2015年7月1日我国新疆哈密地区发生了世界首例由直驱风机风电场造成的次同步振荡事故，含直驱风机风电场的电力系统次同步振荡问题成为研究热点。传统分析方法难以计及风电场并网系统中数量众多的参数，尤其是各种随机性参数对系统动态特性的影响，因此亟需拓展现有方法，为含风电场的电力系统次同步振荡研究提供可靠的理论基础。本文主要工作如下：针对多种影响因素下如何快速准确地描述系统动态特性的问题，本文提出了一种考虑复杂影响因素的电力系统特征量计算方法。该方法解决了高维参数空间中的抽样问题，并给出了求取特征量在整个参数空间中近似分布的方法。该方法不仅可以计算单个系统内的特征量概率分布，还可对不同系统间的差异进行定量分析，从而为本文后续研究提供了理论工具。针对风电场等值模型能否应用于次同步振荡研究的问题，本文提出了评判等值模型是否适用于次同步振荡分析的指标及其计算方法。通过将等值模型的参数求解问题建模为考虑风电随机性的优化问题，提出以等值模型与详细模型对应特征值在复平面上的平均欧氏距离作为检验等值方法有效性的指标，并给出了考虑随机影响因素时该指标的计算方法。此外，本文还提供了一种评估风电汇集区域内多个风电场进行进一步等值的可行性的方法。针对多种影响因素下如何复现与实际事故场景相一致的等效系统场景问题，本文提出了快速生成存在次同步振荡风险的系统场景并定量描述振荡频率分布范围的方法。通过评价各种工况下系统的阻尼特性，选取了具有次同步振荡风险的系统场景。之后，计算了振荡特征量在随机风电场工况下概率分布。结果表明，多种存在振荡风险的场景中，具有宽频率分布特性，且阻尼比可能为负的场景才能解释实际系统中的次同步电流频率变化特性。针对风电场内其他设备对次同步振荡的影响，本文提出了一种考虑SVG控制作用的次同步振荡机理。讨论了弱电网下不同控制特性逆变器的相互作用，指出不同风电场内SVG可能由于控制目标冲突导致系统出现次同步振荡。通过对双风电场模型的分析说明了各类因素对系统次同步振荡模式的影响，并验证了其与所提机理预测相符。通过计算该机理下系统振荡频率分布，说明了所提机理可以解释实际系统中的次同步电流频率变化特性。
The sub-synchronous oscillation incident on July 1, 2015 in the Hami region of Xinjiang was the world’s first subsynchronous oscillation caused by direct-driven wind generators (DDWG) based wind farms. Since then the sub-synchronous oscillation involved with wind farms has become a research hotspot. Conventional analysis methods for sub-synchronous oscillation can hardly take into account the effects of numerous factors of grid-connected wind farms, especially various random factors. Therefore, it is urgent to improve the existing analysis methods to provide a reliable theoretical basis for the study of the sub-synchronous oscillation in power systems involved with wind farms. The main work of this thesis are as follows:Subject to the problem of how to fast and accurately describe system dynamics under various influencing factors, this thesis proposes a power system feature calculation method that takes into account the complex influencing factors. This method figures out the sampling problem in the high-dimensional parameter space and calculates the approximative distribution of the feature quantity in the entire parameter space. This method can not only calculate the distribution of feature quantities in a single system, but also can quantitatively analyze the differences between different systems under random influencing factors, thus providing a theoretical tool for the follow-up study of this thesis. In order to answer whether an equivalent model of wind farm can be applied to the study of sub-synchronous oscillation, this thesis proposes an index and its calculation method to evaluate whether the equivalent model is suitable. By modeling the parameter-solving problem of the equivalent model as an optimization problem, and explicating the significance of the problem considering randomness, the average Euclidean distance in the complex plane between the eigenvalues of the equivalent model and the detailed model is chosen as the index to evaluate the suitability. The calculation method for the index is also given when considering the random factors. In addition, this thesis provides a method for assessing the feasibility of equivalence between different wind farms.In order to answer the problem of how to reproduce an equivalent system scenario that is consistent with the actual system under a variety of influencing factors, this thesis provides a method to rapidly generate system scenarios with subsynchronous oscillation risks and to quantitatively describe the distributing range of the oscillation frequency. By evaluating the damping characteristics of the system under various operating conditions, the scenarios with subsynchronous oscillation risks are selected. After that, the operating conditions of the wind farms are considered as random variables to calculate the distributions of the oscillation characteristics. The results show that the system in which there is a mode with a wide frequency distribution and a certain probability of negative damping ratio is closer to the actual system.In order to study the influence of other equipment in the wind farm on sub-synchronous oscillation, this thesis proposes an oscillation mechanism considering the control of SVG. The interaction of inverters with different control characteristics is discussed. It is proposed that SVGs in different wind farms may cause sub-synchronous oscillations due to conflict of their control targets. The influence of various factors on the sub-synchronous oscillation modes is illustrated, and it is verified that the influence is consistent with the prediction of the proposed mechanism. By calculating the frequency distribution range of the sub-synchronous oscillation under this mechanism, it is shown that the proposed mechanism can explain the variation of the sub-synchronous current frequency that appears in the actual system.
