中文摘要：

鉴于S变换时频分辨率低、计算量大,实际应用受限,该文构建改进S变换算法,提出基于改进S变换的电能质量扰动信号特征提取方法。首先计算采样信号的快速傅里叶变换（fast Fourier transform,FFT）频谱,创立基于包络极值算法的特征频率保留机制与信号无关频率点的自动识别机制,剔除非特征信息;然后依据信号频率与频段分析需求,引入新的窗宽调节尺度因子构成Gauss自适应优化窗,克服传统Gauss窗主瓣较宽、频率分辨率低的局限,构建基于Gauss自适应优化窗的改进S变换;对特征频率点进行变换处理,提取特征向量,实现扰动信号的自适应检测。仿真分析和试验结果表明,本文提出的改进S变换算法提高了电能质量扰动信号的时频分析准确率,计算量小,适用于电力系统扰动信号的准确、快速检测。

英文摘要：

Since the low time-frequency resolution and heavy computation,the traditional S-transform meets difficulties in practice.The modified S-transform（MST） algorithm was established,and the feature extraction method of power quality disturbance signals based on the modified S-transform was proposed.Firstly,FFT of the signal was calculated,the reserved mechanism based on envelope extremum algorithm and the automatic identification mechanism of independent frequency were established to remove the non-feature information.Secondly,according to the analysis requirements of the signal frequency and spectrum,the new window width adjustment factors were introduced to constitute the adaptive optimizing Gauss window,which could overcome the limitations of the wide main lobe of the Gauss window and the low frequency resolution,and then the modified S-transform based on adaptive optimizing Gauss window was established.Finally,the S-transform of these feature frequency points was performed,and then the adaptive detection of the disturbance signals was achieved after the obtained the feature vectors.The simulation and test results show that the proposed method provides results with higher accuracy in time-frequency analysis without heavy computation,and is appropriate for accurate and rapid detection of power system disturbance signals.