Research on ANN-based model of joint collocation of water quantity and quality

Eutrophication has become one of the main problems lakes confront. City lakes are facing serious eutrophication problems. This paper, takes the West Sea Lake as an example, uses joint collocation technique of water quantity and quality to deal with city lakes´ eutrophication problems. Based on analyzing the monitoring data of both source water and target water, it designs several scenarios of water collocation, uses WASP6.0 to simulate each scenario. It chooses TP, TN, Chla and BOD₅ to calculate TSI (Trophic State Index) of source water, target water and also the outcome of each scenario. Then, it uses the neural network toolbox of Matlab6.5 to form and train a BP neural network. Afterward, it takes scenarios as samples to validate the BP network. If expected goal is achieved, the model can be put into use. At last, it uses the finished model to determine a water collocation scheme, specialized in a certain scenario. The result generalized by the model can be a reference for water collation.