Application isssues of the semi-markov reliability model

Predicting the reliability of marine internal combustion engines, for instance, is of particular importance, as it makes it possible to predict their future reliability states based on the information on the past states. Correct reliability prediction is a complex process which consists in processing empirical results obtained from operating practice, complemented by analytical considerations. The process of technical state changes of each mechanical device is stochastic and continuous in states and time, hence the need to divide this infinite set of engine states into a finite number of subsets (classes), which can be clearly and permanently identified using the existing diagnosing system. Using the engine piston-crankshaft system as an example, the article presents a proposal for a mathematical model of reliability which, on the one hand, takes into account random nature of phenomena leading to the damage, and at the same time reveals certain application flexibility and the resultant practical usability.