【作者】 张宏宇；

【导师】 印永华； 申洪；

【作者基本信息】 中国电力科学研究院 ， 电力系统及其自动化， 2013， 博士

【摘要】 能源是人类社会赖以生存和发展的重要物质基础。能源的可持续发展是经济社会可持续发展的前提条件。随着煤炭、石油、天然气等化石能源的日趋枯竭以及环境问题日益突出,能源的可持续性面临巨大挑战。大力发展可再生能源成为人类的必然选择。风能因资源丰富,分布广泛,清洁无污染等优点成为大力开发利用的可再生能源之一。进入21世纪以来,将风能转化为电能的风力发电技术持续快速发展,风电装机容量逐年提高,我国以及欧美许多国家均提出了宏大的风电发展规划目标。但风电出力具有明显的波动性和不确定性,风电大规模并网给电网运行带来诸多问题,主要有调峰、调频、调压、稳定及电能质量等问题。我国电源结构以燃煤火电为主,调峰电源较为匮乏。风电大规模并网加剧了匮乏程度,使调峰问题凸显。风电调峰问题解决的好坏直接关系到风电发展规划目标的实现。本文主要研究与调峰密切相关的风电规划问题,分别对含大规模风电的系统调峰充裕性的概率性指标和容量性指标、考虑风电调峰要求的系统随机生产模拟,以及考虑风电调峰要求的机组检修计划进行了研究。主要研究内容如下：(1)分析了电源结构对风电并网后系统调峰的影响,全面概括总结了目前风电并网后的系统调峰问题研究现状以及含风电的电力系统规划研究现状,主要包括含风电的可靠性、随机生产模拟以及机组检修计划的研究现状。(2)分析了风速的变化特性,总结提出了风速时间序列建模原则和方法。采取基于各月同一钟点对风速进行统计的方法,有效计及风速的季节特性和日时间尺度特性；采用时间序列回归模型考虑风速波动的自相关特性,提出了基于概率测度变换的风速时间序列建模方法,避免基于回归模型的风速模拟过程中出现负风速问题；基于概率测度变换方法,采用自回归滑动平均模型(ARMA)及向量自回归模型(VAR)模型,对单一风电场及多风电场风速进行时间序列建模。(3)提出了含大规模风电的电力系统调峰充裕性评估方法。基于序贯蒙特卡罗方法以日为单位考察调峰问题,提出了调峰充裕性概率性指标——调峰不足概率,将调峰充裕性指标处理为发电充裕时的条件概率,分析了风电预测精度对调峰不足概率指标的影响,对比了发电不足概率与调峰不足概率,探讨了风电并网容量、长期基荷运行机组、风电-负荷相关性、风电场风速相关性以及地区电网外送功率规模对系统调峰不足概率指标的影响。(4)提出了系统调峰充裕性容量性指标评估方法,包括基于简化牛顿迭代法的储能调整需求容量评估方法及外送调整需求容量评估方法,有效弥补了调峰不足概率指标操作性的不足,分析了风电不同时间尺度的相关性等因素对调峰充裕性容量指标的影响。(5)基于序贯蒙特卡罗方法建立了考虑风电调峰要求的电力系统随机生产模拟方法,提出了弃风功率损失期望指标,分析了日前风电预测精度、风电并网容量水平等因素对系统发电成本、常规机组利用小时数和弃风功率损失期望的影响。(6)基于Benders分解法建立了考虑风电调峰要求的电力系统机组检修安排模型,将考虑带有序贯性特点调峰要求的检修问题分解为满足峰荷能力和系统最小出力容量在各个检修时段分布是否合理的问题；修正了基于负荷持续曲线的传统随机生产模拟模型,用虚拟弃风量来表征各个周基荷容量的大小；以虚拟弃风量和系统检修成本最小为目标函数,以各个检修时段发电可靠性为约束条件,通过Benders方法进行分解协调,实现了对电力系统常规机组检修安排。(7)通过IEEE-RTS79系统的仿真算例验证了所提方法的可行性与有效性,并以风电规划并网规模较大的通辽电网为背景,建立工程算例,进一步验证论文提出的模型和方法的实际应用价值。更多还原

【Abstract】 Energy is important material basis for the survival and development of human society. Sustainable energy development is a prerequisite for sustainable economic and social development. With coal, oil, natural gas and other fossil energy resources depleting and environmental issues becoming increasingly prominent, the sustainability of energy is facing enormous challenges. It has become an inevitable choice of mankind to develop renewable energy vigorously. Wind energy becomes the focus of attention, for its advantages of abundance, widely distribution, and free pollution.Since the beginning of the21st century, wind power technology, which converts wind energy to electric power, has been developing rapidly. Installed capacity of wind power has increased annually, meanwhile, great wind power development and planing has been put forward in many countries including China. However, concerning its power fluctuation and uncertainty, large-scale wind power integration has great impacts upon the power system, such as peak-shaving, frequency regulating and voltage regulating issues, stability, power quality and so on. The peak-shaving power resources of China are originally scarce for the power generation structure is mainly based on coal-fired power. With large-scale wind power integration, the scarcity is becoming highlighted. The wind power peak-shaving issues are directly related to the realization of wind power development and planning objectives, especially in China.The research, in this dissertation, is mainly about wind power planning closely related with the peaking, including peak-shaving adequacy probability and capacity indices, probabilistic production simulation and generator maintenance scheduling which can consider the requirement of peak-shaving of wind power integrated systems. The main contents are as follows:(1) The impacts of the power source structure in China on peak-shaving capacity of wind power integrated system are analyzed. The research status of peak-shaving problems and planning problems of wind power integrated systems is summarized.(2) Based on the analysis of variation characteristics of wind, wind speed time series modeling principles are proposed. The actual sampling wind speed series and the time-series of regression analysis models are connected by probability measure transformation in order to avoid negative wind speed in regressive models. Based on the probability measure transformation methods, the models of autoregressive moving average (ARMA) and vector autoregressive (VAR) are used separately to build wind speed time-series models for single wind farm and multi-wind farms.(3) A peak-shaving adequacy evaluation method associated with large-scale wind power integrated systems is developed. It is based on sequential Monte-Carlo method, using reliability theory. Peak-shaving capacity insufficient probability index and its calculation method are proposed. Peak-shaving adequacy evaluation of IEEE-RTS system demonstrates that the proposed method is feasible and effective.(4) In order to increase the index operability, the peak-shaving capacity index evaluation method is developed. It is based on capacity credit evaluation theory. Two indexes of regulating requirement capacity of peak-valley difference and regulating requirement capacity for transmission are proposed by simplified Newton iteration method. Based on the IEEE-RTS system, the proposed method is verified to be effective.(5) A power system probabilistic product simulation method is built, which can consider the requirement of peak-shaving of wind power integrated systems. Based on the Monte-Carlo method, an expectation index of wind energy loss is proposed. In the probabilistic production method, day-ahead units arrangement is based on stealth enumeration method, and hydro power unit operation positions are determined by an iterative method of variable step size. Probabilistic product simulation of IEEE-RTS system demonstrates that the proposed method is feasible and effective, and the impacts of wind power prediction error and integration scale on system generation cost, unit utilization hours and loss of wind energy expectation are analyzed. (6) Based on Benders decomposition method, generator maintenance scheduling model is built, which can consider the requirement of peak-shaving of wind power integrated systems. The problem is decomposed to two parts:a deterministic multi-objective integer programming master problem and two sub-problems (loss of wind energy calculation and power generation reliability). The results on IEEE RTS system with wind farms demonstrate that the proposed method is feasible and effective.(7) An engineering example of Tongliao grid is built to verify the above methods further.更多还原