提出基于组合滤波和时频原子变换的故障录波数据分析新算法。组合滤波通过推导并分析非同步采样时采样值修正算法对非周期分量的滤波误差公式,提出利用差分法有效消除误差.实现更理想滤波。时频原子变换具有良好的频率特性,能克服非同步采样影响;具有复带通滤波特性.能准确输出基波复相量而不受其他频带分量干扰;利用灵活可调的时频域带宽,能获取较短的计算数据窗。仿真研究表明.所提方法有快速动态特性。能在固定采样率条件下有效滤除非周期分量并实现故障录波数据基波幅值、频率和相位的精确测量.同时具有较好的抗干扰性能。
An analysis algorithm based on CF(Combinational Filter) and TFT(Time-Frequency Translorm) is proposed for fault recording data. The sampling value correction error due to asynchronous sampling is analyzed and it is proposed to use difference method to eliminate the error for achieving ideal filtering. Being immune to the effect of asynchronous sampling,TFT has ideal frequency characteristics. Because of its complex band-pass filter characteristics,it outputs the fundamental complex phasor without the components of other bands. Because the time-frequency bandwidth of TFT can be flexibly adjusted,shorter calculation data window is thus obtained. Simulative results indicate that,with better dynamic performance and anti- interference capability,the proposed algorithm filters effectively the aperiodic components,measures accurately the magnitude ,frequency and phase of fundamental with the fixed sampling rate.