基于BP神经网络-隶属度的河流黑臭评价研究  

Study on Black-Stink Assessment of Rivers Based on BP Neural Network-membership Grade

作　　者：徐明德[1] 阎正坤[1] 朱秋丽[1] 王帆 郑江[1] 

机构地区：[1]太原理工大学环境科学与工程学院,山西太原030024

出　　处：《数学的实践与认识》2012年第17期55-61,共7页Mathematics in Practice and Theory

基　　金：山西省自然科学基金(2006011058)

摘　　要：通过建立BP神经网络-隶属度模型,解决多致黑因子间复杂作用关系,以及网络输出收敛于两等级间的模糊隶属问题,实现对河流黑臭的评价.以太原城区内与汾河交汇的河流为例,根据各条河流的黑臭情况选择黑臭评价指标,拟定河流黑臭分级标准,建立河流黑臭评价模型.从模型计算结果可知,计算结果与各支流实际情况吻合较好,表明所建模型基本合理可行、黑臭水质分级标准拟定合理.The BP neural-membership appraisal model on black-odor status of river was established in the paper. The problems of multiple resulting black-odor factors＇ nonlinear interaction were resolved, and the vague attribution of network＇s outputs was classified by this model. The results of quantitative discrimination for general rivers could be achieved. The .article take Rivers intersect to FenRiver in Taiyuan city as examples, according to the water quality, choose index of black-odor, and formulated the standard of black-stink water quality classification, and setout the assessment model, the monitoring data of the river was used in the verification of the BP neural-membership model. The comparison between the the model＇s outputs and the realistic showed a good conformity, the＇ results have proved that the model and the water quality classification is basically feasible.

关 键 词：河流黑臭 BP神经网络 隶属度 太原城区内与汾河交汇的河流 

分 类 号：X824[环境科学与工程—环境工程]