针对郑州市东周水厂铁锰复合污染条件下的含砷地下原水,经中试和生产性试验研究,在无混凝的曝气/接触氧化过滤除砷技术的基础上,建立了投药量BP神经网络模型。该模型在以出水As<10μg/L为控制条件的前提下,进行不同原水水质条件下药剂投加量的模拟计算,并作为核心模拟算法模块,应用于水厂除砷自动控制加药系统中,完善了水厂当前的自控系统。 Aiming at the arsenic underground raw water under the iron-manganese complex pollution condition of Dongzhou Waterworks in Zhengzhou City, the pilot-scale and productive experimental study was carried out on the basis of no coagulation aeration / contact oxidation filter arsenic removal technology to establish the dosage of BP Neural Network Model. Under the premise of As <10μg / L effluent, the model was used to simulate the dosage of pharmaceuticals under different raw water quality conditions and used as the core simulation algorithm module in the automatic dosing system of arsenic removal in waterworks , Improve the water plant’s current automatic control system.