Research And Application Of Parameter Calibration Of Rain Flood Model Based On SWMM

As an important tool for studying urban waterlogging,urban rainwater model can provide more scientific basis and suggestions for urban waterlogging prevention and control.Among them,SWMM rain and flood models are more common,The parameters in model are generally selected based on the experience of the value range given in the SWMM manual.However,due to the large number of parameters in the model itself,blind parameter adjustment is time-consuming and the simulation calculation results differ greatly from the actual situation.Based on the SWMM model,this paper analyzes and studies the following three aspects based on the above problems:(1)Based on the data collected in the previous research area,preprocess the data,use geographic information systems to extract key information such as elevation,slope,and pipe network,and divide the catchment area of the study area,obtain parameter data of subcatchments and pipe networks,construct a SWMM model for the study area.(2)Based on the local perturbation analysis method and the global partial correlation analysis method,sensitivity analysis is performed on the empirical parameter items that need to be calibrated to obtain:For peak flow,the Manning coefficient in the impervious zone and the Manning coefficient in the pervious zone are sensitive parameters.The maximum infiltration rate and the pipe Manning coefficient are insensitive parameters.For total runoff,the Manning coefficient in the impervious zone,the minimum infiltration rate and the Manning coefficient in the pervious zone are sensitive parameters.The maximum infiltration rate and the pipe Manning coefficient are insensitive parameters.Parameter sensitivity analysis can save time and improve efficiency for the next parameter calibration.(3)Using BP neural network combined with multi-objective particle swarm optimization algorithm to determine the model parameters,the calculation result of the optimized combination after the calibration is in good agreement with the actual situation,and the different of rainfall are used to simulate the obtained optimized combination,and the relative error of the calculation result Small,the relative error of the two calculation results is within 10%,indicating that the calibration method can make the calculation result with high accuracy and good stability.Using thesingle-objective particle swarm optimization algorithm to the model parameters,the relative error of the calculation result is 23.53%,compared with the multi-objective calibration method in this study,the relative error is larger.The results show that the accuracy of this method is accurate and suitable for the calibration of SWMM model parameters in the study area.