The tribodiagnostic measurement and analysis of wear debris from oil - lubricated machinery is used as a prevention of unwanted failure and as a design tool in the development of new machinery. There are many types of wear such as adhesion, abrasion, corrosion, erosion, fretting, cavitation, fatigue, melting and other. Each of these results in its own characteristic form of wear particle, the identification of which is sometimes difficult. There are many methods for identifying particles and for monitoring their development over time. One such method is ferrography a technique for separating wear particles from the lubricant matrix and depositing them on a glass slide, arranged or sorted by particle size.
An important element of a rational approach to tribodiagnostic measurement is a suitable expression of stochastic quantities to describe an examined object. For complex diagnosed objects, e.g., vehicle combustion engines, there are two approaches to describe the technical state of the objects:
- 1
Select a small number of dominant tribodiagnostic parameters and then directly evaluate them by common methods.
- 2
Apply statistical methods to evaluate a comparatively large set of diagnostic parameters. Representative of statistical methods is discriminative analysis. Utilization of discriminative analysis is based on the ability of the method to describe one latent qualitative parameter by means of several quantitative variables.
This paper discusses ways to apply mathematical methods to evaluate the results of tribodiagnostics (ferrography) related to vehicle combustion engines. The idea is based on a discriminative analysis which makes it possible to describe one qualitative parameter, i.e., complex technical state of the engine, by means of several quantitative parameters, i.e., quantity of diagnosed wear particles in used oil. The results have been verified by means of considerable statistical data of T3 - 930 engines made in Czech Republic used in trucks and automobiles.