【作者】 谢洪波；

【导师】 钱壮志；

【作者基本信息】 长安大学 ， 地学信息工程， 2008， 博士

【摘要】 “人口、资源、环境”是当今世界面临的三大难题。人类的生存与发展在相当程度上依赖于水的获取和对水的控制。全世界约有15亿到20亿人口靠地下水生活。尽管地下水资源保护受到了各国政府、联合国、国际组织、官员、专家学者以及社会公众各方面的广泛关注，但是，全世界每年大约有2．5亿人新患上经水传染的疾病，其中大约1000万人死于非命。在中国，2／3城市地下水受到污染，300多个城市由于地下水污染造成供水紧张，有数千万人在饮用不符合饮水标准的地下水，严重危害身体健康。焦作市位于河南省西北部，总面积4071km~2，总人口345．5万，是一座以煤炭为基础发展起来的矿业城市。焦作市正面临由水资源短缺引起的“水资源危机”和由于地下水污染带来的“水质危机”。本文以焦作市地下水为研究对象，通过野外污染源调查、水文地质调查、室内分析，以211个地下水监测井数据为基础，分别用“F分值法”、“内梅罗指数法”、“模糊综合评价法”、“灰色聚类法”和“LVQ人工神经网络法”对焦作市地下水质量进行了综合评价；在需求分析、系统分析、系统设计的基础上开发出了“基于GIS的焦作市地下水污染预警系统”。为焦作市地下水资源规划、合理开发、污染治理提供了可靠的技术支持和管理平台。论文取得了如下主要成果：1．详细调查了焦作市地下水的主要污染源；2．查清了焦作市地下水15种超标因子的空间分布规律；3．对现行地下水质量标准进行了合理细化，给出了“标准值”；4．针对pH值在地下水质量评价中的“双序列”现象，论文给出了pH值“单序列化”的方法；5．用5种方法对焦作市地下水进行了综合评价，认为总体上焦作市岩溶水质量较好，孔隙水污染较严重，并给出了具体的评价结果，绘制了焦作市孔隙水质量等级分区图；6．通过“F分值法”、“内梅罗指数法”、“模糊综合评价法”、“灰色聚类法”和“LVQ神经网络法”在焦作市地下水质量综合评价中的实际应用情况的分析和模型理论分析，得出在焦作市进行地下水质量综合评价中，应首选“LVQ人工神经网络法”，其次是“灰色聚类法”和“模糊综合评价法”；7．将地下水质量综合评价模型及污染预测模型与GIS技术相结合，建立了基于GIS的焦作市地下水污染预警系统，实现了抽象数据可视化、分析评价可视化、地下水质量发展演化可视化以及“可视化污染预警”，用此系统查清了地下水资源中主要污染因子的分布规律及综合污染状况，并绘制了相关图件。论文的创新之处：1．提出了地下水质量评价中pH值存在的的“双序列”问题，并针对该问题造成的在实际地下水质量综合评价中将pH值作为评价因子带来的困难，使多数评价者规避把pH值作为评价因子的现象，论文按照将处于“负序列”的pH值映射到相应级别的“正序列”上的思路，提出了pH值的“单序列化”方法；2．将“学习矢量量化(LVQ)”人工神经网络引入地下水质量综合评价中，并在焦作市地下水质量综合评价中具体应用，通过与其它评价方法评价结果的对比，指出LVQ网络在地下水质量综合评价方面表现出比BP网络更强的容错性、稳定性，也更容易实现，是焦作市地下水质量综合评价的首选方法；3．针对“地下水质量标准”(GB／T14848-93)中的标准限值过粗及“不同类别标准值相同时，从优不从劣”的规定造成的地下水质量综合评价中Ⅱ类水的缺失率42％，Ⅲ类地下水的缺失率26．3％，Ⅳ类水的缺失率15．8％的现象，细化了地下水分类指标标准限值。针对“地下水质量标准”中规定的是“地下水分类指标标准限值”而常用地下水质量综合评价模型中使用的是“标准值”造成的使用不便和混乱，给出了“地下水质量分类标准值”。更多还原

【Abstract】 The population, the resources, the environment are three big difficult problems which theworld faces now. Humanity’s survival and the development fundamentally rely on the watergain and the water control. In the world, there are 1,500,000,000 to 2,000,000,000 populationsdepend on the ground water to make their life. Although the ground water resourcesconservation has received various aspects various levels widespread attention from thecountries’ government, the United Nations, the international organization, the official, theexperts, as well as the public. Every year, the world probably has 250,000,000 people to suffernewly from disease which disseminates by water, in which about 10,000,000 people died. InChina, 2/3 city ground water is polluted generally, more than 300 cities the water supply to betense cause by the ground water polluted, millions people drink the ground water which doesnot conform to the potable water standard, severely impair their health.Jiaozuo is located northeast Henan Province, the total area 4071 square kilometers, thetotal population 3,455,000, is a mining industry city which develops by the coal foundation. Itis facing the water resources crisis which causes by the water resources short and waterquality crisis which causes by the underground water pollution.Taking the Jiaozuo ground water as the research object, through open country pollutionsource investigating and hydrology geological investigating, taking the data of 211 groundwater monitor wells as the foundation, this paper comprehensively appraised the Jiaozuoground water quality using the F Value Method, the Nemerow Index Method, the FuzzyComprehensive Method, the Gray Cluster Method and the LVQ Neural Network Method. Onthe basis of the demand analysis, the system analysis and the system design, a ground waterpollution warning-forecast system based on GIS in Jiaozuo is developed. It provided thereliable technical support and the management platform for the Jiaozuo ground waterresources plan, the reasonable development and pollution government.The main results and conclusions1. Closing investigated the source of ground water pollution in Jiaozuo; 2. Investigated thoroughly 15 factors: spatial distribution rule which exceeding theallowed of the ground water in Jiaozuo;3. Has carded on the reasonable refinement to the present ground water quality standard,and has given the normal value;4. In view of the double sequences which the pH value performance in the ground waterquality appraisal, this paper gave a method by which the pH value can be changed to thesingle sequence;5. Has appraised on the quality synthetic of the Jiaozuo ground water with 5 methods,thought as a whole that the Jiaozuo karst water quality is good, the pore water pollution ispolluted seriously, and has given the concrete appraisal: result, has drawn up the Jiaozuo porewater quality rank district chart;6. Through analysis on the result of the F Value method, the Nemerow Index method, thefuzzy synthesis method, the gray cluster method and the LVQ nerve network method used inthe Jiaozuo ground water quality synthetic appraise and analysis on the model theory, thepaper got a conclusion, in the ground water quality synthetic appraise in Jiaozuo, the LVQneural network method should be the first choice, next be the gray cluster method and thefuzzy synthesis method;7. Through coupling the ground water quality synthetic appraise model and the pollutionforecast: model to the GIS technology, developed the ground water pollution warning-forecastsystem based on GIS in Jiaozuo, realized the abstract data visualization, the analysis appraisalvisualization, the ground water quality development appraise visualization and thevisualization pollution warning-forecast. Using this system, investigated thoroughly the spatildistributed rule of the main pollution factors and the ground water pollution condition inJiaozuo, and has drawn up the relation maps;The innovations1. Proposed the double sequence problem that the pH value performance in ground waterquality appraising, aims at this problem make the difficulty which bring the pH value as theappraisal factor in the actual ground water quality synthetic appraise, and causes the most valuers to dodge taking the pH value as the appraisal factor, according to the mentality whichwill map the negative sequence of the pH value to the positive sequence in the correspondingrank, the paper proposed the pH value list serializing method;2. Introduced the Learning Vector Quantization Artificial Neural Network (LVQ-ANN)to the ground water quality synthetic appraising, and used it in Jiaozuo ground water qualitysynthetic appraise, after contrasting the appraise result to the other method, pointed out thatthe LVQ-ANN compared to BP network, it performance stronger fault tolerance, the morestability, also easier to realize, and it is the first choice method in the ground water qualitysynthetic appraise in Jiaozuo;3. In view of the stipulation that the thick standard limiting value in the ground waterquality standard, and the stipulation that the different category normal value same liberallynot from poor, and these Stipulations making theⅡwater flaw rate 42%, theⅢwater flawrate 26.3%, theⅣwater flaw rate 15.8%, the paper refined the ground water classification.In view of it is the standard limiting value rather than the normal value in the ground waterquality classification standard, and the stipulation causing inconvenient and chaotic inapplication, this paper putout the ground water quality classification normal value.更多还原