Research On Constructing Of Asynchronous Sensor System In Water Distribution Networks For Hydraulic Model Calibration

In recent years,with the increasing scale of urban water distrubution networks,the intelligent monitoring and management of water distrubution networks has become a major trend.In order to manage the water distrubution network system,the real-time hydraulic model is widely used in operation,and the calibration of the network model needs to rely on the monitoring data provided by the sensors in the network.However,limited by hardware performance,power supply and others,the low upload frequency of existing sensor systems can not meet the needs of high-frequency hydraulic model calibration.The most direct approach is to add a large number of new sensors,but this approach ignores the existing sensors and is uneconomical.Based on the existing sensor system monitoring redundancy,it is possible to structure an asynchronous sensor system so that the old low-frequency sensors can also be used in the high-frequency hydraulic model calibration.At the same time,some new sensors will be placed in the network to further improve the calibration accuracy in synergy with the old sensors.This study focuses on node water demand calibration of hydraulic model based on asynchronous sensor system and sensor placement based on asynchronous sensor system.Firstly,this study proposes a method to structure an asynchronous sensor system.The existing system is converted into an asynchronous sensor system by grouping old sensors and uploading monitoring data asynchronously,which improves the overall data uploading frequency and calibration frequency of the monitoring system.To ensure the calibration accuracy,a hierarchical cluster algorithm is used to group the sensors,which improves the representativeness of the monitoring data in each time period and thus improves the accuracy of the online model.This method is applied to a small virtual pipe network and a large actual pipe network,and the results show that the sensor grouping obtained by clustering analysis can guarantee the calibration accuracy,and the method is not only applicable to small pipe networks,but also can be applied to actual large pipe networks.However,asynchronous uploading of monitoring data will inevitably reduce the calibration accuracy,so it is necessary to arrange high-frequency sensors to improve the calibration accuracy.Then,this study proposes an sensor placement method based on asynchronous sensor system.How to couple the new high-frequency and existing asynchronous lowfrequency sensors is the difficulty of the study.To solve this difficulty,based on the principle of reducing the interference of nodeless points in monitoring similarity assessment,the hydraulic distance between nodes is calculated,the set of measurement nodes is constructed,and the monitoring effect difference(MED)is defined accordingly.Based on the MED,the objective function of sensor placement is constructed.Besides,the characteristics of the unfavorable hydraulic calibration nodes are analyzed,and these two types of nodes are removed from the set of alternative nodes according to the corresponding indicators.A case study of an actual large urban water distrubution network shows that the sensor placement method proposed in this study can take into account the monitoring capability of the original monitoring points,so that the newly sensors can collaborate with the original sensor system for hydraulic model calibration,make up for the shortcomings of the original system,and improve the calibration accuracy of the overall network as well as specific areas.