基于小波变换的光纤光栅测频辨识

Study on Frequency Identification on Wavelet Transform Using Fiber Bragg Grating Measuring Technique

针对光纤光栅测量振动模态频率噪声较大的问题,提出采用小波变换去噪。同时获取时频信号,判断光纤光栅测量模态频率结果的有效性,对模态频率进行辨识。从直接快速傅里叶变换方法获得的频域数据分析来看,振动模态频率为7阶,而利用有限元分析和CRAS模态测试平台的结果均为5阶,因此对光纤光栅测频结果存在误判。用小波变换进行去噪处理后,所得测试结果为5阶主要振动模态频率,与有限元分析和CRAS模态系统测试结果相近,分别为:27.25、232.00、423.25、692.50、1 390.25 Hz。其信噪比依次约为65、35、30、23、20 dB,信号的信噪比随着模态频率阶次的增加而降低。因此,采用小波分析可以有效防止光纤光栅测频误判。

Wavelet transform has been proposed to de-noise because high frequency noise was very strong and unavoidable when modal frequencies were measured based on fiber Bragg grating（FBG） test bench.Simultaneously,time-frequency signals were obtained to testify the validity of experimental results.According to frequency domain data analyzed on the basis of direct fast Fourier transform（FFT）,seven stages of dominant modal frequencies were obtained.However,the results based on finite element analysis and computer random signal and vibration analysis system（CRAS） modal test bench demonstrated that five stages of dominant modal frequencies were obtained.Accordingly,dominant modal frequencies using FBG test bench were misjudged.After high frequency noise was removed using wavelet transform,five stages of dominant modal frequencies using FBG test bench were obtained,i.e.27.25Hz,232.00Hz,423.25Hz,692.50Hz and 1390.25Hz,respectively,which yielded good agreement with those based on finite element analysis and CRAS modal test bench.Signal-to-noise ratios（SNR） were of 65dB,35dB,30dB,23dB and 20dB,respectively.Besides,the amplitude decreased against the frequency increment.Therefore,wavelet transform is an effective approach to avoiding misjudging the modal frequencies based on FBG measuring technique.