Research On Application Of Particle Swarm Neural Network In Corrosion Prediction Of Water Supply Pipeline

Water is an essential resource for people in normal life,ensuring the urban water supply system work safely is an important basis to guarantee people’s normal life and work.Internal corrosion is an important cause of underground pipe network malfunctions as rupture and other troubles,and it is harmful and difficult to find out.Domestic water supply pipe networks are mostly in service for more than five years.The lack of various factors,such as the increasing time of use and improper management measures in the operation and maintenance process,underground water supply pipe network systems often have hidden dangers of varying degrees,and are accompanied by various problems related to usage of water.It is relatively complex of the factors that affects the internal corrosion of the water supply pipe network,and various factors interact with each other.At present,all the scholars in the industry have not came out a general corrosion rate prediction model for this problem.In this paper,a new artificial neural network model is designed for the calculation of corrosion rate in water supply pipelines and conducted relevant experiments to verify accuracy.The main research content is as follows.(1)Corrosion rate in water supply pipeline is a regression matter under the influence of multiple influencing factors.In the process of regression calculation using ANN,the selection of influencing factors is crucial.In this paper,the relevant data of some gray cast iron pipes collected in the project were collated,and related parameters were selected according to the importance of internal corrosion research experts and scholars on the importance of internal corrosion factors.(2)Using quantum square potential well to improve the PSO algorithm,and according to the algorithm,the weights of artificial neural network model are optimized.The constructed neural network is verified to have fewer control parameters and better convergence effect.Combined with the project data,the corrosion rate was evaluated and predicted.The experimental results confirm that the QPSO-nn model proposed in this paper is suitable for the calculation of the rate of internal corrosion of gray cast iron pipes in selected areas,and has a certain reference value and meaning.