Research On Water Quality Assimilation Simulation Of Taihu Based On Particle Filter

Water resource is the fundamental material of human survival.In recent years,water pollution and eutrophication have become more and more serious,and the security situation of the lake water quality is particularly important,hence,water quality monitoring and protection become urgent.At present,the numerical model simulation is one of the most important ways to study the distribution of water quality parameters,and it has become an effective way to study the movement of lake water and quantitative evaluation of water pollution.As the complexity of lake hydrodynamic,the diversity of the parameters in numerical simulation model,the difficulty of data obtaining,some error may exist in the numerical model simulation results and it is difficult to accurately simulate changes of water quality in detail by the numerical model.On the other hand,the measured data can reflect the "real value" of the water quality at the place where and when the data is acquired,which can monitor the changes of water quality in a high accuracy.Data assimilation methods can adjust simulation process of the water quality numerical model in real-time based on the measured data,which can make it possible reflects the law of hydrodynamic and the distribution of water quality parameters.Based on this idea,the research on data assimilation simulation method is conducted,combined with the hydrodynamic numerical model.With the Taihu Lake as the study area,and chlorophyll as the research object,the research combines the particle filter methods with the FVCOM model to implement the spatiotemporal assimilation simulation analysis of chlorophyll.The main contents and contributions are as follows:(1)Research on assimilation methods based on particle filter.As implemented with a recursive Bayesian filtering algorithm from the time domain by non-parametric Monte Carlo sampling methods,particle filter needn't to meet the system requirement of linear constraints and Gaussian distribution noises,which has a wide range of applications.Therefore,we integrate the particle filter into the 3 dimensional water dynamics and water quality model,and constructed the Taihu assimilation simulation model by aggregating the model-based control simulation and particle filter together,which had made an effective combination of the measured data and the model simulation results and achieved a study of water quality parameters simulation in Taihu Lake.(2)The experimental study of Taihu assimilation simulation.Taking the Taihu Lake as the study area,chlorophyll concentration change as experimental object,based on the FVCOM hydrodynamic model and particle filter method,the assimilation simulation of Taihu chlorophyll concentration variation regularity and detailed features real-time dynamic simulation is conducted,which has taken six aspects:triangle mesh generation,FVCOM influencing parameters,initial background field generation,FVCOM parameter estimation,assimilation simulation factor,multi-threaded computing model.By comparing the assimilation simulation results and statistical error analysis,it turns out that assimilation results are superior to that of FVCOM with a higher simulation accuracy.(3)Taihu Lake assimilation simulation system.With the need of prediction accuracy of simulation model,this paper has implemented a multithreaded parallel computing model by integrate particle filter and FVCOM,based on GIS visualization technology.Besides,Taihu Lake assimilation simulation system is constructed by aggregating the model data entry,data preprocessing,simulation control,simulation and assimilation results visualization together,and it has achieved data simulation assimilation and real-time 3D data visualization,which will provide decision supports for eutrophication and water pollution in Taihu Lake.