湖泊富营养化评价方法研究及其系统设计

发布时间：2018-07-31 18:15

【摘要】：经济发展和工业进步带动了社会发展,促使人类对淡水污染的防治问题越来越重视。采用合理的水质评价方法以及建立稳定可靠的水质评价体系是进行水资源污染监测和治理的重要手段。这就需要针对大面积的湖泊水域,进行水质等级监测方法的研究以及量化的分析软件系统的设计。本课题以此为研究背景,将水污染这个复杂的机理量化到一些与富营养化相关性比较大的水质参数中,通过经验关系选择总氮(TN)、总磷(TP)、叶绿素a(Chl_a)、悬浮物(ss)、高锰酸(CODmn)五个参数作为研究变量,依据国内地表水的评价标准和本课题的研究参数,制定本课题中湖泊富营养评价的参照标准,结合主流的水质评价算法:综合营养状态指数法(TLI)、单因子评价法、模糊综合评价法、BP神经网络算法,分别进行水体质量评估。水质评价体系建立的关键在于水质评价优化模型的构建。从择优的角度出发,对各算法评价结果与实际情况进行对比分析,确立BP算法作为水质评价优化算法建模的原始模型。利用遗传算法全局寻优的能力,弥补BP算法初始权值和阈值不确定带来的精度差异,构建优化的GA-BP模型对水体质量进行评价。与传统的BP神经网络算法仿真对比,显示混合型GA-BP算法在计算效率和评价精度上存在明显的优势,优化后的GA-BP模型评价精度从原来的0.1达到了0.05,且评价速度加快了4倍。为了构建完整的评价体系,本课题设计了水质评价系统,本系统功能包括:系统登陆、带有海量数据的多格式反演影像的读取、数据预处理、数据保存、数据运算、评价算法的稳定嵌入以及评价结果的可视化显示。选取IDL开发平台和IDL编程语言进行系统开发。设计系统的登陆模块与评价模块的流程与结构,详细介绍了水质评价系统的UI界面设计思路以及各个菜单栏布局,同时对系统进行了功能性和非功能性的软件测试。系统能够准确的对以龙泉湖为样本,针对课题中5种评价参数的反演影像浓度获取,进而做水质等级可视化划分,建立量化的水质评价软件系统,有效的对大面积的水域污染进行监测。其研究成果为成都市科技局的水质监测研发项目提供了可靠的保障。
[Abstract]:Economic development and industrial progress have driven social development, prompting human beings to pay more and more attention to the prevention and treatment of freshwater pollution. It is an important means to monitor and control water pollution by adopting reasonable water quality evaluation method and establishing a stable and reliable water quality evaluation system. Therefore, it is necessary to study the method of water quality grade monitoring and the design of quantitative analysis software system for the large area of lake waters. Based on this research background, this paper quantifies the complex mechanism of water pollution into some water quality parameters which have great correlation with eutrophication. Five parameters of total nitrogen (TN), total phosphorus (TP), chlorophyll a (Chl_a) and suspended substance (ss), high manganese acid (CODmn) were selected as the research variables through empirical relationship. The evaluation criteria of domestic surface water and the research parameters of this subject were used. The reference standard of lake eutrophication evaluation was established in this paper. Combined with the mainstream water quality evaluation algorithms: (TLI), single factor evaluation method and fuzzy comprehensive evaluation method BP neural network algorithm were used to evaluate water quality respectively. The key to the establishment of water quality evaluation system lies in the construction of water quality evaluation optimization model. From the point of view of optimal selection, the evaluation results of each algorithm are compared with the actual situation, and the BP algorithm is established as the original model of water quality evaluation optimization algorithm. The global optimization ability of genetic algorithm is used to make up for the accuracy difference caused by the uncertainty of initial weight and threshold of BP algorithm, and an optimized GA-BP model is constructed to evaluate the water quality. Compared with the traditional BP neural network algorithm, it shows that the hybrid GA-BP algorithm has obvious advantages in computational efficiency and evaluation accuracy. The evaluation accuracy of the optimized GA-BP model reaches 0.05 from the original 0.1, and the evaluation speed is 4 times faster. In order to construct a complete evaluation system, a water quality evaluation system is designed. The functions of the system include: system landing, reading of multi-format inversion image with massive data, data preprocessing, data saving, data operation, etc. Stable embedding of evaluation algorithm and visualization of evaluation results. IDL development platform and IDL programming language are selected to develop the system. The process and structure of the landing module and evaluation module of the system are designed. The UI interface design idea of the water quality assessment system and the layout of each menu bar are introduced in detail. At the same time, the functional and non-functional software tests are carried out on the system. The system can accurately take Longquan Lake as a sample to obtain the inversion image concentration of five evaluation parameters in the subject, and then make visual classification of water quality grade, and establish a quantitative water quality evaluation software system. Effective monitoring of large areas of water pollution. The research results provide a reliable guarantee for the R & D project of water quality monitoring in Chengdu Science and Technology Bureau.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：X824

【参考文献】