An inadequate lubricant volume at rolling contact inlets, termed contact starvation, leads to a reduced separation between opposing moving surfaces. Starvation has been linked with various failure mechanisms and can affect rolling element bearings of any size. The level of starvation can be defined by the length of the inlet meniscus, the position where lubricant films from the raceway and roller join at the inlet. Therefore, an in-situ meniscus measuring tool is key for future bearing condition monitoring systems. Currently, no other technology is capable of taking a meniscus measurement from a field bearing.
The inlet meniscus is thin, occurs over a small area, and is hidden deep within the working components of the bearing, meaning such an in-situ measurement has been previously impossible. Ultrasound, which requires no direct access to the contact and has been proven sensitive to films within the micron range in which the meniscus exists, and has been highlighted as a potential technology to take such a measurement.
In this thesis, a novel ultrasonic thin film measurement technique, capable of measuring oil and grease film thickness in-situ on a rolling element bearing raceway, has been developed. The method involves measuring the resonant frequency of the free surface film on the raceway, and using a calibrated acoustic velocity-temperature relationship, calculating the film thickness. From the measured raceway thickness and theoretical roller film, a model was created to determine the meniscus length, and thus starvation level.
The measurement technique was shown to be repeatable in a full scale wind turbine gearbox bearing test rig with both oil and grease lubrication at a range of loads, speeds and over an extended time period. The shape of the film leading into the contact inlet was generally observed to be non-uniform across the rolling axis, with the inlet film thinner at the axis centre and thicker towards the raceway edges to create a 'U' shape distribution. This shows contacts have the potential to be fully flooded and starved at the same time, due to localised lubricant inlet conditions.
Although starvation was not repeatedly recreated in this work, for high viscosity oil lubricated contacts, a pre-starvation pattern was identified, which could be used as an indicator to future contact starvation and premature wear. The calculated starvation ratio during this pattern was of a similar magnitude to alternative starvation work in a deep groove rolling element bearing. When lubricated with grease, evidence of the channelling and clearing sub-phases of churn were observed in-situ. An inlet pattern similar to previous grease starvation and reflow contact patterns was seen, highlighting the dependence of contact separation on the inlet conditions.
A qualitative measurement of roller skew was also performed, based on the cage speed and time interval between contact passes. Sudden step changes in skew magnitude, termed skew impulses, were observed at steady state conditions, but were more common at higher bearing speeds. Results show there is a relationship between skew magnitude and lubrication state.
