【作者】 戴明龙；

【导师】 袁鹏；

【作者基本信息】 四川大学 ， 水文学及水资源， 2004， 硕士

【摘要】 洪水灾害是我国和世界上最常见的自然灾害之一,给人类造成了极大的危害。在流域防洪管理中,水库群作为最重要的防洪工程之一,起着蓄洪、滞洪、错峰、削峰的作用,其联合优化调度具有重要意义。水库群优化调度,就是应用系统工程学的理论方法,将水库群作为一个完整系统,并使用优化方法来解决其调度的优化问题,为水库群防洪效益的进一步发挥创造了广阔的前景。本次论文研究方向为流域防洪实时联合调度优化研究,结合课题的需要,确定将大系统分解协调算法与GM(1,1)模型耦合应用。本次研究,将运用降雨径流模型来预报洪水来水过程,主要研究在已知来水的情况下(预报结果),大系统分解协调算法的简化与GM(1,1)模型的改进,并将它们在防洪调度决策中耦合应用。从整个流域防洪角度,要求流域内水库群的总余留防洪能力最大,以满足调度最优化的要求。本文在大系统优化原理的指导下,以系统学、灰色数学、模糊数学、水资源学等为技术手段,对流域整体的防洪能力进行优化组合。针对南盘江上游段自然地理、社会经济、生态环境及防洪能力的实际状况,对水库群及河道实施统一优化调度。大系统优化理论是20世纪70年代为求解大系统优化决策而发展起来的一门新兴学科,它将大系统分解成若干个子系统之后,形成了递阶结构形式,然后在总体目标和约束调节下进行协调(Coordination)。首先应用现有的优化方<WP=3>法实现各子系统的局部最优,然后再根据大系统的总目标,相互协调各子系统,以获得整个大系统的总体最优。大系统具有高维性、不确定性、规模庞大、结构复杂、功能综合、因素众多等特征,分解协调方法几乎贯穿于大系统理论的所有方面。水库群系统就是一个相互关联的复杂大系统。目前,大系统优化理论在水电站水库群系统优化调度领域渐受重视,在防洪水库群系统中的应用才刚刚起步。本文将从大系统分解协调算法的基本原理入手,研究如何简化改进,并将之合理简便地应用到流域实时防洪调度中来。灰色系统理论主要通过对系统中已知的“部分”信息进行生成、开发,提取有价值的信息,实现对系统运行特征的正确认识和有效控制。其研究思路是:对灰色量不是通过大样本找统计规律,而是用数据生成的方法,将杂乱无章的原始样本序列整理成规律性较强的生成数列,然后进行预测或其他用途。本文将从常规的GM(1,1)模型的精度入手,研究其改进方法,并利用改进后的GM(1,1)模型来预测河道未来的流量过程,为建立河库联合调度模型做准备。本文针对流域实时防洪调度的特点,对常规的大系统分解协调算法在防洪上的一般应用方法进行了探讨,指出了可以简化改进的地方,并将简化后的大系统分解协调算法与改进的GM(1,1)模型耦合应用,建立了河库联合调度模型。研究在确保各水库安全的前提下,水库群之间如何互相错峰、削峰,减轻下游河道的防洪压力,保证河道安全渡汛,并将该河库联合调度模型应用到南盘江上游段,开发了南盘江上段河库联合调度系统软件。系统采用前后台设计思路,前台采用面向对象的编程思想和标准的面向对象化语言Object Pascal,负责发布指令,命令后台进行计算、查询等操作,并使用图形、表格、报表三种方式来显示后台的执行结果;后台(服务器)采用Client/Server(客户/服务器)的结构方式,中央服务器用来存放数据库,该服务器可以被多台客户机访问。数据库应用的处理过程分布在客户机和服务器上,系统中的运算过程全部使用SQL语言编写成存储过程,将大部分功能纳入后台处理,在后台上完成。更多还原

【Abstract】 Flood disaster is the most ordinary natural disaster in the world, it does great harm to the humankind. In the management of drainage basin flood control, multi-reservoirs system is the most important control work, for it can store floodwater、deter flood、stagger flood-peak、lower flood-peak. It is very important for the multi-reservoirs system to schedule combined. Optimization-schedule of multi-reservoirs system is treating the multi-reservoirs system as a whole system, using theory and methods of system-engineering science and optimization method to solve its schedule problem, making a wide foreground for the control flood benefit of multi-reservoirs system.This paper is to study on the real-time optimization-schedule combined of drainage basin, according to the aim of the research, we decide to couple the Large Scale System Decomposition-Coordination Method (LSSDCM) and GM(1,1) Model. Firstly, use Rainfall-Water Model to forecast the incoming floodwater, according the known incoming floodwater, mainly study on how to better LSSDCM and GM(1,1) Model and how to use them in flood control schedule well.From the point of all drainage basin flood control, to reach the optimization schedule, it requires maximum of the total pre-kept ability. This paper uses the Large Scale System Optimization Theory, using system-theory、Gray- Mathematics、Fuzzy- Mathematics、water-resource as technical measure, and tries to find the optimization of the drain basin’s total flood control ability. According to the physical <WP=5>geography、socio-economy、environment and flood control ability and so on natural status and character, schedule the multi-reservoirs system and the river combined.Large Scale System Optimization Theory is a rising subject to settle Large Scale System (LSS) optimization decision-making in 1970s, it decomposes the LSS into some subsystems, forming successively structure, then coordinate them under the strict of total aim and restriction condition. Use exists optimization methods to achieve all subsystems part maximum, coordinate all subsystems to get the whole LSS optimization as a whole according to the LSS’s total aim. LSS has characters of multidimensional、indefinitely、large scale、complicated structure、integrated function、numerous factors, Decomposition-Coordination Method almost represent all aspects of the LSS theory .Multi-reservoirs system is an interactional and complex LSS. Now, LSS optimization theory has been attached great importance to in the area of hydro-power plants reservoirs system optimization, but it is still underway in the flood control reservoirs. This paper will start with the basic theory of LSSDCM, research how to simplify and better it, and apply it in the real time flood control of the drainage basin.Gray system theory achieves correct understand and effective control of the system’s function character, and pick up useful information, mainly through dealing with known “partial” information of the system. Its principle is not to look for statistical rule from many data, but to use measure of generate data, make wild original data become strong well-regulated generate data, then to forecast or other purpose. This paper will start with precision of the routine GM(1,1) model, work on how to better it, and use the bettered GM(1,1) model to forecast the incoming runoff process, make a prepare for set up a model of river and multi-reservoirs schedule combined.This paper aims at the specialty of drainage basin real time flood control, discusses normal use of routine LSSDCM in flood control, brings forward simplified <WP=6>and mended model, and couples it with bettered GM(1,1) model, makes a river and multi-reservoirs schedule combined model. The model works on how the reservoirs stagger flood-peak、lower flood-peak reciprocally to lighten the flood control pressure of the lower reaches and let the river get through flood reason, on the precondition of all reservoirs safe. This paper also uses the model in upper of the NanPanJiang River and gets an excellent resul更多还原