Simulation Of Pollutants At Nest-in Chaohu Lake Based On LSTM And EFDC Coupling Model

In recent years,the state has paid great attention to the ecological environmental protection and sustainable economic development of the Yangtze River Basin.In dealing with the relationship between economic development and environmental protection,it has put forward the development principle based on the premise of protecting the ecological environment,namely,"focusing on the protection together,not engaging in large-scale development".Water resources,the core element of the concept of ecological environment,must be vigorously protected.Chaohu Lake,as the fifth largest freshwater lake in China,is one of the important water sources in Anhui Province.Chaohu lake has rich water resources,fishery resources and shipping resources,the status is very important.Around 2010,chaohu Lake environmental problems became increasingly prominent,cyanobacteria,duckweed,water hyacinth and other aquatic plants unrestrained crazy growth,industrial wastewater,domestic wastewater and agricultural water untreated direct discharge have led to the further deterioration of chaohu Lake water quality.With the strengthening of the overall efforts around Chaohu Lake in recent years,the water quality of Chaohu Lake has returned to the third-class water standard.However,the water flow in Chaohu Lake is less,the ecological environment is fragile,and the water change cycle is long.Therefore,the pollution prevention and control of Chaohu Lake water is still an important task for the local people.（1）Construct the deep learning model of the inflow flow of The South Feihe River,using 2018 a year into the nest traffic data to validate the LSTM model and rate,will be 70%of the amount of data as a training set,30%as a validation set,through the training set the LSTM model for training and tuning,using the remaining 30%of the data to verify this model prediction accuracy.Finally,the LSTM model was used to predict the flow of The South Feihe River in January 2019.Compared with the actual monitoring value,the prediction accuracy of the model could reach 92%,and the error RMSE was 0.015.（2）The EFDC water quality model of the south bank of Chaohu Lake was constructed,and the COD and TP of the nestmouth of Nanfei River and the COD and ammonia nitrogen of the nestmouth of Shiwuli River were simulated respectively.The accuracy of the simulation model was reflected by the accuracy of the simulation.The model was compared with the COD and TP measured data for 20 days at the entrance of the South Feihe River to verify the accuracy of the model simulation.The results showed that the RMSE of COD simulation was 0.27,the Nash coefficient of the model reached 74%,the RMSE of TP simulation was 0.22,the Nash coefficient of the model reached 0.80.The accuracy of simulation is higher than 70%,and the simulation results are more reliable.The model uses the COD and ammonia nitrogen measured data of the Shiwuli River inlet for 20 days as comparison to verify the accuracy of the model simulation.The results show that the RMSE of COD simulated by the model is 0.15,the Nash coefficient of the model is 0.86,the RMSE of ammonia nitrogen simulated by the model is 0.14,the Nash coefficient of the model is 0.85.The simulation accuracy of the model is higher than 80%,better reaction of pollutant diffusion process,the simulation results are reliable.（3）The flow value in the predicted period of LSTM model is converted into a.Dat file that can be read by EFDC model by using Python language,which is used as the hydrodynamic boundary condition of EFDC model to achieve the coupling of the two models.Through the coupling model of chaohu lake basin,three typical river（10 the wulihe river,nanfeihe river,hangzhou port）into the nest section was simulated,respectively select each river ecological water inlet（5m3/s,8m3/s,10 m3/s）three kinds of working condition of filling water,through the simulation value compared with measured values,get the coupling model of three kinds of filling water solution curve of the joint degree is higher.The coupling model can dynamically and intuitively predict the concentration of various substances entering the nest in four rivers at different times,providing technical support for the health of chaohu Lake water quality.