| The development of hydrological models is driven by a variety of factors,one of which is the prediction of model numerical simulation.The main influence of hydrological model numerical simulation is the complexity of model structure and the accuracy of calibration parameters.To make the model can reflect more semi-humid and semi-arid region of small watershed rainfall runoff simulation,and the solution to a single standard is not enough to reflect the different aspects of the measurement system,this study consider the hydrologic cycle,climate,underlying surface,and the influence of artificial factors,on the basis of vertical mixture runoff yield model to join the water conservancy project construction,set up suitable sub-humid northern semi-arid region of vertical hybrid model of qinshui river,the basin rainfall runoff for multi-objective numerical simulation research.It provides an important scientific basis for the development and utilization of regional water resources and has practical significance for the rational allocation of water resources.Research and analysis the structure and function of hydrological models,calculation methods at home and abroad,model characteristics and suitable conditions,through the principle of mathematics and physics theory describe the watershed hydrological processes,combined with the feature of domestic specific watershed hydrologic cycle and hydrological data in qinshui river basin in the lumped hydrological model calibration,adjust the conceptual model parameters to match the runoff simulation and observation.Research in order to overcome the manual test error,choose global optimization single objective SCE UA algorithm and multi-objective MOSCEM-UA and MOPSO algorithm parameters into the model to carry on the numerical simulation,to solve the problem of multi-objective calibration,the introduction of the automatic optimization procedure based on the hybrid complex evolutionary algorithm algorithm,from Pareto inferior solution concentration with the Pareto front solutions,the Pareto front see DRMS and obvious non inferior relation between the LOG target,MOSCEM uniform search can be achieved in the numerical simulation of model space sampling,effectiveness and distribution parameters are better,More suitable for model simulation.After modeling according to the characteristics of the basin,the data of rainfall,evaporation and runoff in the basin from 1981 to 2011 were collected,and applied to the vertical mixing model of qinshui river basin for example simulation.The simulation was carried out by daily model and secondary flood model.The results showed that the daily model could better reflect the overall runoff characteristics of the basin.MOSCEM's NASH efficiency coefficient class a was 74%,with a percentage deviation of 90.9%.Secondary flood simulation more directly reflects the characteristics of flood peak and flood volume under the four targets.The qualified rate of flood peak is 93.8%,81.3%,91.7% and 95.8%,respectively.The qualified rate of relative error of runoff is 62.1%,81.3%,83.3% and 91.7%,respectively.Considering the four target simulation results,DRMS tends to be high flow fitting and LOG tends to be low flow,the two multi-target algorithms can take into account multiple factors at the same time,and MOSCEM process line has a better trend of conformity,which further verifies the effectiveness and advantage of multi-target parameter calibration.Therefore,the vertical mixing model of qinshui river basin is simulated by using multi-objective algorithm,and the accuracy of fitting results is improved by taking many factors into consideration. |
PDF Full Text Request