电力系统网架重构的变时段建模和协调优化研究

发布时间：2018-08-16 11:15

【摘要】：近年来,电力系统大规模互联的趋势日益明显,复杂程度也在不断提高,不仅增加了故障发生的概率,而且影响范围极易扩大,甚至引发恶性连锁反应,酿成大面积停电事故。世界范围内的多次事故表明,大停电是现代电力系统难以完全避免的严重威胁。在这种背景下,作为电力系统安全防御的重要措施之一,可靠的系统恢复方案对于减轻恢复控制负担、加快系统恢复速度、减小大停电事故带来的经济损失及社会影响具有重要意义。因此,深入研究系统恢复过程中的优化控制技术是十分必要的。本文在全面学习和借鉴已有研究工作的基础上,对电力系统恢复过程中的网架重构优化技术进行了重点研究,论文工作主要包括以下几方面:(1)机组启动性能动态评估与排序优化。大停电后电厂热力系统的工作效率会逐步衰减,直接影响机组的启动时间和并网速度。分析此情况对各时段机组启动顺序产生的影响,对电厂关键设备热力系统温度下降趋势进行评估,采用模糊理论对机组各状态进行处理,确定各机组启动效率损失指标,建立机组启动动态优化模型,并应用多种群遗传算法进行求解,重点分析机组恢复顺序差异对各时段长度及总体恢复进程的影响。(2)网架重构变时段建模与求解方法。在“分时段建模,多时段协调优化”的总体策略下,通过研究机组启动特性和设备恢复要求,对各时段恢复时间进行模糊评价,结合恢复过程的串并行机制,确定恢复时段的划分原则和方法,提出网架重构的变时段建模框架。结合近似算法中的限制策略,研究适用于变时段模型的优化方法,以提高求解效率。此方法克服了分时段决策极易陷入局部最优的缺点,为寻求全局最优解提供了思路。(3)网架重构和负荷恢复的协调优化。在确定每一时段的待启动机组之后,通过综合考虑恢复代价、重要性、功率平衡及后效性影响等多方面因素,结合负荷投入特点,从机组、节点和线路三方面提取出反映系统运行状况的指标,建立网架重构和负荷恢复协调优化模型。求解网架重构方案的非支配解集,采用多属性决策方法进行最优解筛选。该方法可以有效改善依据效益型指标制订方案导致系统安全性降低的问题,特别是在系统相对薄弱时期,有计划地恢复一定量的负荷对于电网十分必要,是后续恢复工作顺利进行的重要保证。(4)考虑资源约束的过渡恢复方案优化。针对目前常用的黑启动电源和厂站应急电源存在持续出力受限的情况,提出一种考虑资源约束的受限电源替代方案优化方法。利用受限机组启动电气距离较近、性能较好的非受限机组,在最短时间内实现并网发电,以此代替受限电源进行后续恢复。合理选取优化目标,综合考虑约束条件,建立系统恢复过渡模型,并采用相应算法对模型进行优化求解。应用非受限机组代替受限电源进行后续恢复工作,可以为电网提供更加持久的发电量支持,确保后续恢复方案顺利实施,具有较强的实际意义。
[Abstract]:In recent years, the trend of large-scale interconnection of power systems has become increasingly evident, and the complexity of power systems is also increasing. It not only increases the probability of fault occurrence, but also extends the scope of impact, even triggers a vicious chain reaction, resulting in large-scale blackouts. In this context, as one of the important measures of power system security defense, reliable system restoration scheme is of great significance for reducing the burden of restoration control, speeding up the system recovery and reducing the economic loss and social impact caused by blackouts. Based on the comprehensive study and reference of the existing research work, this paper focuses on the grid reconfiguration optimization technology in the process of power system restoration. The main work of this paper includes the following aspects: (1) dynamic evaluation and sequencing optimization of unit startup performance. Efficiency will gradually decay, directly affecting the start-up time and grid-connected speed of generating units. This paper analyzes the influence of this situation on the start-up sequence of generating units in various periods, evaluates the temperature decline trend of the thermodynamic system of the key equipment in the power plant, adopts fuzzy theory to deal with each state of the generating unit, determines the loss index of starting efficiency of each unit, and establishes the generating unit. Start-up dynamic optimization model and apply multi-population genetic algorithm to solve the problem. Focus on the analysis of the impact of different restoration sequence on the length of each period and the overall recovery process. (2) Variable-time modeling and solution method for grid reconfiguration. In order to improve the efficiency of the solution, a variable-time modeling framework for grid reconfiguration is proposed based on the fuzzy evaluation of the recovery time of each time interval and the serial-parallel mechanism of the recovery process. It overcomes the disadvantage that time-division decision is easy to fall into local optimum and provides a way to seek global optimum. (3) Coordinated optimization of grid reconstruction and load recovery. The coordinated optimization model of grid reconfiguration and load recovery is established. The non-dominated solution set of grid reconfiguration schemes is solved and the optimal solution is screened by multi-attribute decision making method. It is necessary to restore a certain amount of load in a planned way, especially in the relatively weak period of the system. It is also an important guarantee for the smooth progress of the subsequent restoration work. (4) Optimizing the transition recovery scheme considering resource constraints. In the case of constraints, an optimization method for alternative schemes of constrained power supply considering resource constraints is proposed. The non-constrained units with short starting distance and good performance can be used to realize grid-connected power generation in the shortest time to replace the constrained power supply for subsequent restoration. The model is optimized by the corresponding algorithm. The application of unconstrained units instead of constrained power supply for subsequent restoration can provide more lasting power generation support for the grid and ensure the smooth implementation of subsequent restoration schemes.
【学位授予单位】：华北电力大学(北京)
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TM732

【参考文献】