Dynamic reliability modelling — Applications in gas industry

One of the main problems in reliability assessment of a complex dynamic system is to take into account dependencies of the system structure resulting from changes of its physical parameters. The aim of this paper is to provide the mathematical model of actual natural gas compression station which is situated in the Czech Republic. The Czech Republic represents one of the European transit countries for distribution of natural gas in east-west directions. The compression station and its surrounding gas pipeline network undergo performance changes in context with the amount of transported natural gas. The assumed performance changes that are provided and supervised by a human dispatcher, are based on consumer wishes which are changing during time. We describe the industrial multi-state system by the use of Markov process with continuous time. The model parameters are based on historical data, on the list of rules of possible human dispatcher performance changes behaviour and also on expert information from the gas network industry. An effect of change of the hardware configuration of the system by a human dispatcher and its corresponding change of the system availability is presented and discussed with conclusion that the pipeline network is a highly reliable system.