An updating kernel density estimation method for guided wave-based quantitative damage diagnosis under vibration conditions

Abstract

In practical applications, engineering structures are typically subjected to complex vibrational loads during service, resulting in structural failures and accidents. Guided-wave-based structural health monitoring (SHM) is a promising method to ensure safe operations. However, the propagation of guided waves is sensitive to structural vibrations, which reduces the diagnostic accuracy of guided-wave-based SHM methods. This paper addresses the challenges in SHM under vibration conditions using guided waves and proposes an updating kernel density estimation method for quantitative damage diagnosis. First, this study investigated the impact of various vibration conditions on guided waves and introduced a new damage index: the instantaneous phase synchronization damage index. In addition to the Pearson coefficient damage index, a two-dimensional feature vector was constructed. Based on this, the baseline and updating feature vector sets were established, and the corresponding kernel density was estimated online during damage monitoring. Finally, multi-path distance metrics were employed as a quantitative damage index for diagnosis. The method was verified using an aluminum alloy plate with varying degrees of crack damage under vibration conditions. The experimental results demonstrated that the proposed method can quantitatively monitor damage under different vibration conditions, achieving high accuracy in estimating the crack length.