【作者】 杨贺钧；

【导师】 谢开贵；

【作者基本信息】 重庆大学 ， 电气工程， 2014， 博士

【摘要】 随着传统能源资源的日益枯竭及其污染气体的严重排放，人们正在积极寻找能够满足人类能源需求的可替代能源。风能作为一种清洁的可再生能源，是最主要的可替代能源之一。世界上许多国家已出台支持开发风能资源的相应政策，加之风力发电技术的日趋成熟，风能的开发规模也越来越大。风速的随机性和间歇性使得风力发电功率具有波动性和不可控性，大规模风电并网将对电网的安全性和可靠性产生影响。因此，研究风电系统可靠性评估模型、分析风电场对电力系统可靠性的综合影响等，对含风电场的电力系统设计、规划和运行等具有重要的理论指导意义和实际应用价值。随着风电场规模的扩大、数目的增多、集电系统拓扑的扩展以及风电机组密集度的增加，诸多因素，如：尾流效应、单风电场风的条件相依性、多风电场风的多重相关性、集电系统故障、风电场风机布局等，对风电场年发电量及可靠性的影响也更加凸显。基于此，本文结合国家自然科学基金“风电场可靠性评估的概率模型及算法”（51077135）和重庆市杰出青年基金“大规模风电系统可靠性评估及优化配置研究”（CSTC2010BA3006）项目，对计及前述因素的风速模型、风电场可靠性评估、风电场可靠性优化等问题进行了深入研究。风的大小和方向是风速尾流模型的关键影响因素，而尾流效应对风电场年发电量、可靠性等有很大影响。风速与风向并非完全独立的随机变量，它们之间蕴含有一定的条件相依性，因此，在建立风速和风向模型时需要计及它们之间的条件相依性，以提高风电场可靠性评估的精确性。基于此，本文基于风向经验概率分布函数，建立风向样本生成模型；基于尾流效应原理和风速Weibull模型以及风速与风向之间的条件相依性，建立计及尾流效应和风向分布的风速条件相依模型，进而生成具有条件相依性的风速与风向。实测风速和风向数据验证了该模型的有效性和实用性。多个风电场的风速、风向间常存在多重多维相关性关系。当多个风电场接入电网时，模拟具有多重多维相关性的风速和风向样本是进行可靠性评估的先决条件。本文根据Copula理论分别建立风电场风速、风向的正态Copula函数和Archimedean Copula函数，利用经验Copula函数对模型进行选择。采用风向经验分布和风速Weibull分布模型及聚类、变换和插值方法建立风速与风向的条件相依模型，进而建立具有多重多维相关性的风速和风向数据生成模型。通过实际观测风速和风向数据对所提方法和模型进行了验证，结果表明模拟数据能够保持与原始数据相同的相关性、条件相依性、数字特征和分布特性。风电场中除风电机组的故障影响风电场发电量外，其它因素，如：集电线路故障、开关设备的隔离和切换操作、集电系统拓扑结构等，均对风电场发电量有不同程度的影响，现有研究鲜有计及该类因素的综合影响评估风电系统可靠性。本文结合风电场集电系统的结构特点，提出集电系统分块概念和风电场集电系统可靠性的指标。计及集电系统元件的多重故障以及开关设备的故障隔离、切换及隔离范围等因素影响，利用分块枚举法和矩阵运算等建立风电场集电系统的多状态概率模型，进而得到风电场集电系统可靠性评估模型。分别对四种典型结构的风电场集电系统进行可靠性评估，算例验证了方法的可行性和有效性，结论表明该方法对风电场集电系统拓扑结构规划具有重要的参考价值。由于尾流效应影响和风电场地形限制等因素，风电场中不同风机布局对应的风电场发电量不尽相同。在风电场规划阶段，为提高风电场年发电量和系统充裕性需要根据区域内风的分布特性和尾流原理优化风电场最优风机布局。本文根据几何学理论确定风电机组尾流区域和计算风电机组之间的水平距离和垂直距离；计及区域风速与风向的联合分布特性，以最大化风电场发电量为目标函数，利用带收缩因子和变异因子的改进粒子群优化算法优化风电场最优风机布局。将该优化模型应用到实际区域的风电场布局规划中，并将优化后的风电场接入到可靠性测试系统分析风电场布局优化的可靠性效益。算例表明优化后的风电场能有效降低尾流效应的影响，增加风电场的发电量，改善系统的可靠性。风力发电与传统机组发电的主要差别是风电具有波动性和不可调度性。现有含风电场电力系统可靠性的研究能有效计及风电对电网的充裕性贡献，但鲜有能计及风电功率波动对电网可靠性的不利影响。电力系统在稳态运行中，系统既要实时满足负荷功率需求，还要在发电或负荷功率波动时做实时调整，基于电网稳态运行特性，本文计及电网的静态频率特性探讨了能够刻画风电功率波动因素的含风电场发电系统可靠性评估模型。采用负荷等值、机组状态模拟、系统静态频率调节等方法建立评估模型，并模拟系统稳态运行和负荷削减过程。将该模型应用于可靠性测试系统，算例表明该模型不仅能够反映风电功率对电力系统可靠性的功率贡献还能反映风电功率波动对电力系统可靠性的负面影响。更多还原

【Abstract】 With the increasing exhaustion of traditional energy resources and emission ofnon-environmentally friendly gases, people are actively exploring alternative energyresources which can be used to meet the human energy demand. As a clean andrenewable energy, wind energy is one of the most important alternative energy resources.In the world, many countries have made different policies for encouraging thedevelopment of wind energy resources. With the development of wind power generationtechniques, the scale of wind energy was increasing step by step in the last decade.Generally, wind is considered as a fluctuant and uncontrollable energy resouce due tothe stochastic and intermittent natures of wind speed. Integration of large scale windfarms into a power grid has a significant effect on the security and reliability of powergrid. Therefore, it is necessary for the experts to study the reliabliey assessment modelof wind power systems and to analyze the comprehensive impacts of wind farm on thereliability of power systems. These researches have an important theoretical andpractical significance in the designing, planning and operation of power systemcontaining wind farms.With the scale expanding of wind farm, increasing the number of wind farm,extending collector system topology, and increasing wind turbine generator (WTG)intensity in a wind farm, mutilple factors, such as wind speed wake, multi-dimensionalcorrelations, collector system failures, and wind farm layout, have significant impactson energy production of wind farms and reliability of power systems. The papertheoretically researches the impact of above factors on wind energy production andwind power system reliability under the support in part by the National Natural ScienceFoundation of China (No.51077135)“Probabilistic model and algorithm of wind farmreliability evaluation” and the Natural Science Foundation Project of CQ CSTC (No.CSTC2010BA3006)“Reliability evaluation and optimization models and algorithmsfor power system containing large scale wind energies”.Wind speed and wind direction are the critical influence factors of wake effect,which has a very significant effect on wind energy production of wind farm. Wind speedand wind direction are not completely independent variables; on the contrary, they havea certain conditional dependence relationship between them. Therefore, this conditionaldependence should be considered in the proposed wind speed model to improve theaccuracy. According to the above analysis, wind direction sampling model is built using the empirical distribution of wind direction. A conditionally dependent wind speedmodel is established using the wind wake theory, Weibull model of wind speeds, andconditional dependence between wind speeds and wind directions. The proposed modelcan be used to generate wind speed and wind direction data with a given conditionaldependence. The collected wind speed and wind direction data in the practical regionhave been used to verify the validity and applicability of the proposed model.Correlation and conditional dependence of wind speed and wind direction seriesalways appear in multiple wind farms, which should also be considered in reliabilityevaluation models of power systems containing wind farm. Wind speed correlationmodel and wind direction correlation model are built using normal copula andArchimedean copula in this paper. The empirical copula is used to select a propriatecorrelation model. Clustering method, transformation and interpolation techniques areused to build the conditional dependence model of wind speed and wind direction basedon empirical distribution and Weibull distribution models. Then, wind speed and winddirection generation model considering their correlation and conditional dependence isobtained. The proposed model is verified using the observed wind data, and resultsshow that simulated and observed wind data have an identical features in correlation,conditional dependence, statistics characteristics, and distribution characteristics. Inother words, the results have verified the validity and availability of the proposedmodel.Besides WTG failures in a wind farm, other factors, such as the failures ofelectrical collector system (ECS), disconnection and switching operation of switchdevices, and different collector system topologies, also have effects on energyproduction of a wind farm. There are relatively little studies on the comprehensiveimpact on the reliability of wind power system. Novel indices for describing thereliability of wind farm ECS are presented based on the topologies of wind farm in thispaper. The concept of section for a partitioning wind farm ECS is defined. A reliabilityevaluation model of wind farm ECS is built using the state enumeration algorithm andmatrix operations. The proposed model can not only considers the multi-failure of ECScomponents, including failures of cable feeder, WTG and wind turbine transformer, butalso consider the operational states of switching devices in failure, disconnection andswitching processes. Four different wind farm ECS topologies are implemented usingproposed model. Case studies on the reliability evaluation of wind farm ECS are used toverify the feasibility and validity of the proposed technique. The proposed technique isvery significant and valuable on wind farm collector system planning. Energy production of a wind farm is considerably affected by wake effect.Therefore, in wind farm planning, it is essential to optimize the WTGs’ layout fordecreasing the effect of wake effect and increasing the energy production of wind farm.A geometric theory is used to determine the upstream WTGs and calculate thehorizontal and deviation distances between any two WTGs along the wind direction. Aparticle swarm optimization algorithm with the shrinkage factor and mutation factor isused to obtain the optimal solution with maximum energy production of a wind farm.The discrete joint distribution probability of wind direction and wind speed isincorporated into the optimization process. Based on the proposed technique, theoptimal wind farm layout can be obtained using the historical wind speed and winddirection data of the practical region. A modified reliability test system containing awind farm with the optimal layout is used to evaluate the reliability of a generationsystem. The results show that the proposed technique is effective to decrease the wakeeffect, to increase the energy production of wind farm, and to improve the reliability ofpower system.Compared with conventional generators, wind power is fluctuant andnon-dispatched. The existing researches on reliability of power system with wind farmcan always consider the contribution of wind power on power system adequacy, butthese researches cannot consider the negative impact of wind power fluctuation onpower system reliability. When the power system operates in the steady-state, the taskof power system is not only to meet load demand, but also to implement real timeadjustment when generation side or demand side is suffered from a disturbance.Therefore, according to the steady state operation characteristic of power system, thepaper presents a reliability evaluation model of power system with wind farmconsidering steady-state frequency characteristics. The fluctuation of wind power can bereflected by the proposed model. The model is built by the load power equivalence, thestate simulation of conventional generator, and the steady-state frequency adjustment.The proposed model is applied to reliability test system, and the results show that theproposed method can not only reflect the contribution of wind power to the overallpower system reliability, but also reflect the negative impact of wind power fluctuationon power system reliability.更多还原