基于BP神经网络的臭氧精确投加控制系统

自来水厂采用臭氧化工艺时臭氧投加量通常由生产经验判断确定,缺乏一定的准确性和时效性。根据浙江省T水厂150组实际运行样本数据,选用BP神经网络构建臭氧投加系统的前馈控制模型,能够在给定的工艺参数条件下较好地预测出水水质情况,也可根据进水水质情况和预期出水水质目标对所需的臭氧投加量进行预测。结果表明:基于BP神经网络的臭氧投加模型可以满足不同的水质变化,模拟精度较高,具有明显的优越性,对进一步提高供水安全性、节约制水成本具有重要的推动作用,也为臭氧-活性炭深度处理运行的自动化控制提出了新的理论思路。

Accurate Ozone Dosage Control System Based on BP Neural Network

The determination of ozone dosage is usually judged by experiences from the operation of ozonation process in water plant,therefore,causing lack of certain accuracy and timeliness.According to 150 sets of actual sample data from operation in T Water Treatment Plant(WTP)in Zhejiang Province,the BP neural network was used to construct the feedforword control model of ozone dosage,which was able to predict the effluent water quality through the given parameters as well as the required ozone dosage according to influent water quality and the expected quality target of effluent water.The results showed that the ozone dosage model based on BP neural network could meet water quality changes,with high simulative accuracy and obvious advantages.It would play an important role in further improving the safety of water supply and saving the cost of water production.It also put forward a new theoretical idea for the automatic control of ozone and activated carbon advanced integral treatment operation in water plant.