everal researchers showed that the small particles tend to make agglomerations which can be misleadingly considered as large single particles [internal citations omitted]. In order to break these agglomerates, sonication [was] employed.”

    “In order to prepare wear particles samples for further analysis, the following isolation procedure was applied. First, particles were collected with the oil after the wear test. Isopropanol (2-propanol, IPA) was added to the oil holder which was ultrasonically cleaned for 15 min at 40 °C…. IPA was used as a cleaning agent. The mixture of IPA and oil with particles was placed into 40-ml centrifuge tube and centrifuged for 15 h. The upper 30 ml was then discarded, and the tube was filled with deionized water (DI) and centrifuged again for 15 h. During the centrifugation, high-density wear particles… [formed] sediment at the bottom of the tube, whereas low-density oil … forms an upper layer in the tube. By discarding the upper layer, the oil can be eliminated. This procedure of adding DI water and centrifuging is repeated 3 times to make sure that all the oil was eliminated…. The suspension of particles in DI water (3 ml) is then collected for further analysis….”

    In the real world, 3rd-body wear particles in lubricated contacts do not wash themselves in IPA and DI water, then deagglomerate themselves via centrifugation for almost two days.

    In real world lubricated contacts, 3rd-body wear particles are subjected to (and cause) all manners of horribleness. They agglomerate, fuse, and are covered in contaminates. As we travel farther and farther away from reality in these experiments (simulations of reality), we loose sight of what is really happening in lubricated bearing contacts.

    As an alternative, I suggest the following experimental design…

    Take bearing, spin until broken, open bearing, study surfaces.

    “It can be noticed that [the study] model underestimates the friction. This is ascribed to the roughening of the surface due to abrasive wear of [3rd-body] particles re-entering the contact, an effect that is not included in the [presented study] model.”

  2. Rick, as you mentioned, the impact of the wear particles coming back to the contact was noticed based on the friction coefficient level. But it was also noticed that the wear model predicts the size of wear particles quite well. So as the next step, we would introduce the impact of wear particles entrapment on the wear, friction and surface roughness evolution. Step-by-step we will reach the model capable of accurately predicting these quantities depending on the applied conditions. Some day! 🙂

  3. “So as the next step, we would introduce the impact of wear particles entrapment on the wear, friction and surface roughness evolution. Step-by-step we will reach the model capable of accurately predicting these quantities depending on the applied conditions.”

    Aydar, in your future continued efforts to model the size and impact of 3rd-body wear particles in lubricated contacts, are you going to negate the reality of agglomeration?

    Further, when dealing in comparative radii of non-spherical particles, how is the model accounting for the highly irregular shapes of the particles and their differing contributions to friction and surface damage based on morphology?

    My earlier point was related to the tendency of tribological studies to divorce themselves from reality. Continuing the model based upon data achieved from centrifugation of wear particles to release agglomerates is not going to yield accurate conclusions.

  4. We would like to be able to consider agglomeration as well of course. It is not so clear though how largely the wear process, in particularly mixed lubrication conditions, is affected by the agglomeration itself. So the current model would help one to answer this question. We also used sonication to brake possible agglomerations.
    The model does not consider the particles to be spherical. They are actually of irregular shape and in our previous article (DOI: 10.1007/s11249-016-0701-z) we compared calculated length, width, depth of the particles with the AFM data. But for the sake of convenience and also to be able to compare the data with DLS measurements, we introduced an equivalent radius of the particle.
    In my opinion, the developed model represented the testing conditions (so the reality) very nicely. We did not use any fitting parameters, and the results suggest that we describe the reality quite well. So far, we don’t know the degree of agglomeration and also we don’t know how much the agglomeration of the wear particles affects the process, at least in these conditions. So we could use the model to answer these questions and setting up the right experiments.
    I think that we should go from simpler models to more complex. Based on the developed model we can consider all kinds other situations by adding additional features to the theory.

  5. This poor guy (real reality) doesn’t want any more modeling done, he wants somebody to stop the asperities in the lubricated contacts destroying the bearing and producing so much heat that the lubricant boils away. I’ll bet there is agglomeration in his model ;), when the failed bearing starts shooting metal into his engine as the inner race begins to displace laterally from the resulting wear… https://m.youtube.com/watch?v=k5DbwqhFPjY.

    My point here is about departures from reality in modeling and in the lab, and the pressing need to simply address the real world problems in tribological contacts, namely the presence of asperities themselves in such contacts.

    Wear modeling in the pursuit of new AW/EP additives is futile when the failure mechanisms produce sufficient heat to boil away the lubricant and shoot metal into the engine.

    In short, it’s time to just make the asperities disappear.

Leave a Reply

Your email address will not be published.


*