Research On The Optimization Method Of Urban Water Supply Pipe Network Based On Partition

As the basic guarantee of urban production and life,urban water supply system is an indispensable and necessary facility for urban production and development.Along with the accelerated urbanization process,the urban water supply system is also facing new challenges.This research focuses on three elements for optimizing the urban water supply system: urban water supply network partition,urban water supply network water consumption prediction,and pressure optimization scheduling of the urban water supply network.To improve the scheduling of urban water supply networks,the partition method is used to group nodes with similar water consumption patterns into the same partition,effectively eliminating the phenomena of under-and over-pressure at individual nodes while managing pressure in the partition.In this essay,the flow data of the water supply network generated by the actual production and life in a city has been analyzed.And use the Random Forest algorithm to filter the time-domain features of the flow data.Then optimize the water supply network partitioning strategy based on the Hierarchical Clustering algorithm.The results based on the improved Hierarchical Clustering method are compared with the results based on Pearson analysis.The results of the simulation experiments show that the improved hierarchical clustering-based water network partition method gives better partition results than those based on Pearson analysis.It is provided that the subsequent theoretical basis and technical support for the optimal scheduling of the water supply network.For the flow prediction of urban water supply networks,a range of water supply network flow prediction models are evaluated and compared.The flow data was given different weights based on its time series characteristics.And the flow prediction model based on the improved GRU algorithm is proposed,based on the characteristics of water supply network flow data as a time series data,combined with the strategy of local weighted linear regression analysis.The results of the simulation experiments show that the flow prediction model based on the improved GRU algorithm has less error than the flow prediction model based on the GRU algorithm.Finally,based on water supply network partition methods and flow forecasts,a partitionoriented pressure optimization scheduling approach is investigated.To visualize the pressure optimization results,an optimized scheduling model is proposed with the water supply energy consumption of the partition’s regulator nodes as the objective function.The pressure values to be regulated by the regulator nodes are optimized using an improved PSO algorithm,and the optimized energy consumption of the regulator is calculated.Based on the known flow and pressure demand of each node inside each partition,the pressure at the pressure regulating node is optimized with the minimum network pressure requirement and the maximum pressure limit of the network as the constraints,to achieve the flow and pressure demand inside each partition with the minimum energy consumption.