Inversion Of Roughness Parameter Of River Network Based On Genetic Algorithm

As we all know, In order to adapt to the needs of national economic constructions,To solve the problem of river traning in the land management,the computation of water conveyance losses in the water conservacy constrution,the flood computaton in the Flood Prevention,the Hydrological Computation when bridge design of bridge in the highway and railway constructions, and the roughness of the river channel is a important parameter in the Hydraulic calculation, many scholars commtted to the roughness of the river floods.This paper elaborated present situation and significance of inverse analysis of channel friction firstly, summarizes Research Method and the contents of Inverse problem.This paper applied the genetic algorithms to inversion of roughness parameter of river network.And write the interface program coupled the swmm source program with GA program to complete the process of the inversion automatically,And analysis the result.The main contents are as follows:1. Summarized The Research Method of inversion of roughness parameter of river network,GA is Proposed to applying to inversion of roughness parameter in river network. Because GA intergrated the both advantages of global optimization and local optimization,we improved the precision and computational efficiency by using GA. And solve the problems of falling into the local optimum and Reduce the workload of artificial debugging.2. In this paper divided the river network into two types: treelike network and annular network,and use time sequence observations on few observation points as the inversion of information, verified use of genetic algorithm inverse of the entire river network's roughness is feasible.3,Through the comparative study with diffrent number of observation points, found that when the number of observation points in the river network come to the half of the number of river channels, the river network be able to get the optimum combination of roughness,and the roughness accuracy tends to a certain value, with the number of observation points increase,the accuracy of roughness improve not obviously.4,This article also done a preliminary study with the different number of Observing time, and found have a similar law with the number of observation points, the more the number of observation time, the higher the accuracy of the roughness, but after a certain number of precision tends to a certain value.5,In this Paper analysis several different complexity of the river network, found that the more complex the river network, laid on the location of the observation points higher demand,and theannular network high sensitivity than the treelike network in the Observation points layout. The layout of the observation points evenly distributed in the river network, the roughness of the portfolio be better,and The more able to reflect the real distribution of roughness.