Data Assimilation Of Cellular Automata Model For Urban Flood

With the rapid development of urbanization,the urban area has been continuously expanded,the original ecological wetlands and river channels have been covered,and the ground with good permeability has gradually been replaced by impervious cement ground.When heavy rainfall occurs,the surface runoff increases.Due to the limited capacity of the urban drainage system,the drainage pipes have been in disrepair and aging for a long time,and the surface water cannot be discharged in time,which often causes more severe flooding disasters.Urban flooding has become a bottleneck restricting the sustainable development of society and economy.The flood evolution law is complex,and how to deal with the risk of urban flood disasters has become an urgent problem.Effectively conducting early-warning research on urban floods is a frontier topic in hydrological research.Urban flood simulation technology is an important means of urban flood prevention and disaster reduction,which can obtain its impact in time when a rainstorm event occurs,and its simulation accuracy is directly related to the effectiveness of disaster prevention and reduction.In order to construct an accurate urban rain flood model and study the formation mechanism and evolution process of the flood,the accuracy and effectiveness of the model are the key issues to be considered.At present,most urban rain flood models have certain defects,such as high data requirements,Low computational efficiency,unable to obtain the spatiotemporal evolution of floods,etc.In the face of the fast and accurate requirements of urban flood simulation,this paper builds a simplified model for numerical simulation of urban flood based on cellular automata.The main research process and results as follows(1)Construction of urban flood model based on cellular automata.From the aspects of cell space,cell state,neighborhood type,and evolutionary rules,it explains how to combine them with flood process simulation.Among them,evolutionary rules are the core content of model construction,and they are divided into runoff rules and convergence rules The rules need to consider multiple runoff factors of different underlying surfaces.This article mainly considers rainfall losses such as infiltration,plant retention,and gully discharge in urban areas,and formulates corresponding models based on previous research results;The actual situation calculates the water exchange between cells and the effect of flow time.With the help of MATLAB tool,the various modules of the cellular automata model are integrated to realize the data input of the model,the calculation of production and confluence,and the output of results.(2)The assimilation of CA model and observation data.The basic idea of data assimilation is explained by standard Kalman filter and ensemble Kalman filter.The combination of the ensemble Kalman filter and the CA model is explored from the aspects of state variables,set size,and error selection.Comprehensively consider the problems of assimilation effect and calculation amount,and set reasonable values of assimilation algorithm parameters,so as to improve the assimilation effect as much as possible while not causing excessive calculation amount.(3)Case studies and analysis.Using the constructed CA model and data assimilation scheme to conduct a case study in the main urban area of Jinan City.The simulation results of the flood process based on CA show that the cellular automaton can obtain the submerged water depth and flood evolution process in the study area.Compared with the measured maximum water depth of multiple hydrological stations,the model accuracy is better,and the error is about 20%.It is reliable to construct urban rain and flood models by cellular automata.After using data assimilation technology to substitute the measured water depth data of multiple hydrological stations in the study area into the CA model,the model accuracy has been significantly improved,the model trajectory has changed,and the water depth value is closer to the observed data.The use of data assimilation technology can further improve the simulation effect of the CA model.In summary,the cellular automaton model does not have high restrictions on the data like the hydrological and hydrodynamic models.It mainly requires high-precision terrain data,and the modeling idea is simple.It is easy to implement through computer technology.Ensemble Kalman filter can further improve the accuracy of the model,which is an important development direction of future urban flooding models.