Stochastic Models in Acceptance Testing for Complex Machines

Abstract

The stochastic models of complex machines are built and used in acceptance testing to estimate the technical state of the machines after they have been assembled. The modelling method is based on the multivariate analysis of time series values of the machine diagnostic parameters. The working hypothesis is the following. The non-stationary time series of informative diagnostic machine parameters which characterize the working capacity and reliability of the machine are connected with each other by the stationary statistical dependencies. Identification of the changes in the dependencies is the basis for the proposed information technology to check the performance of the tested machines.