Neural networks based optimum coagulation dosing rate control applied to water purification system

By the analysis of coagulant dosing rate and its influencing factors, the neural network predicting theory was introduced into the water treatment technology creatively and a predicting model of coagulant dosing rate was established. The test results obtained indicate that this model is adaptive and its self-learning ability is effective. The prediction results´ accuracy can be markedly improved by the neural network´s online self-learning. The online predictive control of coagulant dosing rates can be achieved by using this model, and presents an effective way for the realization of optimal coagulant dosing rates.