Research On Coagulant Dosing Control Of Water Supply Factory

Actually, coagulant control mostly is coagulant dosing addition control. Coagulant dosing is indispensability and a key role in the process of drinking water treatment, It is sixty-four-dollar question to improve coagulant effect and economic benefits. At present the coagulant dosing control effect of domestic most water supply factories are not satisfying. It is a at large care and desiderate solve problem how to calculate the efficient coagulant dosage according currently raw water qualities in water treatment industry.Based on various factors in the process of coagulation, such as many characteristics of raw water quality, capacity of treating water, coagulant technics and so on; Complex mechanism in research, and coagulant process is a big time-delay, nonlinear and uncertain system, so it is difficult to control and ascertain nicety mathematical model. Grounded on the traditional classical control method, this thesis introduces the theory and idea of neural network, and introduces the structure, theory of modeling, analysis of simulation, forecast of model and realization of BP and GRNN neural network.In this thesis, the background of the thesis choice and the actuality and development of coagulant dosing control methods and manners are first introduced. The thesis analyses the structure and the calculation of BP that properly apply to set up the coagulant process modeling, aimed at the characters of BP, the thesis makes a availability improvement of the calculation of BP, introduces a improvement calculation of BP—SCG. The structure and traits of GRNN was analyzed.By analyzing the characters of coagulant process and the schemes of coagulant dosing control, designs the raw water turbidity, temperature, pH, basicity and flux as the import nerve cell parameters of BP and GRNN neural network model, the export nerve cell parameters is coagulant dosing addition. According to collecting original data, filtering invalid data and using method of neural network model, makes the analysis of simulation, and we validated the extension capacity of the model. For the sake of comparing the superior and inferior of the model, using method of least squares, we set up a multiple linear regression mathematical model with the parameters of raw water qualities and the addition quantity of coagulant, found a function with the addition quantity of coagulant and the raw water turbidity. According to the chart of contrast, correlation, mean-root-square error(σ) and so on, the neural network model is distinct superior to the traditional mathematical model. We empolder a applied interface of the control model analysis by using Visual Basic 6.0 in the interest of using convenience.At last, we combine the control actuality of the water supply factory, by coordinating and optimizing of the system model, the scheme of feed-forward control of coagulant dosage how to practice in engineering is put forward.