基于特征子空间分解的高精度实时电网频率测量方法研究

With the rapid development of smart grid and microgrid, accurate and real time power grid frequency measurement is increasingly important to safe, reliable and efficient performance of power grid. Fully considering the waveform distortions and time-varying characters of power grid electrical signal, this project introduces eigen-subspace decomposition algorithm into power grid frequency measurement. The high frequency resolution character of eigen-subspace decompostion algorithm is used to eliminate frequency measurement errors caused by the interferences of waveform distortions and time-varying disturbances in the electrical measurement signal, thus improving power grid frequency measurement precision. Subspace iteration and update techniques are used to reduce the computation complexity and implement real time tracking and estimation of eigen-subspace, thus improving the response speed of power grid frequency measurement method. The factors, such as measurement signal model, sample data window length and estimation methods for measurement signal order, which affect the frequency measurement precision and response speed are studied, and frequency measurement precision and response speed are improved by the accurate construction of time-varying measurement model, approporiate setting of data window length and precise estimation of signal order. Combining the above work, a high precision, real time power grid frequency measurement method based on eigen-subspace decomposition is built. Simulation experiments and theoretical analysis are developed to analyze and evaluate the proposed method in order to fulfill the accuracy and real time requirements for power grid frequency measurement.

随着智能电网和微电网技术的快速发展，准确、实时的电网频率测量对电网安全、可靠和高效运行的意义越来越重要。本项目针对电气观测信号波形畸变及动态时变的特点，把特征子空间分解方法引入电网频率测量，利用特征子空间分解方法高分辨率特性消除观测信号中各种波形畸变和动态时变干扰对频率测量的影响，提高电网频率测量的精度；通过子空间迭代和更新算法减小特征子空间分解的计算复杂度，实现特征子空间的实时跟踪和估计，提高频率测量算法的响应速度；研究观测信号模型、采样数据窗长度及信号阶次估计方法对频率测量精度和算法响应速度的影响，通过时变观测模型的正确建立、数据窗长度的合理选取、信号阶次的准确估计等方法进一步提高电网频率测量精度和算法响应速度。在上述工作的基础上，建立基于特征子空间分解的高精度实时电网频率测量方法，并通过仿真实验和理论分析的手段评价和比较所建立方法的性能，以满足电网频率测量的准确性和实时性要求。

在国家自然科学基金青年基金项目“基于特征子空间的电网频率高精度在线测量方法研究”（No.51207160）的资助下，项目负责人在国际学术期刊IEEE Transactions上以第一/通讯作者发表论文3篇，申请国家发明专利3项，培养毕业硕士生2名，在读硕士生3名。项目负责人2014年获批国家自然科学基金面上项目1项。项目负责人被评为2013年教育部新世纪优秀人才。.主要工作进展包括：① 研究了低频振荡干扰条件下，频率和相量测量误差，提出了连续点DFT方法以提高低频振荡条件下的频率及相量测量精度。② 研究频率偏移带来的相量测量误差，提出了移动窗平均方法以消除频率偏移时导致的相量测量误差。③ 提出了一种基于FIR滤波器滤波的频率测量方法，该方法具有良好的谐波和负频率分量抑制能力。④ 研究了幅值变动对频率测量的影响，并根据误差分析提出了一种加窗傅立叶算法以提高幅值变动对电网频率测量的影响。⑤ 提出了一种新的窗函数，该窗函数具有更高的负频率抑制能力。该方法可解决现有技术中矩形窗和傅里叶变换中较多的高频干扰和频谱能量泄漏的问题。⑥ 提出了一种电力系统中相量测量方法,该方法通过对采样信号的分析模型进行变形和计算得到采样信号的相量参数,并通过所述相量参数对电力系统中的相量进行分析,所提出方法能够在含噪声、谐波/间谐波和多直流衰减分量的影响下得到精确的相量、谐波/间谐波和直流衰减分量分析。