摘要
电网原始瞬态波形与噪声的离散GABOR分布特征不同,原始瞬态波形的离散GABOR分布集中,幅值较大;而噪声的离散GABOR分布均匀,能量一定,幅值较低,故可通过选择合适的噪声抑制阀值构造时频掩模函数和多次GABOR展开迭代计算,以实现原始瞬态波形和噪声的有效分离。该文利用了电网瞬态信号与噪声GABOR分布特征的不同,构造时频掩模函数,并通过基于离散GABOR变换的迭代算法实现噪声有效滤除。为了验证该算法在去噪的同时能有效保留原信号,文中用IEC1083-2 TDG(Test Data Generator)产生各种具有不同噪声水平的瞬态电压信号,进行去噪和重现原始瞬态信号试验研究,结果表明该算法能有效地消除噪声干扰, 噪声滤除后信号的NRMS误差不超过1%,可准确地重现各种原始瞬态电压信号。且能保持其原始特征。该迭代算法适用于各种电网瞬态信号去噪。
Electric power transient signal & its noise has different GABOR distribution in time-frequency domain. The signals GABOR distribution is concentrated while noise GABOR distribution is evenly scattered in time-frequency domain. By masking the desired signal from its discrete GABOR transform, an effective noise filtering is achieved via suitable noise offset value selection & iterative algorithm based on discrete GABOR expansion. To verify the effectiveness of noise filtering algorithm, IEC1083-2 TDG (test Data Generator) is used to simulate electric power transient signal with different noise level. Calculation result from IEC1083-2 TDG shows that the algorithm is very effective in noise filtering as desired signal NRMS error after noise filtering is less than 1% & original signal features are remain unchanged. This iterative algorithm is applicable for de-noising of various transient signal in electric power system.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2004年第11期69-73,共5页
Proceedings of the CSEE
