风电渗透率的快速增长对电力系统频率稳定性和频率质量产生显著影响，严重时还可能导致系统频率失稳。现有风电机组参与调频控制的设计一般仅考虑孤立的频率大扰动事件，而系统的持续频率调节默认由常规机组提供。本论文考虑面向持续调频的风电机组控制设计与优化问题，围绕调频能力评估、控制设计、性能优化和协调等方面进行了深入研究。论文的主要工作如下：第一，提出了一种改进的风电机组调频能力在线快速评估方法。所提方法考虑风电机组调频期间捕获风能的变化，推导了风电机组调频能力与输入风速关系的二次近似显式表达式，可用于快速、准确地评估风电机组调频能力，避免旋转元件中的动能过度释放，保证风电机组参与持续调频过程中的安全性。第二，提出了风电机组自适应变下垂调频控制策略。论文通过分析风电机组参与系统持续调频的技术特点与要求，根据控制过程中的转子转速恢复条件，设计了自适应变下垂调频控制的基本框架；提出了基于Butterworth函数的自适应变下垂调频控制设计方法，可实现风电机组调频控制与转速恢复控制间的自适应平滑切换，使之适于参与电力系统的持续调频。第三，提出了基于近似动态规划串联附加结构的风电机组调频控制优化方法。论文首先设计了近似动态规划串联附加结构及其在线算法，可对原控制器进行在线优化，易于工程实现；本文进一步将其应用于风电机组调频控制优化，仿真分析表明该方法可有效提升控制器的动态性能。第四，提出了风电场参与持续调频控制的整体协调优化方法。论文通过风电场的等值模型，将单机调频控制优化方法扩展到风电场调频优化中；在此基础上设计了风电场协调优化策略，主要包括风电场集中式优化学习及机组间有功出力的协调分配；仿真分析表明该方法可以有效地改善风电场参与持续调频的性能水平。综上，本文提出了一套风电机组参与系统持续调频的控制与优化方法，可有效解决风电机组难以参与持续调频的问题。
As the penetration of wind power generation dramatically increases in recent years, the frequency stability and quality of the power system have been facing great challenges. Different from the exisiting frequency controllers of wind generators that are designed particularly for isolated and severe frequency disturbance events, this thesis addresses the control design and optimization problem of wind generators for the prupose of participating in sustained frequency regulation of power systems. The main contributions of the thesis are summarized as follows.(1) An improved evaluation method is proposed for online evaluating the frequency requlation capability of wind generators. By analyzing the change of captured wind power by wind turbines during frequency support process, the relationship between the frequency regulation capability of wind generators and the wind speed is revealed. This relationship is further characterized by a closed-form expression using quadratic approximation, enabling an online evaluation of the capability of wind generators participating in frequency regulation. It helps avoid overspeed or stalk of wind generators due to excessive use of the kinetic energy stored in rotating components, and then ensure the security of wind generators during sustained frequency regulation. (2) An adaptive variable-droop frequency control strategy is devised for wind generators. The thesis first analyzes the technical issues and requirements of wind generators when they participate in sustained frequency regulation. Based on the speed recovery condition of wind generators, a framework of adaptive variable-droop frequency control is estabilished for wind generator, naturally merging frequency support and speed recovery (which are two traditionally separated functions). Then a Butterworth-function based implementation method is proposed to smooth the switches between frequency support and speed recovery, while avoiding mechanical wear during frequency regulation. In this way, wind generators are empowered to participate in sustained frequency regulation of power systems. (3) An approximate dynamic programming (ADP) based method is proposed for optimizing the frequency controller of wind generators. Different from the traditional method, a series ADP-based add-on structure is designed with an online learning algorithm. With no need to replace or change the original controller, the ADP based add-on controller can have the original controller online optimized during operation, which is very easy to implement. This method is then applied to the optimization of frequency controller of wind generators, particualry for sustained frequency regulation. Simulation results comfirm that the proposed method can effectively enhance the dynamic performance of the frequency controller of wind generators. (4) A coordination method is proposed for optimizing the frequency control of wind farms for sustained frequency regulation. Based on the equivalent model of wind frams, the frequency controller optimization method for a single generator is extended to wind farms. Then a heirachical coordination and optimization strategy is devised, which includes the centralized optimization learning of wind farm and the coordination and distribution of active power output among wind generators. Simulation results show that the proposed method can effectively improve the pefromance of wind farms participating sustained frequency regulation.To sum up, this thesis comperehensvely addresses the controller design and optimization issues for wind power generation to participating in sustained frequency regulation, which contributes to the development of grid-friendly operation of wind power generation, and many other types of renewable generation.
