基于计算智能的流域污染排放优化模式研究  

Computational intelligence based optimization study on the watershed discharge of sewage

作　　者：王祎[1] 李静文[1] 邵雪[1] 田在兴[1] 郭亮[1] 姜继平[1] 王鹏[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

出　　处：《中国环境科学》2012年第1期173-180,共8页China Environmental Science

基　　金：国家自然科学基金资助项目(50821002);城市水资源与水环境国家重点实验室资助项目(2008TS06)

摘　　要：基于人工神经网络和遗传算法建立了流域排污削减控制的技术框架.通过对排污口和目标断面水质监测数据的模拟与优化提出最优的排污削减控制策略,从而使目标功能区达标,可以间接的实现环境容量总量控制.结合情景分析理论对松花江哈尔滨段的朱顺屯-东江桥(S1)与东江桥-大顶子山(S2)功能区进行了COD的排污优化研究.结果表明,S1区段执行Ⅲ类功能区标准时,何家沟与松北两个排污口平均削减率分别为23%和25%;执行Ⅱ类功能区标准时2个排污平均削减率分别为64%和42%.S2执行Ⅱ类功能区标准,太平,阿什河和呼兰河排污口全年平均削减率分别为18%、53%和25%.基于计算智能的削减控制模式实用可操作性强,可以科学、合理的对各个排污口源强进行优化,给出最优的污染排放策略.A general optimization framework about watershed discharge was established based on artificial neutral network and genetic algorithm.Through simulating and optimizing the sampling data from sewage outlets and monitoring sections,the optimal reducing discharge strategies could be obtained to reach the permitted standards.Then combined with scenario analysis theory,the COD optimization research was studied on Zhushuntun-Dongjiangqiao（S1） and Dongjiangqiao-Dadingzishan（S2） functional areas in Songhua river-Harbin region.The average COD cut rates of Hejiagou and Songbei outlets were 23% and 25% respectively when the S1 was under criterion III for functional areas,while they increased to 64% and 42% when S1 was under criterion II.And when the S2 was under criterion II,the cut rates of Taiping,Ashen River and Hulan River were 18%,53% and 25%,respectively.The computational intelligence based optimization method has high operability and practicality,and it also could get the optimal discharge strategy of each outlet scientifically and reasonably.

关 键 词：人工神经网络 遗传算法 流域管理 情景分析 松花江哈尔滨段 

分 类 号：X703.5[环境科学与工程—环境工程]