Method for determining the number of states of the Markov Model of damage accumulation in predicting the technical condition of a fiber-optic cable

Abstract

Estimation of the residual service life of operating fiber-optic cables is an urgent t ask. Usually this problem is solved based on the use of the Markov chain model. However, due to the nonlinear dependence of the probability of rejection on the rate of gradual failures, the task of selecting the number of states of a Markov chain becomes difficult. The article discusses a technique for determining the required number of states of the Markov model of damage accumulation based on a given value of the modeling accuracy. The characteristic values of the time and the probability of failure are found for the model of the destruction of optical fibers made of silica glass. The determination of the required number of states of a Markov chain in the article is carried out using the Python programming language.