Research on denoising method for low-frequency fiber Bragg grating sensing signal based on ICEEMDAN and SampEn combined with WT

The detection of Fiber Bragg Grating (FBG) sensing signals is of significant importance in fields such as structural health monitoring, environmental monitoring, and earthquake early warning. To effectively filter out the substantial noise that FBG sensing signals encounter in complex environments, a joint denoising method for FBG sensing signals is proposed based on Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) and Sample Entropy (SampEn) combined with Wavelet Thresholding (WT). First, ICEEMDAN is used to decompose the signal into Intrinsic Mode Functions (IMFs). Then, the sample entropy, variance contribution rate, and correlation coefficient of the IMFs are calculated to identify and discard noisy IMFs. Subsequently, the retained IMFs are denoised using the WT method, and finally, a linear reconstruction is performed to obtain the denoised signal. Experimental results show that compared with existing WT thresholding denoising methods and ICEEMDAN-WT denoising methods, this method reduces the denoising error ratio by 3.60 dB and 2.54 dB, respectively, and reduces the root mean square error by 82.94% and 79.26%, respectively. This indicates that the proposed method outperforms the other two methods in terms of denoising effectiveness, smoothness, and stability of the denoised signal. The proposed method demonstrates significant advantages in processing low-frequency signals in structural health monitoring and earthquake early warning systems, improving the reliability and accuracy of the signals, and has broad potential for practical applications.