FORECASTING PROBABILITIES OF GAS PIPELINES TECHNICAL CONDITION ON A BASIS OF GRAPH AND AGGREGATED MODELS

This paper concentrates on an actual problem of forecasting of pipelines technical condition. A new classification offers local, aggregated and graph mathematical models. The forecasting problem has been solved with help of identification and forecasting filters based on aggregated and graph models. A construction principle of a graph model and a differential equations system for four types of damages are given on an example of joined pipeline which passed two intratrumpet inspections. According to the analytical solution of the differential equations system, we received probabilities distribution of technical condition for a pipeline. An optimum model of equivalent probability helps to predict technical conditions and to estimate a pipeline residual resource. The second method of identification and forecasting uses a nonparametric solution of the Wiener-Hopf-Rajbman integrated equation.