Modelling relation between oxidation resistance and tribological properties of non-toxic lubricants with the use of artificial neural networks

Abstract

The lubricants based on vegetable oils, as environmental friendly, are urgently sought. However, in addition to ecological characteristics, the lubricating properties have to be met. To meet these requirements the active additives influencing the lubricating properties and oxidation resistance are used. The useful lifetime of lubricants is determined largely by their abilities to resist oxidation. The article presented the results of new, ecological lubricants development. The oxidation performances of different developed lubricants have been tested. The experimentally determined oxidation stability of the compositions based on vegetable oils are presented. Analysed oxidation onset temperature was obtained from the differential scanning calorimetry (DSC) curves, which provides the rapid prediction of the oxidative stability of lubricants. Besides the lubricating composition based on vegetable oils, the developed greases-based mineral, or synthetic oil were investigated. The properties of these greases were evaluated using the measurement of parameters describing structure (penetration) and resistance to high temperature (dropping point). The lubricating properties of both the greases and vegetable oil compositions were tested on four-ball testing machine. In the results of the modelling of the lubricating properties the neural network models for the both types of the lubricants were developed. A discussion of the research results and analysis of models validity is given below. The experimental results are compared with the calculated using the neural models. An acceptable agreement was achieved.