| Water is an important resource for human development in agricultural production and social production,as well as is essential for human's surviving.Increasing demand on running water has led more investment into the construction of urban water supply network.Without taking into consideration the reliability of the pipe network and security of the water supply,the economy efficiency is the only target when it comes to optimizing the pipe network.This paper mainly studied the necessity of the multi-targeted model established for the water supply pipe network,and also elaborated how to realize it during the design of optimizing the pipe network.Starting from the practical engineering of the design of the water supply pipe network,the multi-targeted model of water supply pipe network was studied based on the theory of multi-targeted optimization,random search algorithm and computer technology.Finally,the multi-targeted model,taking total costs of the pipe network as the economic target,the redundancy of pipe network as the target of the reliable water supply and the node water age as the target of a secure water quality,has been established.In this study of the optimization of the multi-targeted water supply pipe network,EPANET was regarded as the engine for the hydraulic calculation of the pipe network to conduct the hydraulic calculation,and meanwhile,EPANET was used to simulate the note water age in the pipe network to know the features of it's changes,and furtherly to know the changes in water quality.Epanet2.0 was researched and developed by National Institute of Risk Management of United States Environmental Protection Agency,mainly functioning to conduct professional simulation of the water supply pipe network and pressurized system,and conduct the hydraulic calculation and water quality analysis needed by the owners.The software functions comprehensively.It can not only simulate and analyze the water quality of the water supply pipe network,but also handle the compensating computation and operation simulation and so on of the pipe network.While the optimization of the pipe network is using EPANET for hydraulic calculation,it meanwhile can make the optimization by using Genetic Algorithm(hereinafter GA),which is optimization algorithm with natural selection and genetic mechanism based on evolution theory.GA has a strong expansibility and can provide excellent searching results,which guarantee GA to be used in various industries and be promoted for mixture usage.Using GA in the optimization calculation to settle the constraint condition,this study adopted penalty functions and transformed the constraint condition to the unconstrainted model to get the solution.Penalty functions are always being studied significantly.Penalty results in the optimization will vary with the penalty functions.As there are many kinds of penalty function forms,such as static function method and successive function method which are used commonly,it was another key point for this study to ascertain the penalty functions suitable for this study.As for the microcosmic hydraulic model of the pipe network,GA has been adopted to construct optimized scheduling model.In line with the features of the issues,the fitness functions have been designed to ensure that the gap between the fitness of the dominant individual and the non-dominant individual in the early evolution will not be too large.In this way,premature convergence could be avoided,and the gap between the individuals with similar fitness was increased,enabling the superior individuals to flow into the next generation with a greater probability.The algorithms and calculations involved in this study were written in C# language by the author,in which the function library of EPANET was written in C / C ++,so a interface converting for the used functions is necessary when the software function is substituted.However,the GA,which adopted the function library using third-party open source code based on the C #,can be used directly. |
PDF Full Text Request