【作者】 刘强；

【导师】 倪以信；

【作者基本信息】 华北电力大学（河北） ， 电力系统及其自动化， 2009， 博士

【摘要】 现代电力系统随着市场机制的引入而日趋复杂。日益增长的负荷需求与设备的相对老化之间的矛盾导致系统接近其运行稳定极限,易在级联故障条件下引发大停电事故。因此,如何在事故发生后快速、高效地恢复系统运行是电力系统研究领域的重要课题。本文在学习和借鉴已有研究工作的基础上,提出一套全新的现代电力系统恢复控制的优化体系,着重从模型建立、求解策略、协作优化、整体寻优、时步概念、以及交互策略等方面进行研究。提高恢复控制的智能水平,改善总体优化效果,促成由传统的经验性决策方式向规范化、标准化决策方式的转变。论文工作包括以下几方面:(1)按照优化对象的不同,将恢复控制划分为机组启动的优化、网络重构的优化和负荷恢复的优化,研究各自的优化目标和约束条件。提出电力系统恢复控制协调优化策略的总体框架,为恢复控制的优化提供一种新的思路。(2)提出机组启动的优化模型和求解策略。模型以优化时间段内机组发电量最大为全局优化目标,同时计及各种约束条件。通过采用基于数据包络分析模型的相对有效性评估的定性方法与利用回溯算法求解背包问题的定量方法相结合的方式进行求解。所提方法能较好地权衡对所求解问题的快速性和正确性要求。(3)提出最优送电路径的通用模型和相应的智能优化算法解算模式。以寻找最短加权送电路径为优化目标,将网络重构建模为一个寻找图的Steiner树问题。利用遗传算法进行求解,并对算法本身进行优化以提高求解速度、稳定性和寻优效率。(4)将传统意义的负荷恢复扩展到网络框架还未完全恢复的阶段,提出该阶段负荷重要性评价模型和负荷恢复优化模型,以及相应的求解策略。优化模型在数学上构成一个多属性决策问题,利用基于偏好数据包络分析(DEA)的评价方法对负荷进行相对重要性评估,采用基于价值密度的贪心算法确定优先恢复的负荷点。(5)研究机组启动、网络重构和负荷恢复之间的协作机制。以恢复控制优化策略的总体框架为基础,提出一种基于短期目标网的交互式恢复控制策略。利用离散化技术,将恢复控制这一整体问题分解为每一时步的子问题。在恢复过程的每一时步,以前一时步已经恢复的系统为基础,通过整体寻优,确定当前时步优先恢复的机组、线路和负荷点,由此构成综合指标最优或次最优的若干目标网络,逐步引导恢复控制过程。采用交互式技术,保证决策过程与实际恢复控制过程的结合。提出多智能体恢复控制优化决策体系的设计思想,研究各智能体之间的协作机制,提出恢复控制优化决策体系的协作模型。更多还原

【Abstract】 Modern power systems have become more and more complex due to the introduction of the deregulated and unbundled power market operational mechanism. Conflict between increased load demand and the aged equipments has also pushed power systems to be operated close to their limits, which is increasing the risk of large blackouts after the cascading failures. Accordingly, how to restore power systems rapidly and effectively after blackouts has become more and more important issue of interest in power engineering. Based on the existing literature survey investigated thoroughly to power system restoration, a novel global intelligent optimization framework towards modern power system restorative control is proposed in this thesis. The key points of this research involve the construction of optimization model, the solution strategy, the behavior of collaborative optimization, the overall optimization target, the concept of time step and the interactive strategy, etc. This research can enhance the restorative control level and improve the whole optimization results, on the other hand, can implement the transformation of decision-making mode of power system restoration control from the traditional experience-driven strategy to the standardization and normalization strategy. The main achievements are as follows:1. With respect to the specific optimization objective, the whole power system restoration problem can be supposed to be divided into the optimal units start-up, the optimal network reconfiguration, and the optimal load recovery. The corresponding optimization models and the constraint conditions were studied in this thesis. The framework of the optimal strategy for the power system restoration was proposed, which can provide a new way to implement the optimization of the power system restoration.2. An optimal strategy involving the corresponding model and approach to units start-up were presented. The objective of the model is to maximize the total power generation capability (MWh) over a restoration period whilst being subjected to the specific constraints. The proposed model is a typical multi-constraint knapsack problem from the mathematical point of view. The combination of the data envelopment analysis (DEA) method and by solving the knapsack problem was employed to determine the units to be cranked. The proposed method, to some extent, can make the trade off between the simulation accuracy and the computing efforts better.3. An optimal strategy involving the corresponding model and approach for the network reconfiguration were presented. The goal of the proposed model is to find the shortest weighted path for units start-up or load recovery in restoration duration whilst considering all kinds of constraints. The proposed model is considered as a typical Steiner tree problem from the mathematical point of view. The genetic algorithm method with characteristics of global optimization and handling the discrete variables easily and effectively was employed to solve this problem. Furthermore, the performance of genetic algorithm was optimized in order to improve calculation speed, stability and search efficiency further.4. A novel evaluation model for the importance of load and a load recovery optimization model as well as the solution strategy with respect to the period when the network is still reconfiguring, which are different from the traditional concept of the load recovery, were proposed in this thesis. From the mathematical point of view, the proposed model is a multi-attribute decision making problem. The DEA method with preference information and the value of density of pi/wi greedy algorithm were employed to determine the loads to be recovered.5. The interactive and collaborative mechanism during units start-up, network reconfiguration and load recovery was studied in this thesis. An interative strategy with respect to the restorative control based on the short-term-target-network was presented. The whole restorative control problem was discretized as a series of optimization problem in each time step. In each time step, according to the restored system in last time step, the corresponding optimal target networks with the units, networks and loads to be restored at current time step can be determined based on the global optimization technique. These restored short-term-target-networks can guide the restoration process step by step effectively. Furthermore, the decision making process can be combined with the actual restoration process by the interative techniques. A Multi-Agent System based optimal strategy framework towards restoration control was proposed. The corresponding collaborative mechanism among Agents was studied as well. Finally, the coordination model of the optimal strategy towards restorative control was presented.更多还原