摘要
同步发电机实测辨识中次暂态参数的辨识精度与权函数的选择密切相关,然而如何选择目前尚缺乏系统的研究。该文针对基于时频变换的参数辨识方法,提出了以灵敏度分析为依据的权函数选择原则。首先构建了包含机端电气量和待辨识参数的矩阵方程,推导出机端可测电气量对每个参数的灵敏度时变曲线(即轨迹灵敏度)公式。然后,根据轨迹灵敏度的大小确定合适的加权区间和权重系数。最后,对比研究了权函数选择在不同试验工况和样本数据长度下对辨识结果的影响,并定量评估了基于时频变换的同步发电机参数辨识算法的系统误差。实例计算表明,文中提出的权函数选择方法可有效提高发电机参数的辨识精度和稳定性。
The identification accuracy of the sub-transient parameters is closely related to the choice of the weight function in synchronous generator parameter identification under field measured conditions, but it is lack of systematic research on how to choose it. Combined with the timefrequency transform based parameter identification method, we established the criterion of choosing weight functions based on sensitivity analysis. Firstly, matrix equations involving terminal electric quantities and parameters to identify were constructed, and formula of the time varying curves of sensitivity of the measurable electric quantity with respect to each parameter（trajectory sensitivity） were derived. Secondly, the appropriate weighting interval and weight coefficients were determined according to the trajectory sensitivity. Lastly, the influences of weight functions on the identification results under different experimental conditions and sample data length were compared. The example shows that the method can effectively improve accuracy and stability of identification of generator parameters.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2016年第3期828-835,共8页
Proceedings of the CSEE
基金
国家自然科学基金项目(51107122)
国家电网公司2014年基础性前瞻性科技项目(XT71-14-006)~~
关键词
同步发电机
参数辨识
轨迹灵敏度
权函数
synchronous generator
parameters identification
trajectory sensitivity
weighted function
