Research On Air-Entrapment Evaluation Model In Water Pipelines Based On Pressure Signals

With the acceleration of urbanization,the contradictions between urban water supply and demand are increasingly serious.Long distance water transmission projects(LDWTPs)can be used to improve the present situation of the uneven spatial and temporal distribution of water resources and optimize the allocation of water resources from areas with abundant water to cities with water shortage.In the actual long-distance water pipelines,the existence of air is inevitable.However,air entrapment and air pockets accumulation can pose threats to the safe operation of water pipelines.Therefore,air exhaust should be carried out in time in case of the failure of water pipelines.In this study,the air-water two-phase flow patterns are connected to the air-entrapment states of pipelines,so as to realize the effective air-entrapment evaluation based on the accurate flow pattern identification.Based on laboratory experiments,the air-entrapment evaluation model is established to provide potential guidance for engineering applications.The main contents of this study include the following aspects:(1)Research status analysis of two-phase flow patterns.The possible influences of air-entrapment on the operational states of the pipeline are summarized and analyzed,and the definition of two-phase flow and the previous work of flow pattern identification are summarized.Based on the analysis of the domestic and foreign research,the research framework was put forward in this paper to build the air-entrapment evaluation model for water pipelines.(2)Introduction of laboratory experiments and signals data collection process.Based on the laboratory experiments,the downward sloping pipeline is selected as research object which is most likely to accumulate air pockets.And various air-water two-phase flow scenarios in the long-distance transmission pipeline are simulated based on the experiments,and the corresponding signals data are recorded by using the high-frequency pressure sensor.The signals data of typical scenarios are analyzed to prepare for the research on air-entrapment evaluation model for water pipelines.(3)Feature engineering research based on pressure signals.Based on the feature vectors derived from pressure signals,the optimal vector extraction method is analyzed by comparing the performance of various feature vectors on air-water two-phase flow pattern identification.And the optimal feature vectors is used as the input of the air-entrapment model for water pipelines.(4)Research on air-entrapment evaluation model for water pipelines based on supervised learning.Based on the pressure signal derived feature vectors,the performances of the supervised learning models are compared and analyzed.On top of that,the influences of sampling parameters on air-water two-phase flow pattern identification are analyzed for reasonable sampling parameters of pressure signals.(5)Research on air-entrapment evaluation model for water pipelines based on complex networks.Based on the complex network nodes and threshold solutions,the complex network edges are analyzed.The complex network graph topology is composed of nodes and edges.The air-water two-phase flow patterns of different scenarios are identified based on the topology characteristics of the corresponding complex network graphs.