中文摘要：

为了获取类别少、物理含义明确的识别参数，采用具有良好时频域局部化特性的连续复小波变换，并改进现有的复小波窗函数，通过调整窗函数的带宽参数较好地改善了连续复小波变换用于基波检测的动态特性以及高频分析的频率分辨率。采用基于规则匹配的专家系统，以同时识别多重电能扰动类型。MATLAB仿真结果表明，提出的方法能同时准确提取包含谐波、间谐波、频率波动以及电压变动等在内的各电能扰动类型的特征参数，并能高效准确地识别同时发生的多重电能扰动类型。

英文摘要：

To obtain recognition parameters with less categories and more definite physical meanings, continuous complex wavelet with good localization characteristics in time and frequency domain was used, and the complex wavelet＇s window function was improved. By adjusting the window function＇s bandwidth parameters, dynamic characteristics of continuous complex wavelet transform used in fundamental detecting and frequency resolution in the high frequency analysis were greatly improved. To identify multiple power disturbance types simultaneously, an expert system based on rule matching was used. The simulation results from MATLAB indicate that the method can extract accurate characteristic parameters of power disturbances, including harmonics, inter-harmonics, frequency fluctuating and voltage variations. Thus, it can identify concurrent multiple power disturbance types with high efficiency and accuracy.