【作者】 郭丽；

【导师】 袁永生；

【作者基本信息】 河海大学 ， 应用数学， 2007， 硕士

【摘要】 多水库联合调度，追求总体经济效益最大，是一个复杂的动态随机最优控制问题，不仅要解决水库群之间的流量调度问题，还要实现水电站厂内的最优经济运行。本文以上下游两个梯级水库的联合调度为背景，考虑个中的随机特性，及多个因素随时间变化的动态特性，以动态随机最优控制理论和方法，给出调度的最优决策，模型的概化更吻合工程实际，模型的推求严格按照数学理论。其主要内容有： 首先，详细介绍了本文研究问题的物理背景，即水库的基本运行方式和运行特性，说明多水库发电调度涉及到水库蓄水量平衡、发电流量和机组选择等问题，其中的多个参量是动态的，是一个复杂的动态随机最优控制问题。以随机过程刻画影响经济效益的因素(如来水、水库水位、库容等)，使得对工程问题的概化更符合实际。 其次，利用极大似然估计法估计径流量分布，并利用期望值模型建立了下游最佳水位模型、水库调度效益最大模型等。 再次，结合理论分析和水库调度的物理背景，确定目标函数和约束条件，求解非线性的最佳水位模型和最大发电效益模型。采用泛函极值的变分法确定下游最佳水位，以确定上下游水库的放水顺序和上游水库的最佳下泄流量；采用极值原理求解带不等式约束条件的各机组的流量分配，实现厂内经济运行的随机最优控制。 最后，选取一对梯级水库的调度问题，以随机最优控制理论和方法进行了实际求解。更多还原

【Abstract】 During the joint operation of many reservoirs, the pursuit of the biggest overall efficiency is a complex dynamic random optimal control problem, which is not only to solve the problem of flow scheduling between the reservoirs, but also to achieve optimal economic operation of the hydropower station. In this article, the background is of the two cascade reservoirs (one is upper and one is lower) joint operation. The random characteristics and the dynamic time-varying characteristic of many factors are considered, and the optimum scheduling decisions are given by dynamic random optimal control theory and method. The generalization of this model is more realistic, and the model calculation is strictly in accordance with mathematical theory. In view of these goals, the following contents are researched.First, the physical background of this article is introduced in detail, including the basic operational mode and operational characteristics of reservoirs, which is used to explain that the reservoirs’ joint operation involving flow scheduling balance, the selection of power flow and hydraulic units with many dynamic parameters is a complex dynamic random optimal control problem. Then the factors (runoff, reservoir water level, storage, etc.) will affect the efficiency, which are considered as the random process, making the generalization of the project more realistic.Second, runoff is estimated with maximum likelihood estimation method, and the best water level downstream model and the largest power generating efficiency of optimal reservoir scheduling model are established with the estimated model.Third, based on the theoretical analysis and physical background of reservoir scheduling, the objective function and constraints are determined, then the nonlinear best water level model and the nonlinear largest power generating efficiency model are solved. The variation of functional extremeness value method is used to solve the best water level downstream, which is used to determine the laying order of the two reservoirs and the optimal discharge of the upper reservoir. Also, the function extremes principle is used for solving the flow distribution of the hydraulic units with inequality constraints to achieve the random optimal control of economic operation of the inner plant.Finally, one pair of cascade reservoir scheduling problem selected as an example, the practical solution is introduced by random optimal control theory and method.更多还原