CAJ | 학술논문

小波阈值降噪是水电机组振动信号除噪常用的方法，传统降噪算法是软、硬阈值函数，对于硬阈值函数处理过的信号，其重构后的信号在阈值处是间断的，易产生附加振荡；而经软阈值函数降噪后的信号虽然连续性好，但与原始信号之间存在着恒定的偏差，影响重构精度。因此在传统软、硬阈值函数的基础上提出了一种改进阈值函数的小波降噪算法，通过Matlab仿真和电厂采集数据的验证表明，该方法克服了软、硬阈值函数算法的缺点，去噪效果明显。通过比较不同降噪方法对振动信号特征分量的保持程度，说明该方法在各分量保持均优于传统的阈值方法，是一种有效的降噪方法。
소파역치강조시수전궤조진동신호제조상용적방법，전통강조산법시연、경역치함수，대우경역치함수처리과적신호，기중구후적신호재역치처시간단적，역산생부가진탕；이경연역치함수강조후적신호수연련속성호，단여원시신호지간존재착항정적편차，영향중구정도。인차재전통연、경역치함수적기출상제출료일충개진역치함수적소파강조산법，통과Matlab방진화전엄채집수거적험증표명，해방법극복료연、경역치함수산법적결점，거조효과명현。통과비교불동강조방법대진동신호특정분량적보지정도，설명해방법재각분량보지균우우전통적역치방법，시일충유효적강조방법。
Threshold de-noising based on wavelet is a commonly used algorithm in the de-noising of the hydropower units vibration signal .The conventional threshold functions can be divided into hard threshold function and soft threshold func-tion .for hard-threshold function processed signal ,the reconstructed signal is discontinued at the threshold ,which is prone to cause additional oscillations to the reconstructed signal ;whereas the signal de-noised by soft threshold function has a better overall continuity ,but there’s always a constant deviation between the reconstructed signal and the original one ,which is prone to affect the reconstruction accuracy .Based on this situation ,a wavelet de-noising algorithm with an improved threshold function was brought up in this article .The Matlab simulation with the actual collected data indicates that the new method overcomes the drawbacks of the conventional ones ,and it has a good de-noising effect .Comparing with the maintaining level of different de-noising methods on the vibration signal’s characteristic components ,the new method shows a better performance on each component .In summary ,it’s an effective de-noising method .