小波阈值降噪在砌体基本频率识别中的应用

Application of wavelet threshold denoising to basic frequency identification for masonry

为更加准确地获取砌体试件的基本频率,在频谱分析前,对振动信号进行降噪处理.对6个砖砌体试件分别进行环境激励下的动力测试,运用频域峰值点法进行频率识别,采取直接识别和小波阈值降噪后再识别2种手段.比较通过2种途径识别出的基本频率的分布规律,发现运用小波阈值降噪识别的基本频率的变异系数更小.表明在直接识别法识别基本频率的基础上,小波阈值降噪技术能进一步提高频率识别效率,可以应用于砌体结构的现场检测.

To obtain the basic frequency of masonry specimens more accurately, vibration signal can be denoised before spectral analysis. Ambient vibration tests are conducted for six brick masonry specimens; and then basic frequencies of the specimens are identified by peak-picking method （PPM）. The PPM can he conducted in two manners. One is direct identification, and the other is identification after denoising by wavelet threshold method. Comparison of the identification results by the two manners shows that the variation coefficient of the basic frequency identified by the indirect identification is smaller. The wavelet threshold denosing increases the efficiency of frequency identification, and can be used in the inspections for masonry structure in-situ.