Research On Calibrating And Optimizing Dispatching Of Urban Water Supply Network Model Based On Genetic Algorithm

In recent years,with the rapid economic development,more and more attention has been paid to the development of urban water supply network system by our country,a lot of money have been invested for pipe network transformation and expansion each year,the traditional empirical scheduling model has been unable to meet the growing demand for water quality.The overall management of the water supply pipe network in our country is lagging some developed countries,although the data for pipe network is more and more complete,the utilization of these information is very low among the water plant.Therefore,the scientific and reasonable mode of scheduling is the only way for the further development of water supply network system for our country.In this paper,through the study and discussion of related knowledge on water supply network optimization scheduling,the genetic algorithm(GA)is proposed to calibrate the micro-hydraulic model of the water supply network and to establish the optimal scheduling model.The calibrate of the micro-hydraulic model of the pipe network is generally divided into two steps,first,calibrating roughness coefficient of the pipe,second,calibrating water demand of the nodes.The pipe roughness is generally regarded as the “slow variable”,that the pipe roughness coefficient will increase with the increase in pipe age,but this change is slow,usually every 5 years a pipe roughness coefficient is calibrated.In addition,in the establishment of micro-hydraulic model of laboratory pipe network,it is found that changing the roughness coefficient of pipe section has little effect on the node pressure and pipe flow in pipe network,and as the time of the “constant variable”,the nodes water demand need real-time calibrating,so this paper focus on the calibrating of nodes water demand.The water demand of the laboratory pipe network is calibrated by GA,and the node water demand is coded by real number coding.Using adaptive genetic operator,the calibrating effect is good,the algorithm convergence speed is fast,the results of the calibrating have met the calibrating standards proposed by Zhao Hong bin.Using the hydraulic calculation engine of EPANET to carry out the pipe network adjustment calculation to obtain the state pipe network.Compared with the traditional ring equation method and node equation method,EPANET uses mixed node-loop method,based on thegradient information to obtain the state of pipe network,that have a very high computational efficiency.Based on the microscopic hydraulic model of pipe network,GA is used to establish the optimal dispatching model,designing fitness function,according to the characteristics of the problem,so that in the early evolution,the difference dominant individual and non-dominant individual fitness is not too large to avoid premature convergence,in the later stages of evolution,the difference between individuals with similar fitness is increased,so that the dominant individual can enter the next generation with greater probability.Respectively,the establishment of a laboratory pipe network optimization scheduling model and A-city pipe network optimization scheduling model is obtained.The optimal scheduling model based on laboratory pipe network is simple,by defining three different pumps' global efficiency,that is,the efficiency of the set value,the other state conditions are the same,the optimal solution is the opening of the highest efficiency of the pump.During the optimization process,we can see the change of the individual of each generation in the process of GA optimization and the change of the average fitness of the population,also we find that individuals which represent the optimal solution will gradually prevail,eventually "occupation" of the entire population is obtained.Then,based on the microscopic hydraulic model of A-city pipe network,a24-hour optimal scheduling model was established,which is Multi-mode optimization scheduling model,and a day for 24 hours of scheduling scheme is obtained.Compared with the traditional experience scheduling scheme,based on GA optimization of the optimal scheduling program,the pump operation energy cost can be more economic.The algorithm and calculation involved in the text are written by the author based on the C# language.The function library of EPANET is written by C/C++,therefore,when the library is called,it needs to transform the function and genetic algorithm library which is written by C# by using third-party open source code can be directly called.