Modeling And Optimal Control Of Urban Drainage Network System

With the acceleration of urbanization,domestic sewage and industrial wastewater are rising continuously,and urban sewage overflow accidents occurred from time to time.There are many problems in urban drainage network system in China,such as imperfect facilities and backward control methods.In addition to strengthening the infrastructures such as the separation of rain and sewage,pollution source monitoring,sewage transmission and storage,the technical means of production monitoring and operation optimization is urgently needed for safe and efficient drainage.In this thesis,aiming at the problem of the optimization control of pumping stations and the coordination control between two-stage pumping stations,system modeling and optimization control of drainage network systems are studied as follows:1.The level prediction model of sewage pumping station based on LSTM.By analyzing the mechanism of the drainage pipe network,the factors that may cause the change of pump station level are obtained.The gray correlation analysis is carried out for these factors,and the factors with large correlation degree are used as input to establish the prediction model.The experimental results show that the LSTM model has not only higher prediction accuracy,but also better inflection point tracking than the traditional neural network.2.The dynamic hydraulic model of the drainage pipe network based on SWMM.The SWMM model is used to setup a hydraulically model for the actual drainage pipe section.Among them,the important input conditions of the model,the daily change pattern of node flow,are mainly analyzed and studied.In the simulation process,the problem of full pipe flow was overcome.Finally,the model was verified to provide an analytical means for the coordinated control of the sewage pumping station.3.Two application of local optimization control of drainage pipe network.(1)Constant level control method of single pump station.Aiming at the problem of the pumps starting-stopping frequently and large energy consumption,the LSTM prediction model is combined with fuzzy control to be applied to the frequency conversion control in the actual sewage pumping station,compare and analysis the actual operation data before and after transformation,the control effect is good.(2)The coordinated control strategy of the cascade pumping stations.Aiming at the problem of the sewage is easy to overflow in inspection well during the drainage peak period,first use the SWMM model simulate and analyze,a coordinated control strategy was proposed,then combine the strategy with the scheduling software for practical operation guidance,after several months of trial operation,the practical guiding effect is good.