Application Of Genetic Algorithms To Optimizing Neural Networks In Water Supply Network

Water is a precious resource,limited and vulnerable to environmental pollution.China's per capita water consumption is only 25% of the world's per capita water consumption.Since the reform and opening up,in order to achieve economic development,environmental protection has been neglected,and pollution problems have become increasingly serious.Many water sources have been destroyed and are not suitable for residents' drinking water sources.Therefore,the situation of urban water supply is grim.With the environmental pollution and the increasing population of the world,the tension of water resources is deteriorating.Although China is a big water resource country,from the per capita point of view,China is also a serious water shortage country,with a per capita water occupancy of only 225 cubic meters,only 1/4 of the world's per capita water,ranking 110 of the 149 countries in the world.China's water pollution is serious,seven rivers and five lakes have been polluted to varying degrees,so the situation of urban water supply is grim.With the changing situation and tasks of water supply,urban water supply project belongs to a very important infrastructure of a city.Its normal operation ensures the daily work,life and industrial and agricultural output of citizens.With the changing situation and tasks of water supply,the pressure,difficulties and challenges will be greater and more in the process of advancing the key work in depth.At present,there are still some problems to be solved in the water supply system,such as pipe burst,insufficient water supply,insufficient pressure and huge waste of energy.These are due to our inaccurate understanding of the operation conditions of the whole network,so establishing an accurate model of the network operation conditions can help us to understand the operation of the network in real time,facilitate the implementation of drinking water purification project construction,metering system,improve theefficiency of leak repair,improve the user terminal water quality guarantee system,and allow more and more citizens to enjoy safety.High quality tap water.Because the macro-model has the advantages of short modeling time and not needing too many pipe network parameters,this paper chooses to build the macro-model of water supply network,but there are many difficulties in the past macro-model modeling process.When the main factors change greatly,the model will produce some errors.The BP neural network model can deal with the non-linearity and unclear dynamic relationship among the influencing factors,so the BP neural network is chosen to establish the macro-model of the pipeline network.Aiming at the problem that the number of hidden layer neurons of typical BP neural network is not easy to determine and the convergence speed is slow,this paper uses genetic algorithm to optimize the number of hidden layer nodes and improves it by momentum factor and learning rate self-adaptation.Based on the above research,two macro models of Guangzhou pipeline network working conditions are established text different needs and different conditions of water supply network operation.