SBR法曝气量的模糊控制  

Fuzzy control of SBR aeration rate

作　　者：曾薇[1] 彭永臻[1] 张立东 王淑莹[3] 陈韬[1] 

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]大连市建筑科学研究设计院,大连116001 [3]北京工业大学环境与能源工程学院,北京100022

出　　处：《哈尔滨建筑大学学报》2002年第1期53-57,共5页Journal of Harbin University of Civil Engineering and Architecture

基　　金：国家自然科学基金资助项目(59878016);北京市自然科学基金资助项目(8002005)

摘　　要：介绍了SBR法处理啤酒废水的过程中,将在线检测的溶解氧浓度(DO)与人工神经网络系统相结合实现对曝气量的模糊控制的试验研究。结果表明,在同一曝气量下,反应初始阶段(8～10min)溶解氧浓度的大小不仅能够间接的反映进水有机物浓度(COD)的大小,而且可以预测整个反应过程溶解氧浓度高低,而溶解氧的高低还可以由曝气量大小控制。因此,人工神经网络系统可根据初始阶段DO的大小及变化情况预测进水有机物浓度和相应的曝气量,与此同时,以初始阶段的DO作为曝气量的模糊控制参数,实现对曝气量的模糊控制。Experiment results indicated that under the same aeration rate,the DO in the initial reac-tion stage(8～10min)was related to influent COD,i.e.the initial DO declined with the increation of influent COD.Consequently the artificial neural network could forecast the influent COD and the aeration rate suitable to DO and its change in the initial reaction.On the other hand,DO in the ini-tial reaction could reflect DO level in the reaction.If the initial DO was higher,DO level in the re-action would be higher than the proper range (>2～3mg /L).It could not reduce reaction time effective-ly,but would increase the operating cost.On the contrary,lower DO value resulted in longer reaction time and caused sludge bulking easily.So the on-line measurement of DO in the initial reaction could be used as a fuzzy control parameter of SBR aeration rate in orde to keep DO in the proper range (2mg /L).The combination of DO on-line measurement and artificial neural network to realize automatic control of SBR could save air for the same reaction time or reduce reaction time for the same quantity of air.

关 键 词：模糊控制 人工神经网络 曝气量 在线检测 SBR法 

分 类 号：X703[环境科学与工程—环境工程]