【作者】 陆健；

【导师】 陈卫；

【作者基本信息】 河海大学 ， 市政工程， 2007， 硕士

【摘要】 随着生产力水平的提高、城市规模的不断扩大以及社会节能、节水意识的增强，人们对城市供水系统的要求越来越高，供水系统的优化调度与安全运行研究面临着新的挑战和发展。 本文较深入地研究了城市供水系统优化调度的理论、建模方法及相关求解方法，在前人的相关研究基础上，分析了供水系统优化调度存在的问题；以南京市供水系统为研究对象，建立了与之相适应的用水量预测模型、管网分析模型和优化调度决策模型，以期为优化调度与安全运行提供一种切实可行的方法，提高供水企业的经济效益和社会效益，为本领域的理论研究和生产实际提供理论参考和技术指导。 本论文主要工作： (1)综合分析用水量预测方法和各种预测模型的优缺点，针对解释性预测法和时间序列预测法建模难的问题，研究建立基于BP神经网络的解释性预测模型和时间序列预测模型；针对两种预测方法固有的缺点，引入组合权系数优化理论，将两种方法进行组合对城市时用水量进行预测。最后通过对南京市时用水量的预测实例表明所建组合模型的实用性和可行性。 (2)综述国内外建立管网分析模型的常用方法，分析微观模型和宏观模型的特点及适用范围；结合我国的实际情况和南京市供水系统的特性，研究并建立了基于BP神经网络的供水管网分时段宏观模型；通过南京市供水管网的实例验证，说明所建管网分析模型的合理性和可行性。 (3)分析直接优化调度决策模型和两级优化调度决策模型的特点及适用范围，针对我国大中型城市供水系统普遍比较复杂、且多数采用非同步调速运行方式的实际情况，以南京市供水系统为例，建立两级优化调度决策数学模型。 (4)针对优化调度决策模型求解困难的问题，研究基于寻优能力较强的遗传算法对两个模型进行求解。通过优化前后的节能比较，阐述所建两级优化调度模型的适用性及整个优化调度模型的成功性。更多还原

【Abstract】 Along with the improvement of social productivity, the extension of city size and enhance of people’s consciousness to energy saving and water saving, standards of water supply system are becoming higher, so study on optimal and safe operation of water distribution system(WDS) is facing to new challenges and development.Interrelated theories, methods of model building and some equations’ solution are studied firstly in more detail in this paper, then some problems of WDS’ optimal operation at present are summarized according to pioneer studiers. Taking our county’s practice of water supply system into account, this study take water supply system of Nanjing city as subject, and build adaptive models of hourly water consumption forecasting, water distribution network and optimal operation respectively. This study will offer a feasible method for optimal and safe operation that is hopeful for maximizing water supply companies’ economic and social benefits. Furthermore, this paper can also provide reference for other study and guidance for practice in this field.The mainly work of this study are as following aspects:1. Synthetically analyze and compare the methods and models of water consumption forecasting firstly, to explore the merits and setbacks of them. For regression model and time series model have difficulties in equations formulation, developing them based on BP neural network are studied. Because of these two kinds of model’s inherent setbacks in prediction, a theory of optimizing combinatorial coefficients is introduced and a combined model based on them is developed. Furthermore, an example of forecasting hourly water consumption of Nanjing city is used to confirm the combined model’s practicability and feasibility.2. Summarize the common ways of developing equivalent model of water distribution network home and abroad, especially analysis the peculiarities of microcosmic equivalent model and macroscopic equivalent model. Based on our county’s practice of water supply system and characteristics of Nanjing’s, a time-division macroscopic equivalent model on BP network is built, and it’s rationality and feasibility is also confirmed.3. Based on given cases, analyze the peculiarities of the direct optimal operation model and two-stage optimal operation model of water supply systems. Since most of water supply systems in medium and large-scale cities of our county are complex with their pump stations mixed with fixed and variable speed pumps, a two-stage optimal operation model with equations formulation is established based on water supply system of Nanjing city.更多还原