基于小波软硬结合阈值降噪的混凝土坝地震响应信号处理方法

Signal Processing Method for Seismic Response of Concrete Dams Based on Wavelet Soft-Hard Combined Threshold Denoising

摘要: 混凝土坝地震响应信号是开展坝体动力特性分析与抗震安全评估的核心基础，实测信号在采集过程中易受环境干扰、设备噪声等影响混入大量噪声成分，直接制约了信号分析与结构状态识别的精度。小波阈值降噪是信号降噪的常用方法，但传统硬阈值函数存在间断点、软阈值函数存在固定估计偏差，难以满足地震响应信号高精度处理的需求。针对传统小波阈值方法的不足，本文提出基于软硬结合阈值的小波降噪方法，并将其应用于混凝土坝地震响应信号处理。验证结果表明，该方法降噪效果优于传统软、硬阈值方法，能够有效提升信号信噪比与重构精度，可为混凝土坝地震响应信号处理提供可靠的技术支撑。

Abstract: Seismic response signals constitute the fundamental basis for dynamic characteristic analysis and seismic safety evaluation of concrete dams. Field-measured signals are prone to be mixed with substantial noise induced by environmental interference and equipment noise during acquisition, which severely limits the accuracy of signal analysis and structural state identification. Wavelet threshold denoising is a widely adopted technique for signal denoising. However, the conventional hard threshold function suffers from discontinuities, and the soft threshold function has inherent estimation bias. Neither can satisfy the high-precision processing requirements of seismic response signals. To address these drawbacks of traditional wavelet thresholding methods, this paper proposes a wavelet denoising method based on a soft-hard combined threshold and applies it to seismic response signal processing of concrete dams. Verification results indicate that the proposed method achieves superior denoising performance compared with traditional soft and hard threshold approaches. It can effectively increase the signal-to-noise ratio and reconstruction accuracy, thereby offering reliable technical support for seismic response signal processing of concrete dams.