基于BP神经网络的辽河源头区水质评价研究  

Water Quality Assessment in Source Area of Liao River Based on BP Neural Network

作　　者：卞建民[1] 胡昱欣[1] 李育松[1] 马永祥 边静 

机构地区：[1]吉林大学环境与资源学院,长春130021 [2]宁夏回族自治区煤田地质局,银川750004 [3]吉林省地质环境监测总站,长春130021

出　　处：《水土保持研究》2014年第1期147-151,共5页Research of Soil and Water Conservation

基　　金：国家水体污染控制与治理科技专项(2012ZX07202-009)

摘　　要：基于辽河源头区水环境问题日益突出的现状,该文开展了辽河源头区水环境质量的研究,旨在对区内的水体质量进行分析评价。通过资料收集与汇总,基于BP人工神经网络结构的思想和理论,利用研究区内13个控制断面的水质监测数据,建立了包括pH、溶解氧、氨氮、化学需氧量、五日生化需氧量、高锰酸盐指数的水质综合评价模型,并应用训练好的模型进行仿真运算及水质综合评价。结果显示,在选取的13个断面中,约76.92%的断面为Ⅴ类—劣Ⅴ类水质,仅有23.08%的断面水质级别在Ⅱ—Ⅲ类之间,研究区上游断面的水质状况较好,中下游的水质较差。将该结果与《环境公报》公布的主要断面水质结果进行对比,81.25%的评价结果相同,采用BP神经网络对研究区水质进行综合评价具有较强的适用性和可靠性。Based on the present situation that the water environmental issues in source area of Liao River have become increasingly prominent, the research for water environment quality was carried out to evaluate and analyze the regional water quality. Through data collection and summary, a comprehensive water quality e- valuation model was established based on the thought and theory of BP artificial neural network with inclu- ding pH, DO, ammonia nitrogen, COD, BODs, potassium permanganate index, and finished with water quality monitoring data of the 13 sections in the study area. After training well, it can just be applied in mod- el simulating operation and water quality comprehensive evaluation. The results have showed that in the se lected sections, approximately 76.92% of the sections are between class V and worse than class V, leaving only 23.08% of the sections whose water quality levels are between class Ⅱ and class Ⅲ in the selected 13 sections. The sections located in the upper reaches have a better water quality than that in the downstream. Compared the evaluation results with the results of main sections published in Environment Communique, 81.25% of the evaluation results are identical. It has strong applicability and reliability that BP neural net- work was used to comprehensively evaluate water quality in the study area.

关 键 词：辽河源头区 BP神经网络 网络训练 水质评价 

分 类 号：X524[环境科学与工程—环境工程] TP183[自动化与计算机技术—控制理论与控制工程]