Scholarlink

Acoustic emission technology has been gradually applied in the damage diagnosis of concrete beams in engineering, but the acoustic emission signal is susceptible to noise interference, resulting in the accuracy of damage identification can not meet the requirements. In order to reduce the noise level of acoustic emission signals and improve the accuracy of damage identification of concrete beams, a two-stage denoising method (GJO-VMD-WT) based on wavelet denoising (WT) and Golden Jackal Optimization (GJO) to improve variational mode decomposition (VMD) is proposed. Firstly, the initial population of the golden jackal optimization algorithm is set by Tent chaotic map, and the VMD parameters are iteratively optimized to obtain the optimal decomposition level K and penalty factor \(\alpha\) , , so as to establish the golden jackal optimization algorithm improved variational mode decomposition method (GJO-VMD). Then GJO-VMD was used to decompose the original signal into a series of Intrinsic mode functions (IMF), the correlation coefficient r of IMF was calculated, and the IMF was divided into low, medium and high signal components according to the correlation coefficient \(r\) , and the medium correlation signal components were selected for wavelet analysis for denoising. Finally, the low correlation component signal was eliminated, and the denoised medium correlation component signal and the original high correlation component signal were reconstructed to obtain the final acoustic emission denoising signal. In order to verify the effectiveness of the method, the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Wavelet Denoising Method (WT) and CEEMDAN-WT are used to denoise the same set of signals. The results showed that the GJO-VMD-WT retained the effective components of the signal and suppressed the low-frequency noise and high-frequency noise. Compared with WT, CEEMDAN and CEEMDAN-WT, the signal-to-noise ratio before and after denoising increased by 3.6989db,9.7891db and 5.6191db, respectively. Therefore, the denoising method proposed in this paper is obviously superior to the other three methods, and has obvious advantages in acoustic emission signal denoising and enhancement.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Noise Reduction Study of Damage Acoustic Emission Signal from Concrete Beams Based on GJO-VMD-WT

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Wenxin Cai,
  • Xianqiang Wang,
  • Siyu Cai,
  • Yang Yu

摘要

Acoustic emission technology has been gradually applied in the damage diagnosis of concrete beams in engineering, but the acoustic emission signal is susceptible to noise interference, resulting in the accuracy of damage identification can not meet the requirements. In order to reduce the noise level of acoustic emission signals and improve the accuracy of damage identification of concrete beams, a two-stage denoising method (GJO-VMD-WT) based on wavelet denoising (WT) and Golden Jackal Optimization (GJO) to improve variational mode decomposition (VMD) is proposed. Firstly, the initial population of the golden jackal optimization algorithm is set by Tent chaotic map, and the VMD parameters are iteratively optimized to obtain the optimal decomposition level K and penalty factor \(\alpha\) , , so as to establish the golden jackal optimization algorithm improved variational mode decomposition method (GJO-VMD). Then GJO-VMD was used to decompose the original signal into a series of Intrinsic mode functions (IMF), the correlation coefficient r of IMF was calculated, and the IMF was divided into low, medium and high signal components according to the correlation coefficient \(r\) , and the medium correlation signal components were selected for wavelet analysis for denoising. Finally, the low correlation component signal was eliminated, and the denoised medium correlation component signal and the original high correlation component signal were reconstructed to obtain the final acoustic emission denoising signal. In order to verify the effectiveness of the method, the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Wavelet Denoising Method (WT) and CEEMDAN-WT are used to denoise the same set of signals. The results showed that the GJO-VMD-WT retained the effective components of the signal and suppressed the low-frequency noise and high-frequency noise. Compared with WT, CEEMDAN and CEEMDAN-WT, the signal-to-noise ratio before and after denoising increased by 3.6989db,9.7891db and 5.6191db, respectively. Therefore, the denoising method proposed in this paper is obviously superior to the other three methods, and has obvious advantages in acoustic emission signal denoising and enhancement.