Tribotests are used to evaluate the performance of lubricants and surface treatments intended for use in industrial applications. They are invaluable tools for lubricant development since many lubricant parameters can be screened in the laboratory with only the best going on to production trials. Friction force or coefficient of friction is often used as an indicator of lubricant performance with sudden increases in friction coefficient indicating failure through catastrophic adhesion. Under some conditions the identification of the point of failure can be a subjective process. This raises the question:​ Are there better methods for identifying lubricant failure due to catastrophic adhesion that would be beneficial in the evaluation of lubricants? The hypothesis of this research states that a combination of data from various sensors measuring the real-time response of a tribotest provides better detection of adhesive wear than the coefficient of friction alone.

In this investigation an industrial tribotester (the Twist Compression Test) was instrumented with a variety of sensors to record:​ vibrations along two axes, acoustic emissions, electrical resistance, as well as transmitted torsional force and normal force. The signals were collected at 10 kHz for the duration of the tests. In the main study D2 tool steel annular specimens were tested on coldrolled sheet steel at 100 MPa contact pressure in flat sliding at 0.01 m/s. The effects of lubricant viscosity and lubricant chemistry on the adhesive properties of the surface were examined. Tests results were analyzed to establish the apparent point of failure based on the traditional friction criteria. Extended tests of one condition were run to various points up to and after this point and the results analyzed to correlate sensor data with the test specimen surfaces. Sensor data features were used to identify adhesive wear as a continuous process. In particular an increase “friction amplitude” related to a form of stick-slip was used as a key indicator of the occurrence of galling. The findings of this research forms a knowledge base for the development of a decision support system (DSS) to identify lubricant failure based on industrial application requirements.