针对变压器有效振动信号受噪声干扰难以提取的问题,提出一种改进群分解-奇异值分解(Improved Swarm Decomposition-singular Value Decomposition,ISWD-SVD)的变压器振动信号降噪方法.该方法首先将功率谱熵负值作为目标函数,利用麻雀搜索算法(Sparrow Search Algorithm,SSA)寻找群分解算法(Swarm Decomposition,SWD)最优参数;然后,采用最优参数对变压器振动信号进行SWD分解,并剔除剩余分量,得到重构信号;最后,利用SVD去除重构信号中的噪声残留,实现二次降噪.通过对仿真信号与现场信号进行降噪效果验证,并与其他降噪算法进行对比分析.结果表明:ISWD-SVD联合方法对变压器振动信号具有更好的降噪效果,可为变压器机械状态检测和故障诊断提供有力依据.
Transformer Vibration Signal Denoising Based on ISWD-SVD Joint Method
Aiming at the problem that transformer vibration signal disturbed by noise is difficult to be extracted,a method of signal de-noising based on improved swarm decomposition-singular value decomposition (ISWD-SVD) was proposed.Firstly,the negative power spectrum entropy was employed as the objective function,and the optimal parameters of swarm decomposition (SWD) were searched by sparrow search algorithm (SSA).Then,the optimal parameters were used to decompose the transformer vibration signals by SWD,and the residual component was removed to obtain the reconstructed signal.Finally,the singular value decomposition (SVD) was used to remove the noise residue from the reconstructed signal to achieve secondary denoising.The effect of this method for noise reduction was verified by simulation signal and actual signal,and the above method was compared with other noise reduction algorithms.The results show that the ISWD-SVD joint method exhibits a better denoising performance for transformer vibration signals,which provides a strong basis for transformer mechanical state detection and fault diagnosis.
fault diagnosistransformervibration signalimproved swarm decompositionsingular value decompositionnoise reduction
万方数据
维普
