论文部分内容阅读
为了能够为各类电能扰动事件的本质研究和有效治理提供准确、可靠的依据,提出基于时频原子变换(timefrequency transform,TFT)和改进型自组织映射神经网络(improved self-organizing map,ISOM)的电能质量扰动在线监测与识别新方法。TFT具有自适应复带通滤波特性,其频窗中心与频窗半径解耦及频窗宽度不受中心频率的约束,可以灵活调整。通过设置恰当的频域窗口,TFT可有效抑制邻近频率分量的相互干扰,且有较好的动态响应速度。TFT能准确监测电力系统波形中电能质量事件,并为类型识别提供物理意义明确、指标具体的实时模式特征。依据TFT提取的特征向量,ISOM可准确识别单一或同时存在的多重电能质量事件,并对其严重程度进行直观表达,能动态反映电能质量事件各自的发展变化轨迹。仿真验证结果表明了所提出方法的有效性和优越性。
In order to provide an accurate and reliable basis for the essential research and effective governance of various types of electric disturbance events, a time-frequency-based atomic transform (TFT) and an improved self-organizing map (ISOM) A new method of on-line monitoring and identification of power quality disturbance. TFT has adaptive complex band-pass filter characteristics, the center of its frequency window and window frequency decoupling and frequency window width from the center frequency constraints, can be flexibly adjusted. By setting the appropriate frequency domain window, TFT can effectively restrain the mutual interference of adjacent frequency components and have better dynamic response speed. TFT can accurately monitor power quality waveforms in power system waveforms and provide a real-time mode characteristic with definite physical meaning and indicator for type identification. According to the eigenvector extracted by TFT, ISOM can accurately identify multiple or simultaneous multiple power quality events and intuitively express their severity, which can dynamically reflect the respective development trajectories of power quality events. Simulation results show the effectiveness and superiority of the proposed method.