本文提出一种利用S变换局部奇异值的同调机组识别方法。将广域测量系统(Wide Area Measurement System, WAMS)采集到的系统发电机功角信息进行S变换，得到每台发电机的时频信息模值矩阵，将矩阵分割成块，计算各个子块的最大奇异值，利用时频信息模值矩阵中各个子块最大奇异值构造机组特征矩阵，采用分布聚类法对特征矩阵进行聚类分群。IEEE-39节点系统算例表明，该方法能够有效提取功角信息特征，具有很强的抗噪性，能够在不同故障类型下准确识别同调机组。
英文摘要:
A new method of coherency identification using S-transform and local singular value decomposition (SVD) is proposed in this paper. Time-frequency information matrix of each generator is obtained by S-transform of power angle information, which is obtained by WAMS. Calculating the maximum singular value of each sub-block, which is divided by the time-frequency information matrix. The coherency recognition feature matrix is constructed by the maximum singular value of each sub-block. and the coherency grouping is carried out by distributed clustering method. The example of IEEE-39 node system shows that this method can accurately identify coherent generators under different fault types with the characteristic of extracting the valid information of power-angle rocking curve effectively and having strong anti-noise robustness.