【作者】 殷峻暹；

【导师】 王本德；

【作者基本信息】 大连理工大学 ， 水文学及水资源， 2002， 博士

【摘要】 随着我国城市化进程的加快，城市防洪与供水必然成为二十一世纪我国水资源战略调整的重点。本文首先概括总结了国内外城市防洪与供水领域的研究现状，分析了国内城市防洪与供水领域中存在的问题。鉴于城市防洪与供水领域研究内容的广泛性和复杂性，本文针对城市防洪与供水普遍存在的模糊和风险问题开展研究，主要研究内容和研究成果概括如下： (1)针对多目标决策问题，各目标权重合理与否，直接决定评价结果的合理与否。本文以陈守煜建立的模糊优选理论模型为基础，通过确定指标相关系数矩阵来反映指标之间的相关性，从而解决了指标权重的相关性分析问题。在此基础上探讨了基于指标相关性分析的多目标决策模糊优选模型，并为第三章城市防洪标准的确定提供了理论依据，本文也给出了指标相关系数矩阵的推倒过程和求解方法。 (2)城市防洪标准是城市防洪规划、设计、施工和运行管理的一项重要依据，合理确定城市防洪标准有重大现实意义。城市防洪标准的优选指标涉及广泛，包括政治、经济、社会和环境等多方面指标，各个指标之间又具有不可共度性和相关性，因此，本文提出应用基于指标相关性分析的模糊优选模型确定城市防洪标准，并与其它方法进行比较与验证，实例分析表明基于指标相关性分析的模糊优选模型在确定城市防洪标准中具有重要应用价值与现实意义。 (3)针对目前我国城市化地区传统水文分析计算方法存在的不足和由于城市建设日新月异导致城市下垫面的水文特性情况变化比较频繁的特点，本文探讨了利用满宁公式和水量平衡方程建立城市雨洪预报模型的基本方程组，并结合地理信息系统解决城市雨洪预报模型参数的确定问题，以提高模型的预报水平。算例分析表明该预报模型具有较高的预报精度。 (4)针对城市供水领域中供水水源地来水量预测的重要性和复杂性，将模糊数学知识表达性好的特点与人工神经网络模型学习能力强的特点结合起来，探讨了模糊模式识别神经网络模型的预报方法，并建立了模型学习的遗传算法和权重调整即混合算法。预报算例对模型及算法的检验结果说明，模型及其训练算法是可行的，在一定程度上克服了神经网络模型训练中普遍存在的“局部极小点”问题。 (5)随着社会、经济、人口的发展，城市水资源供需矛盾日益加剧，在合理考虑城防水库汛期限制水位动态控制风险的基础上，提高蓄水概率和供水能力就具有重大现实意义。本文从汛期限制水位动态控制研究出发，在王本德教授建立的水库预蓄效益与风险控制模型的基础上，通过科学合理的对效益指标和风险指标进行分析，探讨了在决策后果优选与决策目标权重确定中保持决策者风险偏好一致的方法，并结合碧流河水库的实测资料进行检验，检验结果说明模型是可行的。 最后对全文作了总结，并对有待进一步研究的问题进行了展望。更多还原

【Abstract】 During the expanding of China urban Proceeding, the city flood control and water supply would become the emphases of water resource stratagem regulating in china in 21 century. First, this paper presents the actuality of research field in global city flood control and that of China, and analyses the questions of research field in china city flood control. However, due to the wide-range and complex of its research contents, this paper will mainly deal with the risk and fuzziness of city flood control system. The major contents and research results are as follows:(1) With the problem of the multi-index decision-making, the rationality of evaluating result is decided by the rationality of index weight. Based on fuzzy decision-making analysis theory of professor Chen Shou-yu, the problem of the Index weight relativity-analysis is resolved by confirming the matrix of Index weight relativity-coefficient. Based on this, fuzzy decision-making analysis theory based on index weight relativity-analysis is developed and the means of computing the matrix of Index weight relativity-coefficient is provided.(2) The programming, design, construction and runnable management of city flood control according as the standard of city flood control, so, we can points out that the reasonable standard of city flood control is provided with important realistic signification. Through the analysis of the instance and compare with layer analysis method, we can points out that fuzzy decision-making analysis theory is provided with important application value and realistic signification in making standard of city flood control.(3) An urban storm water model has been established after analysis of the change of hydrological characteristics in urban areas and the deficiencies of traditional methods for hydrological calculation. Based on man-ning formula and equation of water volume balance the model can be used for hydrological and hydraulic analysis and calculation for urbanized areas. Based on GIS, urban storm water model parameter is calculated. The model has been used for an experimental area and satisfactory results have been obtained.(4) Considering the importance and complexity of prediction of hydraulic elements in the field of city water supply, this article applies the neural network with fuzzy pattern recognition model for prediction by combining fuzzy mathematics with neural network, and presents a new methods based on genetic algorithms and BP. With the case study of the examples, the new methods are reasonable, to some extent, this method solves the problem of so called local minimum of neural network model-learning.(5) Through the development of community, economy and population, the contradiction between supply and requirement of water resource is increased. So, the reasonable making level of risk in dynamic-control process of water level of flooding-period is provided with important realistic signification. Through the analysis of Bi Liu-He Teservoir, we can points out that fuzzy risk decision-making analysis theory is provided with important application value and realistic signification in making water line of flooding-period.Finally, a summary is given and some problems to be further studied are discussed.更多还原