MODELING THE DEGRADATION OF VISCOSITY OF ENGINE OILS

Abstract

In this paper, we present an analytical study to describe the monitoring and change in the condition of viscosity of lubricating oil used in combustion engines and exploited under very severe conditions. The objective is to determine the remaining lifetime of this oil. Stochastic methods in general are statistical methods that can be employed in predictive analysis, and they highly depend on probability theory. Markov models, in particular, make it possible to predict the degradation of the oil viscosity after a determined operating time. In these, degradation can be described by transition probabilities from one state to another. Knowing that Markov models are used to describe the transition probability to a new state (i) at time (t) based directly on the probability of the previous state (j) at time (t–1). In this context, we propose a Markov model to study the degradation law for the SAE 40 oil. The latter is designed for the lubrication of locomotives engines under the working conditions of a steel plant in Algeria. We have demonstrated that the proposed model gives detailed results describing the different conditions/states of the analyzed oil.