模糊神经网络在臭氧生物活性炭系统中的应用  

APPLICATION OF FUZZY ARTIFICIAL NEURAL NETWORK IN OZONE BAC SYSTEM

作　　者：段蕾[1] 李伟光[1] 吕炳南[1] 丁驰[1] 韩宏大[2] 

机构地区：[1]哈尔滨工业大学市政环境工程学院,黑龙江哈尔滨150090 [2]天津自来水集团,天津300050

出　　处：《水处理技术》2007年第8期65-67,共3页Technology of Water Treatment

基　　金：国家高技术研究发展计划(863)项目:北方地区安全饮用水保障技术(2002AA601140)

摘　　要：臭氧生物活性炭系统具有非线性和非精确性的特征,为了精确拟合该系统各影响因素之间的内在规律,本文建立了基于BP算法的模糊神经网络定量分析模型。运用该模型,在给定工艺参数条件下精确预测了臭氧生物活性炭出水CODMn,拟合优度达到0.95106,实现了系统的有效预测、增强其可控性;并分析了臭氧投加量与CODMn去除率之间的数量关系,有创见地提出了不同温度条件下的最优臭氧投加量,研究结果表明:该系统中,在保证最低去除率为40%的前提下,温度低于10℃、介于10～23℃以及高于23℃时,臭氧投加量宜分别采用1、2.5～3、0.5～1mg/L,从而使臭氧投加量动态化,有效降低生产的运行成本。Ozone BAC system has an obvious nonlinear and uncertain characteristics. In order to accurately fit in the inherent rule between various affecting factors in this system, a quantitative fuzzy neural network model based on back propagation arithmetic is developed in this paper. Using this model, CODMn in effluents treated by ozone BAC system was accurately estimated on the basis of given parameters. The estimate precision reaches to 0.95106, which makes it viable to achieve the effective predictions and promote the controllability of the system. Meanwhile, the quantitative relationship between CODMn removal efficiency and dosage of ozone was investigated, and optimization of ozone dosage at different experimental temperature was presented innovatively. The results indicated that, on the basis of insuring removal efficiency no less than 40 percent, it is optimum for dosage of ozone to select 1 mg/L, 2.5-3 mg/L or 0.5-1mg/L when temperature was less than 10℃, from 10 to 23℃ or more than 23℃. Accordingly dynamic dosage of ozone was achieved to decrease onerational costs effectively.

关 键 词：神经网络 模糊 臭氧生物活性炭系统 温度 通水倍数 最优臭氧投加量 CODMN 

分 类 号：TQ085.41[化学工程]