Quantitative estimation and prediction of tribological performance of pure additive compounds through computer modelling

Tribologists recognized the importance of the chemical structure of lubricant additives long ago. For many years experimental data for the tribological performance of different pure compounds have been accumulating. However, most of the work done was interpreted only qualitatively. Recently, new opportunities for quantitative estimation of relationships between chemical structure and properties through computer modelling have emerged. The paper describes an algorithm for modelling of tribological properties. It is illustrated with a practical example-a phenomenological and a predictive model of experimental data for four ball machine weld loads of organic sulphides. The chemical structure of the additives is presented numerically with the energy distribution in structures with minimized molecular energy. The phenomenological model, based on tribochemical kinetics, and the predictive model, developed through molecular design with principle components, allow for prediction of the performance of organic sulphides within the experimental error. Opportunities for use of such models for scientific and industrial purposes are outlined. The necessity for compilation of data banks with tribological properties of pure additive compounds, the place of kinetic and adsorption studies, etc. are substantiated. Connection to systematic formulation and design of additive products and modern refinery modelling concepts is also discussed. (C) 1998 Elsevier Science S.A. All rights reserved.