Inversion And Design Optimization Of Hydraulic Parameters Of Water Supply Pipe Network Based On Intelligent Algorithm

With the implementation of the rural revitalization strategy and the promotion of urbanrural integration construction,the centralized water supply rate in rural China has reached 88%currently,and the water supply network has become a lifeline project for the vast rural areas.In addition to the basic guarantee of water quality and quantity,central documents have put forward higher requirements for rural water supply safety and water supply network management.This article first analyzes the current situation of rural water supply networks,points out the main problems currently existing in rural water supply networks,and further proposes the application of intelligent optimization algorithms to the layout of water pressure monitoring points,hydraulic model parameter verification,water quality optimization,and water volume scheduling in the network.This provides a theoretical and practical basis for achieving intelligent management of rural water supply networks.The friction factor is the foundation of hydraulic model calculation for pipeline networks.As the pipeline operation time increases,there may be errors in pressure adjustment based on the design value of the friction factor.Therefore,this article puts forward a method for inverting the friction factor of pipeline segments based on the measured water pressure values at monitoring points.Considering the reality of limited monitoring data for rural pipeline networks,in order to obtain more accurate inversion results through fewer monitoring point data,it is necessary to first select pressure monitoring points.In this paper,the partial derivative of node pressure to friction factor is used to express the node sensitivity function,and the fitness function is the sum of the sensitivity functions at the monitoring points.With the highest value of the global fitness function as the goal,the optimal selection of water pressure monitoring points is carried out by improving the genetic algorithm.In order to verify the impact of different monitoring points on the friction factor inversion results,the inversion results of pipeline friction factor in different monitoring point layout schemes are quantitatively compared and analyzed.On the basis of determining the optimal monitoring point,the node water pressure method is programmed into the fitness function on the basis of the water pressure value at the monitoring point.With the goal of minimizing the average double error of the water pressure at the monitoring point,the dynamic search fireworks algorithm is used to invert the friction factor of the water supply pipe network.The efficiency and accuracy of the dynamic search fireworks algorithm and particle swarm optimization algorithm in the inversion results of the friction factor of the pipe network are compared and analyzed.The construction and operation costs of water towers are also factors that need to be considered in the management of rural water supply networks.This article takes the multi tower water supply network as an example,and obtains the characteristic node water age and comprehensive cost under different tower diameter combinations through uniform design experiments as the training and testing sets.The relationship between water tower diameter,feature node water age,and cost was established using BP neural network and SVM,respectively.A 50 fold cross validation was used to partition suitable training and testing sets,and the results of the two methods were compared.Substitute the training results into the multi-objective genetic algorithm,with the goal of minimizing water age and cost,and propose a design plan for the renovation of water towers in water supply networks.This research got the following results:1.The improved genetic algorithm is suitable for combinatorial optimization problems such as monitoring point optimization.The maximum relative errors of the inversion values of the friction factor of the pipelines in the pipe network before and after the optimal layout of monitoring points are 17.7%and 0.711%respectively,which indicates that the optimal layout of monitoring points improves the accuracy of the inversion results of the friction factor.2.Compared with the results of PSO inversion of the friction factor,the dynamic search fireworks algorithm converges faster and exhibits strong optimization ability in the process of parameter inversion of rural water supply network,providing a feasible method for the inversion of the friction factor of the network.3.In the problem of small data volume and complex calculation process,both SVM and BP neural networks perform well in learning,and SVM has stronger generalization ability than BP.The water tower design scheme optimized by multi-objective genetic algorithm has improved in terms of node water age and cost compared to the current scheme.