Study On Remaining Life Prediction And Maintenance Decision Of Pipeline System Based On Multi-degradation Modeling

The natural gas pipeline system includes several key components such as pipelines and valves.Pipelines play an indispensable role in both transportation and mining links.When the pipelines are in maintenance,the valves can cut off the pipe section to make the upstream and downstream two-way cavitation,and the maintenance can be carried out under the condition of incomplete cavitation.However,due to the interference of soil environment corrosion,its own material and third-party damage,the safe operation of the pipeline system will cause great hidden dangers,and the failure of any kind of components will have a nonnegligible impact on the normal operation of the system.It is difficult to ensure the safe and economic operation of the whole system only considering the reliability model of one kind of components and developing maintenance strategies.Therefore,it is of great significance to conduct reliability modeling and joint maintenance decision-making according to the different degradation characteristics and failure modes of the two heterogeneous units of the pipeline system.In this thesis,the remaining life prediction and maintenance decision of heterogeneous pipelines and valves with different degradation models in pipeline system are studied.The prediction and maintenance decision can not only effectively reduce the probability of system accidents and prolong the normal operation time of the system as much as possible,but also provide theoretical support for the daily maintenance of pipeline maintainers.The main research contents are as follows:(1)To solve the problem of low prediction accuracy due to fixed transition probability assumed by traditional methods,a pipeline remaining life prediction method based on Gamma and time-varying Markov model was proposed.Firstly,continuous stochastic process Gamma was used to describe the pipeline corrosion degradation process.Secondly,the continuous stochastic process is divided into state space and the remaining life prediction model is established by Markov theory.In order to adapt to the actual degradation process of natural gas pipeline,the time-varying transfer probability matrix is constructed to solve the problem of fixed transfer probability.Finally,through the example verification and simulation in MATLAB,the simulation fitting value is compared with the actual value.The results show that the prediction accuracy of the method based on Gamma and time-varying Markov model is high,and the predicted value is smaller than the actual life,which can better guarantee the safe operation of natural gas pipeline.(2)To conduct fault prediction and maintenance decision modeling for heterogeneous units with different degradation modes in the pipeline system to ensure the safety and reliability of the system,multi-degradation models were established for the key units of the pipeline system,pipeline and valve,and an imperfect opportunistic maintenance strategy based on the combination of state and age was proposed according to the pipeline prediction results and different failure modes of valves.Firstly,degradation models were established for different failure modes and degradation characteristics of pipelines and valves with heterogeneous units.Secondly,based on the predicted results of the pipeline and the service age of the valve,the imperfect maintenance strategy was developed.Considering the economic correlation of the two-part system maintenance,the opportunity maintenance strategy was adopted.Then,a maintenance decision optimization model was established with the opportunity maintenance threshold and replacement threshold as optimization variables and the minimum average maintenance cost of the system as objective function.Finally,the effectiveness and economy of this strategy are verified by case experiments and sensitivity analysis,and compared with the maintenance decision without considering opportunity maintenance.