一种新的电能质量扰动信号去噪方法

A New Method of Power Quality Disturb De-noising

针对电能质量扰动信号去噪过程的简化和硬件实现问题,提出基于信号相关性处理的跨小波尺度空间去噪方法。基于电能质量扰动信号不同成分在小波尺度空间上的相关性不同,新方法把原信号分为两个数据长度相等的新信号,分别进行小波变换,跨两个尺度空间对相同分解深度的尺度系数进行相关性处理和软阈值处理,合并后的信号可滤除噪声成分。仿真信号去噪前后的赋范均方误差小且具有更好的检测效果。算法计算复杂度分析结果和实际运行表明,在高速DSP上可以实时实现该算法。

An approach based on correlation processing in two sets of wavelet transform coefficients is proposed for the simplification of power quality （PQ） disturbance de-noising and its application in device. PQ disturbance is split into odd and even sampling so as to get two simultaneous disturbances with the same data length. Wavelet transform, correlation processing have been done between the detailed coefficients of the same scales. Soft threshold processing and wavelet reconstructions get the PQ de-noised disturbance as different content in PQ disturbance results in different correlation. The approach gets small de-noising error and is more effective in PQ disturbance detection. The computing complexity analysis of the algorithm shows that DSP microprocessor is competent for its real time implementation.