【作者】 薛志英；

【导师】 李庚银；

【作者基本信息】 华北电力大学 ， 电力系统及其自动化， 2014， 博士

【摘要】 电力系统充裕性问题是贯穿电力系统规划与运行全过程的重要问题。考虑不确定性因素的影响,对电力系统充裕性进行评估与决策是保障电力系统安全、可靠运行的前提和手段。近年来,大规模风力发电的接入给原本可调、可控的电源出力带来了较大的不确定性,加上用电侧大量电动汽车充、放电的不确定性,使得电力系统运行调控的难度和风险都大大增加。传统的电力系统充裕性评估指标与方法已经难以适应目前电力系统短期运行领域电源与负荷的新特点。在这样的背景下,本论文以含风电电力系统短期充裕性评估与决策研究为题,开展的主要工作和取得的成果如下。针对由风电不确定性引起的备用需求决策问题,在分析不同时间尺度下风电功率的波动特性、风电功率短期预测误差的分布特点的基础上,借鉴风险理论,以风电功率预测误差为风电不确定性的表征,提出了基于时间的和基于功率的风电不确定性度量指标；引入精算学中的Buhlmann信度模型,将2种指标综合起来,形成信度风险度量指标,并将其用于估计由风电不确定性引起的系统运行备用需求。算例分析及指标的有效性检验表明,信度指标继承了基于时间的和基于功率的2种指标的优点,能够恰当地从历史数据信息和未来预测信息中获得某置信度下风电功率预测误差的风险信息,能够在不降低对实际损失覆盖程度的前提下,合理地减少不必要的备用,对电力系统经济运行起到一定的促进作用。针对计及电源与负荷不确定性风险的发电充裕性评估问题,在总结电力系统充裕性的基本知识以及常用的发电系统充裕性评估方法和评估指标的基础上,从短期运行的角度,以系统可用发电容量缺额为充裕性表征函数,对运行失负荷概率(operational loss of load probability, OLOLP)、运行失负荷期望(operational expected load not served, OELNS)指标进行了重新描述,提出了一个新的充裕性度量指标：延迟失负荷概率(buffered loss of load probability, BLOLP);分别分析了充裕性评估指标与负荷需求、负荷预测误差、常规机组容量、常规机组停运率、风电机组停运率和风电功率预测误差等影响因素之间的关系。对指标的定性分析表明,所提出的充裕性评估指标以一种概率性方法将系统运行划分为充裕、警戒和不充裕3种状态,有利于提高系统运行的可靠性。仿真算例表明,所提出的概率评估指标隐含着充裕性函数概率分布的尾部信息,反映的充裕性更全面。针对不确定性对机组组合的影响,基于充裕性约束提出了一种含风电电力系统机组组合模型。在传统的基于机组投运风险(unit commitment risk, UCR)和基于失负荷概率(loss of load probability, LOLP)、基于失负荷期望(expected load not served, ELNS)的充裕性约束的基础上,建立了基于所提出的OLOLP和BLOLP的机组组合模型,研究了其充裕性约束的求解方法。与其它机组组合模型算例的对比表明,本文所提出的基于BLOLP的机组组合模型具有运算速度快、结果可靠性高的优势,还能够给出在不同充裕性水平要求下系统机组的启停方案,实现机组有功出力和提供备用的协调优化,给出运行方案相对应的系统充裕性量化值,为调度工作提供直观参考。以电动汽车作为用电侧主要不确定性的代表,研究了含电动汽车的配电系统短期充裕性动态决策问题。在明确电动汽车工作模式、充放电特性和参与调度的方式的基础上,考虑来自电源侧与负荷侧的双重不确定性因素,以系统运行备用容量为充裕性表征函数,定义了2个充裕性动态度量指标,建立了以最大充裕性为目标的多阶段决策模型。仿真算例结果表明,所提出的充裕性动态度量指标在保证研究期内系统整体充裕性的同时,还能根据负荷大小协调各时段的可调度负荷资源；所建立的多阶段决策模型同时优化购电方案与电动汽车充、放电方案,考虑了研究期内不同时段的优化变量之间的相互关联,体现了动态优化的本质；配电系统的运营部门可以通过适当的引导和管理,减少电动汽车充、放电对系统运行的负面影响,充分发挥其削峰填谷的正面作用。更多还原

【Abstract】 Power system adequacy is a foundmental question throughout system planning and operation. With consideration of uncertainties, power system adequacy evaluation and decision are basic work for system safety and reliability. The integration of large scale wind power generation in recent years has brought great uncertainty to troditional power system. At the same time, the charging and discharging behavior of electric vehicles at demand side are also uncertain. All the above bring lots of difficulties and risks to power system operation and control. Troditional adequacy evaluating indexes and methods can hardly meet the new request from system resource and load. Therefore, short term adequacy evaluation and decision making for power systems with wind power is studied in this dissertation. Main research work and innovative achievements obtained are as follows.To solve the reserve requirement caused by wind power uncertainty, the fluctuations of wind power output under different time scale and distribution characteristics of wind power short term froecast error are analyzed. With reference to the risk theory, wind power uncertainty evaluation indexes in time-dimension and power-dimension are proposed. Buhlmann credibility model is used to bring together the above indexes and establish the credibility indexes which are introduced into reserve requirement evaluation due to wind power uncertainty. Case study and validity test show that credibility indexes inherit the advantages of time-dimension indexes and power-dimension indexes and can obtain the risk information from both historical data and forecast data. The method reduces unreasonable operating reserve without loss of system security, which may be a kind of promotion to system economical operation.Aiming at generation system short term adequacy evaluation with uncertainties from resource and load, traditional reliability indexes operational loss of load probability (OLOLP) and operational expected load not served (OELNS) are redefined and a new index buffered loss of load probability (BLOLP) is proposed. In those definitions, shortage of available capacity (SOAC) is used as the representative function of system adequacy. The relationships between adequacy indexes and load requirement, load forecast error, capacity and forced outrage rate of traditional units, outrage rate of wind turbines and wind power forecast error are studied. Theoretical analysis shows that the system operation mode is divided into3parts:adequate, at risk and inadequate by the proposed index in a probabilistic manner. Simulation tests show that the proposed probability index is comprehensive than the other two because it contains adequacy function tail part information.To solve the unit commitment (UC) problem with adequacy constraint, constraints based on unit commitment risk (UCR), LOLP and ELNS are introduced and their solution methods are discussed. UC models with OLOLP and BLOLP adequacy constrains are established. Their solution methods are studied. By compare with the other UC models in case study, the model based on BLOLP adequacy constraint shows the advantages of fast speed and high reliability. It can solve the operation scheme under different adequacy request, optimize the active power and reserve capacity as well and show the adequacy estimation under certain operation scheme which may be a visualized reference to system operators.Taking plug-in hybrid electric vehicle (PHEV) as an example of demand side uncertainty, the operational adequacy decision making of distribution systems with PHEV is studied. Working mode, charging and discharging characteristics and dispatching manner of PHEV are introduced. Considering uncertainties from resource and load,2operational adequacy measuring indexes are proposed with system operating reserve as adequacy representative function. A multi-stage decision making model with the objective of maximum system adequacy is established. Results of simulation test show that the proposed indexes can coordinate the dispatchable load of different time sections with guarantee of adequacy of the whole study period. The multi-stage decision making model can optimize the purchase scheme of different energy and the dispatch scheme as well. Relationships between decision variables of different time sections are considered and that is the essence of dynamic optimization problem. The study shows a reference to the operation department of a distribution system to properly guide the PHEV charging and discharging behavior in order to reduce their negative effect and play a positive role in power system operation.更多还原