Project in the Spotlight: Grease Life Prognostics: Remaining Grease Life Estimation by Measuring Grease Properties and Using Predictive Models

Key information

Project title: Project in the Spotlight: Grease Life Prognostics: Remaining Grease Life Estimation by Measuring Grease Properties and Using Predictive Models
Project in the Spotlight: T19021
Market: Industrial (railway, wind, etc.)

Funding: This project is co-funded by ‘Holland High Tech, Topsector High Tech Systems and Materials’, with a ‘PPS Innovation Grant Public-Private Partnerships for research and development’

Partner: SKF

Run time: 2021-2025

Written by M2i Program Manager: Franz Bormann

Why predicting grease life matters
Modern machinery consists of a multitude of individual components, where their delicate interplay dictates overall efficiency. One crucial element in many systems is the roller bearing, which keeps rotating parts aligned and running smoothly. These bearings are usually lubricated with grease, not only to reduce wear but also to protect against corrosion and seal out dirt.

Grease is more complex than it appears: roughly eighty to ninety percent base oil is locked inside a sponge-like network of soap or polyurea fibres, supported by additives that fine-tune its performance. During service, this network “bleeds” oil onto the bearing tracks, maintaining a protective film. As more and more oil escapes over time, the grease eventually dries out, the lubricating film can no longer be maintained, and the roller bearing may fail.

For manufacturers and maintenance providers, the ability to predict that moment – the true end of grease life – is both economically and ecologically highly valuable. It enables timely grease replacement, avoiding both premature servicing and the risk of costly bearing failure.

The accuracy of such predictions improves when the condition of the grease is monitored in real time. This means that it is not enough to only predict when grease typically fails, but also to estimate the “remaining grease life” during actual operation.

Understanding how grease degrades
To improve the understanding of the mechanisms behind the mechanical degradation of grease in roller bearings, researchers from the University of Twente worked together with SKF in a four-year PhD project.

The first part of the project focused on studying how grease behaviour evolves inside operating bearings. Using SKF’s R0F+ test rig – a specialised setup designed to assess grease performance under high-speed and high-temperature conditions – the team monitored changes in the grease’s ability to bleed oil over time. These tests showed a clear link: as bearings run, grease loses its capacity to release oil, confirming that bleed is a key parameter for estimating the remaining useful life of grease.

Tracking oil bleed: A closer look
But the key question remained: how does bleed behaviour change as grease ages? To investigate this, the researcher extended a laboratory method based on blotting paper. This technique was originally developed by SKF to assess grease bleed and was later refined during an earlier PhD project. In its basic form, the method involves applying a well-defined quantity of aged grease to the paper under controlled conditions. Because oil has a higher affinity for the paper fibres than for the thickener matrix, it gradually migrates into the substrate.

In the original method, the extent of bleeding was estimated by measuring the size of the resulting oil stain on the paper. However, it became clear that the stain edge was not sharply defined, due to non-homogeneous flow through the paper. This irregularity stems from the local variations in fibre length, density, and structure within the paper itself.

To address this limitation, the researcher extended the method by recording the spreading process over time using a camera setup. Image processing techniques were then applied to analyse the light intensity across the stain area, providing a more detailed measure of the oil distribution. With this improved method, it became possible not only to quantitatively assess how bleed behaviour changes over time, but also to qualitatively track the dynamics of oil diffusion.

By testing grease samples taken both from bearings at different stages of operation and from grease mechanically aged in a grease worker, the evolution of bleed performance was systematically monitored under different conditions. These experiments revealed an important insight: while service ageing clearly changes bleed behaviour, mechanical degradation imposed by the grease worker itself does not significantly influence bleed properties.

Modelling bleed
Capitalising on the knowledge gained, a physical model for bleed – the rate at which oil is released – was developed. The model is based on Darcy’s law and links grease properties, namely permeability, oil–thickener affinity, and grease elasticity, to the oil flow. This modelling approach successfully captured the strong influence of oil concentration on bleed behaviour, while reflecting the negligible contribution of mechanical degradation. Most strikingly, the model was able to reproduce a long-standing rule of thumb in tribology: bleed practically stops once grease has lost about half of its oil content.

Towards a predictive model for grease life
Building on these findings, the next step is the development of a comprehensive model for predicting the remaining useful grease life. This phase of research will go beyond bleed alone and include the physical and chemical processes that determine how lubricating films form and deteriorate inside bearings. The central idea is that grease life is governed by a balance between oil supply and oil loss: oil must bleed from the grease reservoir and reach the contact surfaces, while simultaneously resisting evaporation and oxidation.

Each of these mechanisms – bleed behaviour, oil transport, contact replenishment, and oxidative degradation – will first be modelled individually and then integrated into a single framework. The project, carried out by a postdoctoral researcher at the University of Twente in collaboration with SKF experts, will also build on recent advances in bearing kinematics and starved elastohydrodynamic lubrication (EHL) theory to describe how film thickness evolves both before and after oxidation begins. A probabilistic failure model, based on changes in film thickness and film condition, will then be developed and implemented using a Weibull-based reliability approach.

Ultimately, this integrated model is a crucial step towards enabling predictive maintenance, where grease can be replenished at optimally timed intervals – avoiding both premature replacements that waste material and delayed interventions that risk bearing damage.