辽宁省大伙房水库水源地安全评价研究

发布时间：2018-01-10 12:22

本文关键词：辽宁省大伙房水库水源地安全评价研究　出处：《沈阳农业大学》2015年博士论文　论文类型：学位论文

【摘要】：伴随经济的飞速发展,水环境问题日益显现,恶化的水环境不仅影响居民生活质量,甚至制约经济发展。大伙房水库是辽宁省的重要饮用水源地,近年其水环境质量受到了一定程度的影响。在此背景下,研究其水环境质量现状、预测水质变化趋势,评价水源地安全状况则显得尤为重要,既可以有针对性的提出水环境保护对策,又可以对水源地安全建设提供理论依据。本研究的主要内容及结论如下：(1)在收集了大伙房水库水质监测数据后,建立了基于模糊综合评价法的水质评价模型、基于BP神经网络的水质评价模型、基于灰关联分析的评价模型及基于主成份分析的评价模型,以地表水环境质量标准(GB3838-2002)为依据,水质等级共分为五级。四种方法的评价结果对比分析后得出,BP神经网络评价法和主成分分析法的评价结果基本相同,模糊综合评价法评价结果与他们接近,但评价结果略微偏低,而灰关联评价法的结果与另三种方法相差较大。评价结果表明：浑河是入库河流中污染最为严重的,不同评价方法评价结果分别为Ⅲ级或Ⅳ级,社河、苏子河也存在一定程度的污染,污染源主要为氮素,由化肥、农药、生活污水等非点源污染造成,库区内水质良好,但总氮含量亦较高。(2)应用BP神经网络预测法及灰色理论预测方法建立两个不同的水质预测模型。BP神经网络预测法选择了除抚顺取水口断面以外其它六个监测断面的2005年至2011年的水质污染指标数据作为输入样本建立网络,抚顺取水口断面的水质数据作为输出层神经元,模拟训练网络,2012年的数据用于验证网络。网络训练后,精度较好,相对误差在0.775%～9.07%之间。灰色预测理论是将原始无规律数据通过一定的处理变为较有规律的时间数据序列后,以微分形式建立出表达模型,各监测断面水质预测模型精度均达到90%以上。(3)界定了水源地安全的概念。根据可持续发展理论,以大伙房水库水源地的安全构成为基础,构建了目标层、准则层和指标层三个层次的指标体系。目标层为大伙房水库水源地安全等级；准则层为水质安全、水量安全、生态安全、工程安全、监控体系及管理体系六个方面；指标层为准则层向下的20个具体指标。参考国内外相关学者的研究,结合大伙房水库水源地自身环境状况制定各指标分级标准,并与水源地安全等级相对应的将每个指标分为Ⅰ(很高)、Ⅱ(高)、Ⅲ(中)、Ⅳ(低)、V(很低)五个等级。(4)通过层次分析法确定指标权重,各指标安全等级以隶属度表示,隶属度以定性及定量相结合的方式确定。最后通过综合评价法求得各准则层及水源地总体的安全等级,工程安全为“极高”水平,水质安全为“高”水平,水量安全、管理体系为“中”水平,生态安全、监控系统则为“低”水平,水源地总体安全等级在“高”与“中”之间,属于中等偏上安全水平。
[Abstract]:With the rapid development of economy, water environment problems have become increasingly apparent, the deterioration of water environment not only affects the quality of life of residents, even restricting economic development. The Dahuofang reservoir is an important drinking water source in Liaoning Province in recent years, the quality of the water environment to a certain extent. Under this background, the research status of water environmental quality prediction the change trend of water quality evaluation, water security situation is particularly important, can put forward the Countermeasures of water environment protection, and can provide a theoretical basis for the safety of water source construction. The main contents and conclusions of this study are as follows: (1) in the collection of water quality monitoring data of Dahuofang reservoir, water quality evaluation is established the method of fuzzy comprehensive evaluation model based on the water quality evaluation model based on BP neural network, the evaluation model of grey relational analysis and evaluation model based on principal component analysis based on the surface water ring Environmental quality standard (GB3838-2002) on the basis of water quality is divided into five levels. The four methods of evaluation results of the comparative analysis shows that the BP neural network evaluation method and principal component analysis of the evaluation results are basically the same, the fuzzy comprehensive evaluation method and the results they close, but the evaluation results were slightly lower, and the grey relation the evaluation results and the other three methods are different. The evaluation results show that the Hunhe River is in the most serious pollution in rivers, different evaluation results were grade III or IV, agency Suzihe River, there is a certain degree of pollution, the main pollution sources for nitrogen, by chemical fertilizers, pesticides, life sewage and other non-point source pollution, water quality in the reservoir area is good, but the total nitrogen content was also higher. (2) forecasting method and grey prediction methods two different water quality prediction model of.BP neural network prediction method based on BP neural network The water pollution index data in Fushun water intake section other than the six monitoring sections from 2005 to 2011 as input samples to establish a network of Fushun, the water intake section of the water quality data as output layer neuron network, simulation training, 2012 data is used to verify the network. After the network training, better accuracy, the relative error in 0.775% ~ 9.07% the grey prediction theory. The original data is irregular through certain processing time sequence data into a regular, to establish a differential expression model, each monitoring section water quality prediction model precision reaches more than 90%. (3) defines the concept of water security. According to the theory of sustainable development, in order to water Dahuofang reservoir safety as the basis to construct the target layer, criterion layer and index layer three level index system. The target layer is the Dahuofang reservoir water source safety etc. As the criteria layer level; water quality safety, water safety, ecological safety, the safety of the project, the six aspects of monitoring and management system; the index layer as the criterion layer down 20 specific indicators. The related research reference at home and abroad, combined with the Dahuofang reservoir water environment status develop classification standard of each index, and safety the water level corresponding to each index was divided into 1 (high), II (high), III (in), IV (low), V (low) grade five. (4) to determine the index weight by AHP, the index of the security level to determine the membership, membership in both qualitative and quantitative methods. Finally the criterion layer and water security level are generally obtained through the comprehensive evaluation method for engineering safety, high quality and safety level to high level, water safety management system "in the" level of ecological security monitoring system for At the "low" level, the overall safety level of the water source area is between "high" and "middle", which belongs to the moderate level of safety.

【学位授予单位】：沈阳农业大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TU991.15

【相似文献】